Address/memory/thread sanitizer.

Removing non-present keys is a supported operation on HashMaps,
and most other operations are well-defined on uninitialized
HashMaps.

Currently, removing any key on an uninitialized HashMap will
result in an 'Array index out of bounds' error.

This change guards against such a case.

.gitattributes

@@ -0,0 +1,4 @@
+$ cat .gitattributes
+* text=auto
+*.c3 linguist-language=C
+*.c3t linguist-language=C

.github/FUNDING.yml

@@ -0,0 +1,14 @@
+# These are supported funding model platforms
+
+github: [c3lang]
+patreon: # Replace with a single Patreon username
+open_collective: # Replace with a single Open Collective username
+ko_fi: c3lang
+tidelift: # Replace with a single Tidelift platform-name/package-name e.g., npm/babel
+community_bridge: # Replace with a single Community Bridge project-name e.g., cloud-foundry
+liberapay: # Replace with a single Liberapay username
+issuehunt: # Replace with a single IssueHunt username
+otechie: # Replace with a single Otechie username
+lfx_crowdfunding: # Replace with a single LFX Crowdfunding project-name e.g., cloud-foundry
+polar: # Replace with a single Polar username
+custom: # Replace with up to 4 custom sponsorship URLs e.g., ['link1', 'link2']

.github/workflows/main.yml

@@ -2,70 +2,207 @@ name: CI
 
 on:
   push:
-    branches: [ master, dev, ci_testing ]
+    branches: [ master, dev, ci_testing, experiments ]
   pull_request:
-    branches: [ master ]
+    branches: [ master, dev ]
 
+env:
+  LLVM_RELEASE_VERSION_WINDOWS: 18
+  LLVM_RELEASE_VERSION_MAC: 18
+  LLVM_RELEASE_VERSION_LINUX: 17
+  LLVM_RELEASE_VERSION_UBUNTU20: 17
+  LLVM_DEV_VERSION: 20
 jobs:
-  build-msys2-mingw:
+
+  build-msvc:
     runs-on: windows-latest
     strategy:
+      # Don't abort runners if a single one fails
+      fail-fast: false
+      matrix:
+        build_type: [ Release, Debug ]
+
+    defaults:
+      run:
+        shell: cmd
+    steps:
+      - uses: actions/checkout@v4
+
+      - name: CMake
+        run: |
+          cmake -B build -G "Visual Studio 17 2022" -A x64 -DCMAKE_BUILD_TYPE=${{ matrix.build_type }}
+          cmake --build build --config ${{ matrix.build_type }}
+
+      - name: Compile and run some examples
+        run: |
+          cd resources
+          ..\build\${{ matrix.build_type }}\c3c.exe compile-run -L C:\ --print-linking examples\hello_world_many.c3
+          ..\build\${{ matrix.build_type }}\c3c.exe compile-run --print-linking examples\time.c3
+          ..\build\${{ matrix.build_type }}\c3c.exe compile-run --print-linking examples\fannkuch-redux.c3
+          ..\build\${{ matrix.build_type }}\c3c.exe compile-run examples\contextfree\boolerr.c3
+          ..\build\${{ matrix.build_type }}\c3c.exe compile-run examples\ls.c3
+          ..\build\${{ matrix.build_type }}\c3c.exe compile-run examples\load_world.c3
+          ..\build\${{ matrix.build_type }}\c3c.exe compile-run examples\process.c3
+          ..\build\${{ matrix.build_type }}\c3c.exe compile-run examples\args.c3 -- foo -bar "baz baz"
+          ..\build\${{ matrix.build_type }}\c3c.exe compile --no-entry --test -g -O0 --threads 1 --target macos-x64 examples\constants.c3
+          ..\build\${{ matrix.build_type }}\c3c.exe compile-run msvc_stack.c3
+
+      - name: Build testproject
+        run: |
+          call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64
+          cd resources/testproject
+          ..\..\build\${{ matrix.build_type }}\c3c.exe --debug-log --emit-llvm run hello_world_win32
+          dir build\llvm_ir
+          ..\..\build\${{ matrix.build_type }}\c3c.exe clean
+          dir build\llvm_ir
+
+
+      - name: Build testproject lib
+        run: |
+          cd resources/testproject
+          call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64
+          ..\..\build\${{ matrix.build_type }}\c3c.exe --debug-log build hello_world_win32_lib
+
+      - name: Vendor-fetch
+        run: |
+          build\${{ matrix.build_type }}\c3c.exe vendor-fetch raylib
+
+      - name: Try raylib
+        run: |
+          cd resources
+          ..\build\${{ matrix.build_type }}\c3c.exe vendor-fetch raylib
+          ..\build\${{ matrix.build_type }}\c3c.exe compile --lib raylib --wincrt=none examples\raylib\raylib_arkanoid.c3
+          ..\build\${{ matrix.build_type }}\c3c.exe compile --lib raylib --wincrt=none examples\raylib\raylib_snake.c3
+          ..\build\${{ matrix.build_type }}\c3c.exe compile --lib raylib --wincrt=none examples\raylib\raylib_tetris.c3
+
+      - name: run compiler tests
+        run: |
+          cd test
+          python3.exe src/tester.py ..\build\${{ matrix.build_type }}\c3c.exe test_suite/
+
+      - name: Compile run unit tests
+        run: |
+          cd test
+          ..\build\${{ matrix.build_type }}\c3c.exe compile-test unit -O1
+
+      - name: Test python script
+        run: |
+          py msvc_build_libraries.py --accept-license
+          dir msvc_sdk
+
+      - name: upload artifacts
+        uses: actions/upload-artifact@v3
+        with:
+          name: c3-windows-${{ matrix.build_type }}
+          path: |
+            build\${{ matrix.build_type }}\c3c.exe
+            build\${{ matrix.build_type }}\c3c_rt
+
+  build-msys2-mingw:
+    runs-on: windows-latest
+    # if: ${{ false }}
+    strategy:
+      # Don't abort runners if a single one fails
+      fail-fast: false
       matrix:
         build_type: [Release, Debug]
+
     defaults:
       run:
         shell: msys2 {0}
     steps:
-      - uses: actions/checkout@v2
+      - uses: actions/checkout@v4
 
       - uses: msys2/setup-msys2@v2
         with:
           msystem: MINGW64
           update: true
-          install: git binutils mingw-w64-x86_64-cmake mingw-w64-x86_64-toolchain mingw-w64-x86_64-llvm mingw-w64-x86_64-polly mingw-w64-x86_64-python mingw-w64-x86_64-lld
-
+          install: git binutils mingw-w64-x86_64-clang mingw-w64-x86_64-ninja mingw-w64-x86_64-cmake mingw-w64-x86_64-toolchain mingw-w64-x86_64-python
+      - shell: msys2 {0}
+        run: |
+            pacman --noconfirm -U https://mirror.msys2.org/mingw/mingw64/mingw-w64-x86_64-llvm-18.1.8-1-any.pkg.tar.zst
+            pacman --noconfirm -U https://mirror.msys2.org/mingw/mingw64/mingw-w64-x86_64-lld-18.1.8-1-any.pkg.tar.zst
       - name: CMake
         run: |
-          mkdir build && cd build
-          cmake .. -G "MinGW Makefiles" -DCMAKE_BUILD_TYPE=${{ matrix.build_type }}
-          cmake --build .
+          cmake -B build -G Ninja -DCMAKE_C_COMPILER=clang -DCMAKE_BUILD_TYPE=${{ matrix.build_type }}
+          cmake --build build
+
+      - name: Compile and run some examples
+        run: |
+          cd resources
+          ../build/c3c compile-run --print-linking examples/hello_world_many.c3
+          ../build/c3c compile-run --print-linking examples/time.c3
+          ../build/c3c compile-run --print-linking examples/fannkuch-redux.c3
+          ../build/c3c compile-run --print-linking examples/contextfree/boolerr.c3
+          ../build/c3c compile-run --print-linking examples/load_world.c3
+          ../build/c3c compile-run --print-linking examples/args.c3 -- foo -bar "baz baz"
+          ../build/c3c compile --no-entry --test -g -O0 --threads 1 --target macos-x64 examples/constants.c3
 
       - name: Build testproject
         run: |
           cd resources/testproject
-          ../../build/c3c build
+          ../../build/c3c run --debug-log
+
+      - name: Vendor-fetch
+        run: |
+          ./build/c3c vendor-fetch raylib
+
+      - name: Build testproject lib
+        run: |
+          cd resources/testproject
+          ../../build/c3c build hello_world_lib --cc cc --debug-log
 
       - name: run compiler tests
         run: |
           cd test
           python3 src/tester.py ../build/c3c.exe test_suite/
 
+
   build-msys2-clang:
     runs-on: windows-latest
+    #if: ${{ false }}
     strategy:
+      # Don't abort runners if a single one fails
+      fail-fast: false
       matrix:
         build_type: [Release, Debug]
+
     defaults:
       run:
         shell: msys2 {0}
     steps:
-      - uses: actions/checkout@v2
+      - uses: actions/checkout@v4
 
       - uses: msys2/setup-msys2@v2
         with:
           msystem: CLANG64
-          update: true
+          update: false
           install: git binutils mingw-w64-clang-x86_64-cmake mingw-w64-clang-x86_64-toolchain mingw-w64-clang-x86_64-python
 
       - name: CMake
         run: |
-          mkdir build && cd build
-          cmake .. -G "MinGW Makefiles" -DCMAKE_BUILD_TYPE=${{ matrix.build_type }}
-          cmake --build .
+          cmake -B build -G "MinGW Makefiles" -DCMAKE_BUILD_TYPE=${{ matrix.build_type }}
+          cmake --build build
+
+      - name: Compile and run some examples
+        run: |
+          cd resources
+          ../build/c3c compile-run examples/hello_world_many.c3
+          ../build/c3c compile-run examples/time.c3
+          ../build/c3c compile-run examples/fannkuch-redux.c3
+          ../build/c3c compile-run examples/contextfree/boolerr.c3
+          ../build/c3c compile-run examples/load_world.c3
+          ../build/c3c compile-run examples/args.c3 -- foo -bar "baz baz"
+          ../build/c3c compile --no-entry --test -g -O0 --threads 1 --target macos-x64 examples/constants.c3
       - name: Build testproject
         run: |
           cd resources/testproject
-          ../../build/c3c build
+          ../../build/c3c run --debug-log
+
+      - name: Build testproject lib
+        run: |
+          cd resources/testproject
+          ../../build/c3c build hello_world_lib --debug-log
 
       - name: run compiler tests
         run: |
@@ -75,27 +212,581 @@ jobs:
   build-linux:
     runs-on: ubuntu-latest
     strategy:
+      # Don't abort runners if a single one fails
+      fail-fast: false
       matrix:
         build_type: [Release, Debug]
+        llvm_version: [17, 18, 19, 20]
 
     steps:
-      - uses: actions/checkout@v1
-      - name: (Ubuntu) Download LLVM
+      - uses: actions/checkout@v4
+      - name: Install common deps
         run: |
-          sudo apt-get install zlib1g zlib1g-dev clang-11 libllvm11 llvm-11 llvm-11-dev llvm-11-runtime liblld-11-dev liblld-11 python3
+          sudo apt-get install zlib1g zlib1g-dev python3 ninja-build curl
 
-      - name: CMake
+      - name: Install Clang ${{matrix.llvm_version}}
         run: |
-          mkdir build && cd build
-          cmake .. -DLLVM_DIR=/usr/lib/llvm-11/cmake -DCMAKE_BUILD_TYPE=${{ matrix.build_type }}
-          cmake --build .
+          wget -O - https://apt.llvm.org/llvm-snapshot.gpg.key | sudo apt-key add -
+          if [[ "${{matrix.llvm_version}}" < 18 ]]; then
+            sudo add-apt-repository "deb http://apt.llvm.org/focal/ llvm-toolchain-focal-${{matrix.llvm_version}} main"
+            sudo apt-get update
+            sudo apt-get install -y -t llvm-toolchain-focal-${{matrix.llvm_version}} libpolly-${{matrix.llvm_version}}-dev \
+              clang-${{matrix.llvm_version}} llvm-${{matrix.llvm_version}} llvm-${{matrix.llvm_version}}-dev \
+              lld-${{matrix.llvm_version}} liblld-${{matrix.llvm_version}}-dev libmlir-${{matrix.llvm_version}} \
+              libmlir-${{matrix.llvm_version}}-dev mlir-${{matrix.llvm_version}}-tools
+          else
+            if [[ "${{matrix.llvm_version}}" < "${{env.LLVM_DEV_VERSION}}" ]]; then
+              sudo add-apt-repository "deb http://apt.llvm.org/focal/ llvm-toolchain-focal-${{matrix.llvm_version}} main"
+              sudo apt-get update
+              sudo apt-get install -y -t llvm-toolchain-focal-${{matrix.llvm_version}} libpolly-${{matrix.llvm_version}}-dev \
+                clang-${{matrix.llvm_version}} llvm-${{matrix.llvm_version}} llvm-${{matrix.llvm_version}}-dev \
+                lld-${{matrix.llvm_version}} liblld-${{matrix.llvm_version}}-dev
+            else
+              sudo add-apt-repository "deb http://apt.llvm.org/focal/ llvm-toolchain-focal main"
+              sudo apt-get install -y -t llvm-toolchain-focal libpolly-${{matrix.llvm_version}}-dev \
+                clang-${{matrix.llvm_version}} llvm-${{matrix.llvm_version}} llvm-${{matrix.llvm_version}}-dev \
+                lld-${{matrix.llvm_version}} liblld-${{matrix.llvm_version}}-dev
+            fi
+          fi
+      - name: CMake
+        if: matrix.llvm_version < 18 || matrix.llvm_version == env.LLVM_DEV_VERSION
+        run: |
+          cmake -B build \
+                -G Ninja \
+                -DCMAKE_BUILD_TYPE=${{matrix.build_type}} \
+                -DCMAKE_C_COMPILER=clang-${{matrix.llvm_version}} \
+                -DCMAKE_CXX_COMPILER=clang++-${{matrix.llvm_version}} \
+                -DCMAKE_LINKER=lld-link-${{matrix.llvm_version}} \
+                -DCMAKE_OBJCOPY=llvm-objcopy-${{matrix.llvm_version}} \
+                -DCMAKE_STRIP=llvm-strip-${{matrix.llvm_version}} \
+                -DCMAKE_DLLTOOL=llvm-dlltool-${{matrix.llvm_version}} \
+                -DC3_LLVM_VERSION=${{matrix.llvm_version}}
+          cmake --build build
+      - name: CMake18
+        if: matrix.llvm_version >= 18 && matrix.llvm_version != env.LLVM_DEV_VERSION
+        run: |
+          cmake -B build \
+                -G Ninja \
+                -DCMAKE_BUILD_TYPE=${{matrix.build_type}} \
+                -DCMAKE_C_COMPILER=clang-${{matrix.llvm_version}} \
+                -DCMAKE_CXX_COMPILER=clang++-${{matrix.llvm_version}} \
+                -DCMAKE_LINKER=lld-link-${{matrix.llvm_version}} \
+                -DCMAKE_OBJCOPY=llvm-objcopy-${{matrix.llvm_version}} \
+                -DCMAKE_STRIP=llvm-strip-${{matrix.llvm_version}} \
+                -DCMAKE_DLLTOOL=llvm-dlltool-${{matrix.llvm_version}} \
+                -DC3_LLVM_VERSION=${{matrix.llvm_version}}.1
+          cmake --build build
+
+      - name: Compile and run some examples
+        run: |
+          cd resources
+          ../build/c3c compile examples/base64.c3
+          ../build/c3c compile examples/binarydigits.c3
+          ../build/c3c compile examples/brainfk.c3
+          ../build/c3c compile examples/factorial_macro.c3
+          ../build/c3c compile examples/fasta.c3
+          ../build/c3c compile examples/gameoflife.c3
+          ../build/c3c compile examples/hash.c3
+          ../build/c3c compile-only examples/levenshtein.c3
+          ../build/c3c compile examples/load_world.c3
+          ../build/c3c compile-only examples/map.c3
+          ../build/c3c compile examples/mandelbrot.c3
+          ../build/c3c compile examples/plus_minus.c3
+          ../build/c3c compile examples/nbodies.c3
+          ../build/c3c compile examples/spectralnorm.c3
+          ../build/c3c compile examples/swap.c3
+          ../build/c3c compile examples/contextfree/boolerr.c3
+          ../build/c3c compile examples/contextfree/dynscope.c3
+          ../build/c3c compile examples/contextfree/guess_number.c3
+          ../build/c3c compile examples/contextfree/multi.c3
+          ../build/c3c compile examples/contextfree/cleanup.c3
+          ../build/c3c compile-run examples/hello_world_many.c3
+          ../build/c3c compile-run examples/time.c3
+          ../build/c3c compile-run examples/fannkuch-redux.c3
+          ../build/c3c compile-run examples/contextfree/boolerr.c3
+          ../build/c3c compile-run examples/load_world.c3
+          ../build/c3c compile-run examples/process.c3
+          ../build/c3c compile-run examples/ls.c3
+          ../build/c3c compile-run --linker=builtin linux_stack.c3
+          ../build/c3c compile-run linux_stack.c3
+          ../build/c3c compile-run examples/args.c3 -- foo -bar "baz baz"
+
+      - name: Compile run unit tests
+        run: |
+          cd test
+          ../build/c3c compile-test unit
 
       - name: Build testproject
         run: |
           cd resources/testproject
-          ../../build/c3c build
+          ../../build/c3c run --debug-log
+
+      - name: Test WASM
+        run: |
+          cd resources/testfragments
+          ../../build/c3c compile --reloc=none --target wasm32 -g0 --link-libc=no --no-entry -Os wasm4.c3
+
+      - name: Install QEMU and Risc-V toolchain
+        run: |
+          sudo apt-get install opensbi qemu-system-misc u-boot-qemu gcc-riscv64-unknown-elf
+
+      - name: Compile and run Baremetal Risc-V Example
+        run: |
+          cd resources/examples/embedded/riscv-qemu
+          make C3C_PATH=../../../../build/ run
+
+      - name: Build testproject direct linker
+        run: |
+          cd resources/testproject
+          ../../build/c3c run --debug-log --linker=builtin
+
+      - name: Init a library & a project
+        run: |
+          ./build/c3c init-lib mylib
+          ls mylib.c3l
+          ./build/c3c init myproject
+          ls myproject
 
       - name: run compiler tests
         run: |
           cd test
           python3 src/tester.py ../build/c3c test_suite/
+
+      - name: bundle_output
+        if: matrix.llvm_version == env.LLVM_RELEASE_VERSION_LINUX
+        run: |
+          mkdir c3
+          cp -r lib c3
+          cp msvc_build_libraries.py c3
+          cp build/c3c c3
+          tar czf c3-linux-${{matrix.build_type}}.tar.gz c3
+
+      - name: upload artifacts
+        if: matrix.llvm_version == env.LLVM_RELEASE_VERSION_LINUX
+        uses: actions/upload-artifact@v3
+        with:
+          name: c3-linux-${{matrix.build_type}}
+          path: c3-linux-${{matrix.build_type}}.tar.gz
+
+  build-linux-ubuntu20:
+    runs-on: ubuntu-20.04
+    strategy:
+      # Don't abort runners if a single one fails
+      fail-fast: false
+      matrix:
+        build_type: [Release, Debug]
+        llvm_version: [17, 18, 19]
+    steps:
+      - uses: actions/checkout@v4
+      - name: Install common deps
+        run: |
+          sudo apt-get install zlib1g zlib1g-dev python3 ninja-build curl
+
+      - name: Install Clang ${{matrix.llvm_version}}
+        run: |
+          wget -O - https://apt.llvm.org/llvm-snapshot.gpg.key | sudo apt-key add -
+          if [[ "${{matrix.llvm_version}}" < "${{env.LLVM_DEV_VERSION}}" ]]; then
+            sudo add-apt-repository "deb http://apt.llvm.org/focal/ llvm-toolchain-focal-${{matrix.llvm_version}} main"
+          else
+            sudo add-apt-repository "deb http://apt.llvm.org/focal/ llvm-toolchain-focal main"
+          fi
+          sudo apt-get update
+          sudo apt-get install -y clang-${{matrix.llvm_version}} llvm-${{matrix.llvm_version}} llvm-${{matrix.llvm_version}}-dev lld-${{matrix.llvm_version}} liblld-${{matrix.llvm_version}}-dev
+          sudo apt-get install -y libmlir-${{matrix.llvm_version}} libmlir-${{matrix.llvm_version}}-dev mlir-${{matrix.llvm_version}}-tools
+          sudo apt-get install -y libpolly-${{matrix.llvm_version}}-dev
+      - name: CMake Old
+        if: matrix.llvm_version < 18 || matrix.llvm_version == env.LLVM_DEV_VERSION
+        run: |
+          cmake -B build \
+                -G Ninja \
+                -DCMAKE_BUILD_TYPE=${{matrix.build_type}} \
+                -DCMAKE_C_COMPILER=clang-${{matrix.llvm_version}} \
+                -DCMAKE_CXX_COMPILER=clang++-${{matrix.llvm_version}} \
+                -DCMAKE_LINKER=lld-link-${{matrix.llvm_version}} \
+                -DCMAKE_OBJCOPY=llvm-objcopy-${{matrix.llvm_version}} \
+                -DCMAKE_STRIP=llvm-strip-${{matrix.llvm_version}} \
+                -DCMAKE_DLLTOOL=llvm-dlltool-${{matrix.llvm_version}} \
+                -DC3_LLVM_VERSION=${{matrix.llvm_version}}
+          cmake --build build
+      - name: CMake
+        if: matrix.llvm_version >= 18 && matrix.llvm_version != env.LLVM_DEV_VERSION
+        run: |
+          cmake -B build \
+                -G Ninja \
+                -DCMAKE_BUILD_TYPE=${{matrix.build_type}} \
+                -DCMAKE_C_COMPILER=clang-${{matrix.llvm_version}} \
+                -DCMAKE_CXX_COMPILER=clang++-${{matrix.llvm_version}} \
+                -DCMAKE_LINKER=lld-link-${{matrix.llvm_version}} \
+                -DCMAKE_OBJCOPY=llvm-objcopy-${{matrix.llvm_version}} \
+                -DCMAKE_STRIP=llvm-strip-${{matrix.llvm_version}} \
+                -DCMAKE_DLLTOOL=llvm-dlltool-${{matrix.llvm_version}} \
+                -DC3_LLVM_VERSION=${{matrix.llvm_version}}.1
+          cmake --build build
+      - name: Compile and run some examples
+        run: |
+          cd resources
+          ../build/c3c compile examples/gameoflife.c3
+          ../build/c3c compile-only examples/levenshtein.c3
+          ../build/c3c compile-only examples/map.c3
+          ../build/c3c compile examples/mandelbrot.c3
+          ../build/c3c compile examples/plus_minus.c3
+          ../build/c3c compile examples/spectralnorm.c3
+          ../build/c3c compile examples/swap.c3
+          ../build/c3c compile examples/contextfree/guess_number.c3
+          ../build/c3c compile-run examples/hash.c3
+          ../build/c3c compile-run examples/nbodies.c3
+          ../build/c3c compile-run examples/contextfree/boolerr.c3
+          ../build/c3c compile-run examples/contextfree/dynscope.c3
+          ../build/c3c compile-run examples/contextfree/multi.c3
+          ../build/c3c compile-run examples/contextfree/cleanup.c3
+          ../build/c3c compile-run examples/hello_world_many.c3
+          ../build/c3c compile-run examples/time.c3
+          ../build/c3c compile-run examples/fannkuch-redux.c3
+          ../build/c3c compile-run examples/load_world.c3
+          ../build/c3c compile-run examples/base64.c3
+          ../build/c3c compile-run examples/binarydigits.c3
+          ../build/c3c compile-run examples/brainfk.c3
+          ../build/c3c compile-run examples/factorial_macro.c3
+          ../build/c3c compile-run examples/fasta.c3
+          ../build/c3c compile-run examples/process.c3
+          ../build/c3c compile-run --linker=builtin linux_stack.c3
+          ../build/c3c compile-run linux_stack.c3
+          ../build/c3c compile-run examples/args.c3 -- foo -bar "baz baz"
+
+      - name: Compile run unit tests
+        run: |
+          cd test
+          ../build/c3c compile-test unit
+
+      - name: Build testproject
+        run: |
+          cd resources/testproject
+          ../../build/c3c run --debug-log
+
+      - name: Build testproject direct linker
+        run: |
+          cd resources/testproject
+          ../../build/c3c run --debug-log --linker=builtin
+
+      - name: run compiler tests
+        run: |
+          cd test
+          python3 src/tester.py ../build/c3c test_suite/
+
+      - name: bundle_output
+        if: matrix.llvm_version == env.LLVM_RELEASE_VERSION_UBUNTU20
+        run: |
+          mkdir c3
+          cp -r lib c3
+          cp msvc_build_libraries.py c3
+          cp build/c3c c3
+          tar czf c3-ubuntu-20-${{matrix.build_type}}.tar.gz c3
+
+      - name: upload artifacts
+        if: matrix.llvm_version == env.LLVM_RELEASE_VERSION_UBUNTU20
+        uses: actions/upload-artifact@v3
+        with:
+          name: c3-ubuntu-20-${{matrix.build_type}}
+          path: c3-ubuntu-20-${{matrix.build_type}}.tar.gz
+  
+  build-with-docker:
+    runs-on: ubuntu-latest
+    strategy:
+      fail-fast: false
+      matrix:
+        ubuntu_version: [20.04, 22.04]
+        build_type: [Release, Debug]
+        llvm_version: [17, 18, 19, 20]
+
+    steps:
+      - uses: actions/checkout@v4
+
+      - name: Setup Docker Buildx
+        uses: docker/setup-buildx-action@v3
+
+      - name: Make script executable
+        run: chmod +x ./build-with-docker.sh
+
+      - name: Run build
+        run: |
+          LLVM_VERSION=${{ matrix.llvm_version }} UBUNTU_VERSION=${{ matrix.ubuntu_version }} CMAKE_BUILD_TYPE=${{ matrix.build_type }} ./build-with-docker.sh
+      
+      - name: Compile and run some examples
+        run: |
+          cd resources
+          ../build/c3c compile examples/base64.c3
+          ../build/c3c compile examples/binarydigits.c3
+          ../build/c3c compile examples/brainfk.c3
+          ../build/c3c compile examples/factorial_macro.c3
+          ../build/c3c compile examples/fasta.c3
+          ../build/c3c compile examples/gameoflife.c3
+          ../build/c3c compile examples/hash.c3
+          ../build/c3c compile-only examples/levenshtein.c3
+          ../build/c3c compile examples/load_world.c3
+          ../build/c3c compile-only examples/map.c3
+          ../build/c3c compile examples/mandelbrot.c3
+          ../build/c3c compile examples/plus_minus.c3
+          ../build/c3c compile examples/nbodies.c3
+          ../build/c3c compile examples/spectralnorm.c3
+          ../build/c3c compile examples/swap.c3
+          ../build/c3c compile examples/contextfree/boolerr.c3
+          ../build/c3c compile examples/contextfree/dynscope.c3
+          ../build/c3c compile examples/contextfree/guess_number.c3
+          ../build/c3c compile examples/contextfree/multi.c3
+          ../build/c3c compile examples/contextfree/cleanup.c3
+          ../build/c3c compile-run examples/hello_world_many.c3
+          ../build/c3c compile-run examples/time.c3
+          ../build/c3c compile-run examples/fannkuch-redux.c3
+          ../build/c3c compile-run examples/contextfree/boolerr.c3
+          ../build/c3c compile-run examples/load_world.c3
+          ../build/c3c compile-run examples/process.c3
+          ../build/c3c compile-run examples/ls.c3
+          ../build/c3c compile-run --linker=builtin linux_stack.c3
+          ../build/c3c compile-run linux_stack.c3
+          ../build/c3c compile-run examples/args.c3 -- foo -bar "baz baz"
+
+      - name: Compile run unit tests
+        run: |
+          cd test
+          ../build/c3c compile-test unit
+
+      - name: Build testproject
+        run: |
+          cd resources/testproject
+          ../../build/c3c run --debug-log
+
+      - name: Test WASM
+        run: |
+          cd resources/testfragments
+          ../../build/c3c compile --reloc=none --target wasm32 -g0 --link-libc=no --no-entry -Os wasm4.c3
+
+      - name: Build testproject direct linker
+        run: |
+          cd resources/testproject
+          ../../build/c3c run --debug-log --linker=builtin
+
+      - name: Init a library & a project
+        run: |
+          ./build/c3c init-lib mylib
+          ls mylib.c3l
+          ./build/c3c init myproject
+          ls myproject
+
+      - name: run compiler tests
+        run: |
+          cd test
+          python3 src/tester.py ../build/c3c test_suite/
+
+  build-mac:
+    runs-on: macos-latest
+    strategy:
+      # Don't abort runners if a single one fails
+      fail-fast: false
+      matrix:
+        build_type: [Release, Debug]
+        llvm_version: [17, 18]
+    steps:
+      - uses: actions/checkout@v4
+      - name: Download LLVM
+        run: |
+          brew install llvm@${{ matrix.llvm_version }} ninja curl
+          echo "/opt/homebrew/opt/llvm@${{ matrix.llvm_version }}/bin" >> $GITHUB_PATH
+          TMP_PATH=$(xcrun --show-sdk-path)/user/include
+          echo "CPATH=$TMP_PATH" >> $GITHUB_ENV
+
+      - name: CMake
+        if: matrix.llvm_version < 18
+        run: |
+          cmake -B build -G Ninja -DC3_LLVM_VERSION=${{matrix.llvm_version}} -DCMAKE_BUILD_TYPE=${{ matrix.build_type }}
+          cmake --build build
+      - name: CMake18
+        if: matrix.llvm_version >= 18
+        run: |
+          cmake -B build -G Ninja -DC3_LLVM_VERSION=${{matrix.llvm_version}}.1 -DCMAKE_BUILD_TYPE=${{ matrix.build_type }}
+          cmake --build build
+
+      - name: Vendor-fetch
+        run: |
+          ./build/c3c vendor-fetch raylib
+
+      - name: Compile and run some examples
+        run: |
+          cd resources
+          ../build/c3c compile-run examples/hello_world_many.c3
+          ../build/c3c compile-run examples/time.c3
+          ../build/c3c compile-run examples/fannkuch-redux.c3
+          ../build/c3c compile-run examples/contextfree/boolerr.c3
+          ../build/c3c compile-run examples/process.c3
+          ../build/c3c compile-run examples/load_world.c3
+          ../build/c3c compile-run -O5 examples/load_world.c3
+          ../build/c3c compile-run examples/args.c3 -- foo -bar "baz baz"
+
+      - name: Compile run unit tests
+        run: |
+          cd test
+          ../build/c3c compile-test unit
+
+      - name: Build testproject
+        run: |
+          cd resources/testproject
+          ../../build/c3c run --debug-log
+
+      - name: Build testproject direct linker
+        run: |
+          cd resources/testproject
+          ../../build/c3c run --debug-log --linker=builtin
+
+      - name: Build testproject lib
+        run: |
+          cd resources/testproject
+          ../../build/c3c build hello_world_lib --debug-log
+
+      - name: run compiler tests
+        run: |
+          cd test
+          python3 src/tester.py ../build/c3c test_suite/
+
+      - name: bundle_output
+        if: matrix.llvm_version == env.LLVM_RELEASE_VERSION_MAC
+        run: |
+          mkdir macos
+          cp -r lib macos
+          cp msvc_build_libraries.py macos
+          cp build/c3c macos
+          zip -r c3-macos-${{matrix.build_type}}.zip macos
+
+      - name: upload artifacts
+        if: matrix.llvm_version == env.LLVM_RELEASE_VERSION_MAC
+        uses: actions/upload-artifact@v3
+        with:
+          name: c3-macos-${{matrix.build_type}}
+          path: c3-macos-${{matrix.build_type}}.zip
+
+
+  release:
+    runs-on: ubuntu-latest
+    needs: [build-msvc, build-linux, build-mac]
+    if: github.ref == 'refs/heads/master'
+
+    steps:
+      - uses: actions/checkout@v4
+      - name: delete tag
+        continue-on-error: true
+        uses: actions/github-script@v6
+        with:
+          script: |
+            github.rest.git.deleteRef({
+              owner: context.repo.owner,
+              repo: context.repo.repo,
+              ref: 'tags/latest',
+              sha: context.sha
+            })
+      - name: create tag
+        uses: actions/github-script@v6
+        with:
+          script: |
+            github.rest.git.createRef({
+              owner: context.repo.owner,
+              repo: context.repo.repo,
+              ref: 'refs/tags/latest',
+              sha: context.sha
+            })
+
+      - uses: actions/download-artifact@v3
+      - run: cp -r lib c3-windows-Release
+      - run: cp -r lib c3-windows-Debug
+      - run: cp msvc_build_libraries.py c3-windows-Release
+      - run: cp msvc_build_libraries.py c3-windows-Debug
+      - run: zip -r c3-windows-Release.zip c3-windows-Release
+      - run: zip -r c3-windows-Debug.zip c3-windows-Debug
+
+      - id: create_release
+        uses: actions/create-release@v1
+        env:
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+        with:
+          tag_name: latest
+          release_name: latest
+          draft: false
+          prerelease: true
+
+      - name: upload windows
+        uses: actions/upload-release-asset@v1
+        env:
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+        with:
+          upload_url: ${{ steps.create_release.outputs.upload_url }}
+          asset_path: c3-windows-Release.zip
+          asset_name: c3-windows.zip
+          asset_content_type: application/zip
+
+      - name: upload windows debug
+        uses: actions/upload-release-asset@v1
+        env:
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+        with:
+          upload_url: ${{ steps.create_release.outputs.upload_url }}
+          asset_path: c3-windows-Debug.zip
+          asset_name: c3-windows-debug.zip
+          asset_content_type: application/zip
+
+      - name: upload linux
+        uses: actions/upload-release-asset@v1
+        env:
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+        with:
+          upload_url: ${{ steps.create_release.outputs.upload_url }}
+          asset_path: c3-linux-Release/c3-linux-Release.tar.gz
+          asset_name: c3-linux.tar.gz
+          asset_content_type: application/gzip
+
+      - name: upload linux debug
+        uses: actions/upload-release-asset@v1
+        env:
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+        with:
+          upload_url: ${{ steps.create_release.outputs.upload_url }}
+          asset_path: c3-linux-Debug/c3-linux-Debug.tar.gz
+          asset_name: c3-linux-debug.tar.gz
+          asset_content_type: application/gzip
+
+      - name: upload ubuntu 20
+        uses: actions/upload-release-asset@v1
+        env:
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+        with:
+          upload_url: ${{ steps.create_release.outputs.upload_url }}
+
+          asset_path: c3-ubuntu-20-Release/c3-ubuntu-20-Release.tar.gz
+          asset_name: c3-ubuntu-20.tar.gz
+          asset_content_type: application/gzip
+
+      - name: upload ubuntu 20 debug
+        uses: actions/upload-release-asset@v1
+        env:
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+        with:
+          upload_url: ${{ steps.create_release.outputs.upload_url }}
+          asset_path: c3-ubuntu-20-Debug/c3-ubuntu-20-Debug.tar.gz
+          asset_name: c3-ubuntu-20-debug.tar.gz
+          asset_content_type: application/gzip
+
+      - name: upload macos
+        uses: actions/upload-release-asset@v1
+        env:
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+        with:
+          upload_url: ${{ steps.create_release.outputs.upload_url }}
+          asset_path: c3-macos-Release/c3-macos-Release.zip
+          asset_name: c3-macos.zip
+          asset_content_type: application/zip
+
+      - name: upload macos debug
+        uses: actions/upload-release-asset@v1
+        env:
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+        with:
+          upload_url: ${{ steps.create_release.outputs.upload_url }}
+          asset_path: c3-macos-Debug/c3-macos-Debug.zip
+          asset_name: c3-macos-debug.zip
+          asset_content_type: application/zip

.gitignore

@@ -52,7 +52,6 @@ Module.symvers
 Mkfile.old
 dkms.conf
 /build/
-/cmake-build-debug/
 .idea/
 /resources/grammar.tab.c
 /resources/grammar.vcg
@@ -60,3 +59,14 @@ dkms.conf
 /resources/y.tab.c
 /resources/y.tab.h
 /bin/
+
+#visual studio files
+.vs/
+.vscode/
+out/
+
+/cmake-build-debug/
+/cmake-build-release/
+
+# Emacs files
+TAGS

.gitmodules

@@ -0,0 +1,3 @@
+[submodule "tilde-backend"]
+	path = tilde-backend
+	url = https://github.com/c3lang/tilde-backend

CMakeLists.txt

@@ -1,173 +1,469 @@
-cmake_minimum_required(VERSION 3.10)
-project(c3c)
+cmake_minimum_required(VERSION 3.20)
 
-SET(CMAKE_FIND_PACKAGE_SORT_ORDER NATURAL)
-SET(CMAKE_FIND_PACKAGE_SORT_DIRECTION DEC)
+# Grab the version
+file(READ "src/version.h" ver)
+if (NOT ${ver} MATCHES "COMPILER_VERSION \"([0-9]+.[0-9]+.[0-9]+)\"")
+    message(FATAL_ERROR "Compiler version could not be parsed from version.h")
+endif()
 
-#set(CMAKE_BUILD_TYPE Release)
-#set(CMAKE_CXX_FLAGS_RELEASE "-O3")
+# Set the project and version
+project(c3c VERSION ${CMAKE_MATCH_1})
+message("C3C version: ${CMAKE_PROJECT_VERSION}")
 
-find_package(LLVM REQUIRED CONFIG)
+# Avoid warning for FetchContent
+if (CMAKE_VERSION VERSION_GREATER_EQUAL "3.24.0")
+    cmake_policy(SET CMP0135 NEW)
+endif()
+
+if (NOT DEFINED CMAKE_INSTALL_LIBDIR)
+    if (MSVC)
+        set(CMAKE_INSTALL_LIBDIR "c:\\c3c\\lib")
+        set(CMAKE_INSTALL_BINDIR "c:\\c3c")
+    else ()
+        set(CMAKE_INSTALL_LIBDIR "/usr/local/lib/c3")
+        set(CMAKE_INSTALL_BINDIR "/usr/local/bin/c3c")
+    endif()
+endif ()
+
+# Enable fetching (for Windows)
+include(FetchContent)
+include(FeatureSummary)
+
+set(CMAKE_FIND_PACKAGE_SORT_ORDER NATURAL)
+set(CMAKE_FIND_PACKAGE_SORT_DIRECTION DEC)
+
+# We use C11 and C++17
+set(CMAKE_C_STANDARD 11)
+set(CMAKE_CXX_STANDARD 17)
+
+if(MSVC)
+    message(STATUS "MSVC version ${MSVC_VERSION}")
+    set(CMAKE_CXX_FLAGS_RELEASE "${CMAKE_CXX_FLAGS_RELEASE} /O2 /EHsc")
+    set(CMAKE_C_FLAGS_RELEASE "${CMAKE_C_FLAGS_RELEASE} /O2 /EHsc")
+    set(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} /Od /Zi /EHa")
+    set(CMAKE_C_FLAGS_DEBUG "${CMAKE_C_FLAGS_DEBUG} /Od /Zi /EHa")
+else()
+    if (true)
+        set(CMAKE_CXX_FLAGS_RELEASE "${CMAKE_CXX_FLAGS_RELEASE} -O3 -fno-exceptions")
+        set(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} -gdwarf-3 -fno-exceptions")
+        set(CMAKE_C_FLAGS_RELEASE "${CMAKE_C_FLAGS_RELEASE} -gdwarf-3 -O3 -fno-exceptions")
+        set(CMAKE_C_FLAGS_DEBUG "${CMAKE_C_FLAGS_DEBUG} -gdwarf-3 -fno-exceptions")
+    else()
+        set(CMAKE_CXX_FLAGS_RELEASE "${CMAKE_CXX_FLAGS_RELEASE} -gdwarf-3 -O3 -fsanitize=undefined,address -fno-exceptions")
+        set(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} -gdwarf-3 -O1 -fsanitize=undefined,address -fno-exceptions")
+        set(CMAKE_C_FLAGS_RELEASE "${CMAKE_C_FLAGS_RELEASE} -gdwarf-3 -O3 -fsanitize=undefined,address -fno-exceptions")
+        set(CMAKE_C_FLAGS_DEBUG "${CMAKE_C_FLAGS_DEBUG} -gdwarf-3 -O1 -fsanitize=undefined,address -fno-exceptions")
+        set(CMAKE_EXE_LINKER_FLAGS "${CMAKE_EXE_LINKER_FLAGS} -fsanitize=undefined,address -fno-exceptions")
+    endif()
+endif()
+
+option(C3_LINK_DYNAMIC "link dynamically with LLVM/LLD libs")
+
+set(C3_LLVM_VERSION "auto" CACHE STRING "Use LLVM version [default: auto]")
+option(C3_USE_MIMALLOC "Use built-in mimalloc" OFF)
+option(C3_USE_TB "Use TB" OFF)
+set(C3_MIMALLOC_TAG "v1.7.3" CACHE STRING "Used version of mimalloc")
+
+set(C3_USE_MIMALLOC OFF)
+if(C3_USE_MIMALLOC)
+    option(MI_BUILD_TESTS OFF)
+    option(MI_BUILD_SHARED OFF)
+    option(MI_PADDING OFF)
+    option(MI_DEBUG_FULL OFF)
+    FetchContent_Declare(
+    mimalloc
+    GIT_REPOSITORY https://github.com/microsoft/mimalloc.git
+    GIT_TAG ${C3_MIMALLOC_TAG}
+    )
+    FetchContent_MakeAvailable(mimalloc)
+endif()
+if (NOT WIN32)
+    find_package(CURL)
+endif()
+if (NOT C3_LLVM_VERSION STREQUAL "auto")
+    if (${C3_LLVM_VERSION} VERSION_LESS 17 OR ${C3_LLVM_VERSION} VERSION_GREATER 20)
+        message(FATAL_ERROR "LLVM ${C3_LLVM_VERSION} is not supported!")
+    endif()
+endif()
+
+find_package(Git QUIET)
+if(C3_USE_TB AND GIT_FOUND AND EXISTS "${CMAKE_SOURCE_DIR}/.git")
+    # Update submodules as needed
+    option(GIT_SUBMODULE "Check submodules during build" ON)
+    if(GIT_SUBMODULE)
+        message(STATUS "Submodule update")
+        execute_process(COMMAND ${GIT_EXECUTABLE} submodule update --init --recursive
+                WORKING_DIRECTORY ${CMAKE_CURRENT_SOURCE_DIR}
+                RESULT_VARIABLE GIT_SUBMOD_RESULT)
+        if(NOT GIT_SUBMOD_RESULT EQUAL "0")
+            message(FATAL_ERROR "git submodule update --init --recursive failed with ${GIT_SUBMOD_RESULT}, please checkout submodules")
+        endif()
+    endif()
+endif()
+
+if(CMAKE_C_COMPILER_ID STREQUAL "MSVC")
+    if (C3_LLVM_VERSION STREQUAL "auto")
+        set(C3_LLVM_VERSION "18")
+    endif()
+    FetchContent_Declare(
+            LLVM_Windows
+            URL https://github.com/c3lang/win-llvm/releases/download/llvm_18_1_8_with_rt/llvm-18.1.8-windows-amd64-msvc17-libcmt.7z
+    )
+    FetchContent_Declare(
+            LLVM_Windows_debug
+            URL https://github.com/c3lang/win-llvm/releases/download/llvm_18_1_8_with_rt/llvm-18.1.8-windows-amd64-msvc17-libcmt-dbg.7z
+    )
+    if(CMAKE_BUILD_TYPE STREQUAL "Debug")
+        message("Loading Windows LLVM debug libraries, this may take a while...")
+        FetchContent_MakeAvailable(LLVM_Windows_debug)
+        set(llvm_dir ${llvm_windows_debug_SOURCE_DIR})
+    else()
+        message("Loading Windows LLVM libraries, this may take a while...")
+        FetchContent_MakeAvailable(LLVM_Windows)
+        set(llvm_dir ${llvm_windows_SOURCE_DIR})
+    endif()
+    set(CMAKE_SYSTEM_PREFIX_PATH ${llvm_dir} ${CMAKE_SYSTEM_PREFIX_PATH})
+
+    find_package(LLVM REQUIRED CONFIG)
+    find_package(LLD REQUIRED CONFIG)
+else()
+    if (NOT C3_LLVM_VERSION STREQUAL "auto")
+        find_package(LLVM ${C3_LLVM_VERSION} REQUIRED CONFIG)
+    else()
+        find_package(LLVM REQUIRED CONFIG)
+    endif()
+endif()
 
 message(STATUS "Found LLVM ${LLVM_PACKAGE_VERSION}")
 message(STATUS "Using LLVMConfig.cmake in: ${LLVM_DIR}")
 message(STATUS "Libraries located in: ${LLVM_LIBRARY_DIRS}")
 
+if (NOT LLVM_PACKAGE_VERSION VERSION_GREATER_EQUAL 15.0)
+    message(FATAL_ERROR "LLVM version 15.0 or later is required.")
+endif()
+
+if(LLVM_ENABLE_RTTI)
+    message(STATUS "LLVM was built with RTTI")
+else()
+    message(STATUS "LLVM was not built with RTTI")
+endif()
+
+string(REPLACE "." ";" VERSION_LIST ${LLVM_PACKAGE_VERSION})
+list(GET VERSION_LIST 0 LLVM_MAJOR_VERSION)
+
 include_directories(${LLVM_INCLUDE_DIRS})
+link_directories(${LLVM_LIBRARY_DIRS})
 add_definitions(${LLVM_DEFINITIONS})
 
-set(CMAKE_C_STANDARD 11)
-set(CMAKE_CXX_STANDARD 17)
+if(NOT C3_LINK_DYNAMIC)
+    set(LLVM_LINK_COMPONENTS
+	    AllTargetsAsmParsers
+	    AllTargetsCodeGens
+	    AllTargetsDescs
+	    AllTargetsDisassemblers
+	    AllTargetsInfos
+	    Analysis
+	    AsmPrinter
+	    BitReader
+	    Core
+	    DebugInfoPDB
+	    InstCombine
+	    IrReader
+	    LibDriver
+	    Linker
+	    LTO
+	    MC
+	    MCDisassembler
+	    native
+	    nativecodegen
+	    Object
+	    Option
+	    ScalarOpts
+	    Support
+	    Target
+	    TransformUtils
+	    WindowsManifest
+	    WindowsDriver
+    )
 
-set(LLVM_LINK_COMPONENTS
-        AllTargetsAsmParsers
-        AllTargetsDescs
-        AllTargetsDisassemblers
-        AllTargetsInfos
-        AllTargetsCodeGens
-        Analysis
-        BitReader
-        Core
-        InstCombine
-        MC
-        MCDisassembler
-        Object
-        ScalarOpts
-        Support
-        Target
-        TransformUtils
-        native
-        nativecodegen
-        AsmPrinter
-        Linker
-        LTO
-        WindowsManifest
-        DebugInfoPDB
-        LibDriver
-        Option
-        IrReader
-        )
+    llvm_map_components_to_libnames(llvm_libs ${LLVM_LINK_COMPONENTS})
 
+    # These don't seem to be reliable on windows.
+    message(STATUS "using find_library")
+    find_library(LLD_COFF NAMES lldCOFF.lib lldCOFF.a liblldCOFF.a PATHS ${LLVM_LIBRARY_DIRS} NO_DEFAULT_PATH)
+    find_library(LLD_COMMON NAMES lldCommon.lib lldCommon.a liblldCommon.a PATHS ${LLVM_LIBRARY_DIRS} NO_DEFAULT_PATH)
+    find_library(LLD_ELF NAMES lldELF.lib lldELF.a liblldELF.a PATHS ${LLVM_LIBRARY_DIRS} NO_DEFAULT_PATH)
+    find_library(LLD_MACHO NAMES lldMachO.lib lldMachO.a liblldMachO.a PATHS ${LLVM_LIBRARY_DIRS} NO_DEFAULT_PATH)
+    find_library(LLD_MINGW NAMES lldMinGW.lib lldMinGW.a liblldMinGW.a PATHS ${LLVM_LIBRARY_DIRS} NO_DEFAULT_PATH)
+    find_library(LLD_WASM NAMES lldWasm.lib lldWasm.a liblldWasm.a PATHS ${LLVM_LIBRARY_DIRS} NO_DEFAULT_PATH)
+else()
+    find_library(LLVM NAMES libLLVM.so PATHS ${LLVM_LIBRARY_DIRS} NO_DEFAULT_PATH)
+    set(llvm_libs ${LLVM})
 
-llvm_map_components_to_libnames(llvm_libs ${LLVM_LINK_COMPONENTS})
+    # These don't seem to be reliable on windows.
+    message(STATUS "using find_library")
+    find_library(LLD_COFF NAMES liblldCOFF.so PATHS ${LLVM_LIBRARY_DIRS} NO_DEFAULT_PATH)
+    find_library(LLD_COMMON NAMES liblldCommon.so PATHS ${LLVM_LIBRARY_DIRS} NO_DEFAULT_PATH)
+    find_library(LLD_ELF NAMES liblldELF.so PATHS ${LLVM_LIBRARY_DIRS} NO_DEFAULT_PATH)
+    find_library(LLD_MACHO NAMES liblldMachO.so PATHS ${LLVM_LIBRARY_DIRS} NO_DEFAULT_PATH)
+    find_library(LLD_MINGW NAMES liblldMinGW.so PATHS ${LLVM_LIBRARY_DIRS} NO_DEFAULT_PATH)
+    find_library(LLD_WASM NAMES liblldWasm.so PATHS ${LLVM_LIBRARY_DIRS} NO_DEFAULT_PATH)
+endif()
 
-file(COPY ${CMAKE_SOURCE_DIR}/resources/lib DESTINATION ${CMAKE_BINARY_DIR})
-
-find_library(LLD_COFF NAMES lldCOFF.a liblldCOFF.a PATHS ${LLVM_LIBRARY_DIRS})
-find_library(LLD_COMMON NAMES lldCommon.a liblldCommon.a PATHS ${LLVM_LIBRARY_DIRS})
-find_library(LLD_CORE NAMES lldCore.a liblldCore.a PATHS ${LLVM_LIBRARY_DIRS})
-find_library(LLD_WASM NAMES lldWasm.a liblldWasm.a PATHS ${LLVM_LIBRARY_DIRS})
-find_library(LLD_MINGW NAMES lldMinGW.a liblldMinGW.a PATHS ${LLVM_LIBRARY_DIRS})
-find_library(LLD_ELF NAMES lldELF.a liblldELF.a PATHS ${LLVM_LIBRARY_DIRS})
-find_library(LLD_DRIVER NAMES lldDriver.a liblldDriver.a PATHS ${LLVM_LIBRARY_DIRS})
-find_library(LLD_READER_WRITER NAMES lldReaderWriter.a liblldReaderWriter.a PATHS ${LLVM_LIBRARY_DIRS})
-find_library(LLD_MACHO NAMES lldMachO.a liblldMachO.a PATHS ${LLVM_LIBRARY_DIRS})
-find_library(LLD_YAML NAMES lldYAML.a liblldYAML.a PATHS ${LLVM_LIBRARY_DIRS})
+if (NOT(${CMAKE_BINARY_DIR} EQUAL ${CMAKE_SOURCE_DIR}))
+    file(REMOVE_RECURSE ${CMAKE_BINARY_DIR}/lib)
+    file(COPY ${CMAKE_SOURCE_DIR}/lib DESTINATION ${CMAKE_BINARY_DIR})
+endif()
 
+find_library(LLD_LOONG NAMES libLLVMLoongArchCodeGen.lib libLLVMLoongArchAsmParser.lib libLLVMLoongArchCodeGen.a libLLVMLoongArchAsmParser.a PATHS ${LLVM_LIBRARY_DIRS} NO_DEFAULT_PATH)
 set(lld_libs
-        ${LLD_COFF}
-        ${LLD_COMMON}
-        ${LLD_WASM}
-        ${LLD_MINGW}
-        ${LLD_ELF}
-        ${LLD_DRIVER}
-        ${LLD_READER_WRITER}
-        ${LLD_MACHO}
-        ${LLD_YAML}
-        ${LLD_CORE}
+    ${LLD_COFF}
+    ${LLD_COMMON}
+    ${LLD_WASM}
+    ${LLD_MINGW}
+    ${LLD_ELF}
+    ${LLD_MACHO}
+    )
+
+if (APPLE)
+    set(lld_libs ${lld_libs} xar)
+    find_file(RT_ASAN_DYNAMIC NAMES libclang_rt.asan_osx_dynamic.dylib PATHS "${LLVM_LIBRARY_DIRS}/clang/${LLVM_MAJOR_VERSION}/lib/darwin")
+    find_file(RT_TSAN_DYNAMIC NAMES libclang_rt.tsan_osx_dynamic.dylib PATHS "${LLVM_LIBRARY_DIRS}/clang/${LLVM_MAJOR_VERSION}/lib/darwin")
+    find_file(RT_UBSAN_DYNAMIC NAMES libclang_rt.ubsan_osx_dynamic.dylib PATHS "${LLVM_LIBRARY_DIRS}/clang/${LLVM_MAJOR_VERSION}/lib/darwin")
+    find_file(RT_LSAN_DYNAMIC NAMES libclang_rt.lsan_osx_dynamic.dylib PATHS "${LLVM_LIBRARY_DIRS}/clang/${LLVM_MAJOR_VERSION}/lib/darwin")
+    set(sanitizer_runtime_libraries
+            ${RT_ASAN_DYNAMIC}
+            ${RT_TSAN_DYNAMIC}
+            # Unused
+            # ${RT_UBSAN_DYNAMIC}
+            # ${RT_LSAN_DYNAMIC}
         )
+endif()
+
+message(STATUS "linking to llvm libs ${lld_libs}")
+message(STATUS "Found lld libs ${lld_libs}")
+
+
 add_library(c3c_wrappers STATIC wrapper/src/wrapper.cpp)
+add_library(miniz STATIC dependencies/miniz/miniz.c)
 
 add_executable(c3c
-        src/main.c
+        src/build/builder.c
         src/build/build_options.c
         src/build/project_creation.c
-        src/utils/errors.c
-        src/utils/file_utils.c
-        src/compiler/lexer.c
-        src/compiler/tokens.c
-        src/compiler/symtab.c
-        src/compiler/parser.c
-        src/compiler_tests/tests.c
-        src/compiler_tests/benchmark.c
-        src/utils/malloc.c
-        src/compiler/compiler.c
-        src/compiler/semantic_analyser.c
-        src/compiler/source_file.c
-        src/compiler/diagnostics.c
+        src/build/project_manipulation.c
+        src/build/libraries.c
         src/compiler/ast.c
         src/compiler/bigint.c
-        src/compiler/context.c
-        src/compiler/sema_expr.c
-        src/compiler/enums.h
-        src/compiler/sema_casts.c
-        src/compiler/target.c
+        src/compiler/codegen_general.c
+        src/compiler/compiler.c
         src/compiler/compiler.h
-        src/compiler/types.c
-        src/compiler/module.c
-        src/compiler/llvm_codegen.c
-        src/utils/stringutils.c
-        src/compiler/dwarf.h
+        src/compiler/subprocess.c
+        src/compiler/subprocess.h
+        src/compiler/context.c
         src/compiler/copying.c
-        src/compiler/llvm_codegen_stmt.c
-        src/compiler/llvm_codegen_expr.c
+        src/compiler/diagnostics.c
+        src/compiler/float.c
+        src/compiler/headers.c
+        src/compiler/json_output.c
+        src/compiler/lexer.c
+        src/compiler/linker.c
+        src/compiler/llvm_codegen.c
+        src/compiler/abi/c_abi_aarch64.c
+        src/compiler/abi/c_abi.c
+        src/compiler/abi/c_abi_riscv.c
+        src/compiler/abi/c_abi_wasm.c
+        src/compiler/abi/c_abi_win64.c
+        src/compiler/abi/c_abi_x64.c
+        src/compiler/abi/c_abi_x86.c
         src/compiler/llvm_codegen_debug_info.c
-        src/compiler/llvm_codegen_module.c
-        src/compiler/llvm_codegen_type.c
+        src/compiler/llvm_codegen_expr.c
         src/compiler/llvm_codegen_function.c
-        src/compiler/llvm_codegen_c_abi.c
-        src/build/builder.c
-        src/utils/toml.c src/build/project.c
-        src/compiler/sema_name_resolution.c
-        src/target_info/target_info.c
+        src/compiler/llvm_codegen_instr.c
+        src/compiler/llvm_codegen_module.c
+        src/compiler/llvm_codegen_stmt.c
+        src/compiler/llvm_codegen_type.c
+        src/compiler/llvm_codegen_value.c
+        src/compiler/module.c
+        src/compiler/number.c
         src/compiler/parse_expr.c
+        src/compiler/parse_global.c
+        src/compiler/parser.c
         src/compiler/parser_internal.h
         src/compiler/parse_stmt.c
-        src/compiler/parse_global.c
-        src/compiler/sema_passes.c
-        src/compiler/sema_internal.h
+        src/compiler/sema_casts.c
         src/compiler/sema_decls.c
-        src/compiler/sema_types.c
+        src/compiler/sema_expr.c
+        src/compiler/sema_internal.h
+        src/compiler/sema_name_resolution.c
+        src/compiler/sema_errors.c
+        src/compiler/sema_builtins.c
+        src/compiler/sema_initializers.c
+        src/compiler/semantic_analyser.c
+        src/compiler/sema_passes.c
         src/compiler/sema_stmts.c
-        src/compiler/number.c
-        src/compiler/linker.c
+        src/compiler/sema_types.c
+        src/compiler/source_file.c
+        src/compiler/symtab.c
+        src/compiler/target.c
+        src/compiler/sema_asm.c
+        src/compiler_tests/benchmark.c
+        src/compiler_tests/tests.c
+        src/compiler/tokens.c
+        src/compiler/types.c
+        src/main.c
+        src/utils/errors.c
+        src/utils/file_utils.c
+        src/utils/find_msvc.c
+        src/utils/malloc.c
+        src/utils/stringutils.c
+        src/utils/taskqueue.c
+        src/utils/json.c
+        src/build/project.c
         src/utils/vmem.c
         src/utils/vmem.h
         src/utils/whereami.c
-        src/utils/taskqueue.c
-        src/compiler/llvm_codegen_c_abi_x86.c
-        src/compiler/c_abi_internal.h
-        src/compiler/llvm_codegen_c_abi_x64.c
-        src/compiler/llvm_codegen_c_abi_win64.c
-        src/compiler/llvm_codegen_c_abi_aarch64.c
-        src/compiler/headers.c
-        src/compiler/llvm_codegen_c_abi_riscv.c
-        src/compiler/llvm_codegen_c_abi_wasm.c)
+        src/utils/cpus.c
+        src/utils/unzipper.c
+        src/compiler/decltable.c
+        src/compiler/mac_support.c
+        src/compiler/llvm_codegen_storeload.c
+        src/compiler/windows_support.c
+        src/compiler/codegen_asm.c
+        src/compiler/asm_target.c
+        src/compiler/llvm_codegen_builtins.c
+        src/compiler/expr.c
+        src/utils/time.c
+        src/utils/http.c
+        src/compiler/sema_liveness.c
+        src/build/common_build.c)
 
-target_compile_options(c3c PRIVATE -Wall -Werror -Wno-unknown-pragmas -Wno-unused-result
-        -Wno-unused-function -Wno-unused-variable -Wno-unused-parameter)
+
+if (C3_USE_TB)
+    file(GLOB tilde-sources
+            tilde-backend/src/tb/*.c
+            tilde-backend/src/tb/codegen/*.c
+            tilde-backend/src/tb/bigint/*.c
+            tilde-backend/src/tb/objects/*.c
+            tilde-backend/src/tb/system/*.c
+            tilde-backend/src/tb/debug/cv/*.c
+            tilde-backend/src/tb/opt/*.c
+            tilde-backend/src/tb/x64/*.c
+            tilde-backend/src/tb/wasm/*.c
+            tilde-backend/src/tb/aarch64/*.c
+            )
+    target_sources(c3c PRIVATE
+            src/compiler/tilde_codegen.c
+            src/compiler/tilde_codegen_instr.c
+            src/compiler/tilde_codegen_value.c
+            src/compiler/tilde_codegen_storeload.c
+            src/compiler/tilde_codegen_expr.c
+            src/compiler/tilde_codegen_stmt.c
+            src/compiler/tilde_codegen_type.c
+            src/compiler/tilde_codegen_abi.c
+            src/compiler/tilde_codegen_storeload.c)
+
+    target_compile_definitions(c3c PUBLIC TB_AVAILABLE=1)
+    target_link_libraries(c3c tilde-backend)
+    add_library(tilde-backend STATIC ${tilde-sources})
+    target_include_directories(tilde-backend PRIVATE
+            "${CMAKE_SOURCE_DIR}/tilde-backend/src/" "${CMAKE_SOURCE_DIR}/tilde-backend/include")
+
+    target_include_directories(c3c PRIVATE
+            "${CMAKE_SOURCE_DIR}/tilde-backend/include/")
+else()
+    target_compile_definitions(c3c PUBLIC TB_AVAILABLE=0)
+endif()
 
 
 target_include_directories(c3c PRIVATE
-        "${CMAKE_SOURCE_DIR}/src/")
+        "${CMAKE_SOURCE_DIR}/src/"
+        "${CMAKE_SOURCE_DIR}/wrapper/include/")
+
 
 target_include_directories(c3c_wrappers PRIVATE
-        "${CMAKE_SOURCE_DIR}/wrapper/src/")
+        "${CMAKE_SOURCE_DIR}/wrapper/include/")
 
-
-message(STATUS "Found LLD ${lld_libs}")
+target_include_directories(miniz PUBLIC
+        "${CMAKE_SOURCE_DIR}/dependencies/miniz/")
 
 target_link_libraries(c3c_wrappers ${llvm_libs} ${lld_libs})
-#target_link_libraries(c3c m ${llvm_libs} c3c_wrappers lldCommon lldCore lldCOFF lldWASM lldMinGW lldELF lldDriver lldReaderWriter lldMachO lldYAML)
-target_link_libraries(c3c m ${llvm_libs} c3c_wrappers ${lld_libs})
+target_link_libraries(c3c ${llvm_libs} miniz c3c_wrappers ${lld_libs})
+
+if(C3_USE_MIMALLOC)
+    target_link_libraries(c3c mimalloc-static)
+endif()
 
 if (WIN32)
-    if (CMAKE_C_COMPILER_ID STREQUAL "Clang" OR CMAKE_C_COMPILE_ID STREQUAL "GNU")
-        target_link_options(c3c PRIVATE -pthread)
-        target_compile_definitions(c3c PRIVATE USE_PTHREAD=1)
-        target_compile_options(c3c PRIVATE -mlong-double-64)
+    target_link_libraries(c3c Winhttp.lib)
+endif()
+
+if (CURL_FOUND)
+    target_link_libraries(c3c ${CURL_LIBRARIES})
+    target_include_directories(c3c PRIVATE ${CURL_INCLUDES})
+    target_compile_definitions(c3c PUBLIC CURL_FOUND=1)
+else()
+    target_compile_definitions(c3c PUBLIC CURL_FOUND=0)
+endif()
+
+if(MSVC)
+    message("Adding MSVC options")
+    target_compile_options(c3c PRIVATE /wd4068 /wd4090 /WX /Wv:18)
+    target_compile_options(c3c_wrappers PUBLIC /wd4624 /wd4267 /wd4244 /WX /Wv:18)
+    if (NOT LLVM_ENABLE_RTTI)
+        target_compile_options(c3c_wrappers PUBLIC /GR-)
     endif()
+    target_link_options(c3c_wrappers PUBLIC /ignore:4099)
+    if(CMAKE_BUILD_TYPE STREQUAL "Debug")
+        target_compile_options(c3c PUBLIC /MTd)
+        target_compile_options(c3c_wrappers PUBLIC /MTd)
+        target_compile_options(miniz PUBLIC /MTd)
+        if (C3_USE_TB)
+            target_compile_options(tilde-backend PUBLIC /MTd)
+        endif()
+    else()
+        target_compile_options(c3c PUBLIC /MT)
+        target_compile_options(c3c_wrappers PUBLIC /MT)
+        target_compile_options(miniz PUBLIC /MT)
+        if (C3_USE_TB)
+            target_compile_options(tilde-backend PUBLIC /MT)
+        endif()
+    endif()
+    set(clang_lib_dir ${llvm_dir}/lib/clang/${C3_LLVM_VERSION}/lib/windows)
+    set(sanitizer_runtime_libraries
+        ${clang_lib_dir}/clang_rt.asan-x86_64.lib
+        ${clang_lib_dir}/clang_rt.asan_dynamic-x86_64.lib
+        ${clang_lib_dir}/clang_rt.asan_dynamic-x86_64.dll
+        ${clang_lib_dir}/clang_rt.asan_dynamic_runtime_thunk-x86_64.lib)
+else()
+    message(STATUS "using gcc/clang warning switches")
+    target_link_options(c3c PRIVATE -pthread)
+    if (NOT LLVM_ENABLE_RTTI)
+        target_compile_options(c3c_wrappers PRIVATE -fno-rtti)
+    endif()
+    target_compile_options(c3c PRIVATE -pthread -Wall -Werror -Wno-unknown-pragmas -Wno-unused-result
+            -Wno-unused-function -Wno-unused-variable -Wno-unused-parameter)
 endif()
 
 install(TARGETS c3c DESTINATION bin)
+install(DIRECTORY lib/ DESTINATION lib/c3)
 
+if (DEFINED sanitizer_runtime_libraries)
+    add_custom_command(TARGET c3c POST_BUILD
+        COMMAND "${CMAKE_COMMAND}" -E rm -rf -- $<TARGET_FILE_DIR:c3c>/c3c_rt
+        COMMAND "${CMAKE_COMMAND}" -E make_directory $<TARGET_FILE_DIR:c3c>/c3c_rt
+        COMMAND "${CMAKE_COMMAND}" -E copy ${sanitizer_runtime_libraries} $<TARGET_FILE_DIR:c3c>/c3c_rt
+        VERBATIM
+        COMMENT "Copying sanitizer runtime libraries to output directory")
+
+    if (APPLE)
+        # Change LC_ID_DYLIB to be rpath-based instead of having an absolute path
+        add_custom_command(TARGET c3c POST_BUILD
+            COMMAND find $<TARGET_FILE_DIR:c3c>/c3c_rt -type f -name "*.dylib" -execdir ${LLVM_TOOLS_BINARY_DIR}/llvm-install-name-tool -id @rpath/{} {} $<SEMICOLON>
+            VERBATIM)
+    endif()
+
+    install(DIRECTORY $<TARGET_FILE_DIR:c3c>/c3c_rt/ DESTINATION bin/c3c_rt)
+endif()
+
+feature_summary(WHAT ALL)

CMakeSettings.json

@@ -0,0 +1,15 @@
+{
+  "configurations": [
+    {
+      "name": "build",
+      "generator": "Ninja",
+      "configurationType": "Release",
+      "inheritEnvironments": [ "msvc_x64_x64" ],
+      "buildRoot": "${projectDir}\\out\\build\\${name}",
+      "installRoot": "${projectDir}\\out\\install\\${name}",
+      "cmakeCommandArgs": "-DLLVM_DIR=C:\\llvm\\llvm\\build\\lib\\cmake\\llvm -DCMAKE_CXX_FLAGS:STRING=\"${CMAKE_CXX_FLAGS} /J\"",
+      "buildCommandArgs": "-v",
+      "ctestCommandArgs": ""
+    }
+  ]
+}

CODESTYLE.md

@@ -116,7 +116,8 @@ more tests as you go along.
 ### Don't bring in dependencies
 
 External libraries has maintainability issues. Try to depend on as few libraries
-as possible. Currently c3c only depends on LLVM and libc.
+as possible. Currently, c3c only depends on LLVM and libc with an optional 
+dependency on libcurl.
 
 Do use rewrites of subsets of other libraries to bring in functionality, but avoid
 copying in libraries that needs to be updated separately.

LICENSE_STDLIB

@@ -0,0 +1,20 @@
+Copyright (c) 2022 Christoffer Lernö and contributors
+
+Permission is hereby granted, free of charge, to any person obtaining
+a copy of this software and associated documentation files (the
+"Software"), to deal in the Software without restriction, including
+without limitation the rights to use, copy, modify, merge, publish,
+distribute, sublicense, and/or sell copies of the Software, and to
+permit persons to whom the Software is furnished to do so, subject to
+the following conditions:
+
+The above copyright notice and this permission notice shall be
+included in all copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE
+LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.

README.md

@@ -1,10 +1,22 @@
 # C3 Language
 
-C3 is a C-like language trying to be "an incremental improvement over C" rather than a whole new language. 
-C3 owes a lot to the ideas of the [C2 language](http://c2lang.org): to iterate on top of C without trying to be a 
-whole new language.
+C3 is a programming language that builds on the syntax and semantics of the C language,
+with the goal of evolving it while still retaining familiarity for C programmers. 
 
-C3 tries to be an alternative in the C/C++ niche: fast and close to the metal.
+It's an evolution, not a revolution: the C-like 
+for programmers who like C.
+
+Precompiled binaries for the following operating systems are available:
+ 
+- Windows x64 [download](https://github.com/c3lang/c3c/releases/download/latest/c3-windows.zip), [install instructions](#installing-on-windows-with-precompiled-binaries).
+- Debian x64 [download](https://github.com/c3lang/c3c/releases/download/latest/c3-linux.tar.gz), [install instructions](#installing-on-debian-with-precompiled-binaries).
+- MacOS x64 [download](https://github.com/c3lang/c3c/releases/download/latest/c3-macos.zip), [install instructions](#installing-on-mac-with-precompiled-binaries).
+
+The manual for C3 can be found at [www.c3-lang.org](http://www.c3-lang.org).
+
+![vkQuake](https://github.com/c3lang/c3c/blob/master/resources/images/vkQuake.png?raw=true)
+
+Thanks to full ABI compatibility with C, it's possible to mix C and C3 in the same project with no effort. As a demonstration, vkQuake was compiled with a small portion of the code converted to C3 and compiled with the c3c compiler. (The fork can be found at https://github.com/c3lang/vkQuake)
 
 ### Design Principles
 - Procedural "get things done"-type of language.
@@ -15,34 +27,102 @@ C3 tries to be an alternative in the C/C++ niche: fast and close to the metal.
 - Avoid "big ideas" & the "more is better" fallacy.
 - Introduce some higher level conveniences where the value is great.
 
+C3 owes its inspiration to the [C2 language](http://c2lang.org): to iterate on top of C without trying to be a
+whole new language.
+
 ### Example code
 
-Create a `main.c3` file with:
-```c++
-module hello_world;
-import std::io;
+The following code shows [generic modules](https://c3-lang.org/references/docs/generics/) (more examples can be found at https://c3-lang.org/references/docs/examples/). 
 
-func void main()
+```c++
+module stack (<Type>);
+// Above: the parameterized type is applied to the entire module.
+
+struct Stack
 {
-   io::printf("Hello, world!\n");
+    usz capacity;
+    usz size;
+    Type* elems;
+}
+
+// The type methods offers dot syntax calls,
+// so this function can either be called 
+// Stack.push(&my_stack, ...) or
+// my_stack.push(...)
+fn void Stack.push(Stack* this, Type element)
+{
+    if (this.capacity == this.size)
+    {
+        this.capacity *= 2;
+		if (this.capacity < 16) this.capacity = 16;
+        this.elems = realloc(this.elems, Type.sizeof * this.capacity);
+    }
+    this.elems[this.size++] = element;
+}
+
+fn Type Stack.pop(Stack* this)
+{
+    assert(this.size > 0);
+    return this.elems[--this.size];
+}
+
+fn bool Stack.empty(Stack* this)
+{
+    return !this.size;
 }
 ```
 
-Make sure you have the standard libraries at either `../lib/std/` or `/lib/std/`.
+Testing it out:
 
-Then run
-```sh
-c3c compile main.c3
+```cpp
+import stack;
+
+// Define our new types, the first will implicitly create 
+// a complete copy of the entire Stack module with "Type" set to "int"
+def IntStack = Stack(<int>);
+// The second creates another copy with "Type" set to "double"
+def DoubleStack = Stack(<double>);
+
+// If we had added "define IntStack2 = Stack(<int>)"
+// no additional copy would have been made (since we already
+// have an parameterization of Stack(<int>)) so it would
+// be same as declaring IntStack2 an alias of IntStack
+
+// Importing an external C function is straightforward
+// here is an example of importing libc's printf:
+extern fn int printf(char* format, ...);
+
+fn void main()
+{
+    IntStack stack;
+    // Note that C3 uses zero initialization by default
+    // so the above is equivalent to IntStack stack = {};
+    
+    stack.push(1);
+    // The above can also be written IntStack.push(&stack, 1); 
+    
+    stack.push(2);
+    
+    // Prints pop: 2
+    printf("pop: %d\n", stack.pop());
+    // Prints pop: 1
+    printf("pop: %d\n", stack.pop());
+    
+    DoubleStack dstack;
+    dstack.push(2.3);
+    dstack.push(3.141);
+    dstack.push(1.1235);
+    // Prints pop: 1.123500
+    printf("pop: %f\n", dstack.pop());
+}
 ```
 
-The generated binary will be called `a.out`.
-
-### In what ways do C3 differ from C?
+### In what ways does C3 differ from C?
 
 - No mandatory header files
 - New semantic macro system
 - Module based name spacing
-- Subarrays (slices)
+- Slices
 - Compile time reflection
 - Enhanced compile time execution
 - Generics based on generic modules
@@ -50,120 +130,108 @@ The generated binary will be called `a.out`.
 - Defer
 - Value methods
 - Associated enum data
-- Built-in hooks for convenient string handling
 - No preprocessor
-- Less undefined behaviour and runtime checks in "safe" mode
+- Less undefined behaviour and added runtime checks in "safe" mode
 - Limited operator overloading to enable userland dynamic arrays
 - Optional pre and post conditions
 
 ### Current status
 
-It's possible to try out the current C3 compiler in the browser: https://ide.judge0.com/ – this is courtesy of the
-developer of Judge0. 
+The current stable version of the compiler is **version 0.6.1**.
 
-Design work is still being done in the design draft here: https://c3lang.github.io/c3docs/. If you have any suggestions, send a mail to [christoffer@aegik.com](mailto:christoffer@aegik.com), [file an issue](https://github.com/c3lang/c3c/issues) or discuss 
-C3 on its dedicated Discord: https://discord.gg/qN76R87
+The upcoming 0.6.x releases will focus on expanding the standard library.
+Follow the issues [here](https://github.com/c3lang/c3c/issues).
 
-The compiler should compile on Linux, Windows (under Mingw or MSYS2) and MacOS, 
-but needs some install documentation for Windows. 
+If you have suggestions on how to improve the language, either [file an issue](https://github.com/c3lang/c3c/issues) 
+or discuss C3 on its dedicated Discord: [https://discord.gg/qN76R87](https://discord.gg/qN76R87).
 
-Due to its ABI compatibility with C, it's possible to mix C and C3 in the same project.
-As a demonstration, vkQuake was compiled with a small portion of the code converted
-to C3 and compiled with the c3c compiler:
+The compiler is currently verified to compile on Linux, Windows and MacOS.
 
-![vkQuake](https://github.com/c3lang/c3c/blob/master/resources/images/vkQuake.png?raw=true)
+**Support matrix**
 
-(The vkFork is at https://github.com/c3lang/vkQuake)
+| Platform                 | Native C3 compiler available? | Target supported        | Stack trace | Threads  | Sockets  | Inline asm |
+|--------------------------|-------------------------------|-------------------------|-------------|----------|----------|------------|
+| Win32 x64                | Yes                           | Yes + cross compilation | Yes         | Yes      | Yes      | Yes*       |
+| Win32 Aarch64            | Untested                      | Untested                | Untested    | Untested | Untested | Yes*       |
+| MacOS x64                | Yes                           | Yes + cross compilation | Yes         | Yes      | Yes      | Yes*       |
+| MacOS Aarch64            | Yes                           | Yes + cross compilation | Yes         | Yes      | Yes      | Yes*       |
+| iOS Aarch64              | No                            | Untested                | Untested    | Yes      | Yes      | Yes*       |
+| Linux x86                | Yes                           | Yes                     | Yes         | Yes      | Yes      | Yes*       |
+| Linux x64                | Yes                           | Yes                     | Yes         | Yes      | Yes      | Yes*       |
+| Linux Aarch64            | Yes                           | Yes                     | Yes         | Yes      | Yes      | Yes*       |
+| Linux Riscv32            | Yes                           | Yes                     | Yes         | Yes      | Yes      | Untested   |
+| Linux Riscv64            | Yes                           | Yes                     | Yes         | Yes      | Yes      | Untested   |
+| ELF freestanding x86     | No                            | Untested                | No          | No       | No       | Yes*       |
+| ELF freestanding x64     | No                            | Untested                | No          | No       | No       | Yes*       |
+| ELF freestanding Aarch64 | No                            | Untested                | No          | No       | No       | Yes*       |
+| ELF freestanding Riscv64 | No                            | Untested                | No          | No       | No       | Untested   |
+| ELF freestanding Riscv32 | No                            | Untested                | No          | No       | No       | Untested   |
+| FreeBSD x86              | Untested                      | Untested                | No          | Yes      | Untested | Yes*       |
+| FreeBSD x64              | Untested                      | Untested                | No          | Yes      | Untested | Yes*       |
+| NetBSD x86               | Untested                      | Untested                | No          | Yes      | Untested | Yes*       |
+| NetBSD x64               | Untested                      | Untested                | No          | Yes      | Untested | Yes*       |
+| OpenBSD x86              | Untested                      | Untested                | No          | Yes      | Untested | Yes*       |
+| OpenBSD x64              | Untested                      | Untested                | No          | Yes      | Untested | Yes*       |
+| MCU x86                  | No                            | Untested                | No          | No       | No       | Yes*       |
+| Wasm32                   | No                            | Yes                     | No          | No       | No       | No         |
+| Wasm64                   | No                            | Untested                | No          | No       | No       | No         |
 
-#### Todo / done
+*\* Inline asm is still a work in progress*
 
-- [x] For/while/do
-- [x] `if`/ternary
-- [x] Structs
-- [x] Union
-- [x] Enums
-- [x] Value methods
-- [x] Compound literals
-- [x] Designated initalizers
-- [x] Slicing syntax
-- [x] Arrays and subarrays
-- [x] Modules
-- [x] `$unreachable`
-- [x] Compile time assert with `$assert`
-- [x] Compiler guiding `assert` 
-- [x] C code calling by declaring methods `extern`
-- [x] Compile time variables
-- [x] Basic macros
-- [x] 4cc, 8cc, 2cc
-- [x] Enum type inference in switch/assignment
-- [x] Integer type inference
-- [x] Error type
-- [x] Failable error handling
-- [x] `try` for conditional execution
-- [x] `catch` for error handling
-- [x] Implicit unwrap after `catch`
-- [x] `sizeof`
-- [x] `typeof`
-- [x] 2s complement wrapping operators
-- [x] Labelled break / continue
-- [x] `nextcase` statement
-- [x] Expression blocks
-- [x] Do-without-while
-- [x] Foreach statement
-- [x] Templates
-- [x] Distinct types
-- [x] Built-in linking
-- [x] CT only macros evaluating to constants
-- [x] range initializers e.g. `{ [1..2] = 2 }`
-- [x] Trailing body macros e.g. `@foo(1, 100; int a) { bar(a); };`
-- [x] Complex macros
-- [x] CT type constants
-- [ ] Anonymous structs
-- [ ] Complete C ABI conformance *in progress*
-- [ ] Debug info *in progress*
-- [ ] Virtual type *in progress*
-- [ ] Enum associated data support  
-- [ ] Windows support *in progress*
-- [ ] All attributes *in progress*
-- [ ] Associative array literals
-- [ ] Reflection methods
-- [ ] LTO/ThinLTO setup
-- [ ] `global` / `shared` for globals 
-- [ ] Escape macros
-- [ ] Implicit capturing macros
-- [ ] Subarray initializers
-- [ ] Bitstructs
-- [ ] `asm` section
-- [ ] `$switch`
-- [ ] `$for`
-- [ ] Pre-post conditions
-- [ ] Stdlib inclusion
-- [ ] String functions
-- [ ] Simd vector types
+More platforms will be supported in the future.
 
 #### What can you help with?
 
-- If you wish to contribute with ideas, please file issues on the c3docs: https://github.com/c3lang/c3docs instead of the compiler.
-- Discuss the language on discord to help iron out syntax.
+- If you wish to contribute with ideas, please file issues or discuss on Discord.
 - Interested in contributing to the stdlib? Please get in touch on Discord.
-- Are you a Windows dev? Please help make the compiler work on Windows!
-- Install instructions for other Linux and Unix variants are appreciated.
+- Compilation instructions for other Linux and Unix variants are appreciated.
+- Would you like to contribute bindings to some library? It would be nice to have support for SDL, Raylib and more.
+- Build something with C3 and show it off and give feedback. The language is still open for significant tweaks.
+- Start work on the C -> C3 converter which takes C code and does a "best effort" to translate it to C3. The first version only needs to work on C headers.
+- Do you have some specific area you have deep knowledge of and could help make C3 even better at doing? File or comment on issues.
 
-#### Installing on Ubuntu 20.10
+### Installing
 
-1. Make sure you have a C compiler that handles C11 and a C++ compiler, such as GCC or Clang. Git also needs to be installed.
-2. Install CMake: `sudo apt install cmake`
-3. Install LLVM 11: `sudo apt-get install clang-11 zlib1g zlib1g-dev libllvm11 llvm-11 llvm-11-dev llvm-11-runtime liblld-11-dev liblld-11`
-4. Clone the C3C github repository: `git clone https://github.com/c3lang/c3c.git`
-5. Enter the C3C directory `cd c3c`.
-6. Create a build directory `mkdir build`
-7. Change directory to the build directory `cd build`
-8. Set up CMake build for debug: `cmake -DLLVM_DIR=/usr/lib/llvm-11/cmake -DCMAKE_BUILD_TYPE=Debug ..`
-9. Build: `cmake --build .`
+#### Installing on Windows with precompiled binaries
+1. Download the zip file: [https://github.com/c3lang/c3c/releases/download/latest/c3-windows.zip](https://github.com/c3lang/c3c/releases/download/latest/c3-windows.zip) 
+(debug version [here](https://github.com/c3lang/c3c/releases/download/latest/c3-windows-debug.zip))
+2. Unzip exe and standard lib.
+3. If you don't have Visual Studio 17 installed you can either do so, or run the `msvc_build_libraries.py` Python script which will download the necessary files to compile on Windows. 
+4. Run `c3c.exe`.
 
-You should now have a `c3c` executable.
+#### Installing on Debian with precompiled binaries
+1. Download tar file: [https://github.com/c3lang/c3c/releases/download/latest/c3-linux.tar.gz](https://github.com/c3lang/c3c/releases/download/latest/c3-linux.tar.gz)
+   (debug version [here](https://github.com/c3lang/c3c/releases/download/latest/c3-linux-debug.tar.gz))
+2. Unpack executable and standard lib.
+3. Run `./c3c`.
 
-You can try it out by running some sample code: `./c3c compile ../resources/examples/hash.c3`
+#### Installing on Mac with precompiled binaries
+1. Make sure you have XCode with command line tools installed.
+2. Download the zip file: [https://github.com/c3lang/c3c/releases/download/latest/c3-macos.zip](https://github.com/c3lang/c3c/releases/download/latest/c3-macos.zip)
+   (debug version [here](https://github.com/c3lang/c3c/releases/download/latest/c3-macos-debug.zip))
+3. Unzip executable and standard lib.
+4. Run `./c3c`.
+
+(*Note that there is a known issue with debug symbol generation on MacOS 13, see [issue #1086](https://github.com/c3lang/c3c/issues/1086))
+
+#### Installing on Arch Linux
+There is an AUR package for the c3c compiler : [c3c-git](https://aur.archlinux.org/packages/c3c-git).
+
+Due to some issues with the LLVM packaged for Arch Linux, the AUR package will download and use LLVM 16 for Ubuntu-23.04 to compile the c3c compiler.
+
+You can use your AUR package manager:
+```sh
+paru -S c3c-git
+# or yay -S c3c-git
+# or aura -A c3c-git
+```
+
+Or clone it manually:
+```sh
+git clone https://aur.archlinux.org/c3c-git.git
+cd c3c-git
+makepkg -si
+```
 
 #### Building via Docker
 
@@ -185,10 +253,119 @@ A `c3c` executable will be found under `bin/`.
 #### Installing on OS X using Homebrew
 
 2. Install CMake: `brew install cmake`
-3. Install LLVM 11: `brew install llvm`
+3. Install LLVM 17+: `brew install llvm`
 4. Clone the C3C github repository: `git clone https://github.com/c3lang/c3c.git`
 5. Enter the C3C directory `cd c3c`.
 6. Create a build directory `mkdir build`
 7. Change directory to the build directory `cd build`
 8. Set up CMake build for debug: `cmake ..`
 9. Build: `cmake --build .`
+
+#### Getting started with a "hello world"
+
+Create a `main.c3` file with:
+```c++
+module hello_world;
+import std::io;
+
+fn void main()
+{
+   io::printn("Hello, world!");
+}
+```
+
+Make sure you have the standard libraries at either `../lib/std/` or `/lib/std/`.
+
+Then run
+```sh
+c3c compile main.c3
+```
+
+The generated binary will by default be named after the module that contains the main
+function. In our case that is `hello_world`, so the resulting binary will be
+called `hello_world` or `hello_world.exe`depending on platform.
+
+### Compiling
+
+#### Compiling on Windows
+
+1. Make sure you have Visual Studio 17 2022 installed or alternatively install the "Buildtools for Visual Studio" (https://aka.ms/vs/17/release/vs_BuildTools.exe) and then select "Desktop development with C++" (there is also `c3c/resources/install_win_reqs.bat` to automate this)
+2. Install CMake
+3. Clone the C3C github repository: `git clone https://github.com/c3lang/c3c.git`
+4. Enter the C3C directory `cd c3c`.
+5. Set up the CMake build `cmake -B build -G "Visual Studio 17 2022" -A x64 -DCMAKE_BUILD_TYPE=Release`
+6. Build: `cmake --build build --config Release`
+7. You should now have the c3c.exe
+
+You should now have a `c3c` executable.
+
+You can try it out by running some sample code: `c3c.exe compile ../resources/examples/hash.c3`
+
+*Note that if you run into linking issues when building, make sure that you are using the latest version of VS17.*
+
+
+#### Compiling on Ubuntu 20.10
+
+1. Make sure you have a C compiler that handles C11 and a C++ compiler, such as GCC or Clang. Git also needs to be installed.
+2. Install CMake: `sudo apt install cmake`
+3. Install LLVM 17+ (or greater: C3C supports LLVM 17+): `sudo apt-get install clang-17 zlib1g zlib1g-dev libllvm17 llvm-17 llvm-17-dev llvm-17-runtime liblld-17-dev liblld-17`
+4. Clone the C3C github repository: `git clone https://github.com/c3lang/c3c.git`
+5. Enter the C3C directory `cd c3c`.
+6. Create a build directory `mkdir build`
+7. Change directory to the build directory `cd build`
+8. Set up CMake build: `cmake ..`
+9. Build: `cmake --build .`
+
+You should now have a `c3c` executable.
+
+You can try it out by running some sample code: `./c3c compile ../resources/examples/hash.c3`
+
+
+#### Compiling on Void Linux
+
+1. As root, ensure that all project dependencies are installed: `xbps-install git cmake llvm17 lld17-devel libcurl-devel ncurses-devel zlib-devel libzstd-devel libxml2-devel`
+2. Clone the C3C repository: `git clone https://github.com/c3lang/c3c.git`
+    - If you only need the latest commit, you may want to make a shallow clone instead: `git clone https://github.com/c3lang/c3c.git --depth=1`
+3. Enter the directory: `cd c3c`
+4. Create a build directory: `mkdir build`
+5. Enter the build directory: `cd build`
+6. Create the CMake build cache: `cmake ..`
+7. Build: `cmake --build .`
+
+Your c3c executable should have compiled properly. You may want to test it: `./c3c compile ../resources/examples/hash.c3`  
+For a sytem-wide installation, run the following as root: `cmake --install .`
+
+
+#### Compiling on other Linux / Unix variants
+
+1. Install CMake.
+2. Install or compile LLVM and LLD *libraries* (version 17+ or higher)
+3. Clone the C3C github repository: `git clone https://github.com/c3lang/c3c.git`
+4. Enter the C3C directory `cd c3c`.
+5. Create a build directory `mkdir build`
+6. Change directory to the build directory `cd build`
+7. Set up CMake build for debug: `cmake ..`. At this point you may need to manually 
+provide the link path to the LLVM CMake directories, e.g. `cmake -DLLVM_DIR=/usr/local/opt/llvm/lib/cmake/llvm/ ..`
+8. Build: `cmake --build .`
+
+*A note on compiling for Linux/Unix/MacOS: to be able to fetch vendor libraries
+libcurl is needed. The CMake script should detect it if it is available. Note that
+this functionality is non-essential and it is perfectly fine to user the compiler without it.*
+
+#### Licensing
+
+The C3 compiler is licensed under LGPL 3.0, the standard library itself is
+MIT licensed.
+
+#### Editor plugins
+
+Editor plugins can be found at https://github.com/c3lang/editor-plugins.
+
+#### Contributing unit tests
+
+1. Write the test, either adding to existing test files in `/test/unit/` or add
+   a new file. (If testing the standard library, put it in the `/test/unit/stdlib/` subdirectory).
+2. Make sure that the test functions have the `@test` attribute. 
+3. Run tests and see that they pass. (Recommended settings: `c3c compile-test -O0 test/unit`.
+   - in this example `test/unit/` is the relative path to the test directory, so adjust as required)
+4. Make a pull request for the new tests.

build-with-docker.sh

@@ -1,42 +1,44 @@
 #!/bin/bash
-## build-with-docker.sh
-## @author gdm85
-##
-## Script to build c3c for either Ubuntu 18 or Ubuntu 20.
-##
-#
 
-if [ $# -ne 1 -a $# -ne 2 ]; then
-	echo "Usage: build-with-docker.sh (18|20) [Debug|Release]" 1>&2
-	exit 1
+: ${DOCKER:=docker}
+: ${IMAGE:="c3c-builder"}
+: ${CMAKE_BUILD_TYPE:=Release}
+: ${LLVM_VERSION:=18}
+: ${UBUNTU_VERSION:="22.04"}
+: ${CMAKE_VERSION:="3.20.0"}
+
+cd docker || exit 1  # Exit if the 'docker' directory doesn't exist
+
+$DOCKER build \
+    --build-arg LLVM_VERSION=$LLVM_VERSION \
+    --build-arg CMAKE_VERSION=$CMAKE_VERSION \
+    --build-arg UBUNTU_VERSION=$UBUNTU_VERSION \
+    -t $IMAGE .
+
+if [ $? -ne 0 ]; then
+    echo "Docker image build failed. Exiting."
+    exit 1
 fi
 
-set -e
-
-if [ -z "$2" ]; then
-	CMAKE_BUILD_TYPE=Debug
-else
-	CMAKE_BUILD_TYPE="$2"
-fi
-
-TAG="$1"
-if [ "$1" = 18 ]; then
-	UBUNTU_VERSION="18.04"
-	DEPS="llvm-10-dev liblld-10-dev libclang-10-dev"
-elif [ "$1" = 20 ]; then
-	UBUNTU_VERSION="20.04"
-	DEPS="llvm-11-dev liblld-11-dev clang-11 libllvm11 llvm-11-runtime"
-else
-	echo "ERROR: expected 18 or 20 as Ubuntu version argument" 1>&2
-	exit 2
-fi
-
-cd docker && docker build -t c3c-builder:$TAG --build-arg UID=$(id -u) --build-arg GID=$(id -g) \
-	--build-arg DEPS="$DEPS" --build-arg UBUNTU_VERSION="$UBUNTU_VERSION" .
 cd ..
 
-rm -rf build bin lib
+rm -rf build bin
 mkdir -p build bin
 
-exec docker run -ti --rm -v "$PWD":/home/c3c/source -w /home/c3c/source c3c-builder:$TAG bash -c \
-	"cd build && cmake -DLLVM_DIR=/usr/lib/llvm-11/cmake -DCMAKE_BUILD_TYPE=$CMAKE_BUILD_TYPE .. && cmake --build . && mv c3c ../bin/ && mv lib ../lib"
+chmod -R 777 build bin
+
+exec $DOCKER run -i --rm \
+    -v "$PWD":/home/c3c/source \
+    -w /home/c3c/source $IMAGE bash -c \
+    "cmake -S . -B build \
+            -G Ninja \
+            -DCMAKE_BUILD_TYPE=$CMAKE_BUILD_TYPE \
+            -DCMAKE_C_COMPILER=clang-$LLVM_VERSION \
+            -DCMAKE_CXX_COMPILER=clang++-$LLVM_VERSION \
+            -DCMAKE_LINKER=lld-$LLVM_VERSION \
+            -DCMAKE_OBJCOPY=llvm-objcopy-$LLVM_VERSION \
+            -DCMAKE_STRIP=llvm-strip-$LLVM_VERSION \
+            -DCMAKE_DLLTOOL=llvm-dlltool-$LLVM_VERSION \
+            -DC3_LLVM_VERSION=auto && \
+    cmake --build build && \
+    cp -r build/c3c build/lib bin"

docker/Dockerfile

@@ -1,16 +1,49 @@
+ARG UBUNTU_VERSION=22.04
+FROM ubuntu:${UBUNTU_VERSION}
 
-ARG UBUNTU_VERSION
-FROM ubuntu:$UBUNTU_VERSION
+ARG LLVM_VERSION=18
+ENV LLVM_DEV_VERSION=20
 
-ARG DEPS
+ARG CMAKE_VERSION=3.20
 
-RUN export DEBIAN_FRONTEND=noninteractive && export TERM=xterm && apt-get update && apt-get install -y build-essential cmake zlib1g zlib1g-dev  \
-        $DEPS && \
-	rm -rf /var/lib/apt/lists/*
+RUN apt-get update && apt-get install -y wget gnupg software-properties-common zlib1g zlib1g-dev python3 ninja-build curl g++ && \
+    wget https://github.com/Kitware/CMake/releases/download/v${CMAKE_VERSION}/cmake-$CMAKE_VERSION-linux-x86_64.sh && \
+    mkdir -p /opt/cmake && \
+    sh cmake-${CMAKE_VERSION}-linux-x86_64.sh --prefix=/opt/cmake --skip-license && \
+    rm cmake-${CMAKE_VERSION}-linux-x86_64.sh && \
+    ln -s /opt/cmake/bin/cmake /usr/local/bin/cmake
 
-ARG GID=1000
-ARG UID=1000
+RUN wget -O - https://apt.llvm.org/llvm-snapshot.gpg.key | apt-key add - && \
+    if [ "${LLVM_VERSION}" -lt 18 ]; then \
+        add-apt-repository "deb http://apt.llvm.org/focal/ llvm-toolchain-focal-${LLVM_VERSION} main" && \
+        apt-get update && \
+        apt-get install -y -t llvm-toolchain-focal-${LLVM_VERSION} \
+          libpolly-${LLVM_VERSION}-dev \
+          clang-${LLVM_VERSION} llvm-${LLVM_VERSION} llvm-${LLVM_VERSION}-dev \
+          lld-${LLVM_VERSION} liblld-${LLVM_VERSION}-dev libmlir-${LLVM_VERSION} \
+          libmlir-${LLVM_VERSION}-dev mlir-${LLVM_VERSION}-tools; \
+    elif [ "${LLVM_VERSION}" -lt "${LLVM_DEV_VERSION}" ]; then \
+        add-apt-repository "deb http://apt.llvm.org/focal/ llvm-toolchain-focal-${LLVM_VERSION} main" && \
+        apt-get update && \
+        apt-get install -y -t llvm-toolchain-focal-${LLVM_VERSION} \
+          libpolly-${LLVM_VERSION}-dev \
+          clang-${LLVM_VERSION} clang++-${LLVM_VERSION} llvm-${LLVM_VERSION} llvm-${LLVM_VERSION}-dev \
+          lld-${LLVM_VERSION} liblld-${LLVM_VERSION}-dev; \
+    else \
+        add-apt-repository "deb http://apt.llvm.org/focal/ llvm-toolchain-focal main" && \
+        apt-get update && \
+        apt-get install -y -t llvm-toolchain-focal \
+          libpolly-${LLVM_VERSION}-dev \
+          clang-${LLVM_VERSION} llvm-${LLVM_VERSION} llvm-${LLVM_VERSION}-dev \
+          lld-${LLVM_VERSION} liblld-${LLVM_VERSION}-dev; \
+    fi && \
+    rm -rf /var/lib/apt/lists/*
 
-RUN groupadd --gid=$GID c3c && useradd --gid=$GID --uid=$GID --create-home --shell /bin/bash c3c
+RUN groupadd -g 1337 c3c && \
+    useradd -m -u 1337 -g c3c c3c
 
+# Add cmake to PATH for user c3c
 USER c3c
+ENV PATH="/opt/cmake/bin:${PATH}"
+
+WORKDIR /home/c3c

lib/std/ascii.c3

@@ -0,0 +1,77 @@
+module std::ascii;
+
+macro bool in_range_m(c, start, len) => (uint)(c - start) < len;
+macro bool is_lower_m(c) => in_range_m(c, 0x61, 26);
+macro bool is_upper_m(c) => in_range_m(c, 0x41, 26);
+macro bool is_digit_m(c) => in_range_m(c, 0x30, 10);
+macro bool is_bdigit_m(c) => in_range_m(c, 0x30, 2);
+macro bool is_odigit_m(c) => in_range_m(c, 0x30, 8);
+macro bool is_xdigit_m(c) => in_range_m(c | 32, 0x61, 6) || is_digit_m(c);
+macro bool is_alpha_m(c) => in_range_m(c | 32, 0x61, 26);
+macro bool is_print_m(c) => in_range_m(c, 0x20, 95);
+macro bool is_graph_m(c) => in_range_m(c, 0x21, 94);
+macro bool is_space_m(c) => in_range_m(c, 0x9, 5) || c == 0x20;
+macro bool is_alnum_m(c) => is_alpha_m(c) || is_digit_m(c);
+macro bool is_punct_m(c) => !is_alnum_m(c) && is_graph_m(c);
+macro bool is_blank_m(c) => c == 0x20 || c == 0x9;
+macro bool is_cntrl_m(c) => c < 0x20 || c == 0x7f;
+macro to_lower_m(c) => is_upper_m(c) ? c + 0x20 : c;
+macro to_upper_m(c) => is_lower_m(c) ? c - 0x20 : c;
+
+fn bool in_range(char c, char start, char len) => in_range_m(c, start, len);
+fn bool is_lower(char c) => is_lower_m(c);
+fn bool is_upper(char c) => is_upper_m(c);
+fn bool is_digit(char c) => is_digit_m(c);
+fn bool is_bdigit(char c) => is_bdigit_m(c);
+fn bool is_odigit(char c) => is_odigit_m(c);
+fn bool is_xdigit(char c) => is_xdigit_m(c);
+fn bool is_alpha(char c) => is_alpha_m(c);
+fn bool is_print(char c) => is_print_m(c);
+fn bool is_graph(char c) => is_graph_m(c);
+fn bool is_space(char c) => is_space_m(c);
+fn bool is_alnum(char c) => is_alnum_m(c);
+fn bool is_punct(char c) => is_punct_m(c);
+fn bool is_blank(char c) => is_blank_m(c);
+fn bool is_cntrl(char c) => is_cntrl_m(c);
+fn char to_lower(char c) => (char)to_lower_m(c);
+fn char to_upper(char c) => (char)to_upper_m(c);
+
+fn bool char.in_range(char c, char start, char len) => in_range_m(c, start, len);
+fn bool char.is_lower(char c) => is_lower_m(c);
+fn bool char.is_upper(char c) => is_upper_m(c);
+fn bool char.is_digit(char c) => is_digit_m(c);
+fn bool char.is_bdigit(char c) => is_bdigit_m(c);
+fn bool char.is_odigit(char c) => is_odigit_m(c);
+fn bool char.is_xdigit(char c) => is_xdigit_m(c);
+fn bool char.is_alpha(char c) => is_alpha_m(c);
+fn bool char.is_print(char c) => is_print_m(c);
+fn bool char.is_graph(char c) => is_graph_m(c);
+fn bool char.is_space(char c) => is_space_m(c);
+fn bool char.is_alnum(char c) => is_alnum_m(c);
+fn bool char.is_punct(char c) => is_punct_m(c);
+fn bool char.is_blank(char c) => is_blank_m(c);
+fn bool char.is_cntrl(char c) => is_cntrl_m(c);
+fn char char.to_lower(char c) => (char)to_lower_m(c);
+fn char char.to_upper(char c) => (char)to_upper_m(c);
+/**
+ * @require c.is_xdigit()
+ **/
+fn char char.from_hex(char c) => c.is_digit() ? c - '0' : 10 + (c | 0x20) - 'a';
+
+fn bool uint.in_range(uint c, uint start, uint len) => in_range_m(c, start, len);
+fn bool uint.is_lower(uint c) => is_lower_m(c);
+fn bool uint.is_upper(uint c) => is_upper_m(c);
+fn bool uint.is_digit(uint c) => is_digit_m(c);
+fn bool uint.is_bdigit(uint c) => is_bdigit_m(c);
+fn bool uint.is_odigit(uint c) => is_odigit_m(c);
+fn bool uint.is_xdigit(uint c) => is_xdigit_m(c);
+fn bool uint.is_alpha(uint c) => is_alpha_m(c);
+fn bool uint.is_print(uint c) => is_print_m(c);
+fn bool uint.is_graph(uint c) => is_graph_m(c);
+fn bool uint.is_space(uint c) => is_space_m(c);
+fn bool uint.is_alnum(uint c) => is_alnum_m(c);
+fn bool uint.is_punct(uint c) => is_punct_m(c);
+fn bool uint.is_blank(uint c) => is_blank_m(c);
+fn bool uint.is_cntrl(uint c) => is_cntrl_m(c);
+fn uint uint.to_lower(uint c) => (uint)to_lower_m(c);
+fn uint uint.to_upper(uint c) => (uint)to_upper_m(c);

lib/std/atomic.c3

@@ -0,0 +1,520 @@
+// Copyright (c) 2023 Eduardo José Gómez Hernández. All rights reserved.
+// Use of this source code is governed by the MIT license
+// a copy of which can be found in the LICENSE_STDLIB file.
+module std::atomic::types(<Type>);
+
+struct Atomic
+{
+	Type data;
+}
+
+/**
+ * Loads data atomically, by default this uses SEQ_CONSISTENT ordering.
+ *
+ * @param ordering "The ordering, cannot be release or acquire-release."
+ * @require ordering != RELEASE && ordering != ACQUIRE_RELEASE : "Release and acquire-release are not valid for load"
+ **/
+macro Type Atomic.load(&self, AtomicOrdering ordering = SEQ_CONSISTENT)
+{
+	Type* data = &self.data;
+	switch(ordering)
+	{
+		case NOT_ATOMIC: return $$atomic_load(data, false, AtomicOrdering.NOT_ATOMIC.ordinal);
+		case UNORDERED: return $$atomic_load(data, false, AtomicOrdering.UNORDERED.ordinal);
+		case RELAXED: return $$atomic_load(data, false, AtomicOrdering.RELAXED.ordinal);
+		case ACQUIRE: return $$atomic_load(data, false, AtomicOrdering.ACQUIRE.ordinal);
+		case SEQ_CONSISTENT: return $$atomic_load(data, false, AtomicOrdering.SEQ_CONSISTENT.ordinal);
+		case ACQUIRE_RELEASE:
+		case RELEASE: unreachable("Invalid ordering.");
+	}
+}
+/**
+ * Stores data atomically, by default this uses SEQ_CONSISTENT ordering.
+ *
+ * @param ordering "The ordering, cannot be acquire or acquire-release."
+ * @require ordering != ACQUIRE && ordering != ACQUIRE_RELEASE : "Acquire and acquire-release are not valid for store"
+ **/
+macro void Atomic.store(&self, Type value, AtomicOrdering ordering = SEQ_CONSISTENT)
+{
+	Type* data = &self.data;
+	switch(ordering) 
+	{
+		case NOT_ATOMIC: $$atomic_store(data, value, false, AtomicOrdering.NOT_ATOMIC.ordinal);
+		case UNORDERED: $$atomic_store(data, value, false, AtomicOrdering.UNORDERED.ordinal);
+		case RELAXED: $$atomic_store(data, value, false, AtomicOrdering.RELAXED.ordinal);
+		case RELEASE: $$atomic_store(data, value, false, AtomicOrdering.RELEASE.ordinal);
+		case SEQ_CONSISTENT: $$atomic_store(data, value, false, AtomicOrdering.SEQ_CONSISTENT.ordinal);
+		case ACQUIRE_RELEASE:
+		case ACQUIRE: unreachable("Invalid ordering.");
+	}
+}
+
+macro Type Atomic.add(&self, Type value, AtomicOrdering ordering = SEQ_CONSISTENT)
+{
+	Type* data = &self.data;
+	return @atomic_exec(atomic::fetch_add, data, value, ordering);
+}
+
+macro Type Atomic.sub(&self, Type value, AtomicOrdering ordering = SEQ_CONSISTENT)
+{
+	Type* data = &self.data;
+	return @atomic_exec(atomic::fetch_sub, data, value, ordering);
+}
+
+macro Type Atomic.mul(&self, Type value, AtomicOrdering ordering = SEQ_CONSISTENT)
+{
+	Type* data = &self.data;
+	return @atomic_exec(atomic::fetch_mul, data, value, ordering);
+}
+
+macro Type Atomic.div(&self, Type value, AtomicOrdering ordering = SEQ_CONSISTENT)
+{
+	Type* data = &self.data;
+	return @atomic_exec(atomic::fetch_div, data, value, ordering);
+}
+
+macro Type Atomic.max(&self, Type value, AtomicOrdering ordering = SEQ_CONSISTENT)
+{
+	Type* data = &self.data;
+	return @atomic_exec(atomic::fetch_div, data, value, ordering);
+}
+
+macro Type Atomic.min(&self, Type value, AtomicOrdering ordering = SEQ_CONSISTENT)
+{
+	Type* data = &self.data;
+	return @atomic_exec(atomic::fetch_min, data, value, ordering);
+}
+
+macro Type Atomic.or(&self, uint value, AtomicOrdering ordering = SEQ_CONSISTENT) @if(!types::is_float(Type))
+{
+	Type* data = &self.data;
+	return @atomic_exec(atomic::fetch_or, data, value, ordering);
+}
+
+fn Type Atomic.xor(&self, uint value, AtomicOrdering ordering = SEQ_CONSISTENT) @if(!types::is_float(Type))
+{
+	Type* data = &self.data;
+	return @atomic_exec(atomic::fetch_xor, data, value, ordering);
+}
+
+macro Type Atomic.and(&self, uint value, AtomicOrdering ordering = SEQ_CONSISTENT) @if(!types::is_float(Type))
+{
+	Type* data = &self.data;
+	return @atomic_exec(atomic::fetch_and, data, value, ordering);
+}
+
+macro Type Atomic.shift_right(&self, uint amount, AtomicOrdering ordering = SEQ_CONSISTENT) @if(!types::is_float(Type))
+{
+	Type* data = &self.data;
+	return @atomic_exec(atomic::fetch_shift_right, data, amount, ordering);
+}
+
+macro Type Atomic.shift_left(&self, uint amount, AtomicOrdering ordering = SEQ_CONSISTENT) @if(!types::is_float(Type))
+{
+	Type* data = &self.data;
+	return @atomic_exec(atomic::fetch_shift_left, data, amount, ordering);
+}
+
+macro @atomic_exec(#func, data, value, ordering) @local
+{
+	switch(ordering) 
+	{
+		case RELAXED: return #func(data, value, RELAXED);
+		case ACQUIRE: return #func(data, value, ACQUIRE);
+		case RELEASE: return #func(data, value, RELEASE);
+		case ACQUIRE_RELEASE: return #func(data, value, ACQUIRE_RELEASE);
+		case SEQ_CONSISTENT: return #func(data, value, SEQ_CONSISTENT);
+		default: unreachable("Ordering may not be non-atomic or unordered.");
+	}
+}
+
+module std::atomic;
+import std::math;
+
+/**
+ * @param [&in] ptr "the variable or dereferenced pointer to the data."
+ * @param [in] y "the value to be added to ptr."
+ * @param $ordering "atomic ordering of the load, defaults to SEQ_CONSISTENT"
+ * @return "returns the old value of ptr"
+ *
+ * @require !$alignment || math::is_power_of_2($alignment) "Alignment must be a power of two."
+ * @require types::is_int($typeof(*ptr)) || types::is_float($typeof(*ptr)) "Only integer/float pointers may be used."
+ * @require $ordering != AtomicOrdering.NOT_ATOMIC && $ordering != AtomicOrdering.UNORDERED "Acquire ordering is not valid."
+ **/
+macro fetch_add(ptr, y, AtomicOrdering $ordering = SEQ_CONSISTENT, bool $volatile = false, usz $alignment = 0)
+{
+	$if $alignment == 0:
+		$alignment = $typeof(*ptr).sizeof;
+	$endif
+	return $$atomic_fetch_add(ptr, y, $volatile, $ordering.ordinal, $alignment);
+}
+
+/**
+ * @param [&in] ptr "the variable or dereferenced pointer to the data."
+ * @param [in] y "the value to be added to ptr."
+ * @param $ordering "atomic ordering of the load, defaults to SEQ_CONSISTENT"
+ * @return "returns the old value of ptr"
+ * 
+ * @require !$alignment || math::is_power_of_2($alignment) "Alignment must be a power of two."
+ * @require types::is_int($typeof(*ptr)) || types::is_float($typeof(*ptr)) "Only integer/float pointers may be used."
+ * @require $ordering != AtomicOrdering.NOT_ATOMIC && $ordering != AtomicOrdering.UNORDERED "Acquire ordering is not valid."
+ **/
+macro fetch_sub(ptr, y, AtomicOrdering $ordering = SEQ_CONSISTENT, bool $volatile = false, usz $alignment = 0)
+{
+	$if $alignment == 0:
+		$alignment = $typeof(*ptr).sizeof;
+	$endif
+	return $$atomic_fetch_sub(ptr, y, $volatile, $ordering.ordinal, $alignment);
+}
+
+/**
+ * @param [&in] ptr "the variable or dereferenced pointer to the data."
+ * @param [in] y "the value to be added to ptr."
+ * @param $ordering "atomic ordering of the load, defaults to SEQ_CONSISTENT"
+ * @return "returns the old value of ptr"
+ * 
+ * @require types::is_int($typeof(*ptr)) || types::is_float($typeof(*ptr)) "Only integer/float pointers may be used."
+ * @require $ordering != AtomicOrdering.NOT_ATOMIC && $ordering != AtomicOrdering.UNORDERED "Acquire ordering is not valid."
+ **/
+macro fetch_mul(ptr, y, AtomicOrdering $ordering = SEQ_CONSISTENT)
+{
+	var $load_ordering = $ordering;
+	$if $ordering == RELEASE || $ordering == ACQUIRE_RELEASE:
+	    $load_ordering = AtomicOrdering.SEQ_CONSISTENT;
+	$endif
+	
+	var $StorageType = $typefrom(types::lower_to_atomic_compatible_type($typeof(*ptr)));
+
+	$StorageType* storage_ptr = ($StorageType*)ptr;
+
+	$typeof(*ptr) old_value;
+	$typeof(*ptr) new_value;
+
+	$StorageType storage_old_value;
+	$StorageType storage_new_value;
+
+
+	do {
+		storage_old_value = $$atomic_load(storage_ptr, false, $load_ordering.ordinal);
+		old_value = bitcast(storage_old_value, $typeof(*ptr));
+		new_value = old_value * y;
+		storage_new_value = bitcast(new_value, $StorageType);
+	} while (mem::compare_exchange(storage_ptr, storage_old_value, storage_new_value, $ordering, $load_ordering) != storage_old_value);
+
+	return old_value;
+}
+
+/**
+ * @param [&in] ptr "the variable or dereferenced pointer to the data."
+ * @param [in] y "the value to be added to ptr."
+ * @param $ordering "atomic ordering of the load, defaults to SEQ_CONSISTENT"
+ * @return "returns the old value of ptr"
+ * 
+* @require types::is_int($typeof(*ptr)) || types::is_float($typeof(*ptr)) "Only integer/float pointers may be used."
+* @require $ordering != AtomicOrdering.NOT_ATOMIC && $ordering != AtomicOrdering.UNORDERED "Acquire ordering is not valid."
+ **/
+macro fetch_div(ptr, y, AtomicOrdering $ordering = SEQ_CONSISTENT)
+{
+	var $load_ordering = $ordering;
+	$if $ordering == RELEASE || $ordering == ACQUIRE_RELEASE:
+	    $load_ordering = AtomicOrdering.SEQ_CONSISTENT;
+	$endif
+	
+	var $StorageType = $typefrom(types::lower_to_atomic_compatible_type($typeof(*ptr)));
+
+	$StorageType* storage_ptr = ($StorageType*)ptr;
+
+	$typeof(*ptr) old_value;
+	$typeof(*ptr) new_value;
+
+	$StorageType storage_old_value;
+	$StorageType storage_new_value;
+	
+	do {
+		storage_old_value = $$atomic_load(storage_ptr, false, $load_ordering.ordinal);
+		old_value = bitcast(storage_old_value, $typeof(*ptr));
+		new_value = old_value / y;
+		storage_new_value = bitcast(new_value, $StorageType);
+	} while (mem::compare_exchange(storage_ptr, storage_old_value, storage_new_value, $ordering, $load_ordering) != storage_old_value);
+
+	return old_value;
+}
+
+/**
+ * @param [&in] ptr "the variable or dereferenced pointer to the data."
+ * @param [in] y "the value to be added to ptr."
+ * @param $ordering "atomic ordering of the load, defaults to SEQ_CONSISTENT"
+ * @return "returns the old value of ptr"
+ * 
+ * @require !$alignment || math::is_power_of_2($alignment) "Alignment must be a power of two."
+ * @require types::is_int($typeof(*ptr)) "Only integer pointers may be used."
+ * @require types::is_int($typeof(y)) "The value for or must be an int"
+ * @require $ordering != AtomicOrdering.NOT_ATOMIC && $ordering != AtomicOrdering.UNORDERED "Acquire ordering is not valid."
+ **/
+macro fetch_or(ptr, y, AtomicOrdering $ordering = SEQ_CONSISTENT, bool $volatile = false, usz $alignment = 0)
+{
+	$if	types::is_int($typeof(*ptr)):
+		return $$atomic_fetch_or(ptr, y, $volatile, $ordering.ordinal, $alignment);
+	$endif
+
+	var $load_ordering = $ordering;
+	$if $ordering == RELEASE || $ordering == ACQUIRE_RELEASE:
+	    $load_ordering = AtomicOrdering.SEQ_CONSISTENT;
+	$endif
+	
+	var $StorageType = $typefrom(types::lower_to_atomic_compatible_type($typeof(*ptr)));
+
+	$StorageType* storage_ptr = ($StorageType*)ptr;
+
+	$typeof(*ptr) old_value;
+	$typeof(*ptr) new_value;
+
+	$StorageType storage_old_value;
+	$StorageType storage_new_value;
+	$StorageType storage_y = ($StorageType)y;
+	
+	do {
+		storage_old_value = $$atomic_load(storage_ptr, false, $load_ordering.ordinal);
+		old_value = bitcast(storage_old_value, $typeof(*ptr));
+		new_value = storage_old_value | storage_y;
+		storage_new_value = bitcast(new_value, $StorageType);
+	} while (mem::compare_exchange(storage_ptr, storage_old_value, storage_new_value, $ordering, $load_ordering) != storage_old_value);
+
+	return old_value;
+}
+
+/**
+ * @param [&in] ptr "the variable or dereferenced pointer to the data."
+ * @param [in] y "the value to be added to ptr."
+ * @param $ordering "atomic ordering of the load, defaults to SEQ_CONSISTENT"
+ * @return "returns the old value of ptr"
+ * 
+ * @require !$alignment || math::is_power_of_2($alignment) "Alignment must be a power of two."
+ * @require types::is_int($typeof(*ptr)) "Only integer pointers may be used."
+ * @require types::is_int($typeof(y)) "The value for or must be an int"
+ * @require $ordering != AtomicOrdering.NOT_ATOMIC && $ordering != AtomicOrdering.UNORDERED "Acquire ordering is not valid."
+ **/
+macro fetch_xor(ptr, y, AtomicOrdering $ordering = SEQ_CONSISTENT, bool $volatile = false, usz $alignment = 0)
+{
+	$if	types::is_int($typeof(*ptr)):
+		return $$atomic_fetch_xor(ptr, y, $volatile, $ordering.ordinal, $alignment);
+	$endif
+
+	var $load_ordering = $ordering;
+	$if $ordering == RELEASE || $ordering == ACQUIRE_RELEASE:
+	    $load_ordering = AtomicOrdering.SEQ_CONSISTENT;
+	$endif
+	
+	var $StorageType = $typefrom(types::lower_to_atomic_compatible_type($typeof(*ptr)));
+
+	$StorageType* storage_ptr = ($StorageType*)ptr;
+
+	$typeof(*ptr) old_value;
+	$typeof(*ptr) new_value;
+
+	$StorageType storage_old_value;
+	$StorageType storage_new_value;
+	$StorageType storage_y = ($StorageType)y;
+	
+	do {
+		storage_old_value = $$atomic_load(storage_ptr, false, $load_ordering.ordinal);
+		old_value = bitcast(storage_old_value, $typeof(*ptr));
+		new_value = storage_old_value ^ storage_y;
+		storage_new_value = bitcast(new_value, $StorageType);
+	} while (mem::compare_exchange(storage_ptr, storage_old_value, storage_new_value, $ordering, $load_ordering) != storage_old_value);
+
+	return old_value;
+}
+
+/**
+ * @param [&in] ptr "the variable or dereferenced pointer to the data."
+ * @param [in] y "the value to be added to ptr."
+ * @param $ordering "atomic ordering of the load, defaults to SEQ_CONSISTENT"
+ * @return "returns the old value of ptr"
+ * 
+ * @require !$alignment || math::is_power_of_2($alignment) "Alignment must be a power of two."
+ * @require types::is_int($typeof(*ptr)) "Only integer pointers may be used."
+ * @require types::is_int($typeof(y)) "The value for or must be an int"
+ * @require $ordering != AtomicOrdering.NOT_ATOMIC && $ordering != AtomicOrdering.UNORDERED "Acquire ordering is not valid."
+ **/
+macro fetch_and(ptr, y, AtomicOrdering $ordering = SEQ_CONSISTENT, bool $volatile = false, usz $alignment = 0)
+{
+	$if	types::is_int($typeof(*ptr)):
+		return $$atomic_fetch_and(ptr, y, $volatile, $ordering.ordinal, $alignment);
+	$endif
+
+	var $load_ordering = $ordering;
+	$if $ordering == RELEASE || $ordering == ACQUIRE_RELEASE:
+	    $load_ordering = AtomicOrdering.SEQ_CONSISTENT;
+	$endif
+	
+	var $StorageType = $typefrom(types::lower_to_atomic_compatible_type($typeof(*ptr)));
+
+	$StorageType* storage_ptr = ($StorageType*)ptr;
+
+	$typeof(*ptr) old_value;
+	$typeof(*ptr) new_value;
+
+	$StorageType storage_old_value;
+	$StorageType storage_new_value;
+	$StorageType storage_y = ($StorageType)y;
+	
+	do {
+		storage_old_value = $$atomic_load(storage_ptr, false, $load_ordering.ordinal);
+		old_value = bitcast(storage_old_value, $typeof(*ptr));
+		new_value = storage_old_value & storage_y;
+		storage_new_value = bitcast(new_value, $StorageType);
+	} while (mem::compare_exchange(storage_ptr, storage_old_value, storage_new_value, $ordering, $load_ordering) != storage_old_value);
+
+	return old_value;
+}
+
+/**
+ * @param [&in] ptr "the variable or dereferenced pointer to the data."
+ * @param [in] y "the value to be added to ptr."
+ * @param $ordering "atomic ordering of the load, defaults to SEQ_CONSISTENT"
+ * @return "returns the old value of ptr"
+ * 
+ * @require types::is_int($typeof(*ptr)) "Only integer pointers may be used."
+ * @require types::is_int($typeof(y)) "The value for or must be an int"
+ * @require $ordering != AtomicOrdering.NOT_ATOMIC && $ordering != AtomicOrdering.UNORDERED "Acquire ordering is not valid."
+ **/
+macro fetch_shift_right(ptr, y, AtomicOrdering $ordering = SEQ_CONSISTENT)
+{
+	var $load_ordering = $ordering;
+	$if $ordering == RELEASE || $ordering == ACQUIRE_RELEASE:
+	    $load_ordering = AtomicOrdering.SEQ_CONSISTENT;
+	$endif
+	
+	var $StorageType = $typefrom(types::lower_to_atomic_compatible_type($typeof(*ptr)));
+
+	$StorageType* storage_ptr = ($StorageType*)ptr;
+
+	$typeof(*ptr) old_value;
+	$typeof(*ptr) new_value;
+
+	$StorageType storage_old_value;
+	$StorageType storage_new_value;
+	$StorageType storage_y = ($StorageType)y;
+	
+	do {
+		storage_old_value = $$atomic_load(storage_ptr, false, $load_ordering.ordinal);
+		old_value = bitcast(storage_old_value, $typeof(*ptr));
+		new_value = storage_old_value >> storage_y;
+		storage_new_value = bitcast(new_value, $StorageType);
+	} while (mem::compare_exchange(storage_ptr, storage_old_value, storage_new_value, $ordering, $load_ordering) != storage_old_value);
+
+	return old_value;
+}
+
+/**
+ * @param [&in] ptr "the variable or dereferenced pointer to the data."
+ * @param [in] y "the value to be added to ptr."
+ * @param $ordering "atomic ordering of the load, defaults to SEQ_CONSISTENT"
+ * @return "returns the old value of ptr"
+ * 
+ * @require types::is_int($typeof(*ptr)) "Only integer pointers may be used."
+ * @require types::is_int($typeof(y)) "The value for or must be an int"
+ * @require $ordering != AtomicOrdering.NOT_ATOMIC && $ordering != AtomicOrdering.UNORDERED "Acquire ordering is not valid."
+ **/
+macro fetch_shift_left(ptr, y, AtomicOrdering $ordering = SEQ_CONSISTENT)
+{
+	var $load_ordering = $ordering;
+	$if $ordering == RELEASE || $ordering == ACQUIRE_RELEASE:
+	    $load_ordering = AtomicOrdering.SEQ_CONSISTENT;
+	$endif
+	
+	var $StorageType = $typefrom(types::lower_to_atomic_compatible_type($typeof(*ptr)));
+
+	$StorageType* storage_ptr = ($StorageType*)ptr;
+
+	$typeof(*ptr) old_value;
+	$typeof(*ptr) new_value;
+
+	$StorageType storage_old_value;
+	$StorageType storage_new_value;
+	$StorageType storage_y = ($StorageType)y;
+	
+	do {
+		storage_old_value = $$atomic_load(storage_ptr, false, $load_ordering.ordinal);
+		old_value = bitcast(storage_old_value, $typeof(*ptr));
+		new_value = storage_old_value << storage_y;
+		storage_new_value = bitcast(new_value, $StorageType);
+	} while (mem::compare_exchange(storage_ptr, storage_old_value, storage_new_value, $ordering, $load_ordering) != storage_old_value);
+
+	return old_value;
+}
+
+/**
+ * @param [&in] ptr "the variable or dereferenced pointer to the data."
+ * @param $ordering "atomic ordering of the load, defaults to SEQ_CONSISTENT"
+ * @return "returns the old value of ptr"
+ * 
+ * @require types::is_int($typeof(*ptr)) "Only integer pointers may be used."
+ * @require $ordering != AtomicOrdering.NOT_ATOMIC && $ordering != AtomicOrdering.UNORDERED "Acquire ordering is not valid."
+ **/
+macro flag_set(ptr, AtomicOrdering $ordering = SEQ_CONSISTENT)
+{
+	$typeof(*ptr) old_value;
+	$typeof(*ptr) new_value = true;
+
+	do {
+		old_value = $$atomic_load(ptr, false, $ordering.ordinal);
+	} while (mem::compare_exchange(ptr, old_value, new_value, $ordering, $load_ordering) != old_value);
+	
+	return old_value;
+}
+
+/**
+ * @param [&in] ptr "the variable or dereferenced pointer to the data."
+ * @param $ordering "atomic ordering of the load, defaults to SEQ_CONSISTENT"
+ * @return "returns the old value of ptr"
+ * 
+ * @require types::is_int($typeof(*ptr)) "Only integer pointers may be used."
+ * @require $ordering != AtomicOrdering.NOT_ATOMIC && $ordering != AtomicOrdering.UNORDERED "Acquire ordering is not valid."
+ **/
+macro flag_clear(ptr, AtomicOrdering $ordering = SEQ_CONSISTENT)
+{
+	$typeof(*ptr) old_value;
+	$typeof(*ptr) new_value = false;
+
+	do {
+		old_value = $$atomic_load(ptr, false, $ordering.ordinal);
+	} while (mem::compare_exchange(ptr, old_value, new_value, $ordering, $load_ordering) != old_value);
+
+	return old_value;
+}
+
+/**
+ * @param [&in] ptr "the variable or dereferenced pointer to the data."
+ * @param [in] y "the value to be added to ptr."
+ * @param $ordering "atomic ordering of the load, defaults to SEQ_CONSISTENT"
+ * @return "returns the old value of ptr"
+ * 
+ * @require types::is_int($typeof(*ptr)) || types::is_float($typeof(*ptr)) "Only integer/float pointers may be used."
+ * @require $ordering != AtomicOrdering.NOT_ATOMIC && $ordering != AtomicOrdering.UNORDERED "Acquire ordering is not valid."
+ **/
+macro fetch_max(ptr, y, AtomicOrdering $ordering = SEQ_CONSISTENT, bool $volatile = false, usz $alignment = 0)
+{
+	$if $alignment == 0:
+		$alignment = $typeof(*ptr).sizeof;
+	$endif
+	return $$atomic_fetch_max(ptr, y, $volatile, $ordering.ordinal, $alignment);
+}
+
+/**
+ * @param [&in] ptr "the variable or dereferenced pointer to the data."
+ * @param [in] y "the value to be added to ptr."
+ * @param $ordering "atomic ordering of the load, defaults to SEQ_CONSISTENT"
+ * @return "returns the old value of ptr"
+ * 
+ * @require types::is_int($typeof(*ptr)) || types::is_float($typeof(*ptr)) "Only integer/float pointers may be used."
+ * @require $ordering != AtomicOrdering.NOT_ATOMIC && $ordering != AtomicOrdering.UNORDERED "Acquire ordering is not valid."
+ **/
+macro fetch_min(ptr, y, AtomicOrdering $ordering = SEQ_CONSISTENT, bool $volatile = false, usz $alignment = 0)
+{
+	$if $alignment == 0:
+		$alignment = $typeof(*ptr).sizeof;
+	$endif
+	return $$atomic_fetch_min(ptr, y, $volatile, $ordering.ordinal, $alignment);
+}

lib/std/atomic_nolibc.c3

@@ -0,0 +1,63 @@
+// Copyright (c) 2023 Eduardo José Gómez Hernández. All rights reserved.
+// Use of this source code is governed by the MIT license
+// a copy of which can be found in the LICENSE_STDLIB file.
+module std::atomic;
+
+macro @__atomic_compare_exchange_ordering_failure(ptr, expected, desired, $success, failure, $alignment) {
+	switch(failure)
+	{
+		case AtomicOrdering.RELAXED.ordinal: return $$compare_exchange(ptr, expected, desired, false, false, $success, AtomicOrdering.RELAXED.ordinal, $alignment);
+		case AtomicOrdering.ACQUIRE.ordinal: return $$compare_exchange(ptr, expected, desired, false, false, $success, AtomicOrdering.ACQUIRE.ordinal, $alignment);
+		case AtomicOrdering.SEQ_CONSISTENT.ordinal: return $$compare_exchange(ptr, expected, desired, false, false, $success, AtomicOrdering.SEQ_CONSISTENT.ordinal, $alignment);
+		default: unreachable("Unrecognized failure ordering");
+	}
+	return 0;
+}
+
+macro @__atomic_compare_exchange_ordering_success(ptr, expected, desired, success, failure, $alignment)
+{
+	switch(success)
+	{
+		case AtomicOrdering.RELAXED.ordinal: return @__atomic_compare_exchange_ordering_failure(ptr, expected, desired, AtomicOrdering.RELAXED.ordinal, failure, $alignment);
+		case AtomicOrdering.ACQUIRE.ordinal: return @__atomic_compare_exchange_ordering_failure(ptr, expected, desired, AtomicOrdering.ACQUIRE.ordinal, failure, $alignment);
+		case AtomicOrdering.RELEASE.ordinal: return @__atomic_compare_exchange_ordering_failure(ptr, expected, desired, AtomicOrdering.RELEASE.ordinal, failure, $alignment);
+		case AtomicOrdering.ACQUIRE_RELEASE.ordinal: return @__atomic_compare_exchange_ordering_failure(ptr, expected, desired, AtomicOrdering.ACQUIRE_RELEASE.ordinal, failure, $alignment);
+		case AtomicOrdering.SEQ_CONSISTENT.ordinal: return @__atomic_compare_exchange_ordering_failure(ptr, expected, desired, AtomicOrdering.SEQ_CONSISTENT.ordinal, failure, $alignment);
+		default: unreachable("Unrecognized success ordering");
+	}
+	return 0;
+}
+
+fn CInt __atomic_compare_exchange(CInt size, any ptr, any expected, any desired, CInt success, CInt failure) @extern("__atomic_compare_exchange") @export
+{
+	switch (size)
+	{
+		case 1:
+			char* pt = (char*)ptr;
+			char ex = *(char*)expected;
+			char de = *(char*)desired;
+			if (ex == @__atomic_compare_exchange_ordering_success(pt, ex, de, success, failure, 1)) return 1;
+		case 2:
+			short* pt = (short*)ptr;
+			short ex = *(short*)expected;
+			short de = *(short*)desired;
+			if (ex == @__atomic_compare_exchange_ordering_success(pt, ex, de, success, failure, 2)) return 1;
+		case 4:
+			int* pt = (int*)ptr;
+			int ex = *(int*)expected;
+			int de = *(int*)desired;
+			if (ex == @__atomic_compare_exchange_ordering_success(pt, ex, de, success, failure, 4)) return 1;
+		case 8:
+			$if iptr.sizeof >= 8:
+				long* pt = (long*)ptr;
+				long ex = *(long*)expected;
+				long de = *(long*)desired;
+				if (ex == @__atomic_compare_exchange_ordering_success(pt, ex, de, success, failure, 8)) return 1;
+			$else
+				nextcase;
+			$endif
+		default:
+			unreachable("Unsuported size (%d) for atomic_compare_exchange", size);
+	}
+	return 0;
+}

lib/std/bits.c3

@@ -0,0 +1,171 @@
+module std::bits;
+
+/**
+ * @require types::is_intlike($typeof(i)) `The input must be an integer or integer vector`
+ **/
+macro reverse(i) => $$bitreverse(i);
+
+/**
+ * @require types::is_intlike($typeof(i)) `The input must be an integer or integer vector`
+ **/
+macro bswap(i) @builtin => $$bswap(i);
+
+macro uint[<*>].popcount(self) => $$popcount(self);
+macro uint[<*>].ctz(self) => $$ctz(self);
+macro uint[<*>].clz(self) => $$clz(self);
+macro uint[<*>] uint[<*>].fshl(hi, uint[<*>] lo, uint[<*>] shift) => $$fshl(hi, lo, shift);
+macro uint[<*>] uint[<*>].fshr(hi, uint[<*>] lo, uint[<*>] shift) => $$fshr(hi, lo, shift);
+macro uint[<*>] uint[<*>].rotl(self, uint[<*>] shift) => $$fshl(self, self, shift);
+macro uint[<*>] uint[<*>].rotr(self, uint[<*>] shift) => $$fshr(self, self, shift);
+
+macro int[<*>].popcount(self) => $$popcount(self);
+macro int[<*>].ctz(self) => $$ctz(self);
+macro int[<*>].clz(self) => $$clz(self);
+macro int[<*>] int[<*>].fshl(hi, int[<*>] lo, int[<*>] shift) => $$fshl(hi, lo, shift);
+macro int[<*>] int[<*>].fshr(hi, int[<*>] lo, int[<*>] shift) => $$fshr(hi, lo, shift);
+macro int[<*>] int[<*>].rotl(self, int[<*>] shift) => $$fshl(self, self, shift);
+macro int[<*>] int[<*>].rotr(self, int[<*>] shift) => $$fshr(self, self, shift);
+
+macro ushort[<*>].popcount(self) => $$popcount(self);
+macro ushort[<*>].ctz(self) => $$ctz(self);
+macro ushort[<*>].clz(self) => $$clz(self);
+macro ushort[<*>] ushort[<*>].fshl(hi, ushort[<*>] lo, ushort[<*>] shift) => $$fshl(hi, lo, shift);
+macro ushort[<*>] ushort[<*>].fshr(hi, ushort[<*>] lo, ushort[<*>] shift) => $$fshr(hi, lo, shift);
+macro ushort[<*>] ushort[<*>].rotl(self, ushort[<*>] shift) => $$fshl(self, self, shift);
+macro ushort[<*>] ushort[<*>].rotr(self, ushort[<*>] shift) => $$fshr(self, self, shift);
+
+macro short[<*>].popcount(self) => $$popcount(self);
+macro short[<*>].ctz(self) => $$ctz(self);
+macro short[<*>].clz(self) => $$clz(self);
+macro short[<*>] short[<*>].fshl(hi, short[<*>] lo, short[<*>] shift) => $$fshl(hi, lo, shift);
+macro short[<*>] short[<*>].fshr(hi, short[<*>] lo, short[<*>] shift) => $$fshr(hi, lo, shift);
+macro short[<*>] short[<*>].rotl(self, short[<*>] shift) => $$fshl(self, self, shift);
+macro short[<*>] short[<*>].rotr(self, short[<*>] shift) => $$fshr(self, self, shift);
+
+macro char[<*>].popcount(self) => $$popcount(self);
+macro char[<*>].ctz(self) => $$ctz(self);
+macro char[<*>].clz(self) => $$clz(self);
+macro char[<*>] char[<*>].fshl(hi, char[<*>] lo, char[<*>] shift) => $$fshl(hi, lo, shift);
+macro char[<*>] char[<*>].fshr(hi, char[<*>] lo, char[<*>] shift) => $$fshr(hi, lo, shift);
+macro char[<*>] char[<*>].rotl(self, char[<*>] shift) => $$fshl(self, self, shift);
+macro char[<*>] char[<*>].rotr(self, char[<*>] shift) => $$fshr(self, self, shift);
+
+macro ichar[<*>].popcount(self) => $$popcount(self);
+macro ichar[<*>].ctz(self) => $$ctz(self);
+macro ichar[<*>].clz(self) => $$clz(self);
+macro ichar[<*>] ichar[<*>].fshl(hi, ichar[<*>] lo, ichar[<*>] shift) => $$fshl(hi, lo, shift);
+macro ichar[<*>] ichar[<*>].fshr(hi, ichar[<*>] lo, ichar[<*>] shift) => $$fshr(hi, lo, shift);
+macro ichar[<*>] ichar[<*>].rotl(self, ichar[<*>] shift) => $$fshl(self, self, shift);
+macro ichar[<*>] ichar[<*>].rotr(self, ichar[<*>] shift) => $$fshr(self, self, shift);
+
+macro ulong[<*>].popcount(self) => $$popcount(self);
+macro ulong[<*>].ctz(self) => $$ctz(self);
+macro ulong[<*>].clz(self) => $$clz(self);
+macro ulong[<*>] ulong[<*>].fshl(hi, ulong[<*>] lo, ulong[<*>] shift) => $$fshl(hi, lo, shift);
+macro ulong[<*>] ulong[<*>].fshr(hi, ulong[<*>] lo, ulong[<*>] shift) => $$fshr(hi, lo, shift);
+macro ulong[<*>] ulong[<*>].rotl(self, ulong[<*>] shift) => $$fshl(self, self, shift);
+macro ulong[<*>] ulong[<*>].rotr(self, ulong[<*>] shift) => $$fshr(self, self, shift);
+
+macro long[<*>].popcount(self) => $$popcount(self);
+macro long[<*>].ctz(self) => $$ctz(self);
+macro long[<*>].clz(self) => $$clz(self);
+macro long[<*>] long[<*>].fshl(hi, long[<*>] lo, long[<*>] shift) => $$fshl(hi, lo, shift);
+macro long[<*>] long[<*>].fshr(hi, long[<*>] lo, long[<*>] shift) => $$fshr(hi, lo, shift);
+macro long[<*>] long[<*>].rotl(self, long[<*>] shift) => $$fshl(self, self, shift);
+macro long[<*>] long[<*>].rotr(self, long[<*>] shift) => $$fshr(self, self, shift);
+
+macro uint128[<*>].popcount(self) => $$popcount(self);
+macro uint128[<*>].ctz(self) => $$ctz(self);
+macro uint128[<*>].clz(self) => $$clz(self);
+macro uint128[<*>] uint128[<*>].fshl(hi, uint128[<*>] lo, uint128[<*>] shift) => $$fshl(hi, lo, shift);
+macro uint128[<*>] uint128[<*>].fshr(hi, uint128[<*>] lo, uint128[<*>] shift) => $$fshr(hi, lo, shift);
+macro uint128[<*>] uint128[<*>].rotl(self, uint128[<*>] shift) => $$fshl(self, self, shift);
+macro uint128[<*>] uint128[<*>].rotr(self, uint128[<*>] shift) => $$fshr(self, self, shift);
+
+macro int128[<*>].popcount(self) => $$popcount(self);
+macro int128[<*>].ctz(self) => $$ctz(self);
+macro int128[<*>].clz(self) => $$clz(self);
+macro int128[<*>] int128[<*>].fshl(hi, int128[<*>] lo, int128[<*>] shift) => $$fshl(hi, lo, shift);
+macro int128[<*>] int128[<*>].fshr(hi, int128[<*>] lo, int128[<*>] shift) => $$fshr(hi, lo, shift);
+macro int128[<*>] int128[<*>].rotl(self, int128[<*>] shift) => $$fshl(self, self, shift);
+macro int128[<*>] int128[<*>].rotr(self, int128[<*>] shift) => $$fshr(self, self, shift);
+
+macro uint.popcount(self) => $$popcount(self);
+macro uint.ctz(self) => $$ctz(self);
+macro uint.clz(self) => $$clz(self);
+macro uint uint.fshl(hi, uint lo, uint shift) => $$fshl(hi, lo, shift);
+macro uint uint.fshr(hi, uint lo, uint shift) => $$fshr(hi, lo, shift);
+macro uint uint.rotl(self, uint shift) => $$fshl(self, self, shift);
+macro uint uint.rotr(self, uint shift) => $$fshr(self, self, shift);
+
+macro int.popcount(self) => $$popcount(self);
+macro int.ctz(self) => $$ctz(self);
+macro int.clz(self) => $$clz(self);
+macro int int.fshl(hi, int lo, int shift) => $$fshl(hi, lo, shift);
+macro int int.fshr(hi, int lo, int shift) => $$fshr(hi, lo, shift);
+macro int int.rotl(self, int shift) => $$fshl(self, self, shift);
+macro int int.rotr(self, int shift) => $$fshr(self, self, shift);
+
+macro ushort.popcount(self) => $$popcount(self);
+macro ushort.ctz(self) => $$ctz(self);
+macro ushort.clz(self) => $$clz(self);
+macro ushort ushort.fshl(hi, ushort lo, ushort shift) => $$fshl(hi, lo, shift);
+macro ushort ushort.fshr(hi, ushort lo, ushort shift) => $$fshr(hi, lo, shift);
+macro ushort ushort.rotl(self, ushort shift) => $$fshl(self, self, shift);
+macro ushort ushort.rotr(self, ushort shift) => $$fshr(self, self, shift);
+
+macro short.popcount(self) => $$popcount(self);
+macro short.ctz(self) => $$ctz(self);
+macro short.clz(self) => $$clz(self);
+macro short short.fshl(hi, short lo, short shift) => $$fshl(hi, lo, shift);
+macro short short.fshr(hi, short lo, short shift) => $$fshr(hi, lo, shift);
+macro short short.rotl(self, short shift) => $$fshl(self, self, shift);
+macro short short.rotr(self, short shift) => $$fshr(self, self, shift);
+
+macro char.popcount(self) => $$popcount(self);
+macro char.ctz(self) => $$ctz(self);
+macro char.clz(self) => $$clz(self);
+macro char char.fshl(hi, char lo, char shift) => $$fshl(hi, lo, shift);
+macro char char.fshr(hi, char lo, char shift) => $$fshr(hi, lo, shift);
+macro char char.rotl(self, char shift) => $$fshl(self, self, shift);
+macro char char.rotr(self, char shift) => $$fshr(self, self, shift);
+
+macro ichar.popcount(self) => $$popcount(self);
+macro ichar.ctz(self) => $$ctz(self);
+macro ichar.clz(self) => $$clz(self);
+macro ichar ichar.fshl(hi, ichar lo, ichar shift) => $$fshl(hi, lo, shift);
+macro ichar ichar.fshr(hi, ichar lo, ichar shift) => $$fshr(hi, lo, shift);
+macro ichar ichar.rotl(self, ichar shift) => $$fshl(self, self, shift);
+macro ichar ichar.rotr(self, ichar shift) => $$fshr(self, self, shift);
+
+macro ulong.popcount(self) => $$popcount(self);
+macro ulong.ctz(self) => $$ctz(self);
+macro ulong.clz(self) => $$clz(self);
+macro ulong ulong.fshl(hi, ulong lo, ulong shift) => $$fshl(hi, lo, shift);
+macro ulong ulong.fshr(hi, ulong lo, ulong shift) => $$fshr(hi, lo, shift);
+macro ulong ulong.rotl(self, ulong shift) => $$fshl(self, self, shift);
+macro ulong ulong.rotr(self, ulong shift) => $$fshr(self, self, shift);
+
+macro long.popcount(self) => $$popcount(self);
+macro long.ctz(self) => $$ctz(self);
+macro long.clz(self) => $$clz(self);
+macro long long.fshl(hi, long lo, long shift) => $$fshl(hi, lo, shift);
+macro long long.fshr(hi, long lo, long shift) => $$fshr(hi, lo, shift);
+macro long long.rotl(self, long shift) => $$fshl(self, self, shift);
+macro long long.rotr(self, long shift) => $$fshr(self, self, shift);
+
+macro uint128.popcount(self) => $$popcount(self);
+macro uint128.ctz(self) => $$ctz(self);
+macro uint128.clz(self) => $$clz(self);
+macro uint128 uint128.fshl(hi, uint128 lo, uint128 shift) => $$fshl(hi, lo, shift);
+macro uint128 uint128.fshr(hi, uint128 lo, uint128 shift) => $$fshr(hi, lo, shift);
+macro uint128 uint128.rotl(self, uint128 shift) => $$fshl(self, self, shift);
+macro uint128 uint128.rotr(self, uint128 shift) => $$fshr(self, self, shift);
+
+macro int128.popcount(self) => $$popcount(self);
+macro int128.ctz(self) => $$ctz(self);
+macro int128.clz(self) => $$clz(self);
+macro int128 int128.fshl(hi, int128 lo, int128 shift) => $$fshl(hi, lo, shift);
+macro int128 int128.fshr(hi, int128 lo, int128 shift) => $$fshr(hi, lo, shift);
+macro int128 int128.rotl(self, int128 shift) => $$fshl(self, self, shift);
+macro int128 int128.rotr(self, int128 shift) => $$fshr(self, self, shift);

lib/std/collections/anylist.c3

@@ -0,0 +1,458 @@
+// Copyright (c) 2024 Christoffer Lerno. All rights reserved.
+// Use of self source code is governed by the MIT license
+// a copy of which can be found in the LICENSE_STDLIB file.
+module std::collections::anylist;
+import std::io,std::math;
+
+def AnyPredicate = fn bool(any value);
+def AnyTest = fn bool(any type, any context);
+
+struct AnyList (Printable)
+{
+	usz size;
+	usz capacity;
+	Allocator allocator;
+	any* entries;
+}
+
+
+/**
+ * @param initial_capacity "The initial capacity to reserve"
+ * @param [&inout] allocator "The allocator to use, defaults to the heap allocator"
+ **/
+fn AnyList* AnyList.new_init(&self, usz initial_capacity = 16, Allocator allocator = allocator::heap())
+{
+	self.allocator = allocator;
+	self.size = 0;
+	if (initial_capacity > 0)
+	{
+		initial_capacity = math::next_power_of_2(initial_capacity);
+		self.entries = allocator::alloc_array(allocator, any, initial_capacity);
+	}
+	else
+	{
+		self.entries = null;
+	}
+	self.capacity = initial_capacity;
+	return self;
+}
+
+/**
+ * Initialize the list using the temp allocator.
+ *
+ * @param initial_capacity "The initial capacity to reserve"
+ **/
+fn AnyList* AnyList.temp_init(&self, usz initial_capacity = 16)
+{
+	return self.new_init(initial_capacity, allocator::temp()) @inline;
+}
+
+fn usz! AnyList.to_format(&self, Formatter* formatter) @dynamic
+{
+	switch (self.size)
+	{
+		case 0:
+			return formatter.print("[]")!;
+		case 1:
+			return formatter.printf("[%s]", self.entries[0])!;
+		default:
+			usz n = formatter.print("[")!;
+			foreach (i, element : self.entries[:self.size])
+			{
+				if (i != 0) formatter.print(", ")!;
+				n += formatter.printf("%s", element)!;
+			}
+			n += formatter.print("]")!;
+			return n;
+	}
+}
+
+fn String AnyList.to_new_string(&self, Allocator allocator = allocator::heap()) @dynamic
+{
+	return string::new_format("%s", *self, .allocator = allocator);
+}
+
+fn String AnyList.to_tstring(&self)
+{
+	return string::tformat("%s", *self);
+}
+
+/**
+ * Push an element on the list by cloning it.
+ **/
+macro void AnyList.push(&self, element)
+{
+	if (!self.allocator) self.allocator = allocator::heap();
+	self.append_internal(allocator::clone(self.allocator, element));
+}
+
+fn void AnyList.append_internal(&self, any element) @local
+{
+	self.ensure_capacity();
+	self.entries[self.size++] = element;
+}
+
+/**
+ * Free a retained element removed using *_retained.
+ **/
+fn void AnyList.free_element(&self, any element) @inline
+{
+	allocator::free(self.allocator, element.ptr);
+}
+
+/**
+ * Pop a value who's type is known. If the type is incorrect, this
+ * will still pop the element.
+ *
+ * @return! CastResult.TYPE_MISMATCH, IteratorResult.NO_MORE_ELEMENT
+ **/
+macro AnyList.pop(&self, $Type)
+{
+	if (!self.size) return IteratorResult.NO_MORE_ELEMENT?;
+	defer self.free_element(self.entries[self.size]);
+	return *anycast(self.entries[--self.size], $Type);
+}
+
+/**
+ * Pop the last value and allocate the copy using the given allocator.
+ * @return! IteratorResult.NO_MORE_ELEMENT
+ **/
+fn any! AnyList.new_pop(&self, Allocator allocator = allocator::heap())
+{
+	if (!self.size) return IteratorResult.NO_MORE_ELEMENT?;
+	defer self.free_element(self.entries[self.size]);
+	return allocator::clone_any(allocator, self.entries[--self.size]);
+}
+
+/**
+ * Pop the last value and allocate the copy using the temp allocator
+ * @return! IteratorResult.NO_MORE_ELEMENT
+ **/
+fn any! AnyList.temp_pop(&self) => self.new_pop(allocator::temp());
+
+/**
+ * Pop the last value. It must later be released using list.free_element()
+ * @return! IteratorResult.NO_MORE_ELEMENT
+ **/
+fn any! AnyList.pop_retained(&self)
+{
+	if (!self.size) return IteratorResult.NO_MORE_ELEMENT?;
+	return self.entries[--self.size];
+}
+
+fn void AnyList.clear(&self)
+{
+	for (usz i = 0; i < self.size; i++)
+	{
+		self.free_element(self.entries[i]);
+	}
+	self.size = 0;
+}
+
+/**
+ * Same as pop() but pops the first value instead.
+ **/
+macro AnyList.pop_first(&self, $Type)
+{
+	if (!self.size) return IteratorResult.NO_MORE_ELEMENT?;
+	defer self.remove_at(0);
+	return *anycast(self.entries[0], $Type);
+}
+
+/**
+ * Same as pop_retained() but pops the first value instead.
+ **/
+fn any! AnyList.pop_first_retained(&self)
+{
+	if (!self.size) return IteratorResult.NO_MORE_ELEMENT?;
+	defer self.remove_at(0);
+	return self.entries[0];
+}
+
+/**
+ * Same as new_pop() but pops the first value instead.
+ **/
+fn any! AnyList.new_pop_first(&self, Allocator allocator = allocator::heap())
+{
+	if (!self.size) return IteratorResult.NO_MORE_ELEMENT?;
+	defer self.free_element(self.entries[self.size]);
+	defer self.remove_at(0);
+	return allocator::clone_any(allocator, self.entries[0]);
+}
+
+/**
+ * Same as temp_pop() but pops the first value instead.
+ **/
+fn any! AnyList.temp_pop_first(&self) => self.new_pop_first(allocator::temp());
+
+/**
+ * @require index < self.size
+ **/
+fn void AnyList.remove_at(&self, usz index)
+{
+	if (!--self.size || index == self.size) return;
+	self.free_element(self.entries[index]);
+	self.entries[index .. self.size - 1] = self.entries[index + 1 .. self.size];
+}
+
+fn void AnyList.add_all(&self, AnyList* other_list)
+{
+	if (!other_list.size) return;
+	self.reserve(other_list.size);
+	foreach (value : other_list)
+	{
+		self.entries[self.size++] = allocator::clone_any(self.allocator, value);
+	}
+}
+
+/**
+ * Reverse the elements in a list.
+ **/
+fn void AnyList.reverse(&self)
+{
+	if (self.size < 2) return;
+	usz half = self.size / 2U;
+	usz end = self.size - 1;
+	for (usz i = 0; i < half; i++)
+	{
+		self.swap(i, end - i);
+	}
+}
+
+fn any[] AnyList.array_view(&self)
+{
+	return self.entries[:self.size];
+}
+
+/**
+ * Push an element to the front of the list.
+ **/
+macro void AnyList.push_front(&self, type)
+{
+	self.insert_at(0, type);
+}
+
+/**
+ * @require index < self.size
+ **/
+macro void AnyList.insert_at(&self, usz index, type) @local
+{
+	any value = allocator::copy(self.allocator, type);
+	self.insert_at_internal(self, index, value);
+}
+
+/**
+ * @require index < self.size
+ **/
+fn void AnyList.insert_at_internal(&self, usz index, any value) @local
+{
+	self.ensure_capacity();
+	for (usz i = self.size; i > index; i--)
+	{
+		self.entries[i] = self.entries[i - 1];
+	}
+	self.size++;
+	self.entries[index] = value;
+}
+
+
+/**
+ * @require self.size > 0
+ **/
+fn void AnyList.remove_last(&self)
+{
+	self.free_element(self.entries[--self.size]);
+}
+
+/**
+ * @require self.size > 0
+ **/
+fn void AnyList.remove_first(&self)
+{
+	self.remove_at(0);
+}
+
+macro AnyList.first(&self, $Type)
+{
+	return *anycast(self.first_any(), $Type);
+}
+
+fn any! AnyList.first_any(&self) @inline
+{
+	return self.size ? self.entries[0] : IteratorResult.NO_MORE_ELEMENT?;
+}
+
+macro AnyList.last(&self, $Type)
+{
+	return *anycast(self.last_any(), $Type);
+}
+
+fn any! AnyList.last_any(&self) @inline
+{
+	return self.size ? self.entries[self.size - 1] : IteratorResult.NO_MORE_ELEMENT?;
+}
+
+fn bool AnyList.is_empty(&self) @inline
+{
+	return !self.size;
+}
+
+fn usz AnyList.len(&self) @operator(len) @inline
+{
+	return self.size;
+}
+
+/**
+ * @require index < self.size "Index out of range"
+ **/
+macro AnyList.get(&self, usz index, $Type)
+{
+	return *anycast(self.entries[index], $Type);
+}
+
+/**
+ * @require index < self.size "Index out of range"
+ **/
+fn any AnyList.get_any(&self, usz index) @inline
+{
+	return self.entries[index];
+}
+
+fn void AnyList.free(&self)
+{
+	if (!self.allocator) return;
+	self.clear();
+	allocator::free(self.allocator, self.entries);
+	self.capacity = 0;
+	self.entries = null;
+}
+
+fn void AnyList.swap(&self, usz i, usz j)
+{
+	any temp = self.entries[i];
+	self.entries[i] = self.entries[j];
+	self.entries[j] = temp;
+}
+
+/**
+ * @param filter "The function to determine if it should be removed or not"
+ * @return "the number of deleted elements"
+ **/
+fn usz AnyList.remove_if(&self, AnyPredicate filter)
+{
+	return self._remove_if(filter, false);
+}
+
+/**
+ * @param selection "The function to determine if it should be kept or not"
+ * @return "the number of deleted elements"
+ **/
+fn usz AnyList.retain_if(&self, AnyPredicate selection)
+{
+	return self._remove_if(selection, true);
+}
+
+macro usz AnyList._remove_if(&self, AnyPredicate filter, bool $invert) @local
+{
+	usz size = self.size;
+	for (usz i = size, usz k = size; k > 0; k = i)
+	{
+		// Find last index of item to be deleted.
+		$if $invert:
+			while (i > 0 && !filter(&self.entries[i - 1])) i--;
+		$else
+			while (i > 0 && filter(&self.entries[i - 1])) i--;
+		$endif
+		// Remove the items from this index up to the one not to be deleted.
+		usz n = self.size - k;
+		for (usz j = i; j < k; j++) self.free_element(self.entries[j]);
+		self.entries[i:n] = self.entries[k:n];
+		self.size -= k - i;
+		// Find last index of item not to be deleted.
+		$if $invert:
+			while (i > 0 && filter(&self.entries[i - 1])) i--;
+		$else
+			while (i > 0 && !filter(&self.entries[i - 1])) i--;
+		$endif
+	}
+	return size - self.size;
+}
+
+fn usz AnyList.remove_using_test(&self, AnyTest filter, any context)
+{
+	return self._remove_using_test(filter, false, context);
+}
+
+fn usz AnyList.retain_using_test(&self, AnyTest filter, any context)
+{
+	return self._remove_using_test(filter, true, context);
+}
+
+macro usz AnyList._remove_using_test(&self, AnyTest filter, bool $invert, ctx) @local
+{
+	usz size = self.size;
+	for (usz i = size, usz k = size; k > 0; k = i)
+	{
+		// Find last index of item to be deleted.
+		$if $invert:
+			while (i > 0 && !filter(&self.entries[i - 1], ctx)) i--;
+		$else
+			while (i > 0 && filter(&self.entries[i - 1], ctx)) i--;
+		$endif
+		// Remove the items from this index up to the one not to be deleted.
+		usz n = self.size - k;
+		for (usz j = i; j < k; j++) self.free_element(self.entries[j]);
+		self.entries[i:n] = self.entries[k:n];
+		self.size -= k - i;
+		// Find last index of item not to be deleted.
+		$if $invert:
+			while (i > 0 && filter(&self.entries[i - 1], ctx)) i--;
+		$else
+			while (i > 0 && !filter(&self.entries[i - 1], ctx)) i--;
+		$endif
+	}
+	return size - self.size;
+}
+
+/**
+ * Reserve at least min_capacity
+ **/
+fn void AnyList.reserve(&self, usz min_capacity)
+{
+	if (!min_capacity) return;
+	if (self.capacity >= min_capacity) return;
+	if (!self.allocator) self.allocator = allocator::heap();
+	min_capacity = math::next_power_of_2(min_capacity);
+	self.entries = allocator::realloc(self.allocator, self.entries, any.sizeof * min_capacity);
+	self.capacity = min_capacity;
+}
+
+macro any AnyList.@item_at(&self, usz index) @operator([])
+{
+	return self.entries[index];
+}
+
+/**
+ * @require index <= self.size "Index out of range"
+ **/
+macro void AnyList.set(&self, usz index, value)
+{
+	if (index == self.size)
+	{
+		self.push(value);
+		return;
+	}
+	self.free_element(self.entries[index]);
+	self.entries[index] = allocator::copy(self.allocator, value);
+}
+
+fn void AnyList.ensure_capacity(&self, usz added = 1) @inline @private
+{
+	usz new_size = self.size + added;
+	if (self.capacity >= new_size) return;
+
+	assert(new_size < usz.max / 2U);
+	usz new_capacity = self.capacity ? 2U * self.capacity : 16U;
+	while (new_capacity < new_size) new_capacity *= 2U;
+	self.reserve(new_capacity);
+}

lib/std/collections/bitset.c3

@@ -0,0 +1,160 @@
+/**
+ * @require SIZE > 0
+ **/
+module std::collections::bitset(<SIZE>);
+
+def Type = uint;
+
+const BITS = Type.sizeof * 8;
+const SZ = (SIZE + BITS - 1) / BITS;
+
+struct BitSet
+{
+	Type[SZ] data;
+}
+
+fn usz BitSet.cardinality(&self)
+{
+	usz n;
+	foreach (x : self.data)
+	{
+		n += x.popcount();
+	}
+	return n;
+}
+
+/**
+ * @require i < SIZE
+ **/
+fn void BitSet.set(&self, usz i)
+{
+	usz q = i / BITS;
+	usz r = i % BITS;
+	self.data[q] |= 1 << r;
+}
+
+/**
+ * @require i < SIZE
+ **/
+fn void BitSet.unset(&self, usz i)
+{
+	usz q = i / BITS;
+	usz r = i % BITS;
+	self.data[q] &= ~(1 << r);
+}
+
+/**
+ * @require i < SIZE
+ **/
+fn bool BitSet.get(&self, usz i) @operator([]) @inline
+{
+	usz q = i / BITS;
+	usz r = i % BITS;
+	return self.data[q] & (1 << r) != 0;
+}
+
+fn usz BitSet.len(&self) @operator(len) @inline
+{
+	return SZ * BITS;
+}
+
+/**
+ * @require i < SIZE
+ **/
+fn void BitSet.set_bool(&self, usz i, bool value) @operator([]=) @inline
+{
+	if (value) return self.set(i);
+	self.unset(i);
+}
+
+/**
+ * @require Type.kindof == UNSIGNED_INT
+ **/
+module std::collections::growablebitset(<Type>);
+import std::collections::list;
+
+const BITS = Type.sizeof * 8;
+
+def GrowableBitSetList = List(<Type>);
+
+struct GrowableBitSet
+{
+	GrowableBitSetList data;
+}
+
+/**
+ * @param initial_capacity
+ * @param [&inout] allocator "The allocator to use, defaults to the heap allocator"
+ **/
+fn GrowableBitSet* GrowableBitSet.new_init(&self, usz initial_capacity = 1, Allocator allocator = allocator::heap())
+{
+	self.data.new_init(initial_capacity, allocator);
+	return self;
+}
+
+fn GrowableBitSet* GrowableBitSet.temp_init(&self, usz initial_capacity = 1)
+{
+	return self.new_init(initial_capacity, allocator::temp()) @inline;
+}
+
+fn void GrowableBitSet.free(&self)
+{
+	self.data.free();
+}
+
+fn usz GrowableBitSet.cardinality(&self)
+{
+	usz n;
+	foreach (x : self.data)
+	{
+		n += x.popcount();
+	}
+	return n;
+}
+
+fn void GrowableBitSet.set(&self, usz i)
+{
+	usz q = i / BITS;
+	usz r = i % BITS;
+	usz current_len = self.data.len();
+	if (q >= current_len)
+	{
+		usz n = q + 1;
+		self.data.reserve(n);
+		if (n - 1 >= current_len)
+		{
+			self.data.entries[current_len .. (n - 1)] = 0;
+		}
+		self.data.size = n;
+	}
+	self.data.set(q, self.data[q] | (1 << r));
+}
+
+fn void GrowableBitSet.unset(&self, usz i)
+{
+	usz q = i / BITS;
+	usz r = i % BITS;
+	if (q >= self.data.len()) return;
+	self.data.set(q, self.data[q] &~ (1 << r));
+}
+
+fn bool GrowableBitSet.get(&self, usz i) @operator([]) @inline
+{
+	usz q = i / BITS;
+	usz r = i % BITS;
+	if (q >= self.data.len()) return false;
+	return self.data[q] & (1 << r) != 0;
+}
+
+fn usz GrowableBitSet.len(&self) @operator(len)
+{
+	usz n = self.data.len() * BITS;
+	if (n > 0) n -= (usz)self.data[^1].clz();
+	return n;
+}
+
+fn void GrowableBitSet.set_bool(&self, usz i, bool value) @operator([]=) @inline
+{
+	if (value) return self.set(i);
+	self.unset(i);
+}

lib/std/collections/elastic_array.c3

@@ -0,0 +1,434 @@
+// Copyright (c) 2021-2024 Christoffer Lerno. All rights reserved.
+// Use of self source code is governed by the MIT license
+// a copy of which can be found in the LICENSE_STDLIB file.
+/**
+ * @require MAX_SIZE >= 1 `The size must be at least 1 element big.`
+ **/
+module std::collections::elastic_array(<Type, MAX_SIZE>);
+import std::io, std::math, std::collections::list_common;
+
+def ElementPredicate = fn bool(Type *type);
+def ElementTest = fn bool(Type *type, any context);
+const ELEMENT_IS_EQUATABLE = types::is_equatable_type(Type);
+const ELEMENT_IS_POINTER = Type.kindof == POINTER;
+macro type_is_overaligned() => Type.alignof > mem::DEFAULT_MEM_ALIGNMENT;
+
+struct ElasticArray (Printable)
+{
+	usz size;
+	Type[MAX_SIZE] entries;
+}
+
+fn usz! ElasticArray.to_format(&self, Formatter* formatter) @dynamic
+{
+	switch (self.size)
+	{
+		case 0:
+			return formatter.print("[]")!;
+		case 1:
+			return formatter.printf("[%s]", self.entries[0])!;
+		default:
+			usz n = formatter.print("[")!;
+			foreach (i, element : self.entries[:self.size])
+			{
+				if (i != 0) formatter.print(", ")!;
+				n += formatter.printf("%s", element)!;
+			}
+			n += formatter.print("]")!;
+			return n;
+	}
+}
+
+fn String ElasticArray.to_string(&self, Allocator allocator) @dynamic
+{
+	return string::new_format("%s", *self, .allocator = allocator);
+}
+
+fn String ElasticArray.to_tstring(&self)
+{
+	return string::tformat("%s", *self);
+}
+
+fn void! ElasticArray.push_try(&self, Type element) @inline
+{
+	if (self.size == MAX_SIZE) return AllocationFailure.OUT_OF_MEMORY?;
+	self.entries[self.size++] = element;
+}
+
+/**
+ * @require self.size < MAX_SIZE `Tried to exceed the max size`
+ **/
+fn void ElasticArray.push(&self, Type element) @inline
+{
+	self.entries[self.size++] = element;
+}
+
+fn Type! ElasticArray.pop(&self)
+{
+	if (!self.size) return IteratorResult.NO_MORE_ELEMENT?;
+	return self.entries[--self.size];
+}
+
+fn void ElasticArray.clear(&self)
+{
+	self.size = 0;
+}
+
+/**
+ * @require self.size > 0
+ **/
+fn Type! ElasticArray.pop_first(&self)
+{
+	if (!self.size) return IteratorResult.NO_MORE_ELEMENT?;
+	defer self.remove_at(0);
+	return self.entries[0];
+}
+
+/**
+ * @require index < self.size
+ **/
+fn void ElasticArray.remove_at(&self, usz index)
+{
+	if (!--self.size || index == self.size) return;
+	self.entries[index .. self.size - 1] = self.entries[index + 1 .. self.size];
+}
+
+/**
+ * @require other_list.size + self.size <= MAX_SIZE
+ **/
+fn void ElasticArray.add_all(&self, ElasticArray* other_list)
+{
+	if (!other_list.size) return;
+	foreach (&value : other_list)
+	{
+		self.entries[self.size++] = *value;
+	}
+}
+
+/**
+ * Add as many elements as possible to the new array,
+ * returning the number of elements that didn't fit.
+ **/
+fn usz ElasticArray.add_all_to_limit(&self, ElasticArray* other_list)
+{
+	if (!other_list.size) return 0;
+	foreach (i, &value : other_list)
+	{
+		if (self.size == MAX_SIZE) return other_list.size - i;
+		self.entries[self.size++] = *value;
+	}
+	return 0;
+}
+
+/**
+ * Add as many values from this array as possible, returning the
+ * number of elements that didn't fit.
+ *
+ * @param [in] array
+ **/
+fn usz ElasticArray.add_array_to_limit(&self, Type[] array)
+{
+	if (!array.len) return 0;
+	foreach (i, &value : array)
+	{
+		if (self.size == MAX_SIZE) return array.len - i;
+		self.entries[self.size++] = *value;
+	}
+	return 0;
+}
+
+/**
+ * Add the values of an array to this list.
+ *
+ * @param [in] array
+ * @require array.len + self.size <= MAX_SIZE `Size would exceed max.`
+ * @ensure self.size >= array.len
+ **/
+fn void ElasticArray.add_array(&self, Type[] array)
+{
+	if (!array.len) return;
+	foreach (&value : array)
+	{
+		self.entries[self.size++] = *value;
+	}
+}
+
+
+
+/**
+ * IMPORTANT The returned array must be freed using free_aligned.
+ **/
+fn Type[] ElasticArray.to_new_aligned_array(&self, Allocator allocator = allocator::heap())
+{
+	return list_common::list_to_new_aligned_array(Type, self, allocator);
+}
+
+/**
+ * @require !type_is_overaligned() : "This function is not available on overaligned types"
+ **/
+macro Type[] ElasticArray.to_new_array(&self, Allocator allocator = allocator::heap())
+{
+	return list_common::list_to_new_array(Type, self, allocator);
+}
+
+fn Type[] ElasticArray.to_tarray(&self)
+{
+	$if type_is_overaligned():
+	return self.to_new_aligned_array(allocator::temp());
+	$else
+	return self.to_new_array(allocator::temp());
+	$endif;
+}
+
+/**
+ * Reverse the elements in a list.
+ **/
+fn void ElasticArray.reverse(&self)
+{
+	list_common::list_reverse(self);
+}
+
+fn Type[] ElasticArray.array_view(&self)
+{
+	return self.entries[:self.size];
+}
+
+/**
+ * @require self.size < MAX_SIZE `List would exceed max size`
+ **/
+fn void ElasticArray.push_front(&self, Type type) @inline
+{
+	self.insert_at(0, type);
+}
+
+/**
+ * @require self.size < MAX_SIZE `List would exceed max size`
+ **/
+fn void! ElasticArray.push_front_try(&self, Type type) @inline
+{
+	return self.insert_at_try(0, type);
+}
+
+/**
+ * @require index <= self.size
+ **/
+fn void! ElasticArray.insert_at_try(&self, usz index, Type value)
+{
+	if (self.size == MAX_SIZE) return AllocationFailure.OUT_OF_MEMORY?;
+	self.insert_at(index, value);
+}
+
+/**
+ * @require self.size < MAX_SIZE `List would exceed max size`
+ * @require index <= self.size
+ **/
+fn void ElasticArray.insert_at(&self, usz index, Type type)
+{
+	for (usz i = self.size; i > index; i--)
+	{
+		self.entries[i] = self.entries[i - 1];
+	}
+	self.size++;
+	self.entries[index] = type;
+}
+
+/**
+ * @require index < self.size
+ **/
+fn void ElasticArray.set_at(&self, usz index, Type type)
+{
+	self.entries[index] = type;
+}
+
+fn void! ElasticArray.remove_last(&self) @maydiscard
+{
+	if (!self.size) return IteratorResult.NO_MORE_ELEMENT?;
+	self.size--;
+}
+
+fn void! ElasticArray.remove_first(&self) @maydiscard
+{
+	if (!self.size) return IteratorResult.NO_MORE_ELEMENT?;
+	self.remove_at(0);
+}
+
+fn Type! ElasticArray.first(&self)
+{
+	if (!self.size) return IteratorResult.NO_MORE_ELEMENT?;
+	return self.entries[0];
+}
+
+fn Type! ElasticArray.last(&self)
+{
+	if (!self.size) return IteratorResult.NO_MORE_ELEMENT?;
+	return self.entries[self.size - 1];
+}
+
+fn bool ElasticArray.is_empty(&self) @inline
+{
+	return !self.size;
+}
+
+fn usz ElasticArray.byte_size(&self) @inline
+{
+	return Type.sizeof * self.size;
+}
+
+fn usz ElasticArray.len(&self) @operator(len) @inline
+{
+	return self.size;
+}
+
+fn Type ElasticArray.get(&self, usz index) @inline
+{
+	return self.entries[index];
+}
+
+fn void ElasticArray.swap(&self, usz i, usz j)
+{
+	@swap(self.entries[i], self.entries[j]);
+}
+
+/**
+ * @param filter "The function to determine if it should be removed or not"
+ * @return "the number of deleted elements"
+ **/
+fn usz ElasticArray.remove_if(&self, ElementPredicate filter)
+{
+	return list_common::list_remove_if(self, filter, false);
+}
+
+/**
+ * @param selection "The function to determine if it should be kept or not"
+ * @return "the number of deleted elements"
+ **/
+fn usz ElasticArray.retain_if(&self, ElementPredicate selection)
+{
+	return list_common::list_remove_if(self, selection, true);
+}
+
+fn usz ElasticArray.remove_using_test(&self, ElementTest filter, any context)
+{
+	return list_common::list_remove_using_test(self, filter, false, context);
+}
+
+fn usz ElasticArray.retain_using_test(&self, ElementTest filter, any context)
+{
+	return list_common::list_remove_using_test(self, filter, true, context);
+}
+
+
+macro Type ElasticArray.@item_at(&self, usz index) @operator([])
+{
+	return self.entries[index];
+}
+
+fn Type* ElasticArray.get_ref(&self, usz index) @operator(&[]) @inline
+{
+	return &self.entries[index];
+}
+
+fn void ElasticArray.set(&self, usz index, Type value) @operator([]=)
+{
+	self.entries[index] = value;
+}
+
+
+// Functions for equatable types
+fn usz! ElasticArray.index_of(&self, Type type) @if(ELEMENT_IS_EQUATABLE)
+{
+	foreach (i, v : self)
+	{
+		if (equals(v, type)) return i;
+	}
+	return SearchResult.MISSING?;
+}
+
+fn usz! ElasticArray.rindex_of(&self, Type type) @if(ELEMENT_IS_EQUATABLE)
+{
+	foreach_r (i, v : self)
+	{
+		if (equals(v, type)) return i;
+	}
+	return SearchResult.MISSING?;
+}
+
+fn bool ElasticArray.equals(&self, ElasticArray other_list) @if(ELEMENT_IS_EQUATABLE)
+{
+	if (self.size != other_list.size) return false;
+	foreach (i, v : self)
+	{
+		if (!equals(v, other_list.entries[i])) return false;
+	}
+	return true;
+}
+
+/**
+ * Check for presence of a value in a list.
+ *
+ * @param [&in] self "the list to find elements in"
+ * @param value "The value to search for"
+ * @return "True if the value is found, false otherwise"
+ **/
+fn bool ElasticArray.contains(&self, Type value) @if(ELEMENT_IS_EQUATABLE)
+{
+	foreach (i, v : self)
+	{
+		if (equals(v, value)) return true;
+	}
+	return false;
+}
+
+/**
+ * @param [&inout] self "The list to remove elements from"
+ * @param value "The value to remove"
+ * @return "true if the value was found"
+ **/
+fn bool ElasticArray.remove_last_item(&self, Type value) @if(ELEMENT_IS_EQUATABLE)
+{
+	return @ok(self.remove_at(self.rindex_of(value)));
+}
+
+/**
+ * @param [&inout] self "The list to remove elements from"
+ * @param value "The value to remove"
+ * @return "true if the value was found"
+ **/
+fn bool ElasticArray.remove_first_item(&self, Type value) @if(ELEMENT_IS_EQUATABLE)
+{
+	return @ok(self.remove_at(self.index_of(value)));
+}
+
+/**
+ * @param [&inout] self "The list to remove elements from"
+ * @param value "The value to remove"
+ * @return "the number of deleted elements."
+ **/
+fn usz ElasticArray.remove_item(&self, Type value) @if(ELEMENT_IS_EQUATABLE)
+{
+	return list_common::list_remove_item(self, value);
+}
+
+
+
+fn void ElasticArray.remove_all_from(&self, ElasticArray* other_list) @if(ELEMENT_IS_EQUATABLE)
+{
+	if (!other_list.size) return;
+	foreach (v : other_list) self.remove_item(v);
+}
+
+/**
+ * @param [&in] self
+ * @return "The number non-null values in the list"
+ **/
+fn usz ElasticArray.compact_count(&self) @if(ELEMENT_IS_POINTER)
+{
+	usz vals = 0;
+	foreach (v : self) if (v) vals++;
+	return vals;
+}
+
+fn usz ElasticArray.compact(&self) @if(ELEMENT_IS_POINTER)
+{
+	return list_common::list_compact(self);
+}

lib/std/collections/enummap.c3

@@ -0,0 +1,63 @@
+module std::collections::enummap(<Enum, ValueType>);
+import std::io;
+struct EnumMap (Printable)
+{
+	ValueType[Enum.len] values;
+}
+
+fn void EnumMap.init(&self, ValueType init_value)
+{
+	foreach (&a : self.values)
+	{
+		*a = init_value;
+	}
+}
+
+fn usz! EnumMap.to_format(&self, Formatter* formatter) @dynamic
+{
+	usz n = formatter.print("{ ")!;
+	foreach (i, &value : self.values)
+	{
+		if (i != 0) formatter.print(", ")!;
+		n += formatter.printf("%s: %s", (Enum)i, *value)!;
+	}
+	n += formatter.print(" }")!;
+	return n;
+}
+
+fn String EnumMap.to_new_string(&self, Allocator allocator = allocator::heap()) @dynamic
+{
+	return string::new_format("%s", *self, .allocator = allocator);
+}
+
+fn String EnumMap.to_tstring(&self) @dynamic
+{
+	return string::tformat("%s", *self);
+}
+
+/**
+ * @return "The total size of this map, which is the same as the number of enum values"
+ * @pure
+ **/
+fn usz EnumMap.len(&self) @operator(len) @inline
+{
+	return self.values.len;
+}
+
+/**
+ * @return "Retrieve a value given the underlying enum, if there is no entry, then the zero value for the value is returned."
+ **/
+fn ValueType EnumMap.get(&self, Enum key) @operator([]) @inline
+{
+	return self.values[key.ordinal];
+}
+
+fn ValueType* EnumMap.get_ref(&self, Enum key) @operator(&[]) @inline
+{
+	return &self.values[key.ordinal];
+}
+
+fn void EnumMap.set(&self, Enum key, ValueType value) @operator([]=) @inline
+{
+	self.values[key.ordinal] = value;
+}

lib/std/collections/enumset.c3

@@ -0,0 +1,172 @@
+// Copyright (c) 2021-2024 Christoffer Lerno. All rights reserved.
+// Use of self source code is governed by the MIT license
+// a copy of which can be found in the LICENSE_STDLIB file.
+
+/**
+ * @require Enum.kindof == TypeKind.ENUM : "Only enums maybe be used with an enumset"
+ **/
+module std::collections::enumset(<Enum>);
+import std::io;
+
+def EnumSetType = $typefrom(private::type_for_enum_elements(Enum.elements)) @private;
+
+const IS_CHAR_ARRAY = Enum.elements > 128;
+distinct EnumSet (Printable) = EnumSetType;
+
+fn void EnumSet.add(&self, Enum v)
+{
+	$if IS_CHAR_ARRAY:
+		(*self)[(usz)v / 8] |= (char)(1u << ((usz)v % 8));
+	$else
+		*self = (EnumSet)((EnumSetType)*self | 1u << (EnumSetType)v);
+	$endif
+}
+
+fn void EnumSet.clear(&self)
+{
+	$if IS_CHAR_ARRAY:
+		*self = {};
+	$else
+		*self = 0;
+	$endif
+}
+
+fn bool EnumSet.remove(&self, Enum v)
+{
+	$if IS_CHAR_ARRAY:
+		if (!self.has(v) @inline) return false;
+		(*self)[(usz)v / 8] &= (char)~(1u << ((usz)v % 8));
+		return true;
+	$else
+		EnumSetType old = (EnumSetType)*self;
+		EnumSetType new = old & ~(1u << (EnumSetType)v);
+		*self = (EnumSet)new;
+		return old != new;
+	$endif
+}
+
+fn bool EnumSet.has(&self, Enum v)
+{
+	$if IS_CHAR_ARRAY:
+		return (bool)(((*self)[(usz)v / 8] << ((usz)v % 8)) & 0x01);
+	$else
+		return ((EnumSetType)*self & (1u << (EnumSetType)v)) != 0;
+	$endif
+}
+
+fn void EnumSet.add_all(&self, EnumSet s)
+{
+	$if IS_CHAR_ARRAY:
+		foreach (i, c : s) (*self)[i] |= c;
+	$else
+		*self = (EnumSet)((EnumSetType)*self | (EnumSetType)s);
+	$endif
+}
+
+fn void EnumSet.retain_all(&self, EnumSet s)
+{
+	$if IS_CHAR_ARRAY:
+		foreach (i, c : s) (*self)[i] &= c;
+	$else
+		*self = (EnumSet)((EnumSetType)*self & (EnumSetType)s);
+	$endif
+}
+
+fn void EnumSet.remove_all(&self, EnumSet s)
+{
+	$if IS_CHAR_ARRAY:
+		foreach (i, c : s) (*self)[i] &= ~c;
+	$else
+		*self = (EnumSet)((EnumSetType)*self & ~(EnumSetType)s);
+	$endif
+}
+
+fn EnumSet EnumSet.and_of(&self, EnumSet s)
+{
+	$if IS_CHAR_ARRAY:
+		EnumSet copy = *self;
+		copy.retain_all(s);
+		return copy;
+	$else
+		return (EnumSet)((EnumSetType)*self & (EnumSetType)s);
+	$endif
+}
+
+fn EnumSet EnumSet.or_of(&self, EnumSet s)
+{
+	$if IS_CHAR_ARRAY:
+		EnumSet copy = *self;
+		copy.add_all(s);
+		return copy;
+	$else
+		return (EnumSet)((EnumSetType)*self | (EnumSetType)s);
+	$endif
+}
+
+
+fn EnumSet EnumSet.diff_of(&self, EnumSet s)
+{
+	$if IS_CHAR_ARRAY:
+		EnumSet copy = *self;
+		copy.remove_all(s);
+		return copy;
+	$else
+		return (EnumSet)((EnumSetType)*self & ~(EnumSetType)s);
+	$endif
+}
+
+fn EnumSet EnumSet.xor_of(&self, EnumSet s)
+{
+	$if IS_CHAR_ARRAY:
+		EnumSet copy = *self;
+		foreach (i, c : s) copy[i] ^= c;
+		return copy;
+	$else
+		return (EnumSet)((EnumSetType)*self ^ (EnumSetType)s);
+	$endif
+}
+
+fn usz! EnumSet.to_format(&set, Formatter* formatter) @dynamic
+{
+	usz n = formatter.print("[")!;
+	bool found;
+	foreach (value : Enum.values)
+	{
+		if (!set.has(value)) continue;
+		if (found) n += formatter.print(", ")!;
+		found = true;
+		n += formatter.printf("%s", value)!;
+	}
+	n += formatter.print("]")!;
+	return n;
+}
+
+fn String EnumSet.to_new_string(&set, Allocator allocator = allocator::heap()) @dynamic
+{
+	return string::new_format("%s", *set, .allocator = allocator);
+}
+
+fn String EnumSet.to_tstring(&set) @dynamic
+{
+	return string::tformat("%s", *set);
+}
+
+module std::collections::enumset::private;
+
+macro typeid type_for_enum_elements(usz $elements)
+{
+	$switch
+		$case ($elements > 128):
+			return char[($elements + 7) / 8].typeid;
+		$case ($elements > 64):
+			return uint128.typeid;
+		$case ($elements > 32 || $$C_INT_SIZE > 32):
+			return ulong.typeid;
+		$case ($elements > 16 || $$C_INT_SIZE > 16):
+			return uint.typeid;
+		$case ($elements > 8 || $$C_INT_SIZE > 8):
+			return ushort.typeid;
+		$default:
+			return char.typeid;
+	$endswitch
+}

lib/std/collections/hashmap.c3

@@ -0,0 +1,400 @@
+// Copyright (c) 2023 Christoffer Lerno. All rights reserved.
+// Use of this source code is governed by the MIT license
+// a copy of which can be found in the LICENSE_STDLIB file.
+module std::collections::map(<Key, Value>);
+import std::math;
+
+struct HashMap
+{
+	Entry*[] table;
+	Allocator allocator;
+	uint count; // Number of elements
+	uint threshold; // Resize limit
+	float load_factor;
+}
+
+/**
+ * @param [&inout] allocator "The allocator to use"
+ * @require capacity > 0 "The capacity must be 1 or higher"
+ * @require load_factor > 0.0 "The load factor must be higher than 0"
+ * @require !self.allocator "Map was already initialized"
+ * @require capacity < MAXIMUM_CAPACITY "Capacity cannot exceed maximum"
+ **/
+fn HashMap* HashMap.new_init(&self, uint capacity = DEFAULT_INITIAL_CAPACITY, float load_factor = DEFAULT_LOAD_FACTOR, Allocator allocator = allocator::heap())
+{
+	capacity = math::next_power_of_2(capacity);
+	self.allocator = allocator;
+	self.load_factor = load_factor;
+	self.threshold = (uint)(capacity * load_factor);
+	self.table = allocator::new_array(allocator, Entry*, capacity);
+	return self;
+}
+
+/**
+ * @require capacity > 0 "The capacity must be 1 or higher"
+ * @require load_factor > 0.0 "The load factor must be higher than 0"
+ * @require !self.allocator "Map was already initialized"
+ * @require capacity < MAXIMUM_CAPACITY "Capacity cannot exceed maximum"
+ **/
+fn HashMap* HashMap.temp_init(&self, uint capacity = DEFAULT_INITIAL_CAPACITY, float load_factor = DEFAULT_LOAD_FACTOR)
+{
+	return self.new_init(capacity, load_factor, allocator::temp()) @inline;
+}
+
+
+/**
+ * Has this hash map been initialized yet?
+ *
+ * @param [&in] map "The hash map we are testing"
+ * @return "Returns true if it has been initialized, false otherwise"
+ **/
+fn bool HashMap.is_initialized(&map)
+{
+	return (bool)map.allocator;
+}
+
+/**
+ * @param [&inout] allocator "The allocator to use"
+ * @param [&in] other_map "The map to copy from."
+ **/
+fn HashMap* HashMap.new_init_from_map(&self, HashMap* other_map, Allocator allocator = allocator::heap())
+{
+	self.new_init(other_map.table.len, other_map.load_factor, allocator);
+	self.put_all_for_create(other_map);
+	return self;
+}
+
+/**
+ * @param [&in] other_map "The map to copy from."
+ **/
+fn HashMap* HashMap.temp_init_from_map(&map, HashMap* other_map)
+{
+	return map.new_init_from_map(other_map, allocator::temp()) @inline;
+}
+
+fn bool HashMap.is_empty(&map) @inline
+{
+	return !map.count;
+}
+
+fn usz HashMap.len(&map) @inline
+{
+	return map.count;
+}
+
+fn Value*! HashMap.get_ref(&map, Key key)
+{
+	if (!map.count) return SearchResult.MISSING?;
+	uint hash = rehash(key.hash());
+	for (Entry *e = map.table[index_for(hash, map.table.len)]; e != null; e = e.next)
+	{
+		if (e.hash == hash && equals(key, e.key)) return &e.value;
+	}
+	return SearchResult.MISSING?;
+}
+
+fn Entry*! HashMap.get_entry(&map, Key key)
+{
+	if (!map.count) return SearchResult.MISSING?;
+	uint hash = rehash(key.hash());
+	for (Entry *e = map.table[index_for(hash, map.table.len)]; e != null; e = e.next)
+	{
+		if (e.hash == hash && equals(key, e.key)) return e;
+	}
+	return SearchResult.MISSING?;
+}
+
+/**
+ * Get the value or update and
+ * @require $assignable(#expr, Value)
+ **/
+macro Value HashMap.@get_or_set(&map, Key key, Value #expr)
+{
+	if (!map.count)
+	{
+		Value val = #expr;
+		map.set(key, val);
+		return val;
+	}
+	uint hash = rehash(key.hash());
+	uint index = index_for(hash, map.table.len);
+	for (Entry *e = map.table[index]; e != null; e = e.next)
+	{
+		if (e.hash == hash && equals(key, e.key)) return e.value;
+	}
+	Value val = #expr;
+	map.add_entry(hash, key, val, index);
+	return val;
+}
+
+fn Value! HashMap.get(&map, Key key) @operator([])
+{
+	return *map.get_ref(key) @inline;
+}
+
+fn bool HashMap.has_key(&map, Key key)
+{
+	return @ok(map.get_ref(key));
+}
+
+fn bool HashMap.set(&map, Key key, Value value) @operator([]=)
+{
+	// If the map isn't initialized, use the defaults to initialize it.
+	if (!map.allocator)
+	{
+		map.new_init();
+	}
+	uint hash = rehash(key.hash());
+	uint index = index_for(hash, map.table.len);
+	for (Entry *e = map.table[index]; e != null; e = e.next)
+	{
+		if (e.hash == hash && equals(key, e.key))
+		{
+			e.value = value;
+			return true;
+		}
+	}
+	map.add_entry(hash, key, value, index);
+	return false;
+}
+
+fn void! HashMap.remove(&map, Key key) @maydiscard
+{
+	if (!map.remove_entry_for_key(key)) return SearchResult.MISSING?;
+}
+
+fn void HashMap.clear(&map)
+{
+	if (!map.count) return;
+	foreach (Entry** &entry_ref : map.table)
+	{
+		Entry* entry = *entry_ref;
+		if (!entry) continue;
+		Entry *next = entry.next;
+		while (next)
+		{
+			Entry *to_delete = next;
+			next = next.next;
+			map.free_entry(to_delete);
+		}
+		map.free_entry(entry);
+		*entry_ref = null;
+	}
+	map.count = 0;
+}
+
+fn void HashMap.free(&map)
+{
+	if (!map.allocator) return;
+	map.clear();
+	map.free_internal(map.table.ptr);
+	map.table = {};
+}
+
+fn Key[] HashMap.key_tlist(&map)
+{
+	return map.key_new_list(allocator::temp()) @inline;
+}
+
+fn Key[] HashMap.key_new_list(&map, Allocator allocator = allocator::heap())
+{
+	if (!map.count) return {};
+
+	Key[] list = allocator::alloc_array(allocator, Key, map.count);
+	usz index = 0;
+	foreach (Entry* entry : map.table)
+	{
+		while (entry)
+		{
+			list[index++] = entry.key;
+			entry = entry.next;
+		}
+	}
+	return list;
+}
+
+macro HashMap.@each(map; @body(key, value))
+{
+	map.@each_entry(; Entry* entry) {
+		@body(entry.key, entry.value);
+	};
+}
+
+macro HashMap.@each_entry(map; @body(entry))
+{
+	if (map.count)
+	{
+		foreach (Entry* entry : map.table)
+		{
+			while (entry)
+			{
+				@body(entry);
+				entry = entry.next;
+			}
+		}
+	}
+}
+
+fn Value[] HashMap.value_tlist(&map)
+{
+	return map.value_new_list(allocator::temp()) @inline;
+}
+
+fn Value[] HashMap.value_new_list(&map, Allocator allocator = allocator::heap())
+{
+	if (!map.count) return {};
+	Value[] list = allocator::alloc_array(allocator, Value, map.count);
+	usz index = 0;
+	foreach (Entry* entry : map.table)
+	{
+		while (entry)
+		{
+			list[index++] = entry.value;
+			entry = entry.next;
+		}
+	}
+	return list;
+}
+
+fn bool HashMap.has_value(&map, Value v) @if(VALUE_IS_EQUATABLE)
+{
+	if (!map.count) return false;
+	foreach (Entry* entry : map.table)
+	{
+		while (entry)
+		{
+			if (equals(v, entry.value)) return true;
+			entry = entry.next;
+		}
+	}
+	return false;
+}
+
+// --- private methods
+
+fn void HashMap.add_entry(&map, uint hash, Key key, Value value, uint bucket_index) @private
+{
+	$if COPY_KEYS:
+	key = key.copy(map.allocator);
+	$endif
+	Entry* entry = allocator::new(map.allocator, Entry, { .hash = hash, .key = key, .value = value, .next = map.table[bucket_index] });
+	map.table[bucket_index] = entry;
+	if (map.count++ >= map.threshold)
+	{
+		map.resize(map.table.len * 2);
+	}
+}
+
+fn void HashMap.resize(&map, uint new_capacity) @private
+{
+	Entry*[] old_table = map.table;
+	uint old_capacity = old_table.len;
+	if (old_capacity == MAXIMUM_CAPACITY)
+	{
+		map.threshold = uint.max;
+		return;
+	}
+	Entry*[] new_table = allocator::new_array(map.allocator, Entry*, new_capacity);
+	map.transfer(new_table);
+	map.table = new_table;
+	map.free_internal(old_table.ptr);
+	map.threshold = (uint)(new_capacity * map.load_factor);
+}
+
+fn void HashMap.transfer(&map, Entry*[] new_table) @private
+{
+	Entry*[] src = map.table;
+	uint new_capacity = new_table.len;
+	foreach (uint j, Entry *e : src)
+	{
+		if (!e) continue;
+		do
+		{
+			Entry* next = e.next;
+			uint i = index_for(e.hash, new_capacity);
+			e.next = new_table[i];
+			new_table[i] = e;
+			e = next;
+		}
+		while (e);
+	}
+}
+
+fn void HashMap.put_all_for_create(&map, HashMap* other_map) @private
+{
+	if (!other_map.count) return;
+	foreach (Entry *e : other_map.table)
+	{
+		if (!e) continue;
+		map.put_for_create(e.key, e.value);
+	}
+}
+
+fn void HashMap.put_for_create(&map, Key key, Value value) @private
+{
+	uint hash = rehash(key.hash());
+	uint i = index_for(hash, map.table.len);
+	for (Entry *e = map.table[i]; e != null; e = e.next)
+	{
+		if (e.hash == hash && equals(key, e.key))
+		{
+			e.value = value;
+			return;
+		}
+	}
+	map.create_entry(hash, key, value, i);
+}
+
+fn void HashMap.free_internal(&map, void* ptr) @inline @private
+{
+	allocator::free(map.allocator, ptr);
+}
+
+fn bool HashMap.remove_entry_for_key(&map, Key key) @private
+{
+	uint hash = rehash(key.hash());
+	uint i = index_for(hash, map.table.len);
+	Entry* prev = map.table[i];
+	Entry* e = prev;
+	while (e)
+	{
+		Entry *next = e.next;
+		if (e.hash == hash && equals(key, e.key))
+		{
+			map.count--;
+			if (prev == e)
+			{
+				map.table[i] = next;
+			}
+			else
+			{
+				prev.next = next;
+			}
+			map.free_entry(e);
+			return true;
+		}
+		prev = e;
+		e = next;
+	}
+	return false;
+}
+
+fn void HashMap.create_entry(&map, uint hash, Key key, Value value, int bucket_index) @private
+{
+	Entry *e = map.table[bucket_index];
+	$if COPY_KEYS:
+	key = key.copy(map.allocator);
+	$endif
+	Entry* entry = allocator::new(map.allocator, Entry, { .hash = hash, .key = key, .value = value, .next = map.table[bucket_index] });
+	map.table[bucket_index] = entry;
+	map.count++;
+}
+
+fn void HashMap.free_entry(&self, Entry *entry) @local
+{
+	$if COPY_KEYS:
+	allocator::free(self.allocator, entry.key);
+	$endif
+	self.free_internal(entry);
+}
+

lib/std/collections/linkedlist.c3

@@ -0,0 +1,330 @@
+// Copyright (c) 2021-2024 Christoffer Lerno. All rights reserved.
+// Use of self source code is governed by the MIT license
+// a copy of which can be found in the LICENSE_STDLIB file.
+module std::collections::linkedlist(<Type>);
+
+const ELEMENT_IS_EQUATABLE = types::is_equatable_type(Type);
+
+struct Node @private
+{
+	Node *next;
+	Node *prev;
+	Type value;
+}
+
+struct LinkedList
+{
+	Allocator allocator;
+	usz size;
+	Node *_first;
+	Node *_last;
+}
+
+/**
+ * @param [&inout] allocator "The allocator to use, defaults to the heap allocator"
+ * @return "the initialized list"
+ **/
+fn LinkedList* LinkedList.new_init(&self, Allocator allocator = allocator::heap())
+{
+	*self = { .allocator = allocator };
+	return self;
+}
+
+
+fn LinkedList* LinkedList.temp_init(&self)
+{
+	return self.new_init(allocator::temp()) @inline;
+}
+
+/**
+ * @require self.allocator
+ **/
+macro void LinkedList.free_node(&self, Node* node) @private
+{
+	allocator::free(self.allocator, node);
+}
+
+macro Node* LinkedList.alloc_node(&self) @private
+{
+	if (!self.allocator) self.allocator = allocator::heap();
+	return allocator::alloc(self.allocator, Node);
+}
+
+fn void LinkedList.push_front(&self, Type value)
+{
+	Node *first = self._first;
+	Node *new_node = self.alloc_node();
+	*new_node = { .next = first, .value = value };
+	self._first = new_node;
+	if (!first)
+	{
+		self._last = new_node;
+	}
+	else
+	{
+		first.prev = new_node;
+	}
+	self.size++;
+}
+
+fn void LinkedList.push(&self, Type value)
+{
+	Node *last = self._last;
+	Node *new_node = self.alloc_node();
+	*new_node = { .prev = last, .value = value };
+	self._last = new_node;
+	if (!last)
+	{
+		self._first = new_node;
+	}
+	else
+	{
+		last.next = new_node;
+	}
+	self.size++;
+}
+
+fn Type! LinkedList.peek(&self) => self.first() @inline;
+fn Type! LinkedList.peek_last(&self) => self.last() @inline;
+
+fn Type! LinkedList.first(&self)
+{
+	if (!self._first) return IteratorResult.NO_MORE_ELEMENT?;
+	return self._first.value;
+}
+
+fn Type! LinkedList.last(&self)
+{
+	if (!self._last) return IteratorResult.NO_MORE_ELEMENT?;
+	return self._last.value;
+}
+
+fn void LinkedList.free(&self) => self.clear() @inline;
+
+fn void LinkedList.clear(&self)
+{
+	for (Node* node = self._first; node != null;)
+	{
+		Node* next = node.next;
+		self.free_node(node);
+		node = next;
+	}
+	self._first = null;
+	self._last = null;
+	self.size = 0;
+}
+
+fn usz LinkedList.len(&self) @inline => self.size;
+
+/**
+ * @require index < self.size
+ **/
+macro Node* LinkedList.node_at_index(&self, usz index)
+{
+	if (index * 2 >= self.size)
+	{
+		Node* node = self._last;
+		index = self.size - index - 1;
+		while (index--) node = node.prev;
+		return node;
+	}
+	Node* node = self._first;
+	while (index--) node = node.next;
+	return node;
+}
+/**
+ * @require index < self.size
+ **/
+fn Type LinkedList.get(&self, usz index)
+{
+	return self.node_at_index(index).value;
+}
+
+/**
+ * @require index < self.size
+ **/
+fn void LinkedList.set(&self, usz index, Type element)
+{
+	self.node_at_index(index).value = element;
+}
+
+/**
+ * @require index < self.size
+ **/
+fn void LinkedList.remove_at(&self, usz index)
+{
+	self.unlink(self.node_at_index(index));
+}
+
+/**
+ * @require index <= self.size
+ **/
+fn void LinkedList.insert_at(&self, usz index, Type element)
+{
+	switch (index)
+	{
+		case 0:
+			self.push_front(element);
+		case self.size:
+			self.push(element);
+		default:
+			self.link_before(self.node_at_index(index), element);
+	}
+}
+/**
+ * @require succ != null
+ **/
+fn void LinkedList.link_before(&self, Node *succ, Type value) @private
+{
+	Node* pred = succ.prev;
+	Node* new_node = self.alloc_node();
+	*new_node = { .prev = pred, .next = succ, .value = value };
+	succ.prev = new_node;
+	if (!pred)
+	{
+		self._first = new_node;
+	}
+	else
+	{
+		pred.next = new_node;
+	}
+	self.size++;
+}
+
+/**
+ * @require self._first
+ **/
+fn void LinkedList.unlink_first(&self) @private
+{
+	Node* f = self._first;
+	Node* next = f.next;
+	self.free_node(f);
+	self._first = next;
+	if (!next)
+	{
+		self._last = null;
+	}
+	else
+	{
+		next.prev = null;
+	}
+	self.size--;
+}
+
+fn usz LinkedList.remove(&self, Type t) @if(ELEMENT_IS_EQUATABLE)
+{
+	usz start = self.size;
+	Node* node = self._first;
+	while (node)
+	{
+		switch
+		{
+			case equals(node.value, t):
+				Node* next = node.next;
+				self.unlink(node);
+				node = next;
+			default:
+				node = node.next;
+		}
+	}
+	return start - self.size;
+}
+
+fn Type! LinkedList.pop(&self)
+{
+	if (!self._last) return IteratorResult.NO_MORE_ELEMENT?;
+	defer self.unlink_last();
+	return self._last.value;
+}
+
+fn Type! LinkedList.pop_front(&self)
+{
+	if (!self._first) return IteratorResult.NO_MORE_ELEMENT?;
+	defer self.unlink_first();
+	return self._first.value;
+}
+
+fn void! LinkedList.remove_last(&self) @maydiscard
+{
+	if (!self._first) return IteratorResult.NO_MORE_ELEMENT?;
+	self.unlink_last();
+}
+
+fn void! LinkedList.remove_first(&self) @maydiscard
+{
+	if (!self._first) return IteratorResult.NO_MORE_ELEMENT?;
+	self.unlink_first();
+}
+
+
+fn bool LinkedList.remove_first_match(&self, Type t) @if(ELEMENT_IS_EQUATABLE)
+{
+	for (Node* node = self._first; node != null; node = node.next)
+	{
+		if (node.value == t)
+		{
+			self.unlink(node);
+			return true;
+		}
+	}
+	return false;
+}
+
+fn bool LinkedList.remove_last_match(&self, Type t)  @if(ELEMENT_IS_EQUATABLE)
+{
+	for (Node* node = self._last; node != null; node = node.prev)
+	{
+		if (node.value == t)
+		{
+			self.unlink(node);
+			return true;
+		}
+	}
+	return false;
+}
+/**
+ * @require self._last
+ **/
+fn void LinkedList.unlink_last(&self) @inline @private
+{
+	Node* l = self._last;
+	Node* prev = l.prev;
+	self._last = prev;
+	self.free_node(l);
+	if (!prev)
+	{
+		self._first = null;
+	}
+	else
+	{
+		prev.next = null;
+	}
+	self.size--;
+}
+
+/**
+ * @require x != null
+ **/
+fn void LinkedList.unlink(&self, Node* x) @private
+{
+	Node* next = x.next;
+	Node* prev = x.prev;
+	if (!prev)
+	{
+		self._first = next;
+	}
+	else
+	{
+		prev.next = next;
+	}
+	if (!next)
+	{
+		self._last = prev;
+	}
+	else
+	{
+		next.prev = prev;
+	}
+	self.free_node(x);
+	self.size--;
+}

lib/std/collections/list.c3

@@ -0,0 +1,566 @@
+// Copyright (c) 2021-2024 Christoffer Lerno. All rights reserved.
+// Use of self source code is governed by the MIT license
+// a copy of which can be found in the LICENSE_STDLIB file.
+module std::collections::list(<Type>);
+import std::io, std::math, std::collections::list_common;
+
+def ElementPredicate = fn bool(Type *type);
+def ElementTest = fn bool(Type *type, any context);
+const ELEMENT_IS_EQUATABLE = types::is_equatable_type(Type);
+const ELEMENT_IS_POINTER = Type.kindof == POINTER;
+
+macro type_is_overaligned() => Type.alignof > mem::DEFAULT_MEM_ALIGNMENT;
+
+struct List (Printable)
+{
+	usz size;
+	usz capacity;
+	Allocator allocator;
+	Type *entries;
+}
+
+/**
+ * @param initial_capacity "The initial capacity to reserve"
+ * @param [&inout] allocator "The allocator to use, defaults to the heap allocator"
+ **/
+fn List* List.new_init(&self, usz initial_capacity = 16, Allocator allocator = allocator::heap())
+{
+	self.allocator = allocator;
+	self.size = 0;
+	self.capacity = 0;
+	self.entries = null;
+	self.reserve(initial_capacity);
+	return self;
+}
+
+/**
+ * Initialize the list using the temp allocator.
+ *
+ * @param initial_capacity "The initial capacity to reserve"
+ **/
+fn List* List.temp_init(&self, usz initial_capacity = 16)
+{
+	return self.new_init(initial_capacity, allocator::temp()) @inline;
+}
+
+/**
+ * Initialize a new list with an array.
+ *
+ * @param [in] values `The values to initialize the list with.`
+ * @require self.size == 0 "The List must be empty"
+ **/
+fn List* List.new_init_with_array(&self, Type[] values, Allocator allocator = allocator::heap())
+{
+	self.new_init(values.len, allocator) @inline;
+	self.add_array(values) @inline;
+	return self;
+}
+
+/**
+ * Initialize a temporary list with an array.
+ *
+ * @param [in] values `The values to initialize the list with.`
+ * @require self.size == 0 "The List must be empty"
+ **/
+fn List* List.temp_init_with_array(&self, Type[] values)
+{
+	self.temp_init(values.len) @inline;
+	self.add_array(values) @inline;
+	return self;
+}
+
+/**
+ * @require self.capacity == 0 "The List must not be allocated"
+ **/
+fn void List.init_wrapping_array(&self, Type[] types, Allocator allocator = allocator::heap())
+{
+	self.allocator = allocator;
+	self.capacity = types.len;
+	self.entries = types.ptr;
+	self.set_size(types.len);
+}
+
+fn usz! List.to_format(&self, Formatter* formatter) @dynamic
+{
+	switch (self.size)
+	{
+		case 0:
+			return formatter.print("[]")!;
+		case 1:
+			return formatter.printf("[%s]", self.entries[0])!;
+		default:
+			usz n = formatter.print("[")!;
+			foreach (i, element : self.entries[:self.size])
+			{
+				if (i != 0) formatter.print(", ")!;
+				n += formatter.printf("%s", element)!;
+			}
+			n += formatter.print("]")!;
+			return n;
+	}
+}
+
+fn String List.to_new_string(&self, Allocator allocator = allocator::heap()) @dynamic
+{
+	return string::new_format("%s", *self, .allocator = allocator);
+}
+
+fn String List.to_tstring(&self)
+{
+	return string::tformat("%s", *self);
+}
+
+fn void List.push(&self, Type element) @inline
+{
+	self.reserve(1);
+	self.entries[self.set_size(self.size + 1)] = element;
+}
+
+fn Type! List.pop(&self)
+{
+	if (!self.size) return IteratorResult.NO_MORE_ELEMENT?;
+	defer self.set_size(self.size - 1);
+	return self.entries[self.size - 1];
+}
+
+fn void List.clear(&self)
+{
+	self.set_size(0);
+}
+
+/**
+ * @require self.size > 0
+ **/
+fn Type! List.pop_first(&self)
+{
+	if (!self.size) return IteratorResult.NO_MORE_ELEMENT?;
+	defer self.remove_at(0);
+	return self.entries[0];
+}
+
+/**
+ * @require index < self.size
+ **/
+fn void List.remove_at(&self, usz index)
+{
+	self.set_size(self.size - 1);
+	if (!self.size || index == self.size) return;
+	self.entries[index .. self.size - 1] = self.entries[index + 1 .. self.size];
+}
+
+fn void List.add_all(&self, List* other_list)
+{
+	if (!other_list.size) return;
+	self.reserve(other_list.size);
+	usz index = self.set_size(self.size + other_list.size);
+	foreach (&value : other_list)
+	{
+		self.entries[index++] = *value;
+	}
+}
+
+
+/**
+ * IMPORTANT The returned array must be freed using free_aligned.
+ **/
+fn Type[] List.to_new_aligned_array(&self, Allocator allocator = allocator::heap())
+{
+	return list_common::list_to_new_aligned_array(Type, self, allocator);
+}
+
+/**
+ * @require !type_is_overaligned() : "This function is not available on overaligned types"
+ **/
+macro Type[] List.to_new_array(&self, Allocator allocator = allocator::heap())
+{
+	return list_common::list_to_new_array(Type, self, allocator);
+}
+
+fn Type[] List.to_tarray(&self)
+{
+	$if type_is_overaligned():
+	return self.to_new_aligned_array(allocator::temp());
+	$else
+	return self.to_new_array(allocator::temp());
+	$endif;
+}
+
+/**
+ * Reverse the elements in a list.
+ **/
+fn void List.reverse(&self)
+{
+	list_common::list_reverse(self);
+}
+
+fn Type[] List.array_view(&self)
+{
+	return self.entries[:self.size];
+}
+
+/**
+ * Add the values of an array to this list.
+ *
+ * @param [in] array
+ * @ensure self.size >= array.len
+ **/
+fn void List.add_array(&self, Type[] array)
+{
+	if (!array.len) return;
+	self.reserve(array.len);
+	usz index = self.set_size(self.size + array.len);
+	self.entries[index : array.len] = array[..];
+}
+
+fn void List.push_front(&self, Type type) @inline
+{
+	self.insert_at(0, type);
+}
+
+/**
+ * @require index <= self.size
+ **/
+fn void List.insert_at(&self, usz index, Type type)
+{
+	self.reserve(1);
+	for (usz i = self.size; i > index; i--)
+	{
+		self.entries[i] = self.entries[i - 1];
+	}
+	self.set_size(self.size + 1);
+	self.entries[index] = type;
+}
+
+/**
+ * @require index < self.size
+ **/
+fn void List.set_at(&self, usz index, Type type)
+{
+	self.entries[index] = type;
+}
+
+fn void! List.remove_last(&self) @maydiscard
+{
+	if (!self.size) return IteratorResult.NO_MORE_ELEMENT?;
+	self.set_size(self.size - 1);
+}
+
+fn void! List.remove_first(&self) @maydiscard
+{
+	if (!self.size) return IteratorResult.NO_MORE_ELEMENT?;
+	self.remove_at(0);
+}
+
+fn Type! List.first(&self)
+{
+	if (!self.size) return IteratorResult.NO_MORE_ELEMENT?;
+	return self.entries[0];
+}
+
+fn Type! List.last(&self)
+{
+	if (!self.size) return IteratorResult.NO_MORE_ELEMENT?;
+	return self.entries[self.size - 1];
+}
+
+fn bool List.is_empty(&self) @inline
+{
+	return !self.size;
+}
+
+fn usz List.byte_size(&self) @inline
+{
+	return Type.sizeof * self.size;
+}
+
+fn usz List.len(&self) @operator(len) @inline
+{
+	return self.size;
+}
+
+fn Type List.get(&self, usz index) @inline
+{
+	return self.entries[index];
+}
+
+fn void List.free(&self)
+{
+	if (!self.allocator || !self.capacity) return;
+
+	self.pre_free(); // Remove sanitizer annotation
+
+	$if type_is_overaligned():
+		allocator::free_aligned(self.allocator, self.entries);
+	$else
+		allocator::free(self.allocator, self.entries);
+	$endif;
+	self.capacity = 0;
+	self.size = 0;
+	self.entries = null;
+}
+
+fn void List.swap(&self, usz i, usz j)
+{
+	@swap(self.entries[i], self.entries[j]);
+}
+
+/**
+ * @param filter "The function to determine if it should be removed or not"
+ * @return "the number of deleted elements"
+ **/
+fn usz List.remove_if(&self, ElementPredicate filter)
+{
+	return list_common::list_remove_if(self, filter, false);
+}
+
+/**
+ * @param selection "The function to determine if it should be kept or not"
+ * @return "the number of deleted elements"
+ **/
+fn usz List.retain_if(&self, ElementPredicate selection)
+{
+	return list_common::list_remove_if(self, selection, true);
+}
+
+fn usz List.remove_using_test(&self, ElementTest filter, any context)
+{
+	usz old_size = self.size;
+	defer {
+		if (old_size != self.size) self._update_size_change(old_size, self.size);
+	}
+	return list_common::list_remove_using_test(self, filter, false, context);
+}
+
+
+
+fn usz List.retain_using_test(&self, ElementTest filter, any context)
+{
+	usz old_size = self.size;
+	defer {
+		if (old_size != self.size) self._update_size_change(old_size, self.size);
+	}
+	return list_common::list_remove_using_test(self, filter, true, context);
+}
+
+fn void List.ensure_capacity(&self, usz min_capacity) @local
+{
+	if (!min_capacity) return;
+	if (self.capacity >= min_capacity) return;
+	if (!self.allocator) self.allocator = allocator::heap();
+
+	self.pre_free(); // Remove sanitizer annotation
+
+	min_capacity = math::next_power_of_2(min_capacity);
+	$if type_is_overaligned():
+		self.entries = allocator::realloc_aligned(self.allocator, self.entries, Type.sizeof * min_capacity, .alignment = Type[1].alignof)!!;
+	$else
+		self.entries = allocator::realloc(self.allocator, self.entries, Type.sizeof * min_capacity);
+	$endif;
+	self.capacity = min_capacity;
+
+	self.post_alloc(); // Add sanitizer annotation
+}
+
+macro Type List.@item_at(&self, usz index) @operator([])
+{
+	return self.entries[index];
+}
+
+fn Type* List.get_ref(&self, usz index) @operator(&[]) @inline
+{
+	return &self.entries[index];
+}
+
+fn void List.set(&self, usz index, Type value) @operator([]=)
+{
+	self.entries[index] = value;
+}
+
+fn void List.reserve(&self, usz added)
+{
+	usz new_size = self.size + added;
+	if (self.capacity >= new_size) return;
+
+	assert(new_size < usz.max / 2U);
+	usz new_capacity = self.capacity ? 2U * self.capacity : 16U;
+	while (new_capacity < new_size) new_capacity *= 2U;
+	self.ensure_capacity(new_capacity);
+}
+
+fn void List._update_size_change(&self,usz old_size, usz new_size)
+{
+	if (old_size == new_size) return;
+	sanitizer::annotate_contiguous_container(self.entries,
+	                                         &self.entries[self.capacity],
+	                                         &self.entries[old_size],
+	                                         &self.entries[new_size]);
+}
+/**
+ * @require new_size == 0 || self.capacity != 0
+ **/
+fn usz List.set_size(&self, usz new_size) @inline @private
+{
+	usz old_size = self.size;
+	self._update_size_change(old_size, new_size);
+	self.size = new_size;
+	return old_size;
+}
+
+macro void List.pre_free(&self) @private
+{
+	if (!self.capacity) return;
+	self._update_size_change(self.size, self.capacity);
+}
+
+/**
+ * @require self.capacity
+ **/
+macro void List.post_alloc(&self) @private
+{
+	self._update_size_change(self.capacity, self.size);
+}
+
+// Functions for equatable types
+
+
+fn usz! List.index_of(&self, Type type) @if(ELEMENT_IS_EQUATABLE)
+{
+	foreach (i, v : self)
+	{
+		if (equals(v, type)) return i;
+	}
+	return SearchResult.MISSING?;
+}
+
+fn usz! List.rindex_of(&self, Type type) @if(ELEMENT_IS_EQUATABLE)
+{
+	foreach_r (i, v : self)
+	{
+		if (equals(v, type)) return i;
+	}
+	return SearchResult.MISSING?;
+}
+
+fn bool List.equals(&self, List other_list) @if(ELEMENT_IS_EQUATABLE)
+{
+	if (self.size != other_list.size) return false;
+	foreach (i, v : self)
+	{
+		if (!equals(v, other_list.entries[i])) return false;
+	}
+	return true;
+}
+
+/**
+ * Check for presence of a value in a list.
+ *
+ * @param [&in] self "the list to find elements in"
+ * @param value "The value to search for"
+ * @return "True if the value is found, false otherwise"
+ **/
+fn bool List.contains(&self, Type value) @if(ELEMENT_IS_EQUATABLE)
+{
+	foreach (i, v : self)
+	{
+		if (equals(v, value)) return true;
+	}
+	return false;
+}
+
+/**
+ * @param [&inout] self "The list to remove elements from"
+ * @param value "The value to remove"
+ * @return "true if the value was found"
+ **/
+fn bool List.remove_last_item(&self, Type value) @if(ELEMENT_IS_EQUATABLE)
+{
+	return @ok(self.remove_at(self.rindex_of(value)));
+}
+
+/**
+ * @param [&inout] self "The list to remove elements from"
+ * @param value "The value to remove"
+ * @return "true if the value was found"
+ **/
+fn bool List.remove_first_item(&self, Type value) @if(ELEMENT_IS_EQUATABLE)
+{
+	return @ok(self.remove_at(self.index_of(value)));
+}
+/**
+ * @param [&inout] self "The list to remove elements from"
+ * @param value "The value to remove"
+ * @return "the number of deleted elements."
+ **/
+fn usz List.remove_item(&self, Type value) @if(ELEMENT_IS_EQUATABLE)
+{
+	usz old_size = self.size;
+	defer {
+		if (old_size != self.size) self._update_size_change(old_size, self.size);
+	}
+	return list_common::list_remove_item(self, value);
+}
+
+
+
+fn void List.remove_all_from(&self, List* other_list) @if(ELEMENT_IS_EQUATABLE)
+{
+	if (!other_list.size) return;
+	usz old_size = self.size;
+	defer {
+		if (old_size != self.size) self._update_size_change(old_size, self.size);
+	}
+	foreach (v : other_list) self.remove_item(v);
+}
+
+/**
+ * @param [&in] self
+ * @return "The number non-null values in the list"
+ **/
+fn usz List.compact_count(&self) @if(ELEMENT_IS_POINTER)
+{
+	usz vals = 0;
+	foreach (v : self) if (v) vals++;
+	return vals;
+}
+
+fn usz List.compact(&self) @if(ELEMENT_IS_POINTER)
+{
+	usz old_size = self.size;
+	defer {
+		if (old_size != self.size) self._update_size_change(old_size, self.size);
+	}
+	return list_common::list_compact(self);
+}
+
+// --> Deprecated
+
+/**
+ * @param [&inout] self "The list to remove elements from"
+ * @param value "The value to remove"
+ * @return "true if the value was found"
+ **/
+fn bool List.remove_last_match(&self, Type value) @if(ELEMENT_IS_EQUATABLE) @deprecated
+{
+	return self.remove_last_item(value) @inline;
+}
+
+/**
+ * @param [&inout] self "The list to remove elements from"
+ * @param value "The value to remove"
+ * @return "true if the value was found"
+ **/
+fn bool List.remove_first_match(&self, Type value) @if(ELEMENT_IS_EQUATABLE) @deprecated
+{
+	return self.remove_first_item(value) @inline;
+}
+
+
+/**
+ * @param [&inout] self "The list to remove elements from"
+ * @param value "The value to remove"
+ * @return "the number of deleted elements."
+ **/
+fn usz List.remove_all_matches(&self, Type value) @if(ELEMENT_IS_EQUATABLE) @deprecated
+{
+	return self.remove_item(value) @inline;
+}

lib/std/collections/list_common.c3

@@ -0,0 +1,112 @@
+module std::collections::list_common;
+
+/**
+ * IMPORTANT The returned array must be freed using free_aligned.
+ **/
+macro list_to_new_aligned_array($Type, self, Allocator allocator)
+{
+	if (!self.size) return $Type[] {};
+	$Type[] result = allocator::alloc_array_aligned(allocator, $Type, self.size);
+	result[..] = self.entries[:self.size];
+	return result;
+}
+
+macro list_to_new_array($Type, self, Allocator allocator)
+{
+	if (!self.size) return $Type[] {};
+	$Type[] result = allocator::alloc_array(allocator, $Type, self.size);
+	result[..] = self.entries[:self.size];
+	return result;
+}
+
+macro void list_reverse(self)
+{
+	if (self.size < 2) return;
+	usz half = self.size / 2U;
+	usz end = self.size - 1;
+	for (usz i = 0; i < half; i++)
+	{
+		@swap(self.entries[i], self.entries[end - i]);
+	}
+}
+
+macro usz list_remove_using_test(self, filter, bool $invert, ctx)
+{
+	usz size = self.size;
+	for (usz i = size, usz k = size; k > 0; k = i)
+	{
+		// Find last index of item to be deleted.
+		$if $invert:
+			while (i > 0 && !filter(&self.entries[i - 1], ctx)) i--;
+		$else
+			while (i > 0 && filter(&self.entries[i - 1], ctx)) i--;
+		$endif
+		// Remove the items from this index up to the one not to be deleted.
+		usz n = self.size - k;
+		self.entries[i:n] = self.entries[k:n];
+		self.size -= k - i;
+		// Find last index of item not to be deleted.
+		$if $invert:
+			while (i > 0 && filter(&self.entries[i - 1], ctx)) i--;
+		$else
+			while (i > 0 && !filter(&self.entries[i - 1], ctx)) i--;
+		$endif
+	}
+	return size - self.size;
+}
+
+macro usz list_compact(self)
+{
+	usz size = self.size;
+	for (usz i = size; i > 0; i--)
+	{
+		if (self.entries[i - 1]) continue;
+		for (usz j = i; j < size; j++)
+		{
+			self.entries[j - 1] = self.entries[j];
+		}
+		self.size--;
+	}
+	return size - self.size;
+}
+
+macro usz list_remove_item(self, value)
+{
+	usz size = self.size;
+	for (usz i = size; i > 0; i--)
+	{
+		if (!equals(self.entries[i - 1], value)) continue;
+		for (usz j = i; j < self.size; j++)
+		{
+			self.entries[j - 1] = self.entries[j];
+		}
+		self.size--;
+	}
+	return size - self.size;
+}
+
+
+macro usz list_remove_if(self, filter, bool $invert)
+{
+	usz size = self.size;
+	for (usz i = size, usz k = size; k > 0; k = i)
+	{
+		// Find last index of item to be deleted.
+		$if $invert:
+			while (i > 0 && !filter(&self.entries[i - 1])) i--;
+		$else
+			while (i > 0 && filter(&self.entries[i - 1])) i--;
+		$endif
+		// Remove the items from this index up to the one not to be deleted.
+		usz n = self.size - k;
+		self.entries[i:n] = self.entries[k:n];
+		self.size -= k - i;
+		// Find last index of item not to be deleted.
+		$if $invert:
+			while (i > 0 && filter(&self.entries[i - 1])) i--;
+		$else
+			while (i > 0 && !filter(&self.entries[i - 1])) i--;
+		$endif
+	}
+	return size - self.size;
+}

lib/std/collections/map.c3

@@ -0,0 +1,434 @@
+// Copyright (c) 2023 Christoffer Lerno. All rights reserved.
+// Use of this source code is governed by the MIT license
+// a copy of which can be found in the LICENSE_STDLIB file.
+module std::collections::map(<Key, Value>);
+import std::math;
+
+const uint DEFAULT_INITIAL_CAPACITY = 16;
+const uint MAXIMUM_CAPACITY = 1u << 31;
+const float DEFAULT_LOAD_FACTOR = 0.75;
+const VALUE_IS_EQUATABLE = Value.is_eq;
+const bool COPY_KEYS = types::implements_copy(Key);
+
+distinct Map = void*;
+
+struct MapImpl
+{
+	Entry*[] table;
+	Allocator allocator;
+	uint count; // Number of elements
+	uint threshold; // Resize limit
+	float load_factor;
+}
+
+/**
+ * @require capacity > 0 "The capacity must be 1 or higher"
+ * @require load_factor > 0.0 "The load factor must be higher than 0"
+ * @require capacity < MAXIMUM_CAPACITY "Capacity cannot exceed maximum"
+ **/
+fn Map new(uint capacity = DEFAULT_INITIAL_CAPACITY, float load_factor = DEFAULT_LOAD_FACTOR, Allocator allocator = allocator::heap())
+{
+	MapImpl* map = allocator::alloc(allocator, MapImpl);
+	_init(map, capacity, load_factor, allocator);
+	return (Map)map;
+}
+
+/**
+ * @require capacity > 0 "The capacity must be 1 or higher"
+ * @require load_factor > 0.0 "The load factor must be higher than 0"
+ * @require capacity < MAXIMUM_CAPACITY "Capacity cannot exceed maximum"
+ **/
+fn Map temp(uint capacity = DEFAULT_INITIAL_CAPACITY, float load_factor = DEFAULT_LOAD_FACTOR)
+{
+	MapImpl* map = mem::temp_alloc(MapImpl);
+	_init(map, capacity, load_factor, allocator::temp());
+	return (Map)map;
+}
+
+/**
+ * @param [&in] other_map "The map to copy from."
+ **/
+fn Map new_from_map(Map other_map, Allocator allocator = null)
+{
+	MapImpl* other_map_impl = (MapImpl*)other_map;
+	if (!other_map_impl)
+	{
+		if (allocator) return new(.allocator = allocator);
+		return null;
+	}
+	MapImpl* map = (MapImpl*)new(other_map_impl.table.len, other_map_impl.load_factor, allocator ?: allocator::heap());
+	if (!other_map_impl.count) return (Map)map;
+	foreach (Entry *e : other_map_impl.table)
+    {
+    	if (!e) continue;
+    	map._put_for_create(e.key, e.value);
+    }
+	return (Map)map;
+}
+
+/**
+ * @param [&in] other_map "The map to copy from."
+ **/
+fn Map temp_from_map(Map other_map)
+{
+	return new_from_map(other_map, allocator::temp());
+}
+
+fn bool Map.is_empty(map) @inline
+{
+	return !map || !((MapImpl*)map).count;
+}
+
+fn usz Map.len(map) @inline
+{
+	return map ? ((MapImpl*)map).count : 0;
+}
+
+fn Value*! Map.get_ref(self, Key key)
+{
+	MapImpl *map = (MapImpl*)self;
+	if (!map || !map.count) return SearchResult.MISSING?;
+	uint hash = rehash(key.hash());
+	for (Entry *e = map.table[index_for(hash, map.table.len)]; e != null; e = e.next)
+	{
+		if (e.hash == hash && equals(key, e.key)) return &e.value;
+	}
+	return SearchResult.MISSING?;
+}
+
+fn Entry*! Map.get_entry(map, Key key)
+{
+	MapImpl *map_impl = (MapImpl*)map;
+	if (!map_impl || !map_impl.count) return SearchResult.MISSING?;
+	uint hash = rehash(key.hash());
+	for (Entry *e = map_impl.table[index_for(hash, map_impl.table.len)]; e != null; e = e.next)
+	{
+		if (e.hash == hash && equals(key, e.key)) return e;
+	}
+	return SearchResult.MISSING?;
+}
+
+/**
+ * Get the value or update and
+ * @require $assignable(#expr, Value)
+ **/
+macro Value Map.@get_or_set(&self, Key key, Value #expr)
+{
+	MapImpl *map = (MapImpl*)*self;
+	if (!map || !map.count)
+	{
+		Value val = #expr;
+		map.set(key, val);
+		return val;
+	}
+	uint hash = rehash(key.hash());
+	uint index = index_for(hash, map.table.len);
+	for (Entry *e = map.table[index]; e != null; e = e.next)
+	{
+		if (e.hash == hash && equals(key, e.key)) return e.value;
+	}
+	Value val = #expr;
+	map.add_entry(hash, key, val, index);
+	return val;
+}
+
+fn Value! Map.get(map, Key key) @operator([])
+{
+	return *map.get_ref(key) @inline;
+}
+
+fn bool Map.has_key(map, Key key)
+{
+	return @ok(map.get_ref(key));
+}
+
+macro Value Map.set_value_return(&map, Key key, Value value) @operator([]=)
+{
+	map.set(key, value);
+	return value;
+}
+
+fn bool Map.set(&self, Key key, Value value)
+{
+	// If the map isn't initialized, use the defaults to initialize it.
+	if (!*self) *self = new();
+	MapImpl* map = (MapImpl*)*self;
+	uint hash = rehash(key.hash());
+	uint index = index_for(hash, map.table.len);
+	for (Entry *e = map.table[index]; e != null; e = e.next)
+	{
+		if (e.hash == hash && equals(key, e.key))
+		{
+			e.value = value;
+			return true;
+		}
+	}
+	map._add_entry(hash, key, value, index);
+	return false;
+}
+
+fn void! Map.remove(map, Key key) @maydiscard
+{
+	if (!map || !((MapImpl*)map)._remove_entry_for_key(key)) return SearchResult.MISSING?;
+}
+
+fn void Map.clear(self)
+{
+	MapImpl* map = (MapImpl*)self;
+	if (!map || !map.count) return;
+	foreach (Entry** &entry_ref : map.table)
+	{
+		Entry* entry = *entry_ref;
+		if (!entry) continue;
+		Entry *next = entry.next;
+		while (next)
+		{
+			Entry *to_delete = next;
+			next = next.next;
+			map._free_entry(to_delete);
+		}
+		map._free_entry(entry);
+		*entry_ref = null;
+	}
+	map.count = 0;
+}
+
+fn void Map.free(self)
+{
+	if (!self) return;
+	MapImpl* map = (MapImpl*)self;
+	self.clear();
+	map._free_internal(map.table.ptr);
+	map.table = {};
+	allocator::free(map.allocator, map);
+}
+
+fn Key[] Map.temp_keys_list(map)
+{
+	return map.new_keys_list(allocator::temp()) @inline;
+}
+
+fn Key[] Map.new_keys_list(self, Allocator allocator = allocator::heap())
+{
+	MapImpl* map = (MapImpl*)self;
+	if (!map || !map.count) return {};
+
+	Key[] list = allocator::alloc_array(allocator, Key, map.count);
+	usz index = 0;
+	foreach (Entry* entry : map.table)
+	{
+		while (entry)
+		{
+			list[index++] = entry.key;
+			entry = entry.next;
+		}
+	}
+	return list;
+}
+
+macro Map.@each(map; @body(key, value))
+{
+	map.@each_entry(; Entry* entry) {
+		@body(entry.key, entry.value);
+	};
+}
+
+macro Map.@each_entry(self; @body(entry))
+{
+	MapImpl *map = (MapImpl*)self;
+	if (!map || !map.count) return;
+	foreach (Entry* entry : map.table)
+	{
+		while (entry)
+		{
+			@body(entry);
+			entry = entry.next;
+		}
+	}
+}
+
+fn Value[] Map.temp_values_list(map)
+{
+	return map.new_values_list(allocator::temp()) @inline;
+}
+
+fn Value[] Map.new_values_list(self, Allocator allocator = allocator::heap())
+{
+	MapImpl* map = (MapImpl*)self;
+	if (!map || !map.count) return {};
+	Value[] list = allocator::alloc_array(allocator, Value, map.count);
+	usz index = 0;
+	foreach (Entry* entry : map.table)
+	{
+		while (entry)
+		{
+			list[index++] = entry.value;
+			entry = entry.next;
+		}
+	}
+	return list;
+}
+
+fn bool Map.has_value(self, Value v) @if(VALUE_IS_EQUATABLE)
+{
+	MapImpl* map = (MapImpl*)self;
+	if (!map || !map.count) return false;
+	foreach (Entry* entry : map.table)
+	{
+		while (entry)
+		{
+			if (equals(v, entry.value)) return true;
+			entry = entry.next;
+		}
+	}
+	return false;
+}
+
+// --- private methods
+
+fn void MapImpl._add_entry(&map, uint hash, Key key, Value value, uint bucket_index) @private
+{
+	$if COPY_KEYS:
+	key = key.copy(map.allocator);
+	$endif
+	Entry* entry = allocator::new(map.allocator, Entry, { .hash = hash, .key = key, .value = value, .next = map.table[bucket_index] });
+	map.table[bucket_index] = entry;
+	if (map.count++ >= map.threshold)
+	{
+		map._resize(map.table.len * 2);
+	}
+}
+
+fn void MapImpl._resize(&map, uint new_capacity) @private
+{
+	Entry*[] old_table = map.table;
+	uint old_capacity = old_table.len;
+	if (old_capacity == MAXIMUM_CAPACITY)
+	{
+		map.threshold = uint.max;
+		return;
+	}
+	Entry*[] new_table = allocator::new_array(map.allocator, Entry*, new_capacity);
+	map._transfer(new_table);
+	map.table = new_table;
+	map._free_internal(old_table.ptr);
+	map.threshold = (uint)(new_capacity * map.load_factor);
+}
+
+fn uint rehash(uint hash) @inline @private
+{
+	hash ^= (hash >> 20) ^ (hash >> 12);
+	return hash ^ ((hash >> 7) ^ (hash >> 4));
+}
+
+macro uint index_for(uint hash, uint capacity) @private
+{
+	return hash & (capacity - 1);
+}
+
+fn void MapImpl._transfer(&map, Entry*[] new_table) @private
+{
+	Entry*[] src = map.table;
+	uint new_capacity = new_table.len;
+	foreach (uint j, Entry *e : src)
+	{
+		if (!e) continue;
+		do
+		{
+			Entry* next = e.next;
+			uint i = index_for(e.hash, new_capacity);
+			e.next = new_table[i];
+			new_table[i] = e;
+			e = next;
+		}
+		while (e);
+	}
+}
+
+fn void _init(MapImpl* impl, uint capacity, float load_factor, Allocator allocator) @private
+{
+	capacity = math::next_power_of_2(capacity);
+	*impl = {
+		.allocator = allocator,
+    	.load_factor = load_factor,
+    	.threshold = (uint)(capacity * load_factor),
+    	.table = allocator::new_array(allocator, Entry*, capacity)
+	};
+}
+
+fn void MapImpl._put_for_create(&map, Key key, Value value) @private
+{
+	uint hash = rehash(key.hash());
+	uint i = index_for(hash, map.table.len);
+	for (Entry *e = map.table[i]; e != null; e = e.next)
+	{
+		if (e.hash == hash && equals(key, e.key))
+		{
+			e.value = value;
+			return;
+		}
+	}
+	map._create_entry(hash, key, value, i);
+}
+
+fn void MapImpl._free_internal(&map, void* ptr) @inline @private
+{
+	allocator::free(map.allocator, ptr);
+}
+
+fn bool MapImpl._remove_entry_for_key(&map, Key key) @private
+{
+	if (!map.count) return false;
+	uint hash = rehash(key.hash());
+	uint i = index_for(hash, map.table.len);
+	Entry* prev = map.table[i];
+	Entry* e = prev;
+	while (e)
+	{
+		Entry *next = e.next;
+		if (e.hash == hash && equals(key, e.key))
+		{
+			map.count--;
+			if (prev == e)
+			{
+				map.table[i] = next;
+			}
+			else
+			{
+				prev.next = next;
+			}
+			map._free_entry(e);
+			return true;
+		}
+		prev = e;
+		e = next;
+	}
+	return false;
+}
+
+fn void MapImpl._create_entry(&map, uint hash, Key key, Value value, int bucket_index) @private
+{
+	Entry *e = map.table[bucket_index];
+	$if COPY_KEYS:
+	key = key.copy(map.allocator);
+	$endif
+	Entry* entry = allocator::new(map.allocator, Entry, { .hash = hash, .key = key, .value = value, .next = map.table[bucket_index] });
+	map.table[bucket_index] = entry;
+	map.count++;
+}
+
+fn void MapImpl._free_entry(&self, Entry *entry) @local
+{
+	$if COPY_KEYS:
+	allocator::free(self.allocator, entry.key);
+	$endif
+	self._free_internal(entry);
+}
+
+struct Entry
+{
+	uint hash;
+	Key key;
+	Value value;
+	Entry* next;
+}

lib/std/collections/maybe.c3

@@ -0,0 +1,19 @@
+module std::collections::maybe(<Type>);
+
+struct Maybe
+{
+	Type value;
+	bool has_value;
+}
+
+fn Maybe value(Type val)
+{
+	return { .value = val, .has_value = true };
+}
+
+const Maybe EMPTY = { };
+
+macro Type! Maybe.get(self)
+{
+	return self.has_value ? self.value : SearchResult.MISSING?;
+}

lib/std/collections/object.c3

@@ -0,0 +1,469 @@
+// Copyright (c) 2023 Christoffer Lerno. All rights reserved.
+// Use of this source code is governed by the MIT license
+// a copy of which can be found in the LICENSE_STDLIB file.
+module std::collections::object;
+import std::collections::map, std::collections::list, std::io;
+
+const Object TRUE_OBJECT = { .b = true, .type = bool.typeid };
+const Object FALSE_OBJECT = { .b = false, .type = bool.typeid };
+const Object NULL_OBJECT = { .type = void*.typeid };
+
+struct Object (Printable)
+{
+	typeid type;
+	Allocator allocator;
+	union
+	{
+		uint128 i;
+		double f;
+		bool b;
+		String s;
+		void* other;
+		ObjectInternalList array;
+		ObjectInternalMap map;
+	}
+}
+
+
+fn usz! Object.to_format(&self, Formatter* formatter) @dynamic
+{
+	switch (self.type)
+	{
+		case void:
+			return formatter.printf("{}")!;
+		case void*:
+			return formatter.printf("null")!;
+		case String:
+			return formatter.printf(`"%s"`, self.s)!;
+		case bool:
+			return formatter.printf(self.b ? "true" : "false")!;
+		case ObjectInternalList:
+			usz n = formatter.printf("[")!;
+			foreach (i, ol : self.array)
+			{
+				if (i > 0) n += formatter.printf(",")!;
+				n += ol.to_format(formatter)!;
+			}
+			n += formatter.printf("]")!;
+			return n;
+		case ObjectInternalMap:
+			usz n = formatter.printf("{")!;
+			@stack_mem(1024; Allocator mem)
+			{
+				foreach (i, key : self.map.key_new_list(mem))
+				{
+					if (i > 0) n += formatter.printf(",")!;
+					n += formatter.printf(`"%s":`, key)!;
+					n += self.map.get(key).to_format(formatter)!;
+				}
+			};
+			n += formatter.printf("}")!;
+			return n;
+		default:
+			switch (self.type.kindof)
+			{
+				case SIGNED_INT:
+					return formatter.printf("%d", self.i)!;
+				case UNSIGNED_INT:
+					return formatter.printf("%d", (uint128)self.i)!;
+				case FLOAT:
+					return formatter.printf("%d", self.f)!;
+				case ENUM:
+					return formatter.printf("%d", self.i)!;
+				default:
+					return formatter.printf("<>")!;
+			}
+	}
+}
+
+fn Object* new_obj(Allocator allocator)
+{
+	return allocator::new(allocator, Object, { .allocator = allocator, .type = void.typeid });
+}
+
+fn Object* new_null()
+{
+	return &NULL_OBJECT;
+}
+
+fn Object* new_int(int128 i, Allocator allocator)
+{
+	return allocator::new(allocator, Object, { .i = i, .allocator = allocator, .type = int128.typeid });
+}
+
+macro Object* new_enum(e, Allocator allocator)
+{
+	return allocator::new(allocator, Object, { .i = (int128)e, .allocator = allocator, .type = @typeid(e) });
+}
+
+fn Object* new_float(double f, Allocator allocator)
+{
+	return allocator::new(allocator, Object, { .f = f, .allocator = allocator, .type = double.typeid });
+}
+
+fn Object* new_string(String s, Allocator allocator)
+{
+	return allocator::new(allocator, Object, { .s = s.copy(allocator), .allocator = allocator, .type = String.typeid });
+}
+
+
+fn Object* new_bool(bool b)
+{
+	return b ? &TRUE_OBJECT : &FALSE_OBJECT;
+}
+
+fn void Object.free(&self)
+{
+	switch (self.type)
+	{
+		case void:
+			break;
+		case String:
+			allocator::free(self.allocator, self.s);
+		case ObjectInternalList:
+			foreach (ol : self.array)
+			{
+				ol.free();
+			}
+			self.array.free();
+		case ObjectInternalMap:
+			self.map.@each_entry(; ObjectInternalMapEntry* entry) {
+				allocator::free(self.allocator, entry.key);
+				entry.value.free();
+			};
+		default:
+			break;
+	}
+	if (self.allocator) allocator::free(self.allocator, self);
+}
+
+fn bool Object.is_null(&self) @inline => self == &NULL_OBJECT;
+fn bool Object.is_empty(&self) @inline => self.type == void.typeid;
+fn bool Object.is_map(&self) @inline => self.type == ObjectInternalMap.typeid;
+fn bool Object.is_array(&self) @inline => self.type == ObjectInternalList.typeid;
+fn bool Object.is_bool(&self) @inline => self.type == bool.typeid;
+fn bool Object.is_string(&self) @inline => self.type == String.typeid;
+fn bool Object.is_float(&self) @inline => self.type == double.typeid;
+fn bool Object.is_int(&self) @inline => self.type == int128.typeid;
+fn bool Object.is_keyable(&self) => self.is_empty() || self.is_map();
+fn bool Object.is_indexable(&self) => self.is_empty() || self.is_array();
+
+/**
+ * @require self.is_keyable()
+ **/
+fn void Object.init_map_if_needed(&self) @private
+{
+	if (self.is_empty())
+	{
+		self.type = ObjectInternalMap.typeid;
+		self.map.new_init(.allocator = self.allocator);
+	}
+}
+
+/**
+ * @require self.is_indexable()
+ **/
+fn void Object.init_array_if_needed(&self) @private
+{
+	if (self.is_empty())
+	{
+		self.type = ObjectInternalList.typeid;
+		self.array.new_init(.allocator = self.allocator);
+	}
+}
+
+/**
+ * @require self.is_keyable()
+ **/
+fn void Object.set_object(&self, String key, Object* new_object) @private
+{
+	self.init_map_if_needed();
+	ObjectInternalMapEntry*! entry = self.map.get_entry(key);
+	defer
+	{
+		(void)allocator::free(self.allocator, entry.key);
+		(void)entry.value.free();
+	}
+	self.map.set(key.copy(self.map.allocator), new_object);
+}
+
+macro Object* Object.object_from_value(&self, value) @private
+{
+	var $Type = $typeof(value);
+	$switch
+		$case types::is_int($Type):
+			return new_int(value, self.allocator);
+		$case types::is_float($Type):
+			return new_float(value, self.allocator);
+		$case $Type.typeid == String.typeid:
+			return new_string(value, self.allocator);
+		$case $Type.typeid == bool.typeid:
+			return new_bool(value);
+		$case $Type.typeid == Object*.typeid:
+			return value;
+		$case $Type.typeid == void*.typeid:
+			if (value != null) return CastResult.TYPE_MISMATCH?;
+			return &NULL_OBJECT;
+		$case $assignable(value, String):
+			return new_string(value, self.allocator);
+		$default:
+			$error "Unsupported object type.";
+	$endswitch
+
+}
+
+macro Object* Object.set(&self, String key, value)
+{
+	Object* val = self.object_from_value(value);
+	self.set_object(key, val);
+	return val;
+}
+
+/**
+ * @require self.is_indexable()
+ **/
+macro Object* Object.set_at(&self, usz index, String key, value)
+{
+	Object* val = self.object_from_value(value);
+	self.set_object_at(key, index, val);
+	return val;
+}
+
+/**
+ * @require self.is_indexable()
+ * @ensure return != null
+ **/
+macro Object* Object.push(&self, value)
+{
+	Object* val = self.object_from_value(value);
+	self.push_object(val);
+	return val;
+}
+
+/**
+ * @require self.is_keyable()
+ **/
+fn Object*! Object.get(&self, String key) => self.is_empty() ? SearchResult.MISSING? : self.map.get(key);
+
+
+fn bool Object.has_key(&self, String key) => self.is_map() && self.map.has_key(key);
+
+/**
+ * @require self.is_indexable()
+ **/
+fn Object* Object.get_at(&self, usz index)
+{
+	return self.array.get(index);
+}
+
+/**
+ * @require self.is_indexable()
+ **/
+fn usz Object.get_len(&self)
+{
+	return self.array.len();
+}
+
+/**
+ * @require self.is_indexable()
+ **/
+fn void Object.push_object(&self, Object* to_append)
+{
+	self.init_array_if_needed();
+	self.array.push(to_append);
+}
+
+/**
+ * @require self.is_indexable()
+ **/
+fn void Object.set_object_at(&self, usz index, Object* to_set)
+{
+	self.init_array_if_needed();
+	while (self.array.len() < index)
+	{
+		self.array.push(&NULL_OBJECT);
+	}
+	if (self.array.len() == index)
+	{
+		self.array.push(to_set);
+		return;
+	}
+	self.array.get(index).free();
+	self.array.set_at(index, to_set);
+}
+
+/**
+ * @require $Type.kindof.is_int() "Expected an integer type."
+ **/
+macro get_integer_value(Object* value, $Type)
+{
+	if (value.is_float())
+	{
+		return ($Type)value.f;
+	}
+	if (value.is_string())
+	{
+		$if $Type.kindof == TypeKind.SIGNED_INT:
+			return ($Type)value.s.to_int128();
+		$else
+			return ($Type)value.s.to_uint128();
+		$endif
+	}
+	if (!value.is_int()) return NumberConversion.MALFORMED_INTEGER?;
+	return ($Type)value.i;
+}
+
+
+/**
+ * @require self.is_indexable()
+ * @require $Type.kindof.is_int() : "Expected an integer type"
+ **/
+macro Object.get_integer_at(&self, $Type, usz index) @private
+{
+	return get_integer_value(self.get_at(index), $Type);
+}
+
+/**
+ * @require self.is_keyable()
+ * @require $Type.kindof.is_int() : "Expected an integer type"
+ **/
+macro Object.get_integer(&self, $Type, String key) @private
+{
+	return get_integer_value(self.get(key), $Type);
+}
+
+fn ichar! Object.get_ichar(&self, String key) => self.get_integer(ichar, key);
+fn short! Object.get_short(&self, String key) => self.get_integer(short, key);
+fn int! Object.get_int(&self, String key) => self.get_integer(int, key);
+fn long! Object.get_long(&self, String key) => self.get_integer(long, key);
+fn int128! Object.get_int128(&self, String key) => self.get_integer(int128, key);
+
+fn ichar! Object.get_ichar_at(&self, usz index) => self.get_integer_at(ichar, index);
+fn short! Object.get_short_at(&self, usz index) => self.get_integer_at(short, index);
+fn int! Object.get_int_at(&self, usz index) => self.get_integer_at(int, index);
+fn long! Object.get_long_at(&self, usz index) => self.get_integer_at(long, index);
+fn int128! Object.get_int128_at(&self, usz index) => self.get_integer_at(int128, index);
+
+fn char! Object.get_char(&self, String key) => self.get_integer(ichar, key);
+fn short! Object.get_ushort(&self, String key) => self.get_integer(ushort, key);
+fn uint! Object.get_uint(&self, String key) => self.get_integer(uint, key);
+fn ulong! Object.get_ulong(&self, String key) => self.get_integer(ulong, key);
+fn uint128! Object.get_uint128(&self, String key) => self.get_integer(uint128, key);
+
+fn char! Object.get_char_at(&self, usz index) => self.get_integer_at(char, index);
+fn ushort! Object.get_ushort_at(&self, usz index) => self.get_integer_at(ushort, index);
+fn uint! Object.get_uint_at(&self, usz index) => self.get_integer_at(uint, index);
+fn ulong! Object.get_ulong_at(&self, usz index) => self.get_integer_at(ulong, index);
+fn uint128! Object.get_uint128_at(&self, usz index) => self.get_integer_at(uint128, index);
+
+/**
+ * @require self.is_keyable()
+ **/
+fn String! Object.get_string(&self, String key)
+{
+	Object* value = self.get(key)!;
+	if (!value.is_string()) return CastResult.TYPE_MISMATCH?;
+	return value.s;
+}
+
+/**
+ * @require self.is_indexable()
+ **/
+fn String! Object.get_string_at(&self, usz index)
+{
+	Object* value = self.get_at(index);
+	if (!value.is_string()) return CastResult.TYPE_MISMATCH?;
+	return value.s;
+}
+
+/**
+ * @require self.is_keyable()
+ **/
+macro String! Object.get_enum(&self, $EnumType, String key)
+{
+	Object value = self.get(key)!;
+	if ($EnumType.typeid != value.type) return CastResult.TYPE_MISMATCH?;
+	return ($EnumType)value.i;
+}
+
+/**
+ * @require self.is_indexable()
+ **/
+macro String! Object.get_enum_at(&self, $EnumType, usz index)
+{
+	Object value = self.get_at(index);
+	if ($EnumType.typeid != value.type) return CastResult.TYPE_MISMATCH?;
+	return ($EnumType)value.i;
+}
+
+/**
+ * @require self.is_keyable()
+ **/
+fn bool! Object.get_bool(&self, String key)
+{
+	Object* value = self.get(key)!;
+	if (!value.is_bool()) return CastResult.TYPE_MISMATCH?;
+	return value.b;
+}
+
+
+/**
+ * @require self.is_indexable()
+ **/
+fn bool! Object.get_bool_at(&self, usz index)
+{
+	Object* value = self.get_at(index);
+	if (!value.is_bool()) return CastResult.TYPE_MISMATCH?;
+	return value.b;
+}
+
+/**
+ * @require self.is_keyable()
+ **/
+fn double! Object.get_float(&self, String key)
+{
+	Object* value = self.get(key)!;
+	switch (value.type.kindof)
+	{
+		case SIGNED_INT:
+			return (double)value.i;
+		case UNSIGNED_INT:
+			return (double)(uint128)value.i;
+		case FLOAT:
+			return value.f;
+		default:
+			return CastResult.TYPE_MISMATCH?;
+	}
+}
+
+/**
+ * @require self.is_indexable()
+ **/
+fn double! Object.get_float_at(&self, usz index)
+{
+	Object* value = self.get_at(index);
+	switch (value.type.kindof)
+	{
+		case SIGNED_INT:
+			return (double)value.i;
+		case UNSIGNED_INT:
+			return (double)(uint128)value.i;
+		case FLOAT:
+			return value.f;
+		default:
+			return CastResult.TYPE_MISMATCH?;
+	}
+}
+
+fn Object* Object.get_or_create_obj(&self, String key)
+{
+	if (try obj = self.get(key) && !obj.is_null()) return obj;
+	Object* container = new_obj(self.allocator);
+	self.set(key, container);
+	return container;
+}
+
+def ObjectInternalMap = HashMap(<String, Object*>) @private;
+def ObjectInternalList = List(<Object*>) @private;
+def ObjectInternalMapEntry = Entry(<String, Object*>) @private;
+

lib/std/collections/priorityqueue.c3

@@ -0,0 +1,148 @@
+//  priorityqueue.c3
+//  A priority queue using a classic binary heap for C3.
+//
+//  Copyright (c) 2022 David Kopec
+//
+//  Permission is hereby granted, free of charge, to any person obtaining a copy
+//  of this software and associated documentation files (the "Software"), to deal
+//  in the Software without restriction, including without limitation the rights
+//  to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+//  copies of the Software, and to permit persons to whom the Software is
+//  furnished to do so, subject to the following conditions:
+//
+//  The above copyright notice and this permission notice shall be included in all
+//  copies or substantial portions of the Software.
+//
+//  THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+//  IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+//  FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+//  AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+//  LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+//  OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+//  SOFTWARE.
+module std::collections::priorityqueue(<Type>);
+import std::collections::priorityqueue::private;
+
+distinct PriorityQueue = inline PrivatePriorityQueue(<Type, false>);
+distinct PriorityQueueMax = inline PrivatePriorityQueue(<Type, true>);
+
+module std::collections::priorityqueue::private(<Type, MAX>);
+import std::collections::list, std::io;
+
+def Heap = List(<Type>);
+
+struct PrivatePriorityQueue (Printable)
+{
+	Heap heap;
+}
+
+fn void PrivatePriorityQueue.new_init(&self, usz initial_capacity = 16, Allocator allocator = allocator::heap()) @inline
+{
+	self.heap.new_init(initial_capacity, allocator);
+}
+
+fn void PrivatePriorityQueue.temp_init(&self, usz initial_capacity = 16) @inline
+{
+	self.heap.new_init(initial_capacity, allocator::temp()) @inline;
+}
+
+fn void PrivatePriorityQueue.push(&self, Type element)
+{
+	self.heap.push(element);
+	usz i = self.heap.len() - 1;
+	while (i > 0)
+	{
+		usz parent = (i - 1) / 2;
+		Type item = self.heap[i];
+		Type parent_item = self.heap[parent];
+		$if MAX:
+			bool ok = greater(item, parent_item);
+		$else
+			bool ok = less(item, parent_item);
+		$endif
+		if (!ok) break;
+		self.heap.swap(i, parent);
+		i = parent;
+	}
+}
+
+/**
+* @require self != null
+*/
+fn Type! PrivatePriorityQueue.pop(&self)
+{
+	usz i = 0;
+	usz len = self.heap.len();
+	if (!len) return IteratorResult.NO_MORE_ELEMENT?;
+	usz new_count = len - 1;
+	self.heap.swap(0, new_count);
+	while OUTER: ((2 * i + 1) < new_count)
+	{
+		usz j = 2 * i + 1;
+		Type left = self.heap[j];
+		Type item = self.heap[i];
+		switch
+		{
+			case j + 1 < new_count:
+				Type right = self.heap[j + 1];
+				$if MAX:
+				if (!greater(right, left)) nextcase;
+				if (!greater(right, item)) break OUTER;
+				$else
+				if (!greater(left, right)) nextcase;
+				if (!greater(item, right)) break OUTER;
+				$endif
+				j++;
+			default:
+				$if MAX:
+				if (!greater(left, item)) break OUTER;
+				$else
+				if (!greater(item, left)) break OUTER;
+				$endif
+		}
+		self.heap.swap(i, j);
+		i = j;
+	}
+
+	return self.heap.pop();
+}
+
+fn Type! PrivatePriorityQueue.first(&self)
+{
+	if (!self.len()) return IteratorResult.NO_MORE_ELEMENT?;
+	return self.heap.get(0);
+}
+
+fn void PrivatePriorityQueue.free(&self)
+{
+	self.heap.free();
+}
+
+fn usz PrivatePriorityQueue.len(&self) @operator(len)
+{
+	return self.heap.len();
+}
+
+fn bool PrivatePriorityQueue.is_empty(&self)
+{
+	return self.heap.is_empty();
+}
+
+/**
+ * @require index < self.len()
+ */
+fn Type PrivatePriorityQueue.get(&self, usz index) @operator([])
+{
+	return self.heap[index];
+}
+
+fn usz! PrivatePriorityQueue.to_format(&self, Formatter* formatter) @dynamic
+{
+	return self.heap.to_format(formatter);
+}
+
+fn String PrivatePriorityQueue.to_new_string(&self, Allocator allocator = allocator::heap()) @dynamic
+{
+	return self.heap.to_new_string(allocator);
+}
+

lib/std/collections/range.c3

@@ -0,0 +1,85 @@
+/**
+ * @require Type.is_ordered : "The type must be ordered"
+ **/
+module std::collections::range(<Type>);
+import std::io;
+
+struct Range (Printable)
+{
+	Type start;
+	Type end;
+}
+
+fn usz Range.len(&self) @operator(len)
+{
+	if (self.end < self.start) return 0;
+	return (usz)(self.end - self.start) + 1;
+}
+
+fn bool Range.contains(&self, Type value) @inline
+{
+	return value >= self.start && value <= self.end;
+}
+
+/**
+ * @require index < self.len() : "Can't index into an empty range"
+ **/
+fn Type Range.get(&self, usz index) @operator([])
+{
+	return (Type)(self.start + (usz)index);
+}
+
+fn String Range.to_new_string(&self, Allocator allocator = allocator::heap()) @dynamic
+{
+	return string::new_format("[%s..%s]", self.start, self.end, .allocator = allocator);
+}
+
+fn String Range.to_tstring(&self)
+{
+	return self.to_new_string(allocator::temp());
+}
+
+fn usz! Range.to_format(&self, Formatter* formatter) @dynamic
+{
+	return formatter.printf("[%s..%s]", self.start, self.end)!;
+}
+
+struct ExclusiveRange (Printable)
+{
+	Type start;
+	Type end;
+}
+
+fn usz ExclusiveRange.len(&self) @operator(len)
+{
+	if (self.end < self.start) return 0;
+	return (usz)(self.end - self.start);
+}
+
+fn bool ExclusiveRange.contains(&self, Type value) @inline
+{
+	return value >= self.start && value < self.end;
+}
+
+fn usz! ExclusiveRange.to_format(&self, Formatter* formatter) @dynamic
+{
+	return formatter.printf("[%s..<%s]", self.start, self.end)!;
+}
+
+fn String ExclusiveRange.to_new_string(&self, Allocator allocator = allocator::heap()) @dynamic
+{
+	return string::new_format("[%s..<%s]", self.start, self.end, .allocator = allocator);
+}
+
+fn String ExclusiveRange.to_tstring(&self)
+{
+	return self.to_new_string(allocator::temp());
+}
+
+/**
+ * @require index < self.len() : "Can't index into an empty range"
+ **/
+fn Type ExclusiveRange.get(&self, usz index) @operator([])
+{
+	return (Type)(self.start + index);
+}

lib/std/collections/ringbuffer.c3

@@ -0,0 +1,103 @@
+module std::collections::ringbuffer(<Type, SIZE>);
+
+struct RingBuffer
+{
+	Type[SIZE] buf;
+	usz written;
+	usz head;
+}
+
+fn void RingBuffer.init(&self) @inline
+{
+	*self = {};
+}
+
+fn void RingBuffer.push(&self, Type c)
+{
+	if (self.written < SIZE)
+	{
+		self.buf[self.written] = c;
+		self.written++;
+	}
+	else
+	{
+		self.buf[self.head] = c;
+		self.head = (self.head + 1) % SIZE;
+	}
+}
+
+fn Type RingBuffer.get(&self, usz index) @operator([])
+{
+	index %= SIZE;
+	usz avail = SIZE - self.head;
+	if (index < avail)
+	{
+		return self.buf[self.head + index];
+	}
+	return self.buf[index - avail];
+}
+
+fn Type! RingBuffer.pop(&self)
+{
+	switch
+	{
+		case self.written == 0:
+			return SearchResult.MISSING?;
+		case self.written < SIZE:
+			self.written--;
+			return self.buf[self.written];
+		default:
+			self.head = (self.head - 1) % SIZE;
+			return self.buf[self.head];
+	}
+}
+
+fn usz RingBuffer.read(&self, usz index, Type[] buffer)
+{
+	index %= SIZE;
+	if (self.written < SIZE)
+	{
+		if (index >= self.written) return 0;
+		usz end = self.written - index;
+		usz n = min(end, buffer.len);
+		buffer[:n] = self.buf[index:n];
+		return n;
+	}
+	usz end = SIZE - self.head;
+	if (index >= end)
+	{
+		index -= end;
+		if (index >= self.head) return 0;
+		usz n = min(self.head - index, buffer.len);
+		buffer[:n] = self.buf[index:n];
+		return n;
+	}
+	if (buffer.len <= SIZE - index)
+	{
+		usz n = buffer.len;
+		buffer[:n] = self.buf[self.head + index:n];
+		return n;
+	}
+	usz n1 = SIZE - index;
+	buffer[:n1] = self.buf[self.head + index:n1];
+	buffer = buffer[n1..];
+	index -= n1;
+	usz n2 = min(self.head - index, buffer.len);
+	buffer[:n2] = self.buf[index:n2];
+	return n1 + n2;
+}
+
+fn void RingBuffer.write(&self, Type[] buffer)
+{
+	usz i;
+	while (self.written < SIZE && i < buffer.len)
+	{
+		self.buf[self.written] = buffer[i++];
+		self.written++;
+	}
+	foreach (c : buffer[i..])
+	{
+		self.buf[self.head] = c;
+		self.head = (self.head + 1) % SIZE;
+	}
+}

lib/std/collections/tuple.c3

@@ -0,0 +1,16 @@
+module std::collections::tuple(<Type1, Type2>);
+
+struct Tuple
+{
+	Type1 first;
+	Type2 second;
+}
+
+module std::collections::triple(<Type1, Type2, Type3>);
+
+struct Triple
+{
+	Type1 first;
+	Type2 second;
+	Type3 third;
+}

lib/std/core/allocators/arena_allocator.c3

@@ -0,0 +1,106 @@
+// Copyright (c) 2023 Christoffer Lerno. All rights reserved.
+// Use of this source code is governed by the MIT license
+// a copy of which can be found in the LICENSE_STDLIB file.
+module std::core::mem::allocator;
+import std::math;
+
+struct ArenaAllocator (Allocator)
+{
+	char[] data;
+	usz used;
+}
+
+/**
+ * Initialize a memory arena for use using the provided bytes.
+ **/
+fn void ArenaAllocator.init(&self, char[] data)
+{
+	self.data = data;
+	self.used = 0;
+}
+
+fn void ArenaAllocator.clear(&self)
+{
+	self.used = 0;
+}
+
+struct ArenaAllocatorHeader @local
+{
+	usz size;
+	char[*] data;
+}
+
+/*
+ * @require ptr != null
+ **/
+fn void ArenaAllocator.release(&self, void* ptr, bool) @dynamic
+{
+	assert((uptr)ptr >= (uptr)self.data.ptr, "Pointer originates from a different allocator.");
+	ArenaAllocatorHeader* header = ptr - ArenaAllocatorHeader.sizeof;
+	// Reclaim memory if it's the last element.
+	if (ptr + header.size == &self.data[self.used])
+	{
+		self.used -= header.size + ArenaAllocatorHeader.sizeof;
+	}
+}
+
+fn usz ArenaAllocator.mark(&self) @dynamic => self.used;
+fn void ArenaAllocator.reset(&self, usz mark) @dynamic => self.used = mark;
+
+/**
+ * @require !alignment || math::is_power_of_2(alignment)
+ * @require alignment <= mem::MAX_MEMORY_ALIGNMENT `alignment too big`
+ * @require size > 0
+ **/
+fn void*! ArenaAllocator.acquire(&self, usz size, AllocInitType init_type, usz alignment) @dynamic
+{
+	alignment = alignment_for_allocation(alignment);
+	usz total_len = self.data.len;
+	if (size > total_len) return AllocationFailure.CHUNK_TOO_LARGE?;
+	void* start_mem = self.data.ptr;
+	void* unaligned_pointer_to_offset = start_mem + self.used + ArenaAllocatorHeader.sizeof;
+	void* mem = mem::aligned_pointer(unaligned_pointer_to_offset, alignment);
+	usz end = (usz)(mem - self.data.ptr) + size;
+	if (end > total_len) return AllocationFailure.OUT_OF_MEMORY?;
+	self.used = end;
+	ArenaAllocatorHeader* header = mem - ArenaAllocatorHeader.sizeof;
+	header.size = size;
+	if (init_type == ZERO) mem::clear(mem, size, mem::DEFAULT_MEM_ALIGNMENT);
+	return mem;
+}
+
+/**
+ * @require !alignment || math::is_power_of_2(alignment)
+ * @require alignment <= mem::MAX_MEMORY_ALIGNMENT `alignment too big`
+ * @require old_pointer != null
+ * @require size > 0
+ **/
+fn void*! ArenaAllocator.resize(&self, void *old_pointer, usz size, usz alignment) @dynamic
+{
+	alignment = alignment_for_allocation(alignment);
+	assert(old_pointer >= self.data.ptr, "Pointer originates from a different allocator.");
+	usz total_len = self.data.len;
+	if (size > total_len) return AllocationFailure.CHUNK_TOO_LARGE?;
+	ArenaAllocatorHeader* header = old_pointer - ArenaAllocatorHeader.sizeof;
+	usz old_size = header.size;
+	// Do last allocation and alignment match?
+	if (&self.data[self.used] == old_pointer + old_size && mem::ptr_is_aligned(old_pointer, alignment))
+	{
+		if (old_size >= size)
+		{
+			self.used -= old_size - size;
+		}
+		else
+		{
+			usz new_used = self.used + size - old_size;
+			if (new_used > total_len) return AllocationFailure.OUT_OF_MEMORY?;
+			self.used = new_used;
+		}
+		header.size = size;
+		return old_pointer;
+	}
+	// Otherwise just allocate new memory.
+	void* mem = self.acquire(size, NO_ZERO, alignment)!;
+	mem::copy(mem, old_pointer, old_size, mem::DEFAULT_MEM_ALIGNMENT, mem::DEFAULT_MEM_ALIGNMENT);
+	return mem;
+}

lib/std/core/allocators/dynamic_arena.c3

@@ -0,0 +1,203 @@
+// Copyright (c) 2021-2024 Christoffer Lerno. All rights reserved.
+// Use of this source code is governed by the MIT license
+// a copy of which can be found in the LICENSE_STDLIB file.
+module std::core::mem::allocator;
+import std::math;
+
+struct DynamicArenaAllocator (Allocator)
+{
+	Allocator backing_allocator;
+	DynamicArenaPage* page;
+	DynamicArenaPage* unused_page;
+	usz page_size;
+}
+
+/**
+ * @param [&inout] allocator
+ * @require page_size >= 128
+ **/
+fn void DynamicArenaAllocator.init(&self, usz page_size, Allocator allocator)
+{
+	self.page = null;
+	self.unused_page = null;
+	self.page_size = page_size;
+	self.backing_allocator = allocator;
+}
+
+import std::io;
+fn void DynamicArenaAllocator.free(&self)
+{
+	DynamicArenaPage* page = self.page;
+	while (page)
+	{
+		DynamicArenaPage* next_page = page.prev_arena;
+		allocator::free(self.backing_allocator, page.memory);
+		allocator::free(self.backing_allocator, page);
+		page = next_page;
+	}
+	page = self.unused_page;
+	while (page)
+	{
+		DynamicArenaPage* next_page = page.prev_arena;
+		allocator::free(self.backing_allocator, page.memory);
+		allocator::free(self.backing_allocator, page);
+		page = next_page;
+	}
+	self.page = null;
+	self.unused_page = null;
+}
+
+struct DynamicArenaPage @local
+{
+	void* memory;
+	void* prev_arena;
+	usz total;
+	usz used;
+	void* current_stack_ptr;
+}
+
+struct DynamicArenaChunk @local
+{
+	usz size;
+}
+
+/**
+ * @require ptr
+ * @require self.page `tried to free pointer on invalid allocator`
+ */
+fn void DynamicArenaAllocator.release(&self, void* ptr, bool) @dynamic
+{
+	DynamicArenaPage* current_page = self.page;
+	if (ptr == current_page.current_stack_ptr)
+	{
+		current_page.used = (usz)((ptr - DEFAULT_SIZE_PREFIX) - current_page.memory);
+	}
+	current_page.current_stack_ptr = null;
+}
+
+/**
+ * @require size > 0 `Resize doesn't support zeroing`
+ * @require old_pointer != null `Resize doesn't handle null pointers`
+ * @require self.page `tried to realloc pointer on invalid allocator`
+ */
+fn void*! DynamicArenaAllocator.resize(&self, void* old_pointer, usz size, usz alignment) @dynamic
+{
+	DynamicArenaPage* current_page = self.page;
+	alignment = alignment_for_allocation(alignment);
+	usz* old_size_ptr = old_pointer - DEFAULT_SIZE_PREFIX;
+	usz old_size = *old_size_ptr;
+	// We have the old pointer and it's correctly aligned.
+	if (old_size >= size && mem::ptr_is_aligned(old_pointer, alignment))
+	{
+		*old_size_ptr = size;
+		if (current_page.current_stack_ptr == old_pointer)
+		{
+			current_page.used = (usz)((old_pointer - DEFAULT_SIZE_PREFIX) - current_page.memory);
+		}
+		return old_pointer;
+	}
+	if REUSE: (current_page.current_stack_ptr == old_pointer && mem::ptr_is_aligned(old_pointer, alignment))
+	{
+		assert(size > old_size);
+		usz add_size = size - old_size;
+		if (add_size + current_page.used > current_page.total) break REUSE;
+		*old_size_ptr = size;
+		current_page.used += add_size;
+		return old_pointer;
+	}
+	void* new_mem = self.acquire(size, NO_ZERO, alignment)!;
+	mem::copy(new_mem, old_pointer, old_size, mem::DEFAULT_MEM_ALIGNMENT);
+	return new_mem;
+}
+
+fn void DynamicArenaAllocator.reset(&self, usz mark = 0) @dynamic
+{
+	assert(mark == 0, "Unexpectedly reset dynamic arena allocator with mark %d", mark);
+	DynamicArenaPage* page = self.page;
+	DynamicArenaPage** unused_page_ptr = &self.unused_page;
+	while (page)
+	{
+		DynamicArenaPage* next_page = page.prev_arena;
+		page.used = 0;
+		DynamicArenaPage* prev_unused = *unused_page_ptr;
+		*unused_page_ptr = page;
+		page.prev_arena = prev_unused;
+		page = next_page;
+	}
+	self.page = page;
+}
+
+/**
+ * @require math::is_power_of_2(alignment)
+ * @require size > 0
+ */
+fn void*! DynamicArenaAllocator._alloc_new(&self, usz size, usz alignment) @local
+{
+	// First, make sure that we can align it, extending the page size if needed.
+	usz page_size = max(self.page_size, mem::aligned_offset(size + DynamicArenaChunk.sizeof + alignment, alignment));
+	assert(page_size > size + DynamicArenaChunk.sizeof);
+	// Grab the page without alignment (we do it ourselves)
+	void* mem = allocator::malloc_try(self.backing_allocator, page_size)!;
+	DynamicArenaPage*! page = allocator::new_try(self.backing_allocator, DynamicArenaPage);
+	if (catch err = page)
+	{
+		allocator::free(self.backing_allocator, mem);
+		return err?;
+	}
+	page.memory = mem;
+	void* mem_start = mem::aligned_pointer(mem + DynamicArenaChunk.sizeof, alignment);
+	assert(mem_start + size < mem + page_size);
+	DynamicArenaChunk* chunk = (DynamicArenaChunk*)mem_start - 1;
+	chunk.size = size;
+	page.prev_arena = self.page;
+	page.total = page_size;
+	page.used = mem_start + size - page.memory;
+	self.page = page;
+	page.current_stack_ptr = mem_start;
+	return mem_start;
+}
+
+/**
+ * @require size > 0 `acquire expects size > 0`
+ * @require !alignment || math::is_power_of_2(alignment)
+ */
+fn void*! DynamicArenaAllocator.acquire(&self, usz size, AllocInitType init_type, usz alignment) @dynamic
+{
+	alignment = alignment_for_allocation(alignment);
+	DynamicArenaPage* page = self.page;
+	void* ptr = {|
+		if (!page && self.unused_page)
+		{
+			self.page = page = self.unused_page;
+			self.unused_page = page.prev_arena;
+			page.prev_arena = null;
+		}
+		if (!page) return self._alloc_new(size, alignment);
+		void* start = mem::aligned_pointer(page.memory + page.used + DynamicArenaChunk.sizeof, alignment);
+		usz new_used = start - page.memory + size;
+		if ALLOCATE_NEW: (new_used > page.total)
+		{
+			if ((page = self.unused_page))
+			{
+				start = mem::aligned_pointer(page.memory + page.used + DynamicArenaChunk.sizeof, alignment);
+				new_used = start + size - page.memory;
+				if (page.total >= new_used)
+				{
+					self.unused_page = page.prev_arena;
+					page.prev_arena = self.page;
+					self.page = page;
+					break ALLOCATE_NEW;
+				}
+			}
+			return self._alloc_new(size, alignment);
+		}
+		page.used = new_used;
+		assert(start + size == page.memory + page.used);
+		void* mem = start;
+		DynamicArenaChunk* chunk = (DynamicArenaChunk*)mem - 1;
+		chunk.size = size;
+		return mem;
+	|}!;
+	if (init_type == ZERO) mem::clear(ptr, size, mem::DEFAULT_MEM_ALIGNMENT);
+	return ptr;
+}

lib/std/core/allocators/heap_allocator.c3

@@ -0,0 +1,209 @@
+// Copyright (c) 2021-2024 Christoffer Lerno. All rights reserved.
+// Use of this source code is governed by the MIT license
+// a copy of which can be found in the LICENSE_STDLIB file.
+
+module std::core::mem::allocator;
+import std::math;
+
+struct SimpleHeapAllocator (Allocator)
+{
+	MemoryAllocFn alloc_fn;
+	Header* free_list;
+}
+
+/**
+ * @require allocator "An underlying memory provider must be given"
+ * @require !self.free_list "The allocator may not be already initialized"
+ **/
+fn void SimpleHeapAllocator.init(&self, MemoryAllocFn allocator)
+{
+	self.alloc_fn = allocator;
+	self.free_list = null;
+}
+
+fn void*! SimpleHeapAllocator.acquire(&self, usz size, AllocInitType init_type, usz alignment) @dynamic
+{
+	if (init_type == ZERO)
+	{
+		return alignment > 0 ? @aligned_alloc(self._calloc, size, alignment) : self._calloc(size);
+	}
+	return alignment > 0 ? @aligned_alloc(self._alloc, size, alignment) : self._alloc(size);
+}
+
+fn void*! SimpleHeapAllocator.resize(&self, void* old_pointer, usz size, usz alignment) @dynamic
+{
+	return alignment > 0
+		? @aligned_realloc(self._calloc, self._free, old_pointer, size, alignment)
+		:  self._realloc(old_pointer, size);
+}
+
+fn void SimpleHeapAllocator.release(&self, void* old_pointer, bool aligned) @dynamic
+{
+	if (aligned)
+	{
+		@aligned_free(self._free, old_pointer)!!;
+	}
+	else
+	{
+		self._free(old_pointer);
+	}
+}
+
+/**
+ * @require old_pointer && bytes > 0
+ **/
+fn void*! SimpleHeapAllocator._realloc(&self, void* old_pointer, usz bytes) @local
+{
+	// Find the block header.
+	Header* block = (Header*)old_pointer - 1;
+	if (block.size >= bytes) return old_pointer;
+	void* new = self._alloc(bytes)!;
+	usz max_to_copy = math::min(block.size, bytes);
+	mem::copy(new, old_pointer, max_to_copy);
+	self._free(old_pointer);
+	return new;
+}
+
+fn void*! SimpleHeapAllocator._calloc(&self, usz bytes) @local
+{
+	void* data = self._alloc(bytes)!;
+	mem::clear(data, bytes, mem::DEFAULT_MEM_ALIGNMENT);
+	return data;
+}
+
+fn void*! SimpleHeapAllocator._alloc(&self, usz bytes) @local
+{
+	usz aligned_bytes = mem::aligned_offset(bytes, mem::DEFAULT_MEM_ALIGNMENT);
+	if (!self.free_list)
+	{
+		self.add_block(aligned_bytes)!;
+	}
+
+	Header* current = self.free_list;
+	Header* previous = current;
+	while (current)
+	{
+		switch
+		{
+			case current.size >= aligned_bytes && current.size <= aligned_bytes + Header.sizeof + 64:
+				if (current == previous)
+				{
+					self.free_list = current.next;
+				}
+				else
+				{
+					previous.next = current.next;
+				}
+				current.next = null;
+				return current + 1;
+			case current.size > aligned_bytes:
+				Header* unallocated = (Header*)((char*)current + aligned_bytes + Header.sizeof);
+				unallocated.size = current.size - aligned_bytes;
+				unallocated.next = current.next;
+				if (current == self.free_list)
+				{
+					self.free_list = unallocated;
+				}
+				else
+				{
+					previous.next = unallocated;
+				}
+				current.size = aligned_bytes;
+				current.next = null;
+				return current + 1;
+			default:
+				previous = current;
+				current = current.next;
+		}
+	}
+	self.add_block(aligned_bytes)!;
+	return self._alloc(aligned_bytes);
+}
+
+fn void! SimpleHeapAllocator.add_block(&self, usz aligned_bytes) @local
+{
+	assert(mem::aligned_offset(aligned_bytes, mem::DEFAULT_MEM_ALIGNMENT) == aligned_bytes);
+	char[] result = self.alloc_fn(aligned_bytes + Header.sizeof)!;
+	Header* new_block = (Header*)result.ptr;
+	new_block.size = result.len - Header.sizeof;
+	new_block.next = null;
+	self._free(new_block + 1);
+}
+
+
+fn void SimpleHeapAllocator._free(&self, void* ptr) @local
+{
+	// Empty ptr -> do nothing.
+	if (!ptr) return;
+
+	// Find the block header.
+	Header* block = (Header*)ptr - 1;
+
+	// No free list? Then just return self.
+	if (!self.free_list)
+	{
+		self.free_list = block;
+		return;
+	}
+
+	// Find where in the list it should be inserted.
+	Header* current = self.free_list;
+	Header* prev = current;
+	while (current)
+	{
+		if (block < current)
+		{
+			// Between prev and current
+			if (block > prev) break;
+			// Before current
+			if (current == prev) break;
+		}
+		prev = current;
+		current = prev.next;
+	}
+	if (current)
+	{
+		// Insert after the current block.
+		// Are the blocks adjacent?
+		if (current == (Header*)((char*)(block + 1) + block.size))
+		{
+			// Merge
+			block.size += current.size + Header.sizeof;
+			block.next = current.next;
+		}
+		else
+		{
+			// Chain to current
+			block.next = current;
+		}
+	}
+	if (prev == current)
+	{
+		// Swap new start of free list
+		self.free_list = block;
+	}
+	else
+	{
+		// Prev adjacent?
+		if (block == (Header*)((char*)(prev + 1) + prev.size))
+		{
+			prev.size += block.size + Header.sizeof;
+			prev.next = block.next;
+		}
+		else
+		{
+			// Link prev to block
+			prev.next = block;
+		}
+	}
+}
+
+union Header @local
+{
+	struct
+	{
+		Header* next;
+		usz size;
+	}
+	usz align;
+}

lib/std/core/allocators/libc_allocator.c3

@@ -0,0 +1,156 @@
+// Copyright (c) 2021-2024 Christoffer Lerno. All rights reserved.
+// Use of this source code is governed by the MIT license
+// a copy of which can be found in the LICENSE_STDLIB file.
+
+module std::core::mem::allocator;
+import libc;
+
+const LibcAllocator LIBC_ALLOCATOR = {};
+distinct LibcAllocator (Allocator) = uptr;
+
+module std::core::mem::allocator @if(env::POSIX);
+import std::os;
+import libc;
+
+fn void*! LibcAllocator.acquire(&self, usz bytes, AllocInitType init_type, usz alignment) @dynamic
+{
+	if (init_type == ZERO)
+	{
+		void* data @noinit;
+		if (alignment > mem::DEFAULT_MEM_ALIGNMENT)
+		{
+			if (posix::posix_memalign(&data, alignment, bytes)) return AllocationFailure.OUT_OF_MEMORY?;
+			mem::clear(data, bytes, mem::DEFAULT_MEM_ALIGNMENT);
+			return data;
+		}
+        return libc::calloc(1, bytes) ?: AllocationFailure.OUT_OF_MEMORY?;
+	}
+	else
+	{
+		void* data @noinit;
+		if (alignment > mem::DEFAULT_MEM_ALIGNMENT)
+		{
+			if (posix::posix_memalign(&data, alignment, bytes)) return AllocationFailure.OUT_OF_MEMORY?;
+		}
+		else
+		{
+			if (!(data = libc::malloc(bytes))) return AllocationFailure.OUT_OF_MEMORY?;
+		}
+        $if env::TESTING:
+		    for (usz i = 0; i < bytes; i++) ((char*)data)[i] = 0xAA;
+        $endif
+        return data;
+	}
+}
+
+fn void*! LibcAllocator.resize(&self, void* old_ptr, usz new_bytes, usz alignment) @dynamic
+{
+	if (alignment <= mem::DEFAULT_MEM_ALIGNMENT) return libc::realloc(old_ptr, new_bytes) ?: AllocationFailure.OUT_OF_MEMORY?;
+    void* new_ptr;
+	if (posix::posix_memalign(&new_ptr, alignment, new_bytes)) return AllocationFailure.OUT_OF_MEMORY?;
+
+	$switch
+	$case env::DARWIN:
+		usz old_usable_size = darwin::malloc_size(old_ptr);
+	$case env::LINUX:
+		usz old_usable_size = linux::malloc_usable_size(old_ptr);
+	$default:
+		usz old_usable_size = new_bytes;
+	$endswitch
+
+	usz copy_size = new_bytes < old_usable_size ? new_bytes : old_usable_size;
+	mem::copy(new_ptr, old_ptr, copy_size, mem::DEFAULT_MEM_ALIGNMENT, mem::DEFAULT_MEM_ALIGNMENT);
+	libc::free(old_ptr);
+	return new_ptr;
+}
+
+fn void LibcAllocator.release(&self, void* old_ptr, bool aligned) @dynamic
+{
+	libc::free(old_ptr);
+}
+
+module std::core::mem::allocator @if(env::WIN32);
+import std::os::win32;
+import libc;
+
+fn void*! LibcAllocator.acquire(&self, usz bytes, AllocInitType init_type, usz alignment) @dynamic
+{
+	if (init_type == ZERO)
+	{
+		if (alignment > 0)
+		{
+			return win32::_aligned_recalloc(null, 1, bytes, alignment) ?: AllocationFailure.OUT_OF_MEMORY?;
+		}
+		return libc::calloc(1, bytes) ?: AllocationFailure.OUT_OF_MEMORY?;
+	}
+	void* data = alignment > 0 ? win32::_aligned_malloc(bytes, alignment) : libc::malloc(bytes);
+	if (!data) return AllocationFailure.OUT_OF_MEMORY?;
+    $if env::TESTING:
+		for (usz i = 0; i < bytes; i++) ((char*)data)[i] = 0xAA;
+	$endif
+    return data;
+}
+
+fn void*! LibcAllocator.resize(&self, void* old_ptr, usz new_bytes, usz alignment) @dynamic
+{
+	if (alignment)
+	{
+		return win32::_aligned_realloc(old_ptr, new_bytes, alignment) ?: AllocationFailure.OUT_OF_MEMORY?;
+	}
+	return libc::realloc(old_ptr, new_bytes) ?: AllocationFailure.OUT_OF_MEMORY?;
+}
+
+fn void LibcAllocator.release(&self, void* old_ptr, bool aligned) @dynamic
+{
+	if (aligned)
+	{
+    	win32::_aligned_free(old_ptr);
+    	return;
+	}
+	libc::free(old_ptr);
+}
+
+module std::core::mem::allocator @if(!env::WIN32 && !env::POSIX);
+import libc;
+
+fn void*! LibcAllocator.acquire(&self, usz bytes, AllocInitType init_type, usz alignment) @dynamic
+{
+	if (init_type == ZERO)
+	{
+		void* data = alignment ? @aligned_alloc(fn void*(usz bytes) => libc::calloc(bytes, 1), bytes, alignment)!! : libc::calloc(bytes, 1);
+        return data ?: AllocationFailure.OUT_OF_MEMORY?;
+	}
+	else
+	{
+		void* data = alignment ? @aligned_alloc(libc::malloc, bytes, alignment)!! : libc::malloc(bytes);
+        if (!data) return AllocationFailure.OUT_OF_MEMORY?;
+        $if env::TESTING:
+		    for (usz i = 0; i < bytes; i++) ((char*)data)[i] = 0xAA;
+        $endif
+        return data;
+	}
+}
+
+
+fn void*! LibcAllocator.resize(&self, void* old_ptr, usz new_bytes, usz alignment) @dynamic
+{
+	if (alignment)
+	{
+		void* data = @aligned_realloc(fn void*(usz bytes) => libc::malloc(bytes), libc::free, old_ptr, new_bytes, alignment)!!;
+		return data ?: AllocationFailure.OUT_OF_MEMORY?;
+	}
+	return libc::realloc(old_ptr, new_bytes) ?: AllocationFailure.OUT_OF_MEMORY?;
+}
+
+
+fn void LibcAllocator.release(&self, void* old_ptr, bool aligned) @dynamic
+{
+	if (aligned)
+	{
+		@aligned_free(libc::free, old_ptr)!!;
+	}
+	else
+	{
+		libc::free(old_ptr);
+	}
+}

lib/std/core/allocators/on_stack_allocator.c3

@@ -0,0 +1,149 @@
+module std::core::mem::allocator;
+
+struct OnStackAllocator (Allocator)
+{
+	Allocator backing_allocator;
+	char[] data;
+	usz used;
+	OnStackAllocatorExtraChunk* chunk;
+}
+
+
+struct OnStackAllocatorExtraChunk @local
+{
+	bool is_aligned;
+	OnStackAllocatorExtraChunk* prev;
+	void* data;
+}
+
+/**
+ * @param [&inout] allocator
+ * Initialize a memory arena for use using the provided bytes.
+ **/
+fn void OnStackAllocator.init(&self, char[] data, Allocator allocator)
+{
+	self.data = data;
+	self.backing_allocator = allocator;
+	self.used = 0;
+}
+
+fn void OnStackAllocator.free(&self)
+{
+	OnStackAllocatorExtraChunk* chunk = self.chunk;
+	while (chunk)
+	{
+		if (chunk.is_aligned)
+		{
+			allocator::free_aligned(self.backing_allocator, chunk.data);
+		}
+		else
+		{
+			allocator::free(self.backing_allocator, chunk.data);
+		}
+		void* old = chunk;
+		chunk = chunk.prev;
+		allocator::free(self.backing_allocator, old);
+	}
+	self.chunk = null;
+	self.used = 0;
+}
+
+struct OnStackAllocatorHeader
+{
+	usz size;
+	char[*] data;
+}
+
+/**
+ * @require old_pointer
+ **/
+fn void OnStackAllocator.release(&self, void* old_pointer, bool aligned) @dynamic
+{
+	if (allocation_in_stack_mem(self, old_pointer)) return;
+	on_stack_allocator_remove_chunk(self, old_pointer);
+	self.release(old_pointer, aligned);
+}
+
+fn bool allocation_in_stack_mem(OnStackAllocator* a, void* ptr) @local
+{
+	return ptr >= a.data.ptr && ptr <= &a.data[^1];
+}
+
+fn void on_stack_allocator_remove_chunk(OnStackAllocator* a, void* ptr) @local
+{
+	OnStackAllocatorExtraChunk* chunk = a.chunk;
+	OnStackAllocatorExtraChunk** addr = &a.chunk;
+	while (chunk)
+	{
+		if (chunk.data == ptr)
+		{
+			*addr = chunk.prev;
+			allocator::free(a.backing_allocator, chunk);
+			return;
+		}
+		addr = &chunk.prev;
+		chunk = *addr;
+	}
+	unreachable("Missing chunk");
+}
+
+fn OnStackAllocatorExtraChunk* on_stack_allocator_find_chunk(OnStackAllocator* a, void* ptr) @local
+{
+	OnStackAllocatorExtraChunk* chunk = a.chunk;
+	while (chunk)
+	{
+		if (chunk.data == ptr) return chunk;
+		chunk = chunk.prev;
+	}
+	return null;
+}
+
+/**
+ * @require size > 0
+ * @require old_pointer != null
+ * @require alignment <= mem::MAX_MEMORY_ALIGNMENT `alignment too big`
+ **/
+fn void*! OnStackAllocator.resize(&self, void* old_pointer, usz size, usz alignment) @dynamic
+{
+	if (!allocation_in_stack_mem(self, old_pointer))
+	{
+		OnStackAllocatorExtraChunk* chunk = on_stack_allocator_find_chunk(self, old_pointer);
+		assert(chunk, "Tried to realloc pointer not belonging to the allocator");
+		return chunk.data = self.backing_allocator.resize(old_pointer, size, alignment)!;
+	}
+
+	OnStackAllocatorHeader* header = old_pointer - OnStackAllocatorHeader.sizeof;
+	usz old_size = header.size;
+	void* mem = self.acquire(size, NO_ZERO, alignment)!;
+	mem::copy(mem, old_pointer, old_size, mem::DEFAULT_MEM_ALIGNMENT, mem::DEFAULT_MEM_ALIGNMENT);
+	return mem;
+}
+
+/**
+ * @require alignment <= mem::MAX_MEMORY_ALIGNMENT `alignment too big`
+ * @require size > 0
+ **/
+fn void*! OnStackAllocator.acquire(&self, usz size, AllocInitType init_type, usz alignment) @dynamic
+{
+	bool aligned = alignment > 0;
+	alignment = alignment_for_allocation(alignment);
+	usz total_len = self.data.len;
+	void* start_mem = self.data.ptr;
+	void* unaligned_pointer_to_offset = start_mem + self.used + OnStackAllocatorHeader.sizeof ;
+	void* mem = mem::aligned_pointer(unaligned_pointer_to_offset, alignment);
+	usz end = (usz)(mem - self.data.ptr) + size;
+	Allocator backing_allocator = self.backing_allocator;
+
+	if (end > total_len)
+	{
+		OnStackAllocatorExtraChunk* chunk = allocator::alloc_try(backing_allocator, OnStackAllocatorExtraChunk)!;
+		defer catch allocator::free(backing_allocator, chunk);
+		defer try self.chunk = chunk;
+		*chunk = { .prev = self.chunk, .is_aligned = aligned };
+		return chunk.data = backing_allocator.acquire(size, init_type, aligned ? alignment : 0)!;
+	}
+	self.used = end;
+	OnStackAllocatorHeader* header = mem - OnStackAllocatorHeader.sizeof;
+	header.size = size;
+	return mem;
+}

lib/std/core/allocators/temp_allocator.c3

@@ -0,0 +1,227 @@
+module std::core::mem::allocator;
+import std::io, std::math;
+
+struct TempAllocatorChunk @local
+{
+	usz size;
+	char[*] data;
+}
+
+struct TempAllocator (Allocator)
+{
+	Allocator backing_allocator;
+	TempAllocatorPage* last_page;
+	usz used;
+	usz capacity;
+	char[*] data;
+}
+
+const usz PAGE_IS_ALIGNED @private = (usz)isz.max + 1u;
+
+
+struct TempAllocatorPage
+{
+	TempAllocatorPage* prev_page;
+	void* start;
+	usz mark;
+	usz size;
+	usz ident;
+	char[*] data;
+}
+
+macro usz TempAllocatorPage.pagesize(&self) => self.size & ~PAGE_IS_ALIGNED;
+macro bool TempAllocatorPage.is_aligned(&self) => self.size & PAGE_IS_ALIGNED == PAGE_IS_ALIGNED;
+
+/**
+ * @require size >= 16
+ **/
+fn TempAllocator*! new_temp_allocator(usz size, Allocator allocator)
+{
+	TempAllocator* temp = allocator::alloc_with_padding(allocator, TempAllocator, size)!;
+	temp.last_page = null;
+	temp.backing_allocator = allocator;
+	temp.used = 0;
+	temp.capacity = size;
+	return temp;
+}
+
+fn void TempAllocator.destroy(&self)
+{
+	self.reset(0);
+	if (self.last_page) (void)self._free_page(self.last_page);
+	allocator::free(self.backing_allocator, self);
+}
+
+fn usz TempAllocator.mark(&self) @dynamic => self.used;
+
+fn void TempAllocator.release(&self, void* old_pointer, bool) @dynamic
+{
+	usz old_size = *(usz*)(old_pointer - DEFAULT_SIZE_PREFIX);
+	if (old_pointer + old_size == &self.data[self.used])
+	{
+		self.used -= old_size;
+		asan::poison_memory_region(&self.data[self.used], old_size);
+	}
+}
+fn void TempAllocator.reset(&self, usz mark) @dynamic
+{
+	TempAllocatorPage *last_page = self.last_page;
+	while (last_page && last_page.mark > mark)
+	{
+		TempAllocatorPage *to_free = last_page;
+		last_page = last_page.prev_page;
+		self._free_page(to_free)!!;
+	}
+	self.last_page = last_page;
+	$if env::COMPILER_SAFE_MODE || env::ADDRESS_SANITIZER:
+		if (!last_page)
+		{
+			usz cleaned = self.used - mark;
+			if (cleaned > 0)
+			{
+				$if env::COMPILER_SAFE_MODE:
+					self.data[mark : cleaned] = 0xAA;
+				$endif
+				asan::poison_memory_region(&self.data[mark], cleaned);
+			}
+		}
+	$endif
+	self.used = mark;
+}
+
+fn void! TempAllocator._free_page(&self, TempAllocatorPage* page) @inline @local
+{
+	void* mem = page.start;
+	return self.backing_allocator.release(mem, page.is_aligned());
+}
+
+fn void*! TempAllocator._realloc_page(&self, TempAllocatorPage* page, usz size, usz alignment) @inline @local
+{
+	// Then the actual start pointer:
+	void* real_pointer = page.start;
+
+	// Walk backwards to find the pointer to this page.
+	TempAllocatorPage **pointer_to_prev = &self.last_page;
+	// Remove the page from the list
+	while (*pointer_to_prev != page)
+	{
+		pointer_to_prev = &((*pointer_to_prev).prev_page);
+	}
+	*pointer_to_prev = page.prev_page;
+	usz page_size = page.pagesize();
+	// Clear on size > original size.
+	void* data = self.acquire(size, NO_ZERO, alignment)!;
+	mem::copy(data, &page.data[0], page_size, mem::DEFAULT_MEM_ALIGNMENT, mem::DEFAULT_MEM_ALIGNMENT);
+	self.backing_allocator.release(real_pointer, page.is_aligned());
+	return data;
+}
+
+fn void*! TempAllocator.resize(&self, void* pointer, usz size, usz alignment) @dynamic
+{
+	TempAllocatorChunk *chunk = pointer - TempAllocatorChunk.sizeof;
+	if (chunk.size == (usz)-1)
+	{
+		assert(self.last_page, "Realloc of non temp pointer");
+		// First grab the page
+		TempAllocatorPage *page = pointer - TempAllocatorPage.sizeof;
+		return self._realloc_page(page, size, alignment);
+	}
+
+	TempAllocatorChunk* data = self.acquire(size, NO_ZERO, alignment)!;
+	mem::copy(data, pointer, chunk.size, mem::DEFAULT_MEM_ALIGNMENT, mem::DEFAULT_MEM_ALIGNMENT);
+
+	return data;
+}
+
+/**
+ * @require size > 0
+ * @require !alignment || math::is_power_of_2(alignment)
+ * @require alignment <= mem::MAX_MEMORY_ALIGNMENT `alignment too big`
+ **/
+fn void*! TempAllocator.acquire(&self, usz size, AllocInitType init_type, usz alignment) @dynamic
+{
+	alignment = alignment_for_allocation(alignment);
+	void* start_mem = &self.data;
+	void* starting_ptr = start_mem + self.used;
+	void* aligned_header_start = mem::aligned_pointer(starting_ptr, TempAllocatorChunk.alignof);
+	void* mem = aligned_header_start + TempAllocatorChunk.sizeof;
+	if (alignment > TempAllocatorChunk.alignof)
+	{
+		mem = mem::aligned_pointer(mem, alignment);
+	}
+	usz new_usage = (usz)(mem - start_mem) + size;
+
+	// Arena allocation, simple!
+	if (new_usage <= self.capacity)
+	{
+		asan::unpoison_memory_region(starting_ptr, new_usage - self.used);
+		TempAllocatorChunk* chunk_start = mem - TempAllocatorChunk.sizeof;
+		chunk_start.size = size;
+		self.used = new_usage;
+		if (init_type == ZERO) mem::clear(mem, size, mem::DEFAULT_MEM_ALIGNMENT);
+		return mem;
+	}
+
+	// Fallback to backing allocator
+	TempAllocatorPage* page;
+
+	// We have something we need to align.
+	if (alignment > mem::DEFAULT_MEM_ALIGNMENT)
+	{
+		// This is actually simpler, since it will create the offset for us.
+		usz total_alloc_size = mem::aligned_offset(TempAllocatorPage.sizeof + size, alignment);
+		if (init_type == ZERO)
+		{
+			mem = allocator::calloc_aligned(self.backing_allocator, total_alloc_size, alignment)!;
+		}
+		else
+		{
+			mem = allocator::malloc_aligned(self.backing_allocator, total_alloc_size, alignment)!;
+		}
+		page = (TempAllocatorPage*)mem - 1;
+		page.start = mem;
+		page.size = size | PAGE_IS_ALIGNED;
+	}
+	else
+	{
+		// Here we might need to pad
+		usz padded_header_size = mem::aligned_offset(TempAllocatorPage.sizeof, mem::DEFAULT_MEM_ALIGNMENT);
+		usz total_alloc_size = padded_header_size + size;
+		void* alloc = self.backing_allocator.acquire(total_alloc_size, init_type, 0)!;
+
+		// Find the page.
+		page = alloc + padded_header_size - TempAllocatorPage.sizeof;
+		assert(mem::ptr_is_aligned(page, TempAllocator.alignof));
+		assert(mem::ptr_is_aligned(&page.data[0], mem::DEFAULT_MEM_ALIGNMENT));
+		page.start = alloc;
+		page.size = size;
+	}
+
+	// Mark it as a page
+	page.ident = ~(usz)0;
+	// Store when it was created
+	page.mark = ++self.used;
+	// Hook up the page.
+	page.prev_page = self.last_page;
+	self.last_page = page;
+	return &page.data[0];
+}
+
+fn void! TempAllocator.print_pages(&self, File* f)
+{
+	TempAllocatorPage *last_page = self.last_page;
+	if (!last_page)
+	{
+		io::fprintf(f, "No pages.\n")!;
+		return;
+	}
+	io::fprintf(f, "---Pages----\n")!;
+	uint index = 0;
+	while (last_page)
+	{
+		bool is_not_aligned = !(last_page.size & (1u64 << 63));
+		io::fprintf(f, "%d. Alloc: %d %d at %p%s\n", ++index,
+				last_page.size & ~(1u64 << 63), last_page.mark, &last_page.data[0], is_not_aligned ? "" : " [aligned]")!;
+		last_page = last_page.prev_page;
+	}
+}

lib/std/core/allocators/tracking_allocator.c3

@@ -0,0 +1,217 @@
+// Copyright (c) 2021-2024 Christoffer Lerno. All rights reserved.
+// Use of this source code is governed by the MIT license
+// a copy of which can be found in the LICENSE_STDLIB file.
+
+module std::core::mem::allocator;
+import std::collections, std::io, std::os::backtrace;
+
+const MAX_BACKTRACE = 16;
+struct Allocation
+{
+	void* ptr;
+	usz size;
+	void*[MAX_BACKTRACE] backtrace;
+}
+
+def AllocMap = HashMap(<uptr, Allocation>);
+
+// A simple tracking allocator.
+// It tracks allocations using a hash map but
+// is not compatible with allocators that uses mark()
+struct TrackingAllocator (Allocator)
+{
+	Allocator inner_allocator;
+	AllocMap map;
+	usz mem_total;
+	usz allocs_total;
+}
+
+/**
+ * Initialize a tracking allocator to wrap (and track) another allocator.
+ *
+ * @param [&inout] allocator "The allocator to track"
+ **/
+fn void TrackingAllocator.init(&self, Allocator allocator)
+{
+	*self = { .inner_allocator = allocator };
+	self.map.new_init(.allocator = allocator);
+}
+
+/**
+ * Free this tracking allocator.
+ **/
+fn void TrackingAllocator.free(&self)
+{
+	self.map.free();
+	*self = {};
+}
+
+/**
+ * @return "the total allocated memory not yet freed."
+ **/
+fn usz TrackingAllocator.allocated(&self)
+{
+	usz allocated = 0;
+	@pool()
+	{
+		foreach (&allocation : self.map.value_tlist()) allocated += allocation.size;
+	};
+	return allocated;
+}
+
+/**
+ * @return "the total memory allocated (freed or not)."
+ **/
+fn usz TrackingAllocator.total_allocated(&self) => self.mem_total;
+
+/**
+ * @return "the total number of allocations (freed or not)."
+ **/
+fn usz TrackingAllocator.total_allocation_count(&self) => self.allocs_total;
+
+fn Allocation[] TrackingAllocator.allocations_tlist(&self, Allocator allocator)
+{
+	return self.map.value_tlist();
+}
+
+/**
+ * @return "the number of non-freed allocations."
+ **/
+fn usz TrackingAllocator.allocation_count(&self) => self.map.count;
+
+fn void*! TrackingAllocator.acquire(&self, usz size, AllocInitType init_type, usz alignment) @dynamic
+{
+	void* data = self.inner_allocator.acquire(size, init_type, alignment)!;
+	self.allocs_total++;
+	void*[MAX_BACKTRACE] bt;
+	backtrace::capture_current(&bt);
+	self.map.set((uptr)data, { data, size, bt });
+    self.mem_total += size;
+	self.allocs_total++;
+	return data;
+}
+
+fn void*! TrackingAllocator.resize(&self, void* old_pointer, usz size, usz alignment) @dynamic
+{
+	void* data = self.inner_allocator.resize(old_pointer, size, alignment)!;
+	self.map.remove((uptr)old_pointer);
+	void*[MAX_BACKTRACE] bt;
+	backtrace::capture_current(&bt);
+	self.map.set((uptr)data, { data, size, bt });
+    self.mem_total += size;
+	self.allocs_total++;
+	return data;
+}
+
+fn void TrackingAllocator.release(&self, void* old_pointer, bool is_aligned) @dynamic
+{
+	if (catch self.map.remove((uptr)old_pointer))
+    {
+        unreachable("Attempt to release untracked pointer %p, this is likely a bug.", old_pointer);
+    }
+	self.inner_allocator.release(old_pointer, is_aligned);
+}
+
+fn void TrackingAllocator.clear(&self)
+{
+	self.map.clear();
+}
+
+fn void TrackingAllocator.print_report(&self) => self.fprint_report(io::stdout())!!;
+
+fn void! TrackingAllocator.fprint_report(&self, OutStream out)
+{
+
+	usz total = 0;
+	usz entries = 0;
+	bool leaks = false;
+	@pool()
+	{
+		Allocation[] allocs = self.map.value_tlist();
+		if (allocs.len)
+		{
+			if (!allocs[0].backtrace[0])
+			{
+				io::fprintn(out, "======== Memory Report ========")!;
+            	io::fprintn(out, "Size in bytes   Address")!;
+            	foreach (i, &allocation : allocs)
+            	{
+            	    entries++;
+            		total += allocation.size;
+            		io::fprintfn(out, "%13s   %p", allocation.size, allocation.ptr)!;
+            	}
+            	io::fprintn(out, "===============================")!;
+
+			}
+			else
+			{
+				io::fprintn(out, "================================== Memory Report ==================================")!;
+				io::fprintn(out, "Size in bytes   Address          Function                       ")!;
+				foreach (i, &allocation : allocs)
+				{
+					entries++;
+					total += allocation.size;
+					BacktraceList backtraces = {};
+					Backtrace trace = backtrace::BACKTRACE_UNKNOWN;
+					if (allocation.backtrace[3])
+					{
+						trace = backtrace::symbolize_backtrace(allocation.backtrace[3:1], allocator::temp()).get(0) ?? backtrace::BACKTRACE_UNKNOWN;
+					}
+					if (trace.function.len)	leaks = true;
+					io::fprintfn(out, "%13s   %p   %s:%d", allocation.size,
+						allocation.ptr, trace.function.len ? trace.function : "???",
+						trace.line ? trace.line : 0)!;
+				}
+				io::fprintn(out, "===================================================================================")!;
+			}
+		}
+		else
+		{
+			io::fprintn(out, "* NO ALLOCATIONS FOUND *")!;
+		}
+		io::fprintfn(out, "- Total currently allocated memory:            %d", total)!;
+        io::fprintfn(out, "- Total current allocations:                   %d", entries)!;
+        io::fprintfn(out, "- Total allocations (freed and retained):      %d", self.allocs_total)!;
+        io::fprintfn(out, "- Total allocated memory (freed and retained): %d", self.mem_total)!;
+        if (leaks)
+        {
+			io::fprintn(out)!;
+			io::fprintn(out, "Full leak report:")!;
+            foreach (i, &allocation : allocs)
+            {
+                if (!allocation.backtrace[3])
+                {
+                    io::fprintfn(out, "Allocation %d (%d bytes) - no backtrace available.", i + 1, allocation.size)!;
+                	continue;
+				}
+                BacktraceList backtraces = {};
+                usz end = MAX_BACKTRACE;
+                foreach (j, val : allocation.backtrace)
+                {
+                    if (!val)
+                	{
+                	    end = j;
+                	    break;
+					}
+				}
+        		BacktraceList list = backtrace::symbolize_backtrace(allocation.backtrace[3..(end - 1)], allocator::temp())!;
+                io::fprintfn(out, "Allocation %d (%d bytes): ", i + 1, allocation.size)!;
+        		foreach (trace : list)
+        		{
+        			if (trace.has_file())
+                	{
+                	    io::fprintfn(out, "   %s (in %s:%d)", trace.function, trace.file, trace.line);
+                	    continue;
+                	}
+                	if (trace.is_unknown())
+                	{
+                	    io::fprintfn(out, "   ??? (in unknown)");
+                	    continue;
+                	}
+					io::fprintfn(out, "   %s (source unavailable)", trace.function);
+				}
+        	}
+        }
+	};
+
+}

lib/std/core/array.c3

@@ -0,0 +1,147 @@
+module std::core::array;
+import std::core::array::slice;
+
+/**
+ * @param [in] array
+ * @param [in] element
+ * @return "the first index of the element"
+ * @return! SearchResult.MISSING
+ **/
+macro index_of(array, element)
+{
+	foreach (i, &e : array)
+	{
+		if (*e == element) return i;
+	}
+	return SearchResult.MISSING?;
+}
+
+/**
+ * @require @typekind(array) == VECTOR || @typekind(array) == ARRAY
+ * @require @typekind(array[0]) == VECTOR || @typekind(array[0]) == ARRAY
+ **/
+macro slice2d(array, x = 0, xlen = 0, y = 0, ylen = 0)
+{
+	if (xlen < 1) xlen = $typeof(array[0]).len + xlen;
+	if (ylen < 1) ylen = $typeof(array).len + ylen;
+	var $ElementType = $typeof(array[0][0]);
+	return Slice2d(<$ElementType>) { ($ElementType*)&array, $typeof(array[0]).len, y, ylen, x, xlen };
+}
+
+
+/**
+ * @param [in] array
+ * @param [in] element
+ * @return "the last index of the element"
+ * @return! SearchResult.MISSING
+ **/
+macro rindex_of(array, element)
+{
+	foreach_r (i, &e : array)
+	{
+		if (*e == element) return i;
+	}
+	return SearchResult.MISSING?;
+}
+
+/**
+ * Concatenate two arrays or slices, returning a slice containing the concatenation of them.
+ *
+ * @param [in] arr1
+ * @param [in] arr2
+ * @param [&inout] allocator "The allocator to use, default is the heap allocator"
+ * @require @typekind(arr1) == SLICE || @typekind(arr1) == ARRAY
+ * @require @typekind(arr2) == SLICE || @typekind(arr2) == ARRAY
+ * @require @typeis(arr1[0], $typeof(arr2[0])) "Arrays must have the same type"
+ * @ensure result.len == arr1.len + arr2.len
+ **/
+macro concat_new(arr1, arr2, Allocator allocator = allocator::heap())
+{
+	var $Type = $typeof(arr1[0]);
+	$Type[] result = allocator::alloc_array(allocator, $Type, arr1.len + arr2.len);
+	if (arr1.len > 0)
+	{
+		mem::copy(result.ptr, &arr1[0], arr1.len * $Type.sizeof, $Type.alignof, $Type.alignof);
+	}
+	if (arr2.len > 0)
+	{
+		mem::copy(&result[arr1.len], &arr2[0], arr2.len * $Type.sizeof, $Type.alignof, $Type.alignof);
+	}
+	return result;
+}
+
+/**
+ * Concatenate two arrays or slices, returning a slice containing the concatenation of them,
+ * allocated using the temp allocator.
+ *
+ * @param [in] arr1
+ * @param [in] arr2
+ * @require @typekind(arr1) == SLICE || @typekind(arr1) == ARRAY
+ * @require @typekind(arr2) == SLICE || @typekind(arr2) == ARRAY
+ * @require @typeis(arr1[0], $typeof(arr2[0])) "Arrays must have the same type"
+ * @ensure result.len == arr1.len + arr2.len
+ **/
+macro tconcat(arr1, arr2) => concat(arr1, arr2, allocator::temp());
+
+module std::core::array::slice(<Type>);
+
+struct Slice2d
+{
+	Type* ptr;
+	usz inner_len;
+	usz ystart;
+	usz ylen;
+	usz xstart;
+	usz xlen;
+}
+
+fn usz Slice2d.len(&self) @operator(len)
+{
+	return self.ylen;
+}
+
+fn usz Slice2d.count(&self)
+{
+	return self.ylen * self.xlen;
+}
+
+macro void Slice2d.@each(&self; @body(usz[<2>], Type))
+{
+	foreach (y, line : *self)
+	{
+		foreach (x, val : line)
+		{
+			@body({ x, y }, val);
+		}
+	}
+}
+
+macro void Slice2d.@each_ref(&self; @body(usz[<2>], Type*))
+{
+	foreach (y, line : *self)
+	{
+		foreach (x, &val : line)
+		{
+			@body({ x, y }, val);
+		}
+	}
+}
+
+/**
+ * @require idy >= 0 && idy < self.ylen
+ **/
+macro Type[] Slice2d.get(self, usz idy) @operator([])
+{
+	return (self.ptr + self.inner_len * (idy + self.ystart))[self.xstart:self.xlen];
+}
+
+/**
+ * @require y >= 0 && y < self.ylen
+ * @require x >= 0 && x < self.xlen
+ **/
+fn Slice2d Slice2d.slice(&self, isz x = 0, isz xlen = 0, isz y = 0, isz ylen = 0)
+{
+	if (xlen < 1) xlen = self.xlen + xlen;
+	if (ylen < 1) ylen = self.ylen + ylen;
+	return { self.ptr, self.inner_len, y + self.ystart, ylen, x + self.xstart, xlen };
+}

lib/std/core/bitorder.c3

@@ -0,0 +1,180 @@
+// Copyright (c) 2023 Christoffer Lerno and contributors. All rights reserved.
+// Use of this source code is governed by the MIT license
+// a copy of which can be found in the LICENSE_STDLIB file.
+module std::core::bitorder;
+
+// This module contains types of different endianness.
+// *BE types represent big-endian types
+// *LE types represent little-endian types.
+
+bitstruct ShortBE : short @bigendian
+{
+	short val : 0..15;
+}
+
+bitstruct UShortBE : ushort @bigendian
+{
+	ushort val : 0..15;
+}
+
+bitstruct IntBE : int @bigendian
+{
+	int val : 0..31;
+}
+
+bitstruct UIntBE : int @bigendian
+{
+	uint val : 0..31;
+}
+
+bitstruct LongBE : long @bigendian
+{
+	long val : 0..63;
+}
+
+bitstruct ULongBE : ulong @bigendian
+{
+	ulong val : 0..63;
+}
+
+bitstruct Int128BE : int128 @bigendian
+{
+	int128 val : 0..127;
+}
+
+bitstruct UInt128BE : uint128 @bigendian
+{
+	uint128 val : 0..127;
+}
+
+bitstruct ShortLE : short @littleendian
+{
+	short val : 0..15;
+}
+
+bitstruct UShortLE : ushort @littleendian
+{
+	ushort val : 0..15;
+}
+
+bitstruct IntLE : int @littleendian
+{
+	int val : 0..31;
+}
+
+bitstruct UIntLE : int @littleendian
+{
+	uint val : 0..31;
+}
+
+bitstruct LongLE : long @littleendian
+{
+	long val : 0..63;
+}
+
+bitstruct ULongLE : ulong @littleendian
+{
+	ulong val : 0..63;
+}
+
+bitstruct Int128LE : int128 @littleendian
+{
+	int128 val : 0..127;
+}
+
+bitstruct UInt128LE : uint128 @littleendian
+{
+	uint128 val : 0..127;
+}
+
+/**
+ * @require is_array_or_slice_of_char(bytes) "argument must be an array, a pointer to an array or a slice of char"
+ * @require is_bitorder($Type) "type must be a bitorder integer"
+ **/
+macro read(bytes, $Type)
+{
+	char[] s;
+	$switch (@typekind(bytes))
+	$case POINTER:
+		s = (*bytes)[:$Type.sizeof];
+	$default:
+		s = bytes[:$Type.sizeof];
+	$endswitch
+	return bitcast(*(char[$Type.sizeof]*)s.ptr, $Type).val;
+}
+
+/**
+ * @require is_arrayptr_or_slice_of_char(bytes) "argument must be a pointer to an array or a slice of char"
+ * @require is_bitorder($Type) "type must be a bitorder integer"
+ **/
+macro write(x, bytes, $Type)
+{
+	char[] s;
+	$switch (@typekind(bytes))
+	$case POINTER:
+		s = (*bytes)[:$Type.sizeof];
+	$default:
+		s = bytes[:$Type.sizeof];
+	$endswitch
+	*($typeof(x)*)s.ptr = bitcast(x, $Type).val;
+}
+
+macro is_bitorder($Type)
+{
+	$switch ($Type)
+	$case UShortLE:
+	$case ShortLE:
+	$case UIntLE:
+	$case IntLE:
+	$case ULongLE:
+	$case LongLE:
+	$case UInt128LE:
+	$case Int128LE:
+	$case UShortBE:
+	$case ShortBE:
+	$case UIntBE:
+	$case IntBE:
+	$case ULongBE:
+	$case LongBE:
+	$case UInt128BE:
+	$case Int128BE:
+		return true;
+	$default:
+		return false;
+	$endswitch
+}
+
+macro bool is_array_or_slice_of_char(bytes)
+{
+	$switch (@typekind(bytes))
+	$case POINTER:
+		var $Inner = $typefrom($typeof(bytes).inner);
+		$if $Inner.kindof == ARRAY:
+			var $Inner2 = $typefrom($Inner.inner);
+			return $Inner2.typeid == char.typeid;
+		$endif
+	$case ARRAY:
+	$case SLICE:
+		var $Inner = $typefrom($typeof(bytes).inner);
+		return $Inner.typeid == char.typeid;
+	$default:
+		return false;
+	$endswitch
+}
+
+macro bool is_arrayptr_or_slice_of_char(bytes)
+{
+	$switch (@typekind(bytes))
+	$case POINTER:
+		var $Inner = $typefrom($typeof(bytes).inner);
+		$if $Inner.kindof == ARRAY:
+			var $Inner2 = $typefrom($Inner.inner);
+			return $Inner2.typeid == char.typeid;
+		$endif
+	$case SLICE:
+		var $Inner = $typefrom($typeof(bytes).inner);
+		return $Inner.typeid == char.typeid;
+	$default:
+		return false;
+	$endswitch
+}

lib/std/core/builtin.c3

@@ -0,0 +1,684 @@
+// Copyright (c) 2021-2024 Christoffer Lerno and contributors. All rights reserved.
+// Use of this source code is governed by the MIT license
+// a copy of which can be found in the LICENSE_STDLIB file.
+module std::core::builtin;
+import libc, std::hash, std::io, std::os::backtrace;
+
+/**
+ * Use `IteratorResult` when reading the end of an iterator, or accessing a result out of bounds.
+ **/
+fault IteratorResult { NO_MORE_ELEMENT }
+
+/**
+ * Use `SearchResult` when trying to return a value from some collection but the element is missing.
+ **/
+fault SearchResult { MISSING }
+
+/**
+ * Use `CastResult` when an attempt at conversion fails.
+ **/
+fault CastResult { TYPE_MISMATCH }
+
+def VoidFn = fn void();
+
+/**
+ * Stores a variable on the stack, then restores it at the end of the
+ * macro scope.
+ *
+ * @param variable `the variable to store and restore`
+ **/
+macro void @scope(&variable; @body) @builtin
+{
+	var temp = *variable;
+	defer *variable = temp;
+	@body();
+}
+
+/**
+ * Swap two variables
+ * @require $assignable(*b, $typeof(*a)) && $assignable(*a, $typeof(*b))
+ **/
+macro void @swap(&a, &b) @builtin
+{
+	var temp = *a;
+	*a = *b;
+	*b = temp;
+}
+
+/**
+ * Convert an `any` type to a type, returning an failure if there is a type mismatch.
+ *
+ * @param v `the any to convert to the given type.`
+ * @param $Type `the type to convert to`
+ * @return `The any.ptr converted to its type.`
+ * @ensure @typeis(return, $Type*)
+ * @return! CastResult.TYPE_MISMATCH
+ **/
+macro anycast(any v, $Type) @builtin
+{
+	if (v.type != $Type.typeid) return CastResult.TYPE_MISMATCH?;
+	return ($Type*)v.ptr;
+}
+
+fn bool print_backtrace(String message, int backtraces_to_ignore) @if(env::NATIVE_STACKTRACE)
+{
+	@pool()
+	{
+		void*[256] buffer;
+		void*[] backtraces = backtrace::capture_current(&buffer);
+		backtraces_to_ignore++;
+		BacktraceList! backtrace = backtrace::symbolize_backtrace(backtraces, allocator::temp());
+		if (catch backtrace) return false;
+		if (backtrace.len() <= backtraces_to_ignore) return false;
+		io::eprint("\nERROR: '");
+        io::eprint(message);
+        io::eprintn("'");
+		foreach (i, &trace : backtrace)
+        {
+            if (i < backtraces_to_ignore) continue;
+            String inline_suffix = trace.is_inline ? " [inline]" : "";
+            if (trace.is_unknown())
+            {
+				io::eprintfn("  in ???%s", inline_suffix);
+				continue;
+			}
+			if (trace.has_file())
+            {
+                io::eprintfn("  in %s (%s:%d) [%s]%s", trace.function, trace.file, trace.line, trace.object_file, inline_suffix);
+                continue;
+            }
+            io::eprintfn("  in %s (source unavailable) [%s]%s", trace.function, trace.object_file, inline_suffix);
+		}
+		return true;
+	};
+}
+fn void default_panic(String message, String file, String function, uint line) @if(env::NATIVE_STACKTRACE)
+{
+	$if $defined(io::stderr):
+	if (!print_backtrace(message, 2))
+	{
+		io::eprintfn("\nERROR: '%s', in %s (%s:%d)", message, function, file, line);
+		return;
+	}
+	$endif
+	$$trap();
+
+}
+fn void default_panic(String message, String file, String function, uint line) @if(!env::NATIVE_STACKTRACE)
+{
+	$if $defined(io::stderr):
+	io::eprint("\nERROR: '");
+	io::eprint(message);
+	io::eprintfn("', in %s (%s:%d)", function, file, line);
+	$endif
+	$$trap();
+}
+
+def PanicFn = fn void(String message, String file, String function, uint line);
+
+PanicFn panic = &default_panic;
+
+fn void panicf(String fmt, String file, String function, uint line, args...)
+{
+	@stack_mem(512; Allocator allocator)
+	{
+		DString s;
+		s.new_init(.allocator = allocator);
+		s.appendf(fmt, ...args);
+		panic(s.str_view(), file, function, line);
+	};
+}
+
+/**
+ * Marks the path as unreachable. This will panic in safe mode, and in fast will simply be assumed
+ * never happens.
+ * @param [in] string "The panic message or format string"
+ **/
+macro void unreachable(String string = "Unreachable statement reached.", ...) @builtin @noreturn
+{
+	$if env::COMPILER_SAFE_MODE:
+	panicf(string, $$FILE, $$FUNC, $$LINE, $vasplat);
+	$endif;
+	$$unreachable();
+}
+
+/**
+ * Marks the path as unsupported, this is similar to unreachable.
+ * @param [in] string "The error message"
+ **/
+macro void unsupported(String string = "Unsupported function invoked") @builtin @noreturn
+{
+	panicf(string, $$FILE, $$FUNC, $$LINE, $vasplat);
+	$$unreachable();
+}
+
+/**
+ * Unconditionally break into an attached debugger when reached.
+ **/
+macro void breakpoint() @builtin
+{
+	$$breakpoint();
+}
+
+macro any_make(void* ptr, typeid type) @builtin
+{
+	return $$any_make(ptr, type);
+}
+
+macro any.retype_to(&self, typeid type)
+{
+	return $$any_make(self.ptr, type);
+}
+
+macro any.as_inner(&self)
+{
+	return $$any_make(self.ptr, self.type.inner);
+}
+
+/**
+ * @param expr "the expression to cast"
+ * @param $Type "the type to cast to"
+ *
+ * @require $sizeof(expr) == $Type.sizeof  "Cannot bitcast between types of different size."
+ * @ensure @typeis(return, $Type)
+ **/
+macro bitcast(expr, $Type) @builtin
+{
+	$if $Type.alignof <= $alignof(expr):
+		return *($Type*)&expr;
+	$else
+		$Type x @noinit;
+		$$memcpy(&x, &expr, $sizeof(expr), false, $Type.alignof, $alignof(expr));
+		return x;
+	$endif
+}
+
+/**
+ * @param $Type `The type of the enum`
+ * @param [in] enum_name `The name of the enum to search for`
+ * @require $Type.kindof == ENUM `Only enums may be used`
+ * @ensure @typeis(return, $Type)
+ * @return! SearchResult.MISSING
+ **/
+macro enum_by_name($Type, String enum_name) @builtin
+{
+	typeid x = $Type.typeid;
+	foreach (i, name : x.names)
+	{
+		if (name == enum_name) return ($Type)i;
+	}
+	return SearchResult.MISSING?;
+}
+
+/**
+ * Mark an expression as likely to be true
+ *
+ * @param #value "expression to be marked likely"
+ * @param $probability "in the range 0 - 1"
+ * @require $probability >= 0 && $probability <= 1.0
+ **/
+macro bool @likely(bool #value, $probability = 1.0) @builtin
+{
+    $switch
+    $case env::BUILTIN_EXPECT_IS_DISABLED:
+	return #value;
+    $case $probability == 1.0:
+	return $$expect(#value, true);
+    $default:
+	return $$expect_with_probability(#value, true, $probability);
+    $endswitch
+}
+
+/**
+ * Mark an expression as unlikely to be true
+ *
+ * @param #value "expression to be marked unlikely"
+ * @param $probability "in the range 0 - 1"
+ * @require $probability >= 0 && $probability <= 1.0
+ **/
+macro bool @unlikely(bool #value, $probability = 1.0) @builtin
+{
+    $switch
+    $case env::BUILTIN_EXPECT_IS_DISABLED:
+	return #value;
+    $case $probability == 1.0:
+	return $$expect(#value, false);
+    $default:
+	return $$expect_with_probability(#value, false, $probability);
+    $endswitch
+}
+
+/**
+ * @require values::@is_int(#value) || values::@is_bool(#value)
+ * @require $assignable(expected, $typeof(#value))
+ * @require $probability >= 0 && $probability <= 1.0
+ **/
+macro @expect(#value, expected, $probability = 1.0) @builtin
+{
+    $switch
+    $case env::BUILTIN_EXPECT_IS_DISABLED:
+	return #value == expected;
+    $case $probability == 1.0:
+	return $$expect(#value, ($typeof(#value))expected);
+    $default:
+	return $$expect_with_probability(#value, expected, $probability);
+    $endswitch
+}
+
+/**
+ * Locality for prefetch, levels 0 - 3, corresponding
+ * to "extremely local" to "no locality"
+ **/
+enum PrefetchLocality
+{
+	NO_LOCALITY,
+	FAR,
+	NEAR,
+	VERY_NEAR,
+}
+
+/**
+ * Prefetch a pointer.
+
+ * @param [in] ptr `Pointer to prefetch`
+ * @param $locality `Locality ranging from none to extremely local`
+ * @param $write `Prefetch for write, otherwise prefetch for read.`
+ **/
+macro @prefetch(void* ptr, PrefetchLocality $locality = VERY_NEAR, bool $write = false) @builtin
+{
+	$if !env::BUILTIN_PREFETCH_IS_DISABLED:
+		$$prefetch(ptr, $write ? 1 : 0, $locality.ordinal);
+	$endif
+}
+
+macro swizzle(v, ...) @builtin
+{
+	return $$swizzle(v, $vasplat);
+}
+
+macro swizzle2(v, v2, ...) @builtin
+{
+	return $$swizzle2(v, v2, $vasplat);
+}
+
+macro anyfault @catch(#expr) @builtin
+{
+	if (catch f = #expr) return f;
+	return anyfault {};
+}
+
+macro bool @ok(#expr) @builtin
+{
+	if (catch #expr) return false;
+	return true;
+}
+
+macro char[] @as_char_view(&value) @builtin
+{
+	return ((char*)value)[:$sizeof(*value)];
+}
+
+macro isz @str_find(String $string, String $needle) @builtin => $$str_find($string, $needle);
+macro String @str_upper(String $str) @builtin => $$str_upper($str);
+macro String @str_lower(String $str) @builtin => $$str_lower($str);
+macro uint @str_hash(String $str) @builtin => $$str_hash($str);
+
+macro uint int.hash(int i) => i;
+macro uint uint.hash(uint i) => i;
+macro uint short.hash(short s) => s;
+macro uint ushort.hash(ushort s) => s;
+macro uint char.hash(char c) => c;
+macro uint ichar.hash(ichar c) => c;
+macro uint long.hash(long i) => (uint)((i >> 32) ^ i);
+macro uint ulong.hash(ulong i) => (uint)((i >> 32) ^ i);
+macro uint int128.hash(int128 i) => (uint)((i >> 96) ^ (i >> 64) ^ (i >> 32) ^ i);
+macro uint uint128.hash(uint128 i) => (uint)((i >> 96) ^ (i >> 64) ^ (i >> 32) ^ i);
+macro uint bool.hash(bool b) => (uint)b;
+macro uint typeid.hash(typeid t) => ((ulong)(uptr)t).hash();
+macro uint String.hash(String c) => (uint)fnv32a::encode(c);
+macro uint char[].hash(char[] c) => (uint)fnv32a::encode(c);
+macro uint void*.hash(void* ptr) => ((ulong)(uptr)ptr).hash();
+
+distinct EmptySlot = void*;
+const EmptySlot EMPTY_MACRO_SLOT @builtin = null;
+macro @is_empty_macro_slot(#arg) @builtin => @typeis(#arg, EmptySlot);
+macro @is_valid_macro_slot(#arg) @builtin => !@typeis(#arg, EmptySlot);
+
+const MAX_FRAMEADDRESS = 128;
+/**
+ * @require n >= 0
+ **/
+macro void* get_frameaddress(int n)
+{
+	if (n > MAX_FRAMEADDRESS) return null;
+	switch (n)
+	{
+		case 0: return $$frameaddress(0);
+		case 1: return $$frameaddress(1);
+		case 2: return $$frameaddress(2);
+		case 3: return $$frameaddress(3);
+		case 4: return $$frameaddress(4);
+		case 5: return $$frameaddress(5);
+		case 6: return $$frameaddress(6);
+		case 7: return $$frameaddress(7);
+		case 8: return $$frameaddress(8);
+		case 9: return $$frameaddress(9);
+		case 10: return $$frameaddress(10);
+		case 11: return $$frameaddress(11);
+		case 12: return $$frameaddress(12);
+		case 13: return $$frameaddress(13);
+		case 14: return $$frameaddress(14);
+		case 15: return $$frameaddress(15);
+		case 16: return $$frameaddress(16);
+		case 17: return $$frameaddress(17);
+		case 18: return $$frameaddress(18);
+		case 19: return $$frameaddress(19);
+		case 20: return $$frameaddress(20);
+		case 21: return $$frameaddress(21);
+		case 22: return $$frameaddress(22);
+		case 23: return $$frameaddress(23);
+		case 24: return $$frameaddress(24);
+		case 25: return $$frameaddress(25);
+		case 26: return $$frameaddress(26);
+		case 27: return $$frameaddress(27);
+		case 28: return $$frameaddress(28);
+		case 29: return $$frameaddress(29);
+		case 30: return $$frameaddress(30);
+		case 31: return $$frameaddress(31);
+		case 32: return $$frameaddress(32);
+		case 33: return $$frameaddress(33);
+		case 34: return $$frameaddress(34);
+		case 35: return $$frameaddress(35);
+		case 36: return $$frameaddress(36);
+		case 37: return $$frameaddress(37);
+		case 38: return $$frameaddress(38);
+		case 39: return $$frameaddress(39);
+		case 40: return $$frameaddress(40);
+		case 41: return $$frameaddress(41);
+		case 42: return $$frameaddress(42);
+		case 43: return $$frameaddress(43);
+		case 44: return $$frameaddress(44);
+		case 45: return $$frameaddress(45);
+		case 46: return $$frameaddress(46);
+		case 47: return $$frameaddress(47);
+		case 48: return $$frameaddress(48);
+		case 49: return $$frameaddress(49);
+		case 50: return $$frameaddress(50);
+		case 51: return $$frameaddress(51);
+		case 52: return $$frameaddress(52);
+		case 53: return $$frameaddress(53);
+		case 54: return $$frameaddress(54);
+		case 55: return $$frameaddress(55);
+		case 56: return $$frameaddress(56);
+		case 57: return $$frameaddress(57);
+		case 58: return $$frameaddress(58);
+		case 59: return $$frameaddress(59);
+		case 60: return $$frameaddress(60);
+		case 61: return $$frameaddress(61);
+		case 62: return $$frameaddress(62);
+		case 63: return $$frameaddress(63);
+		case 64: return $$frameaddress(64);
+		case 65: return $$frameaddress(65);
+		case 66: return $$frameaddress(66);
+		case 67: return $$frameaddress(67);
+		case 68: return $$frameaddress(68);
+		case 69: return $$frameaddress(69);
+		case 70: return $$frameaddress(70);
+		case 71: return $$frameaddress(71);
+		case 72: return $$frameaddress(72);
+		case 73: return $$frameaddress(73);
+		case 74: return $$frameaddress(74);
+		case 75: return $$frameaddress(75);
+		case 76: return $$frameaddress(76);
+		case 77: return $$frameaddress(77);
+		case 78: return $$frameaddress(78);
+		case 79: return $$frameaddress(79);
+		case 80: return $$frameaddress(80);
+		case 81: return $$frameaddress(81);
+		case 82: return $$frameaddress(82);
+		case 83: return $$frameaddress(83);
+		case 84: return $$frameaddress(84);
+		case 85: return $$frameaddress(85);
+		case 86: return $$frameaddress(86);
+		case 87: return $$frameaddress(87);
+		case 88: return $$frameaddress(88);
+		case 89: return $$frameaddress(89);
+		case 90: return $$frameaddress(90);
+		case 91: return $$frameaddress(91);
+		case 92: return $$frameaddress(92);
+		case 93: return $$frameaddress(93);
+		case 94: return $$frameaddress(94);
+		case 95: return $$frameaddress(95);
+		case 96: return $$frameaddress(96);
+		case 97: return $$frameaddress(97);
+		case 98: return $$frameaddress(98);
+		case 99: return $$frameaddress(99);
+		case 100: return $$frameaddress(100);
+		case 101: return $$frameaddress(101);
+		case 102: return $$frameaddress(102);
+		case 103: return $$frameaddress(103);
+		case 104: return $$frameaddress(104);
+		case 105: return $$frameaddress(105);
+		case 106: return $$frameaddress(106);
+		case 107: return $$frameaddress(107);
+		case 108: return $$frameaddress(108);
+		case 109: return $$frameaddress(109);
+		case 110: return $$frameaddress(110);
+		case 111: return $$frameaddress(111);
+		case 112: return $$frameaddress(112);
+		case 113: return $$frameaddress(113);
+		case 114: return $$frameaddress(114);
+		case 115: return $$frameaddress(115);
+		case 116: return $$frameaddress(116);
+		case 117: return $$frameaddress(117);
+		case 118: return $$frameaddress(118);
+		case 119: return $$frameaddress(119);
+		case 120: return $$frameaddress(120);
+		case 121: return $$frameaddress(121);
+		case 122: return $$frameaddress(122);
+		case 123: return $$frameaddress(123);
+		case 124: return $$frameaddress(124);
+		case 125: return $$frameaddress(125);
+		case 126: return $$frameaddress(126);
+		case 127: return $$frameaddress(127);
+		case 128: return $$frameaddress(128);
+		default: unreachable();
+	}
+}
+
+/**
+ * @require n >= 0
+ **/
+macro void* get_returnaddress(int n)
+{
+	if (n > MAX_FRAMEADDRESS) return null;
+	switch (n)
+	{
+		case 0: return $$returnaddress(0);
+		case 1: return $$returnaddress(1);
+		case 2: return $$returnaddress(2);
+		case 3: return $$returnaddress(3);
+		case 4: return $$returnaddress(4);
+		case 5: return $$returnaddress(5);
+		case 6: return $$returnaddress(6);
+		case 7: return $$returnaddress(7);
+		case 8: return $$returnaddress(8);
+		case 9: return $$returnaddress(9);
+		case 10: return $$returnaddress(10);
+		case 11: return $$returnaddress(11);
+		case 12: return $$returnaddress(12);
+		case 13: return $$returnaddress(13);
+		case 14: return $$returnaddress(14);
+		case 15: return $$returnaddress(15);
+		case 16: return $$returnaddress(16);
+		case 17: return $$returnaddress(17);
+		case 18: return $$returnaddress(18);
+		case 19: return $$returnaddress(19);
+		case 20: return $$returnaddress(20);
+		case 21: return $$returnaddress(21);
+		case 22: return $$returnaddress(22);
+		case 23: return $$returnaddress(23);
+		case 24: return $$returnaddress(24);
+		case 25: return $$returnaddress(25);
+		case 26: return $$returnaddress(26);
+		case 27: return $$returnaddress(27);
+		case 28: return $$returnaddress(28);
+		case 29: return $$returnaddress(29);
+		case 30: return $$returnaddress(30);
+		case 31: return $$returnaddress(31);
+		case 32: return $$returnaddress(32);
+		case 33: return $$returnaddress(33);
+		case 34: return $$returnaddress(34);
+		case 35: return $$returnaddress(35);
+		case 36: return $$returnaddress(36);
+		case 37: return $$returnaddress(37);
+		case 38: return $$returnaddress(38);
+		case 39: return $$returnaddress(39);
+		case 40: return $$returnaddress(40);
+		case 41: return $$returnaddress(41);
+		case 42: return $$returnaddress(42);
+		case 43: return $$returnaddress(43);
+		case 44: return $$returnaddress(44);
+		case 45: return $$returnaddress(45);
+		case 46: return $$returnaddress(46);
+		case 47: return $$returnaddress(47);
+		case 48: return $$returnaddress(48);
+		case 49: return $$returnaddress(49);
+		case 50: return $$returnaddress(50);
+		case 51: return $$returnaddress(51);
+		case 52: return $$returnaddress(52);
+		case 53: return $$returnaddress(53);
+		case 54: return $$returnaddress(54);
+		case 55: return $$returnaddress(55);
+		case 56: return $$returnaddress(56);
+		case 57: return $$returnaddress(57);
+		case 58: return $$returnaddress(58);
+		case 59: return $$returnaddress(59);
+		case 60: return $$returnaddress(60);
+		case 61: return $$returnaddress(61);
+		case 62: return $$returnaddress(62);
+		case 63: return $$returnaddress(63);
+		case 64: return $$returnaddress(64);
+		case 65: return $$returnaddress(65);
+		case 66: return $$returnaddress(66);
+		case 67: return $$returnaddress(67);
+		case 68: return $$returnaddress(68);
+		case 69: return $$returnaddress(69);
+		case 70: return $$returnaddress(70);
+		case 71: return $$returnaddress(71);
+		case 72: return $$returnaddress(72);
+		case 73: return $$returnaddress(73);
+		case 74: return $$returnaddress(74);
+		case 75: return $$returnaddress(75);
+		case 76: return $$returnaddress(76);
+		case 77: return $$returnaddress(77);
+		case 78: return $$returnaddress(78);
+		case 79: return $$returnaddress(79);
+		case 80: return $$returnaddress(80);
+		case 81: return $$returnaddress(81);
+		case 82: return $$returnaddress(82);
+		case 83: return $$returnaddress(83);
+		case 84: return $$returnaddress(84);
+		case 85: return $$returnaddress(85);
+		case 86: return $$returnaddress(86);
+		case 87: return $$returnaddress(87);
+		case 88: return $$returnaddress(88);
+		case 89: return $$returnaddress(89);
+		case 90: return $$returnaddress(90);
+		case 91: return $$returnaddress(91);
+		case 92: return $$returnaddress(92);
+		case 93: return $$returnaddress(93);
+		case 94: return $$returnaddress(94);
+		case 95: return $$returnaddress(95);
+		case 96: return $$returnaddress(96);
+		case 97: return $$returnaddress(97);
+		case 98: return $$returnaddress(98);
+		case 99: return $$returnaddress(99);
+		case 100: return $$returnaddress(100);
+		case 101: return $$returnaddress(101);
+		case 102: return $$returnaddress(102);
+		case 103: return $$returnaddress(103);
+		case 104: return $$returnaddress(104);
+		case 105: return $$returnaddress(105);
+		case 106: return $$returnaddress(106);
+		case 107: return $$returnaddress(107);
+		case 108: return $$returnaddress(108);
+		case 109: return $$returnaddress(109);
+		case 110: return $$returnaddress(110);
+		case 111: return $$returnaddress(111);
+		case 112: return $$returnaddress(112);
+		case 113: return $$returnaddress(113);
+		case 114: return $$returnaddress(114);
+		case 115: return $$returnaddress(115);
+		case 116: return $$returnaddress(116);
+		case 117: return $$returnaddress(117);
+		case 118: return $$returnaddress(118);
+		case 119: return $$returnaddress(119);
+		case 120: return $$returnaddress(120);
+		case 121: return $$returnaddress(121);
+		case 122: return $$returnaddress(122);
+		case 123: return $$returnaddress(123);
+		case 124: return $$returnaddress(124);
+		case 125: return $$returnaddress(125);
+		case 126: return $$returnaddress(126);
+		case 127: return $$returnaddress(127);
+		case 128: return $$returnaddress(128);
+		default: unreachable();
+	}
+}
+
+module std::core::builtin @if((env::LINUX || env::DARWIN) && env::COMPILER_SAFE_MODE && env::DEBUG_SYMBOLS);
+import libc, std::io;
+
+fn void sig_panic(String message)
+{
+	default_panic(message, "???", "???", 0);
+}
+
+SignalFunction old_bus_error;
+SignalFunction old_segmentation_fault;
+
+fn void sig_bus_error(CInt i)
+{
+	$if !env::NATIVE_STACKTRACE:
+		sig_panic("Illegal memory access.");
+	$else
+		$if $defined(io::stderr):
+			if (!print_backtrace("Illegal memory access.", 1))
+			{
+				io::eprintn("\nERROR: 'Illegal memory access'.");
+			}
+		$endif
+	$endif
+	$$trap();
+}
+
+fn void sig_segmentation_fault(CInt i)
+{
+	$if !env::NATIVE_STACKTRACE:
+		sig_panic("Out of bounds memory access.");
+	$else
+		$if $defined(io::stderr):
+			if (!print_backtrace("Out of bounds memory access.", 1))
+			{
+				io::eprintn("\nERROR: Memory error without backtrace, possible stack overflow.");
+			}
+		$endif
+	$endif
+	$$trap();
+}
+
+fn void install_signal_handler(CInt signal, SignalFunction func) @local
+{
+	SignalFunction old = libc::signal(signal, func);
+	// Restore
+	if ((iptr)old > 1024) libc::signal(signal, old);
+}
+
+// Clean this up
+fn void install_signal_handlers() @init(101) @local @if(env::BACKTRACE)
+{
+	install_signal_handler(libc::SIGBUS, &sig_bus_error);
+	install_signal_handler(libc::SIGSEGV, &sig_segmentation_fault);
+}
+

lib/std/core/builtin_comparison.c3

@@ -0,0 +1,128 @@
+// Copyright (c) 2021-2024 Christoffer Lerno and contributors. All rights reserved.
+// Use of this source code is governed by the MIT license
+// a copy of which can be found in the LICENSE_STDLIB file.
+module std::core::builtin;
+
+/**
+ * @require types::@comparable_value(a) && types::@comparable_value(b)
+ **/
+macro less(a, b) @builtin
+{
+	$switch
+	$case $defined(a.less):
+		return a.less(b);
+	$case $defined(a.compare_to):
+		return a.compare_to(b) < 0;
+	$default:
+		return a < b;
+	$endswitch
+}
+
+/**
+ * @require types::@comparable_value(a) && types::@comparable_value(b)
+ **/
+macro less_eq(a, b) @builtin
+{
+	$switch
+	$case $defined(a.less):
+		return !b.less(a);
+	$case $defined(a.compare_to):
+		return a.compare_to(b) <= 0;
+	$default:
+		return a <= b;
+	$endswitch
+}
+
+/**
+ * @require types::@comparable_value(a) && types::@comparable_value(b)
+ **/
+macro greater(a, b) @builtin
+{
+	$switch
+	$case $defined(a.less):
+		return b.less(a);
+	$case $defined(a.compare_to):
+		return a.compare_to(b) > 0;
+	$default:
+		return a > b;
+	$endswitch
+}
+
+/**
+ * @require types::@comparable_value(a) && types::@comparable_value(b)
+ **/
+macro int compare_to(a, b) @builtin
+{
+	$switch
+	$case $defined(a.compare_to):
+		return a.compare_to(b);
+	$case $defined(a.less):
+		return (int)b.less(a) - (int)a.less(b);
+	$default:
+		return (int)(a > b) - (int)(a < b);
+	$endswitch
+}
+/**
+ * @require types::@comparable_value(a) && types::@comparable_value(b)
+ **/
+macro greater_eq(a, b) @builtin
+{
+	$switch
+	$case $defined(a.less):
+		return !a.less(b);
+	$case $defined(a.compare_to):
+		return a.compare_to(b) >= 0;
+	$default:
+		return a >= b;
+	$endswitch
+}
+
+/**
+ * @require types::@equatable_value(a) && types::@equatable_value(b) `values must be equatable`
+ **/
+macro bool equals(a, b) @builtin
+{
+	$switch
+	$case $defined(a.equals, a.equals(b)):
+		return a.equals(b);
+	$case $defined(a.compare_to, a.compare_to(b)):
+		return a.compare_to(b) == 0;
+	$case $defined(a.less):
+		return !a.less(b) && !b.less(a);
+	$default:
+		return a == b;
+	$endswitch
+}
+
+macro min(x, ...) @builtin
+{
+	$if $vacount == 1:
+		return less(x, $vaarg[0]) ? x : $vaarg[0];
+	$else
+		var result = x;
+		$for (var $i = 0; $i < $vacount; $i++)
+			if (less($vaarg[$i], result))
+			{
+				 result = $vaarg[$i];
+			}
+		$endfor
+		return result;
+	$endif
+}
+
+macro max(x, ...) @builtin
+{
+	$if $vacount == 1:
+		return greater(x, $vaarg[0]) ? x : $vaarg[0];
+	$else
+		var result = x;
+		$for (var $i = 0; $i < $vacount; $i++)
+			if (greater($vaarg[$i], result))
+			{
+				 result = $vaarg[$i];
+			}
+		$endfor
+		return result;
+	$endif
+}
+

lib/std/core/cinterop.c3

@@ -0,0 +1,55 @@
+// Copyright (c) 2021-2024 Christoffer Lerno. All rights reserved.
+// Use of this source code is governed by the MIT license
+// a copy of which can be found in the LICENSE_STDLIB file.
+module std::core::cinterop;
+
+const C_INT_SIZE = $$C_INT_SIZE;
+const C_LONG_SIZE = $$C_LONG_SIZE;
+const C_SHORT_SIZE = $$C_SHORT_SIZE;
+const C_LONG_LONG_SIZE = $$C_LONG_LONG_SIZE;
+
+$assert C_SHORT_SIZE < 32;
+$assert C_INT_SIZE < 128;
+$assert C_LONG_SIZE < 128;
+$assert C_LONG_LONG_SIZE <= 128;
+$assert C_SHORT_SIZE <= C_INT_SIZE;
+$assert C_INT_SIZE <= C_LONG_SIZE;
+$assert C_LONG_SIZE <= C_LONG_LONG_SIZE;
+
+def CShort = $typefrom(signed_int_from_bitsize($$C_SHORT_SIZE));
+def CUShort = $typefrom(unsigned_int_from_bitsize($$C_SHORT_SIZE));
+def CInt = $typefrom(signed_int_from_bitsize($$C_INT_SIZE));
+def CUInt = $typefrom(unsigned_int_from_bitsize($$C_INT_SIZE));
+def CLong = $typefrom(signed_int_from_bitsize($$C_LONG_SIZE));
+def CULong = $typefrom(unsigned_int_from_bitsize($$C_LONG_SIZE));
+def CLongLong = $typefrom(signed_int_from_bitsize($$C_LONG_LONG_SIZE));
+def CULongLong = $typefrom(unsigned_int_from_bitsize($$C_LONG_LONG_SIZE));
+def CSChar = ichar;
+def CUChar = char;
+
+def CChar = $typefrom($$C_CHAR_IS_SIGNED ? ichar.typeid : char.typeid);
+
+// Helper macros
+macro typeid signed_int_from_bitsize(usz $bitsize) @private
+{
+	$switch ($bitsize)
+		$case 128: return int128.typeid;
+		$case 64: return long.typeid;
+		$case 32: return int.typeid;
+		$case 16: return short.typeid;
+		$case 8: return ichar.typeid;
+		$default: $error("Invalid bitsize");
+	$endswitch
+}
+
+macro typeid unsigned_int_from_bitsize(usz $bitsize) @private
+{
+	$switch ($bitsize)
+		$case 128: return uint128.typeid;
+		$case 64: return ulong.typeid;
+		$case 32: return uint.typeid;
+		$case 16: return ushort.typeid;
+		$case 8: return char.typeid;
+		$default: $error("Invalid bitsize");
+	$endswitch
+}

lib/std/core/conv.c3

@@ -0,0 +1,416 @@
+module std::core::string::conv;
+
+const uint UTF16_SURROGATE_OFFSET @private = 0x10000;
+const uint UTF16_SURROGATE_GENERIC_MASK @private = 0xF800;
+const uint UTF16_SURROGATE_GENERIC_VALUE @private = 0xD800;
+const uint UTF16_SURROGATE_MASK @private = 0xFC00;
+const uint UTF16_SURROGATE_CODEPOINT_MASK @private = 0x03FF;
+const uint UTF16_SURROGATE_BITS @private = 10;
+const uint UTF16_SURROGATE_LOW_VALUE @private = 0xDC00;
+const uint UTF16_SURROGATE_HIGH_VALUE @private = 0xD800;
+
+/**
+ * @param c `The utf32 codepoint to convert`
+ * @param [out] output `the resulting buffer`
+ **/
+fn usz! char32_to_utf8(Char32 c, char[] output)
+{
+	if (!output.len) return UnicodeResult.CONVERSION_FAILED?;
+	switch (true)
+	{
+		case c <= 0x7f:
+			output[0] = (char)c;
+			return 1;
+		case c <= 0x7ff:
+			if (output.len < 2) return UnicodeResult.CONVERSION_FAILED?;
+			output[0] = (char)(0xC0 | c >> 6);
+			output[1] = (char)(0x80 | (c & 0x3F));
+			return 2;
+		case c <= 0xffff:
+			if (output.len < 3) return UnicodeResult.CONVERSION_FAILED?;
+			output[0] = (char)(0xE0 | c >> 12);
+			output[1] = (char)(0x80 | (c >> 6 & 0x3F));
+			output[2] = (char)(0x80 | (c & 0x3F));
+			return 3;
+		case c <= 0x10ffff:
+			if (output.len < 4) return UnicodeResult.CONVERSION_FAILED?;
+			output[0] = (char)(0xF0 | c >> 18);
+			output[1] = (char)(0x80 | (c >> 12 & 0x3F));
+			output[2] = (char)(0x80 | (c >> 6 & 0x3F));
+			output[3] = (char)(0x80 | (c & 0x3F));
+			return 4;
+		default:
+			// 0x10FFFF and above is not defined.
+			return UnicodeResult.CONVERSION_FAILED?;
+	}
+}
+
+/**
+ * Convert a code pointer into 1-2 UTF16 characters.
+ *
+ * @param c `The character to convert.`
+ * @param [inout] output `the resulting UTF16 buffer to write to.`
+ **/
+fn void char32_to_utf16_unsafe(Char32 c, Char16** output)
+{
+	if (c < UTF16_SURROGATE_OFFSET)
+	{
+		(*output)++[0] = (Char16)c;
+		return;
+	}
+	c -= UTF16_SURROGATE_OFFSET;
+	Char16 low = (Char16)(UTF16_SURROGATE_LOW_VALUE | (c & UTF16_SURROGATE_CODEPOINT_MASK));
+	c >>= UTF16_SURROGATE_BITS;
+	Char16 high = (Char16)(UTF16_SURROGATE_HIGH_VALUE | (c & UTF16_SURROGATE_CODEPOINT_MASK));
+	(*output)++[0] = (Char16)high;
+	(*output)++[0] = (Char16)low;
+}
+
+/**
+ * Convert 1-2 UTF16 data points into UTF8.
+ *
+ * @param [in] ptr `The UTF16 data to convert.`
+ * @param [inout] available `amount of UTF16 data available.`
+ * @param [inout] output `the resulting utf8 buffer to write to.`
+ **/
+fn void! char16_to_utf8_unsafe(Char16 *ptr, usz *available, char** output)
+{
+	Char16 high = *ptr;
+	if (high & UTF16_SURROGATE_GENERIC_MASK != UTF16_SURROGATE_GENERIC_VALUE)
+	{
+		char32_to_utf8_unsafe(high, output);
+		*available = 1;
+		return;
+	}
+	// Low surrogate first is an error
+	if (high & UTF16_SURROGATE_MASK != UTF16_SURROGATE_HIGH_VALUE) return UnicodeResult.INVALID_UTF16?;
+
+	// Unmatched high surrogate is an error
+	if (*available == 1) return UnicodeResult.INVALID_UTF16?;
+
+	Char16 low = ptr[1];
+
+	// Unmatched high surrogate, invalid
+	if (low & UTF16_SURROGATE_MASK != UTF16_SURROGATE_LOW_VALUE) return UnicodeResult.INVALID_UTF16?;
+
+	// The high bits of the codepoint are the value bits of the high surrogate
+	// The low bits of the codepoint are the value bits of the low surrogate
+	Char32 uc = (high & UTF16_SURROGATE_CODEPOINT_MASK) << UTF16_SURROGATE_BITS
+				| (low & UTF16_SURROGATE_CODEPOINT_MASK) + UTF16_SURROGATE_OFFSET;
+	char32_to_utf8_unsafe(uc, output);
+	*available = 2;
+}
+/**
+ * @param c `The utf32 codepoint to convert`
+ * @param [inout] output `the resulting buffer`
+ **/
+fn usz char32_to_utf8_unsafe(Char32 c, char** output)
+{
+	switch
+	{
+		case c <= 0x7f:
+			(*output)++[0] = (char)c;
+			return 1;
+		case c <= 0x7ff:
+			(*output)++[0] = (char)(0xC0 | c >> 6);
+			(*output)++[0] = (char)(0x80 | (c & 0x3F));
+			return 2;
+		case c <= 0xffff:
+			(*output)++[0] = (char)(0xE0 | c >> 12);
+			(*output)++[0] = (char)(0x80 | (c >> 6 & 0x3F));
+			(*output)++[0] = (char)(0x80 | (c & 0x3F));
+			return 3;
+		default:
+			(*output)++[0] = (char)(0xF0 | c >> 18);
+			(*output)++[0] = (char)(0x80 | (c >> 12 & 0x3F));
+			(*output)++[0] = (char)(0x80 | (c >> 6 & 0x3F));
+			(*output)++[0] = (char)(0x80 | (c & 0x3F));
+			return 4;
+	}
+}
+
+/**
+ * @param [in] ptr `pointer to the first character to parse`
+ * @param [inout] size `Set to max characters to read, set to characters read`
+ * @return `the parsed 32 bit codepoint`
+ **/
+fn Char32! utf8_to_char32(char* ptr, usz* size)
+{
+	usz max_size = *size;
+	if (max_size < 1) return UnicodeResult.INVALID_UTF8?;
+	char c = (ptr++)[0];
+
+	if ((c & 0x80) == 0)
+	{
+		*size = 1;
+		return c;
+	}
+	if ((c & 0xE0) == 0xC0)
+	{
+		if (max_size < 2) return UnicodeResult.INVALID_UTF8?;
+		*size = 2;
+		Char32 uc = (c & 0x1F) << 6;
+		c = *ptr;
+		// Overlong sequence or invalid second.
+		if (!uc || c & 0xC0 != 0x80) return UnicodeResult.INVALID_UTF8?;
+		return uc + c & 0x3F;
+	}
+	if ((c & 0xF0) == 0xE0)
+	{
+		if (max_size < 3) return UnicodeResult.INVALID_UTF8?;
+		*size = 3;
+		Char32 uc = (c & 0x0F) << 12;
+		c = ptr++[0];
+		if (c & 0xC0 != 0x80) return UnicodeResult.INVALID_UTF8?;
+		uc += (c & 0x3F) << 6;
+		c = ptr++[0];
+		// Overlong sequence or invalid last
+		if (!uc || c & 0xC0 != 0x80) return UnicodeResult.INVALID_UTF8?;
+		return uc + c & 0x3F;
+	}
+	if (max_size < 4) return UnicodeResult.INVALID_UTF8?;
+	if ((c & 0xF8) != 0xF0) return UnicodeResult.INVALID_UTF8?;
+	*size = 4;
+	Char32 uc = (c & 0x07) << 18;
+	c = ptr++[0];
+	if (c & 0xC0 != 0x80) return UnicodeResult.INVALID_UTF8?;
+	uc += (c & 0x3F) << 12;
+	c = ptr++[0];
+	if (c & 0xC0 != 0x80) return UnicodeResult.INVALID_UTF8?;
+	uc += (c & 0x3F) << 6;
+	c = ptr++[0];
+	// Overlong sequence or invalid last
+	if (!uc || c & 0xC0 != 0x80) return UnicodeResult.INVALID_UTF8?;
+	return uc + c & 0x3F;
+}
+
+/**
+ * @param utf8 `An UTF-8 encoded slice of bytes`
+ * @return `the number of encoded code points`
+ **/
+fn usz utf8_codepoints(String utf8)
+{
+	usz len = 0;
+	foreach (char c : utf8)
+	{
+		if (c & 0xC0 != 0x80) len++;
+	}
+	return len;
+}
+
+/**
+ * Calculate the UTF8 length required to encode an UTF32 array.
+ * @param [in] utf32 `the utf32 data to calculate from`
+ * @return `the length of the resulting UTF8 array`
+ **/
+fn usz utf8len_for_utf32(Char32[] utf32)
+{
+	usz len = 0;
+	foreach (Char32 uc : utf32)
+	{
+		switch (true)
+		{
+			case uc <= 0x7f:
+				len++;
+			case uc <= 0x7ff:
+				len += 2;
+			case uc <= 0xffff:
+				len += 3;
+			default:
+				len += 4;
+		}
+	}
+	return len;
+}
+
+/**
+ * Calculate the UTF8 length required to encode an UTF16 array.
+ * @param [in] utf16 `the utf16 data to calculate from`
+ * @return `the length of the resulting UTF8 array`
+ **/
+fn usz utf8len_for_utf16(Char16[] utf16)
+{
+	usz len = 0;
+	usz len16 = utf16.len;
+	for (usz i = 0; i < len16; i++)
+	{
+		Char16 c = utf16[i];
+		if (c & UTF16_SURROGATE_GENERIC_MASK != UTF16_SURROGATE_GENERIC_VALUE)
+		{
+			if (c <= 0x7f)
+			{
+				len++;
+				continue;
+			}
+			if (c <= 0x7ff)
+			{
+				len += 2;
+				continue;
+			}
+			len += 3;
+			continue;
+		}
+		len += 4;
+	}
+	return len;
+}
+
+/**
+ * Calculate the UTF16 length required to encode a UTF8 array.
+ * @param utf8 `the utf8 data to calculate from`
+ * @return `the length of the resulting UTF16 array`
+ **/
+fn usz utf16len_for_utf8(String utf8)
+{
+	usz len = utf8.len;
+	usz len16 = 0;
+	for (usz i = 0; i < len; i++)
+	{
+		len16++;
+		char c = utf8[i];
+		if (c & 0x80 == 0) continue;
+		i++;
+		if (c & 0xE0 == 0xC0) continue;
+		i++;
+		if (c & 0xF0 == 0xE0) continue;
+		i++;
+		len16++;
+	}
+	return len16;
+}
+
+/**
+ * @param [in] utf32 `the UTF32 array to check the length for`
+ * @return `the required length of an UTF16 array to hold the UTF32 data.`
+ **/
+fn usz utf16len_for_utf32(Char32[] utf32)
+{
+	usz len = utf32.len;
+	foreach (Char32 uc : utf32)
+	{
+		if (uc >= UTF16_SURROGATE_OFFSET) len++;
+	}
+	return len;
+}
+
+/**
+ * Convert an UTF32 array to an UTF8 array.
+ *
+ * @param [in] utf32
+ * @param [out] utf8_buffer
+ * @return `the number of bytes written.`
+ **/
+fn usz! utf32to8(Char32[] utf32, char[] utf8_buffer)
+{
+	char[] buffer = utf8_buffer;
+	foreach (uc : utf32)
+	{
+		usz used = char32_to_utf8(uc, buffer) @inline!;
+		buffer = buffer[used..];
+	}
+	// Zero terminate if there is space.
+	if (buffer.len > 0) buffer[0] = 0;
+	return utf8_buffer.len - buffer.len;
+}
+
+/**
+ * Convert an UTF8 array to an UTF32 array.
+ *
+ * @param [in] utf8
+ * @param [out] utf32_buffer
+ * @return `the number of Char32s written.`
+ **/
+fn usz! utf8to32(String utf8, Char32[] utf32_buffer)
+{
+	usz len = utf8.len;
+	Char32* ptr = utf32_buffer.ptr;
+	usz len32 = 0;
+	usz buf_len = utf32_buffer.len;
+	for (usz i = 0; i < len;)
+	{
+		if (len32 == buf_len) return UnicodeResult.CONVERSION_FAILED?;
+		usz width = len - i;
+		Char32 uc = utf8_to_char32(&utf8[i], &width) @inline!;
+		i += width;
+		ptr[len32++] = uc;
+	}
+	// Zero terminate if possible
+	if (len32 + 1 < buf_len) ptr[len32] = 0;
+	return len32;
+}
+
+/**
+ * Copy an array of UTF16 data into an UTF8 buffer without bounds
+ * checking. This will assume the buffer is sufficiently large to hold
+ * the converted data.
+ *
+ * @param [in] utf16 `The UTF16 array containing the data to convert.`
+ * @param [out] utf8_buffer `the (sufficiently large) buffer to hold the UTF16 data.`
+ **/
+fn void! utf16to8_unsafe(Char16[] utf16, char* utf8_buffer)
+{
+	usz len16 = utf16.len;
+	for (usz i = 0; i < len16;)
+	{
+		usz available = len16 - i;
+		char16_to_utf8_unsafe(&utf16[i], &available, &utf8_buffer) @inline!;
+		i += available;
+	}
+}
+
+/**
+ * Copy an array of UTF8 data into an UTF32 buffer without bounds
+ * checking. This will assume the buffer is sufficiently large to hold
+ * the converted data.
+ *
+ * @param [in] utf8 `The UTF8 buffer containing the data to convert.`
+ * @param [out] utf32_buffer `the (sufficiently large) buffer to hold the UTF8 data.`
+ **/
+fn void! utf8to32_unsafe(String utf8, Char32* utf32_buffer)
+{
+	usz len = utf8.len;
+	for (usz i = 0; i < len;)
+	{
+		usz width = len - i;
+		Char32 uc = utf8_to_char32(&utf8[i], &width) @inline!;
+		i += width;
+		(utf32_buffer++)[0] = uc;
+	}
+}
+
+/**
+ * Copy an array of UTF8 data into an UTF16 buffer without bounds
+ * checking. This will assume the buffer is sufficiently large to hold
+ * the converted data.
+ *
+ * @param [in] utf8 `The UTF8 buffer containing the data to convert.`
+ * @param [out] utf16_buffer `the (sufficiently large) buffer to hold the UTF8 data.`
+ **/
+fn void! utf8to16_unsafe(String utf8, Char16* utf16_buffer)
+{
+	usz len = utf8.len;
+	for (usz i = 0; i < len;)
+	{
+		usz width = len - i;
+		Char32 uc = utf8_to_char32(&utf8[i], &width) @inline!;
+		char32_to_utf16_unsafe(uc, &utf16_buffer) @inline;
+		i += width;
+	}
+}
+
+/**
+ * Copy an array of UTF32 code points into an UTF8 buffer without bounds
+ * checking. This will assume the buffer is sufficiently large to hold
+ * the converted data.
+ *
+ * @param [in] utf32 `The UTF32 buffer containing the data to convert.`
+ * @param [out] utf8_buffer `the (sufficiently large) buffer to hold the UTF8 data.`
+ **/
+fn void utf32to8_unsafe(Char32[] utf32, char* utf8_buffer)
+{
+	char* start = utf8_buffer;
+	foreach (Char32 uc : utf32)
+	{
+		char32_to_utf8_unsafe(uc, &utf8_buffer) @inline;
+	}
+}

lib/std/core/dstring.c3

@@ -0,0 +1,529 @@
+module std::core::dstring;
+import std::io;
+
+distinct DString (OutStream) = void*;
+
+const usz MIN_CAPACITY @private = 16;
+
+/**
+ * @require !self.data() "String already initialized"
+ **/
+fn DString DString.new_init(&self, usz capacity = MIN_CAPACITY, Allocator allocator = allocator::heap())
+{
+	if (capacity < MIN_CAPACITY) capacity = MIN_CAPACITY;
+	StringData* data = allocator::alloc_with_padding(allocator, StringData, capacity)!!;
+	data.allocator = allocator;
+	data.len = 0;
+	data.capacity = capacity;
+	return *self = (DString)data;
+}
+
+/**
+ * @require !self.data() "String already initialized"
+ **/
+fn DString DString.temp_init(&self, usz capacity = MIN_CAPACITY)
+{
+	self.new_init(capacity, allocator::temp()) @inline;
+	return *self;
+}
+
+fn DString new_with_capacity(usz capacity, Allocator allocator = allocator::heap())
+{
+	return DString{}.new_init(capacity, allocator);
+}
+
+fn DString temp_with_capacity(usz capacity) => new_with_capacity(capacity, allocator::temp()) @inline;
+
+fn DString new(String c = "", Allocator allocator = allocator::heap())
+{
+	usz len = c.len;
+	StringData* data = (StringData*)new_with_capacity(len, allocator);
+	if (len)
+	{
+		data.len = len;
+		mem::copy(&data.chars, c.ptr, len);
+	}
+	return (DString)data;
+}
+
+fn DString temp_new(String s = "") => new(s, allocator::temp()) @inline;
+
+
+fn void DString.replace_char(self, char ch, char replacement)
+{
+	StringData* data = self.data();
+	foreach (&c : data.chars[:data.len])
+	{
+		if (*c == ch) *c = replacement;
+	}
+}
+
+fn void DString.replace(&self, String needle, String replacement)
+{
+	StringData* data = self.data();
+	usz needle_len = needle.len;
+	if (!data || data.len < needle_len) return;
+	usz replace_len = replacement.len;
+	if (needle_len == 1 && replace_len == 1)
+	{
+		self.replace_char(needle[0], replacement[0]);
+		return;
+	}
+	@pool(data.allocator) {
+		String str = self.tcopy_str();
+		self.clear();
+		usz len = str.len;
+		usz match = 0;
+		foreach (i, c : str)
+		{
+			if (c == needle[match])
+    	    {
+    	        match++;
+    		    if (match == needle_len)
+    		    {
+    			    self.append_chars(replacement);
+    			    match = 0;
+    			    continue;
+    		    }
+    		    continue;
+    		}
+    		if (match > 0)
+    		{
+    		    self.append_chars(str[i - match:match]);
+    		    match = 0;
+    		}
+    		self.append_char(c);
+    	}
+    	if (match > 0) self.append_chars(str[^match:match]);
+	};
+}
+
+fn DString DString.new_concat(self, DString b, Allocator allocator = allocator::heap())
+{
+	DString string;
+	string.new_init(self.len() + b.len(), allocator);
+	string.append(self);
+	string.append(b);
+	return string;
+}
+
+fn DString DString.temp_concat(self, DString b) => self.new_concat(b, allocator::temp());
+
+fn ZString DString.zstr_view(&self)
+{
+	StringData* data = self.data();
+	if (!data) return "";
+	if (data.capacity == data.len)
+	{
+		self.reserve(1);
+		data = self.data();
+		data.chars[data.len] = 0;
+	}
+	else if (data.chars[data.len] != 0)
+	{
+		data.chars[data.len] = 0;
+	}
+	return (ZString)&data.chars[0];
+}
+
+fn usz DString.capacity(self)
+{
+	if (!self) return 0;
+	return self.data().capacity;
+}
+
+fn usz DString.len(&self) @dynamic
+{
+	if (!*self) return 0;
+	return self.data().len;
+}
+
+/**
+ * @require new_size <= self.len()
+ */
+fn void DString.chop(self, usz new_size)
+{
+	if (!self) return;
+	self.data().len = new_size;
+}
+
+fn String DString.str_view(self)
+{
+	StringData* data = self.data();
+	if (!data) return "";
+	return (String)data.chars[:data.len];
+}
+
+fn void DString.append_utf32(&self, Char32[] chars)
+{
+	self.reserve(chars.len);
+	foreach (Char32 c : chars)
+	{
+		self.append_char32(c);
+	}
+}
+
+/**
+ * @require index < self.len()
+ **/
+fn void DString.set(self, usz index, char c)
+{
+	self.data().chars[index] = c;
+}
+
+fn void DString.append_repeat(&self, char c, usz times)
+{
+	if (times == 0) return;
+	self.reserve(times);
+	StringData* data = self.data();
+	for (usz i = 0; i < times; i++)
+	{
+		data.chars[data.len++] = c;
+	}
+}
+
+/**
+ * @require c <= 0x10ffff
+ */
+fn void DString.append_char32(&self, Char32 c)
+{
+	char[4] buffer @noinit;
+	char* p = &buffer;
+	usz n = conv::char32_to_utf8_unsafe(c, &p);
+	self.reserve(n);
+	StringData* data = self.data();
+	data.chars[data.len:n] = buffer[:n];
+	data.len += n;
+}
+
+fn DString DString.tcopy(&self) => self.copy(allocator::temp());
+
+fn DString DString.copy(self, Allocator allocator = null)
+{
+	if (!self)
+	{
+		if (allocator) return new_with_capacity(0, allocator);
+		return (DString)null;
+	}
+	StringData* data = self.data();
+	if (!allocator) allocator = allocator::heap();
+	DString new_string = new_with_capacity(data.capacity, allocator);
+	mem::copy((char*)new_string.data(), (char*)data, StringData.sizeof + data.len);
+	return new_string;
+}
+
+fn ZString DString.copy_zstr(self, Allocator allocator = allocator::heap())
+{
+	usz str_len = self.len();
+	if (!str_len)
+	{
+		return (ZString)allocator::calloc(allocator, 1);
+	}
+	char* zstr = allocator::malloc(allocator, str_len + 1);
+	StringData* data = self.data();
+	mem::copy(zstr, &data.chars, str_len);
+	zstr[str_len] = 0;
+	return (ZString)zstr;
+}
+
+fn String DString.copy_str(self, Allocator allocator = allocator::heap())
+{
+	return (String)self.copy_zstr(allocator)[:self.len()];
+}
+
+fn String DString.tcopy_str(self) => self.copy_str(allocator::temp()) @inline;
+
+fn bool DString.equals(self, DString other_string)
+{
+	StringData *str1 = self.data();
+	StringData *str2 = other_string.data();
+	if (str1 == str2) return true;
+	if (!str1) return str2.len == 0;
+	if (!str2) return str1.len == 0;
+	usz str1_len = str1.len;
+	if (str1_len != str2.len) return false;
+	for (int i = 0; i < str1_len; i++)
+	{
+		if (str1.chars[i] != str2.chars[i]) return false;
+	}
+	return true;
+}
+
+fn void DString.free(&self)
+{
+	if (!*self) return;
+	StringData* data = self.data();
+	if (!data) return;
+	allocator::free(data.allocator, data);
+	*self = (DString)null;
+}
+
+fn bool DString.less(self, DString other_string)
+{
+	StringData* str1 = self.data();
+	StringData* str2 = other_string.data();
+	if (str1 == str2) return false;
+	if (!str1) return str2.len != 0;
+	if (!str2) return str1.len == 0;
+	usz str1_len = str1.len;
+	usz str2_len = str2.len;
+	if (str1_len != str2_len) return str1_len < str2_len;
+	for (int i = 0; i < str1_len; i++)
+	{
+		if (str1.chars[i] >= str2.chars[i]) return false;
+	}
+	return true;
+}
+
+fn void DString.append_chars(&self, String str)
+{
+	usz other_len = str.len;
+	if (!other_len) return;
+	if (!*self)
+	{
+		*self = new(str);
+		return;
+	}
+	self.reserve(other_len);
+	StringData* data = self.data();
+	mem::copy(&data.chars[data.len], str.ptr, other_len);
+	data.len += other_len;
+}
+
+fn Char32[] DString.copy_utf32(&self, Allocator allocator = allocator::heap())
+{
+	return self.str_view().to_new_utf32(allocator) @inline!!;
+}
+
+fn void DString.append_string(&self, DString str)
+{
+	StringData* other = str.data();
+	if (!other) return;
+	self.append(str.str_view());
+}
+
+fn void DString.clear(self)
+{
+	if (!self) return;
+	self.data().len = 0;
+}
+
+fn usz! DString.write(&self, char[] buffer) @dynamic
+{
+	self.append_chars((String)buffer);
+	return buffer.len;
+}
+
+fn void! DString.write_byte(&self, char c) @dynamic
+{
+	self.append_char(c);
+}
+
+fn void DString.append_char(&self, char c)
+{
+	if (!*self)
+	{
+		*self = new_with_capacity(MIN_CAPACITY);
+	}
+	self.reserve(1);
+	StringData* data = self.data();
+	data.chars[data.len++] = c;
+}
+
+/**
+ * @require start < self.len()
+ * @require end < self.len()
+ * @require end >= start "End must be same or equal to the start"
+ **/
+fn void DString.delete_range(&self, usz start, usz end)
+{
+	self.delete(start, end - start + 1);
+}
+
+/**
+ * @require start < self.len()
+ * @require start + len <= self.len()
+ **/
+fn void DString.delete(&self, usz start, usz len = 1)
+{
+	if (!len) return;
+	StringData* data = self.data();
+	usz new_len = data.len - len;
+	if (new_len == 0)
+	{
+		data.len = 0;
+		return;
+	}
+	usz len_after = data.len - start - len;
+	if (len_after > 0)
+	{
+		data.chars[start:len_after] = data.chars[start + len:len_after];
+	}
+	data.len = new_len;
+}
+
+macro void DString.append(&self, value)
+{
+	var $Type = $typeof(value);
+	$switch ($Type)
+		$case char:
+		$case ichar:
+			self.append_char(value);
+		$case DString:
+			self.append_string(value);
+		$case String:
+			self.append_chars(value);
+		$case Char32:
+			self.append_char32(value);
+		$default:
+			$switch
+				$case $defined((Char32)value):
+					self.append_char32((Char32)value);
+				$case $defined((String)value):
+					self.append_chars((String)value);
+				$default:
+					$error "Unsupported type for append – use appendf instead.";
+			$endswitch
+	$endswitch
+}
+
+fn void DString.insert_at(&self, usz index, String s)
+{
+	if (s.len == 0) return;
+	self.reserve(s.len);
+	StringData* data = self.data();
+	usz len = self.len();
+	if (data.chars[:len].ptr == s.ptr)
+	{
+		// Source and destination are the same: nothing to do.
+		return;
+	}
+	index = min(index, len);
+	data.len += s.len;
+
+	char* start = data.chars[index:s.len].ptr; // area to insert into
+	mem::move(start + s.len, start, len - index); // move existing data
+	switch
+	{
+		case s.ptr <= start && start < s.ptr + s.len:
+			// Overlapping areas.
+			foreach_r (i, c : s)
+			{
+				data.chars[index + i] = c;
+			}
+		case start <= s.ptr && s.ptr < start + len:
+			// Source has moved.
+			mem::move(start, s.ptr + s.len, s.len);
+		default:
+			mem::move(start, s, s.len);
+	}
+}
+
+fn usz! DString.appendf(&self, String format, args...) @maydiscard
+{
+	if (!self.data()) self.new_init(format.len + 20);
+	@pool(self.data().allocator)
+	{
+		Formatter formatter;
+		formatter.init(&out_string_append_fn, self);
+		return formatter.vprintf(format, args);
+	};
+}
+
+fn usz! DString.appendfn(&self, String format, args...) @maydiscard
+{
+	if (!self.data()) self.new_init(format.len + 20);
+	@pool(self.data().allocator)
+	{
+		Formatter formatter;
+		formatter.init(&out_string_append_fn, self);
+		usz len = formatter.vprintf(format, args)!;
+		self.append('\n');
+		return len + 1;
+	};
+}
+
+fn DString new_join(String[] s, String joiner, Allocator allocator = allocator::heap())
+{
+	if (!s.len) return (DString)null;
+	usz total_size = joiner.len * s.len;
+	foreach (String* &str : s)
+	{
+		total_size += str.len;
+	}
+	DString res = new_with_capacity(total_size, allocator);
+	res.append(s[0]);
+	foreach (String* &str : s[1..])
+	{
+		res.append(joiner);
+		res.append(*str);
+	}
+	return res;
+}
+
+fn void! out_string_append_fn(void* data, char c) @private
+{
+	DString* s = data;
+	s.append_char(c);
+}
+
+
+fn StringData* DString.data(self) @inline @private
+{
+	return (StringData*)self;
+}
+
+fn void DString.reserve(&self, usz addition)
+{
+	StringData* data = self.data();
+	if (!data)
+	{
+		*self = dstring::new_with_capacity(addition);
+		return;
+	}
+	usz len = data.len + addition;
+	if (data.capacity >= len) return;
+	usz new_capacity = data.capacity * 2;
+	if (new_capacity < MIN_CAPACITY) new_capacity = MIN_CAPACITY;
+	while (new_capacity < len) new_capacity *= 2;
+	data.capacity = new_capacity;
+	*self = (DString)allocator::realloc(data.allocator, data, StringData.sizeof + new_capacity);
+}
+
+fn usz! DString.read_from_stream(&self, InStream reader)
+{
+	if (&reader.available)
+	{
+		usz total_read = 0;
+		while (usz available = reader.available()!)
+		{
+			self.reserve(available);
+			StringData* data = self.data();
+			usz len = reader.read(data.chars[data.len..(data.capacity - 1)])!;
+			total_read += len;
+			data.len += len;
+		}
+		return total_read;
+	}
+	usz total_read = 0;
+	while (true)
+	{
+		// Reserve at least 16 bytes
+		self.reserve(16);
+		StringData* data = self.data();
+		// Read into the rest of the buffer
+		usz read = reader.read(data.chars[data.len..(data.capacity - 1)])!;
+		data.len += read;
+		// Ok, we reached the end.
+		if (read < 16) return total_read;
+		// Otherwise go another round
+	}
+}
+
+struct StringData @private
+{
+	Allocator allocator;
+	usz len;
+	usz capacity;
+	char[*] chars;
+}

lib/std/core/env.c3

@@ -0,0 +1,194 @@
+// Copyright (c) 2021-2024 Christoffer Lerno. All rights reserved.
+// Use of this source code is governed by the MIT license
+// a copy of which can be found in the LICENSE_STDLIB file.
+module std::core::env;
+
+enum CompilerOptLevel
+{
+	O0,
+	O1,
+	O2,
+	O3
+}
+
+enum MemoryEnvironment
+{
+	NORMAL,
+	SMALL,
+	TINY,
+	NONE
+}
+
+enum OsType
+{
+	UNKNOWN,
+	NONE,
+	ANANAS,
+	CLOUD_ABI,
+	DRAGON_FLY,
+	FREEBSD,
+	FUCHSIA,
+	IOS,
+	KFREEBSD,
+	LINUX,
+	PS3,
+	MACOS,
+	NETBSD,
+	OPENBSD,
+	SOLARIS,
+	WIN32,
+	HAIKU,
+	MINIX,
+	RTEMS,
+	NACL,       // Native Client
+	CNK,        // BG/P Compute-Node Kernel
+	AIX,
+	CUDA,
+	NVOPENCL,
+	AMDHSA,
+	PS4,
+	ELFIAMCU,
+	TVOS,
+	WATCHOS,
+	MESA3D,
+	CONTIKI,
+	AMDPAL,
+	HERMITCORE,
+	HURD,
+	WASI,
+	EMSCRIPTEN,
+}
+
+enum ArchType
+{
+	UNKNOWN,
+	ARM,          // ARM (little endian): arm, armv.*, xscale
+	ARMB,         // ARM (big endian): armeb
+	AARCH64,      // AArch64 (little endian): aarch64
+	AARCH64_BE,   // AArch64 (big endian): aarch64_be
+	AARCH64_32,   // AArch64 (little endian) ILP32: aarch64_32
+	ARC,          // ARC: Synopsys ARC
+	AVR,          // AVR: Atmel AVR microcontroller
+	BPFEL,        // eBPF or extended BPF or 64-bit BPF (little endian)
+	BPFEB,        // eBPF or extended BPF or 64-bit BPF (big endian)
+	HEXAGON,      // Hexagon: hexagon
+	MIPS,         // MIPS: mips, mipsallegrex, mipsr6
+	MIPSEL,       // MIPSEL: mipsel, mipsallegrexe, mipsr6el
+	MIPS64,       // MIPS64: mips64, mips64r6, mipsn32, mipsn32r6
+	MIPS64EL,     // MIPS64EL: mips64el, mips64r6el, mipsn32el, mipsn32r6el
+	MSP430,       // MSP430: msp430
+	PPC,          // PPC: powerpc
+	PPC64,        // PPC64: powerpc64, ppu
+	PPC64LE,      // PPC64LE: powerpc64le
+	R600,         // R600: AMD GPUs HD2XXX - HD6XXX
+	AMDGCN,       // AMDGCN: AMD GCN GPUs
+	RISCV32,      // RISC-V (32-bit): riscv32
+	RISCV64,      // RISC-V (64-bit): riscv64
+	SPARC,        // Sparc: sparc
+	SPARCV9,      // Sparcv9: Sparcv9
+	SPARCEL,      // Sparc: (endianness = little). NB: 'Sparcle' is a CPU variant
+	SYSTEMZ,      // SystemZ: s390x
+	TCE,          // TCE (http://tce.cs.tut.fi/): tce
+	TCELE,        // TCE little endian (http://tce.cs.tut.fi/): tcele
+	THUMB,        // Thumb (little endian): thumb, thumbv.*
+	THUMBEB,      // Thumb (big endian): thumbeb
+	X86,          // X86: i[3-9]86
+	X86_64,       // X86-64: amd64, x86_64
+	XCORE,        // XCore: xcore
+	NVPTX,        // NVPTX: 32-bit
+	NVPTX64,      // NVPTX: 64-bit
+	LE32,         // le32: generic little-endian 32-bit CPU (PNaCl)
+	LE64,         // le64: generic little-endian 64-bit CPU (PNaCl)
+	AMDIL,        // AMDIL
+	AMDIL64,      // AMDIL with 64-bit pointers
+	HSAIL,        // AMD HSAIL
+	HSAIL64,      // AMD HSAIL with 64-bit pointers
+	SPIR,         // SPIR: standard portable IR for OpenCL 32-bit version
+	SPIR64,       // SPIR: standard portable IR for OpenCL 64-bit version
+	KALIMBA,      // Kalimba: generic kalimba
+	SHAVE,        // SHAVE: Movidius vector VLIW processors
+	LANAI,        // Lanai: Lanai 32-bit
+	WASM32,       // WebAssembly with 32-bit pointers
+	WASM64,       // WebAssembly with 64-bit pointers
+	RSCRIPT32,    // 32-bit RenderScript
+	RSCRIPT64,    // 64-bit RenderScript
+	XTENSA,       // Xtensa
+}
+
+const OsType OS_TYPE = (OsType)$$OS_TYPE;
+const ArchType ARCH_TYPE = (ArchType)$$ARCH_TYPE;
+const bool ARCH_32_BIT = $$REGISTER_SIZE == 32;
+const bool ARCH_64_BIT = $$REGISTER_SIZE == 64;
+const bool LIBC = $$COMPILER_LIBC_AVAILABLE;
+const bool NO_LIBC = !$$COMPILER_LIBC_AVAILABLE;
+const CompilerOptLevel COMPILER_OPT_LEVEL = (CompilerOptLevel)$$COMPILER_OPT_LEVEL;
+const bool BIG_ENDIAN = $$PLATFORM_BIG_ENDIAN;
+const bool I128_NATIVE_SUPPORT = $$PLATFORM_I128_SUPPORTED;
+const bool F16_SUPPORT = $$PLATFORM_F16_SUPPORTED;
+const bool F128_SUPPORT = $$PLATFORM_F128_SUPPORTED;
+const REGISTER_SIZE = $$REGISTER_SIZE;
+const bool COMPILER_SAFE_MODE = $$COMPILER_SAFE_MODE;
+const bool DEBUG_SYMBOLS = $$DEBUG_SYMBOLS;
+const bool BACKTRACE = $$BACKTRACE;
+const usz LLVM_VERSION = $$LLVM_VERSION;
+const bool BENCHMARKING = $$BENCHMARKING;
+const bool TESTING = $$TESTING;
+const MemoryEnvironment MEMORY_ENV = (MemoryEnvironment)$$MEMORY_ENVIRONMENT;
+const bool TRACK_MEMORY = DEBUG_SYMBOLS && (COMPILER_SAFE_MODE || TESTING);
+const bool X86_64 = ARCH_TYPE == X86_64;
+const bool X86 = ARCH_TYPE == X86;
+const bool AARCH64 = ARCH_TYPE == AARCH64;
+const bool NATIVE_STACKTRACE = LINUX || DARWIN || WIN32;
+const bool LINUX = LIBC && OS_TYPE == LINUX;
+const bool DARWIN = LIBC && os_is_darwin();
+const bool WIN32 = LIBC && OS_TYPE == WIN32;
+const bool POSIX = LIBC && os_is_posix();
+const bool OPENBSD = LIBC && OS_TYPE == OPENBSD;
+const bool FREEBSD = LIBC && OS_TYPE == FREEBSD;
+const bool NETBSD = LIBC && OS_TYPE == NETBSD;
+const bool WASI = LIBC && OS_TYPE == WASI;
+const bool WASM_NOLIBC @builtin = !LIBC && ARCH_TYPE == ArchType.WASM32 || ARCH_TYPE == ArchType.WASM64;
+const bool ADDRESS_SANITIZER = $$ADDRESS_SANITIZER;
+const bool MEMORY_SANITIZER = $$MEMORY_SANITIZER;
+const bool THREAD_SANITIZER = $$THREAD_SANITIZER;
+const bool ANY_SANITIZER = ADDRESS_SANITIZER || MEMORY_SANITIZER || THREAD_SANITIZER;
+
+macro bool os_is_darwin() @const
+{
+	$switch (OS_TYPE)
+	$case IOS:
+	$case MACOS:
+	$case TVOS:
+	$case WATCHOS:
+		return true;
+	$default:
+		return false;
+	$endswitch
+}
+
+macro bool os_is_posix() @const
+{
+	$switch (OS_TYPE)
+	$case IOS:
+	$case MACOS:
+	$case NETBSD:
+	$case LINUX:
+	$case KFREEBSD:
+	$case FREEBSD:
+	$case OPENBSD:
+	$case SOLARIS:
+	$case TVOS:
+	$case WATCHOS:
+		return true;
+	$case WIN32:
+	$case WASI:
+	$case EMSCRIPTEN:
+		return false;
+	$default:
+		$echo("Assuming non-Posix environment");
+		return false;
+	$endswitch
+}
+
+const BUILTIN_EXPECT_IS_DISABLED = $feature(DISABLE_BUILTIN_EXPECT);
+const BUILTIN_PREFETCH_IS_DISABLED = $feature(DISABLE_BUILTIN_PREFETCH);

lib/std/core/mem.c3

@@ -0,0 +1,753 @@
+// Copyright (c) 2021-2023 Christoffer Lerno. All rights reserved.
+// Use of this source code is governed by the MIT license
+// a copy of which can be found in the LICENSE_STDLIB file.
+module std::core::mem;
+import std::core::mem::allocator @public;
+import std::math;
+
+const MAX_MEMORY_ALIGNMENT = 0x1000_0000;
+const DEFAULT_MEM_ALIGNMENT = (void*.alignof) * 2;
+
+macro bool @constant_is_power_of_2($x) @const @private
+{
+	return $x != 0 && ($x & ($x - 1)) == 0;
+}
+
+/**
+ * Load a vector from memory according to a mask assuming default alignment.
+ *
+ * @param ptr "The pointer address to load from."
+ * @param mask "The mask for the load"
+ * @param passthru "The value to use for non masked values"
+ * @require $assignable(&&passthru, $typeof(ptr)) : "Pointer and passthru must match"
+ * @require @typekind(passthru) == VECTOR : "Expected passthru to be a vector"
+ * @require passthru.len == mask.len : "Mask and passthru must have the same length"
+ *
+ * @return "A vector with the loaded values where the mask is true, passthru where the mask is false"
+ **/
+macro masked_load(ptr, bool[<*>] mask, passthru)
+{
+	return $$masked_load(ptr, mask, passthru, 0);
+}
+
+/**
+ * Load a vector from memory according to a mask.
+ *
+ * @param ptr "The pointer address to load from."
+ * @param mask "The mask for the load"
+ * @param passthru "The value to use for non masked values"
+ * @param $alignment "The alignment to assume for the pointer"
+ *
+ * @require $assignable(&&passthru, $typeof(ptr)) : "Pointer and passthru must match"
+ * @require @typekind(passthru) == VECTOR : "Expected passthru to be a vector"
+ * @require passthru.len == mask.len : "Mask and passthru must have the same length"
+ * @require @constant_is_power_of_2($alignment) : "The alignment must be a power of two"
+ *
+ * @return "A vector with the loaded values where the mask is true, passthru where the mask is false"
+ **/
+macro @masked_load_aligned(ptr, bool[<*>] mask, passthru, usz $alignment)
+{
+	return $$masked_load(ptr, mask, passthru, $alignment);
+}
+
+/**
+ * Load values from a pointer vector, assuming default alignment.
+ *
+ * @param ptrvec "The vector of pointers to load from."
+ * @param mask "The mask for the load"
+ * @param passthru "The value to use for non masked values"
+ *
+ * @require @typekind(ptrvec) == VECTOR : "Expected ptrvec to be a vector"
+ * @require @typekind(passthru) == VECTOR : "Expected passthru to be a vector"
+ * @require $assignable(&&passthru[0], $typeof(ptrvec[0])) : "Pointer and passthru must match"
+ * @require passthru.len == mask.len : "Mask and passthru must have the same length"
+ * @require mask.len == ptrvec.len : "Mask and ptrvec must have the same length"
+ *
+ * @return "A vector with the loaded values where the mask is true, passthru where the mask is false"
+ **/
+macro gather(ptrvec, bool[<*>] mask, passthru)
+{
+	return $$gather(ptrvec, mask, passthru, 0);
+}
+
+
+/**
+ * Load values from a pointer vector.
+ *
+ * @param ptrvec "The vector of pointers to load from."
+ * @param mask "The mask for the load"
+ * @param passthru "The value to use for non masked values"
+ * @param $alignment "The alignment to assume for the pointers"
+ *
+ * @require @typekind(ptrvec) == VECTOR : "Expected ptrvec to be a vector"
+ * @require @typekind(passthru) == VECTOR : "Expected passthru to be a vector"
+ * @require $assignable(&&passthru[0], $typeof(ptrvec[0])) : "Pointer and passthru must match"
+ * @require passthru.len == mask.len : "Mask and passthru must have the same length"
+ * @require mask.len == ptrvec.len : "Mask and ptrvec must have the same length"
+ * @require @constant_is_power_of_2($alignment) : "The alignment must be a power of two"
+ *
+ * @return "A vector with the loaded values where the mask is true, passthru where the mask is false"
+ **/
+macro @gather_aligned(ptrvec, bool[<*>] mask, passthru, usz $alignment)
+{
+	return $$gather(ptrvec, mask, passthru, $alignment);
+}
+
+
+/**
+ * Store parts of a vector according to the mask, assuming default alignment.
+ *
+ * @param ptr "The pointer address to store to."
+ * @param value "The value to store masked"
+ * @param mask "The mask for the store"
+ *
+ * @require $assignable(&&value, $typeof(ptr)) : "Pointer and value must match"
+ * @require @typekind(value) == VECTOR : "Expected value to be a vector"
+ * @require value.len == mask.len : "Mask and value must have the same length"
+ **/
+macro masked_store(ptr, value, bool[<*>] mask)
+{
+	return $$masked_store(ptr, value, mask, 0);
+}
+
+/**
+ * @param ptr "The pointer address to store to."
+ * @param value "The value to store masked"
+ * @param mask "The mask for the store"
+ * @param $alignment "The alignment of the pointer"
+ *
+ * @require $assignable(&&value, $typeof(ptr)) : "Pointer and value must match"
+ * @require @typekind(value) == VECTOR : "Expected value to be a vector"
+ * @require value.len == mask.len : "Mask and value must have the same length"
+ * @require @constant_is_power_of_2($alignment) : "The alignment must be a power of two"
+ *
+ **/
+macro @masked_store_aligned(ptr, value, bool[<*>] mask, usz $alignment)
+{
+	return $$masked_store(ptr, value, mask, $alignment);
+}
+
+/**
+ * @param ptrvec "The vector pointer containing the addresses to store to."
+ * @param value "The value to store masked"
+ * @param mask "The mask for the store"
+ * @require @typekind(ptrvec) == VECTOR : "Expected ptrvec to be a vector"
+ * @require @typekind(value) == VECTOR : "Expected value to be a vector"
+ * @require $assignable(&&value[0], $typeof(ptrvec[0])) : "Pointer and value must match"
+ * @require value.len == mask.len : "Mask and value must have the same length"
+ * @require mask.len == ptrvec.len : "Mask and ptrvec must have the same length"
+ *
+ **/
+macro scatter(ptrvec, value, bool[<*>] mask)
+{
+	return $$scatter(ptrvec, value, mask, 0);
+}
+
+/**
+ * @param ptrvec "The vector pointer containing the addresses to store to."
+ * @param value "The value to store masked"
+ * @param mask "The mask for the store"
+ * @param $alignment "The alignment of the load"
+ *
+ * @require @typekind(ptrvec) == VECTOR : "Expected ptrvec to be a vector"
+ * @require @typekind(value) == VECTOR : "Expected value to be a vector"
+ * @require $assignable(&&value[0], $typeof(ptrvec[0])) : "Pointer and value must match"
+ * @require value.len == mask.len : "Mask and value must have the same length"
+ * @require mask.len == ptrvec.len : "Mask and ptrvec must have the same length"
+ * @require @constant_is_power_of_2($alignment) : "The alignment must be a power of two"
+ **/
+macro @scatter_aligned(ptrvec, value, bool[<*>] mask, usz $alignment)
+{
+	return $$scatter(ptrvec, value, mask, $alignment);
+}
+
+/**
+ * @param [in] x "The variable or dereferenced pointer to load."
+ * @param $alignment "The alignment to assume for the load"
+ * @return "The value of x"
+ *
+ * @require @constant_is_power_of_2($alignment) : "The alignment must be a power of two"
+ **/
+macro @unaligned_load(&x, usz $alignment) @builtin
+{
+	return $$unaligned_load(x, $alignment);
+}
+
+/**
+ * @param [out] x "The variable or dereferenced pointer to store to."
+ * @param value "The value to store."
+ * @param $alignment "The alignment to assume for the store"
+ * @return "The value of x"
+ *
+ * @require $assignable(value, $typeof(*x)) : "The value doesn't match the variable"
+ * @require @constant_is_power_of_2($alignment) : "The alignment must be a power of two"
+ **/
+macro @unaligned_store(&x, value, usz $alignment) @builtin
+{
+	return $$unaligned_store(x, ($typeof(*x))value, $alignment);
+}
+
+macro @volatile_load(&x) @builtin
+{
+	return $$volatile_load(x);
+}
+
+/**
+ * @require $assignable(y, $typeof(*x)) : "The value doesn't match the variable"
+ **/
+macro @volatile_store(&x, y) @builtin
+{
+	return $$volatile_store(x, ($typeof(*x))y);
+}
+
+enum AtomicOrdering : int
+{
+	NOT_ATOMIC,         // Not atomic
+	UNORDERED,          // No lock
+	RELAXED,            // Consistent ordering
+	ACQUIRE,            // Barrier locking load/store
+	RELEASE,            // Barrier releasing load/store
+	ACQUIRE_RELEASE,    // Barrier fence to load/store
+	SEQ_CONSISTENT,     // Acquire semantics, ordered with other seq_consistent
+}
+
+/**
+ * @param [in] x "the variable or dereferenced pointer to load."
+ * @param $ordering "atomic ordering of the load, defaults to SEQ_CONSISTENT"
+ * @param $volatile "whether the load should be volatile, defaults to 'false'"
+ * @return "returns the value of x"
+ *
+ * @require $ordering != AtomicOrdering.RELEASE "Release ordering is not valid for load."
+ * @require $ordering != AtomicOrdering.ACQUIRE_RELEASE "Acquire release is not valid for load."
+ * @require types::may_load_atomic($typeof(x)) "Only integer, float and pointers may be used."
+ * @require @typekind(x) == POINTER "You can only load from a pointer"
+ **/
+macro @atomic_load(&x, AtomicOrdering $ordering = SEQ_CONSISTENT, $volatile = false) @builtin
+{
+	return $$atomic_load(x, $volatile, (int)$ordering);
+}
+
+/**
+ * @param [out] x "the variable or dereferenced pointer to store to."
+ * @param value "the value to store."
+ * @param $ordering "the atomic ordering of the store, defaults to SEQ_CONSISTENT"
+ * @param $volatile "whether the store should be volatile, defaults to 'false'"
+ *
+ * @require $ordering != AtomicOrdering.ACQUIRE "Acquire ordering is not valid for store."
+ * @require $ordering != AtomicOrdering.ACQUIRE_RELEASE "Acquire release is not valid for store."
+ * @require types::may_load_atomic($typeof(x)) "Only integer, float and pointers may be used."
+ **/
+macro void @atomic_store(&x, value, AtomicOrdering $ordering = SEQ_CONSISTENT, $volatile = false) @builtin
+{
+	$$atomic_store(x, value, $volatile, (int)$ordering);
+}
+
+/**
+ * @require $success != AtomicOrdering.NOT_ATOMIC && $success != AtomicOrdering.UNORDERED "Acquire ordering is not valid."
+ * @require $failure != AtomicOrdering.RELEASE && $failure != AtomicOrdering.ACQUIRE_RELEASE "Acquire release is not valid."
+ **/
+macro compare_exchange(ptr, compare, value, AtomicOrdering $success = SEQ_CONSISTENT, AtomicOrdering $failure = SEQ_CONSISTENT, bool $volatile = true, bool $weak = false, usz $alignment = 0)
+{
+	return $$compare_exchange(ptr, compare, value, $volatile, $weak, $success.ordinal, $failure.ordinal, $alignment);
+}
+
+/**
+ * @require $success != AtomicOrdering.NOT_ATOMIC && $success != AtomicOrdering.UNORDERED "Acquire ordering is not valid."
+ * @require $failure != AtomicOrdering.RELEASE && $failure != AtomicOrdering.ACQUIRE_RELEASE "Acquire release is not valid."
+ **/
+macro compare_exchange_volatile(ptr, compare, value, AtomicOrdering $success = SEQ_CONSISTENT, AtomicOrdering $failure = SEQ_CONSISTENT)
+{
+	return compare_exchange(ptr, compare, value, $success, $failure, true);
+}
+
+/**
+ * @require math::is_power_of_2(alignment)
+ **/
+fn usz aligned_offset(usz offset, usz alignment)
+{
+	return alignment * ((offset + alignment - 1) / alignment);
+}
+
+macro void* aligned_pointer(void* ptr, usz alignment)
+{
+	return (void*)(uptr)aligned_offset((uptr)ptr, alignment);
+}
+
+/**
+ * @require math::is_power_of_2(alignment)
+ **/
+fn bool ptr_is_aligned(void* ptr, usz alignment) @inline
+{
+	return (uptr)ptr & ((uptr)alignment - 1) == 0;
+}
+
+macro void clear(void* dst, usz len, usz $dst_align = 0, bool $is_volatile = false, bool $inlined = false)
+{
+	$$memset(dst, (char)0, len, $is_volatile, $dst_align);
+}
+
+macro void clear_inline(void* dst, usz $len, usz $dst_align = 0, bool $is_volatile = false)
+{
+	$$memset_inline(dst, (char)0, $len, $is_volatile, $dst_align);
+}
+
+/**
+ * Copy memory from src to dst efficiently, assuming the memory ranges do not overlap.
+ *
+ * @param [&out] dst "The destination to copy to"
+ * @param [&in] src "The source to copy from"
+ * @param len "The number of bytes to copy"
+ * @param $dst_align "the alignment of the destination if different from the default, 0 assumes the default"
+ * @param $src_align "the alignment of the destination if different from the default, 0 assumes the default"
+ * @param $is_volatile "True if this copy should be treated as volatile, i.e. it can't be optimized away."
+ *
+ * @require len == 0 || dst + len <= src || src + len <= dst : "Ranges may not overlap"
+ **/
+macro void copy(void* dst, void* src, usz len, usz $dst_align = 0, usz $src_align = 0, bool $is_volatile = false, bool $inlined = false)
+{
+	$$memcpy(dst, src, len, $is_volatile, $dst_align, $src_align);
+}
+
+/**
+ * Copy memory from src to dst efficiently, assuming the memory ranges do not overlap, it
+ * will always be inlined and never call memcopy
+ *
+ * @param [&out] dst "The destination to copy to"
+ * @param [&in] src "The source to copy from"
+ * @param $len "The number of bytes to copy"
+ * @param $dst_align "the alignment of the destination if different from the default, 0 assumes the default"
+ * @param $src_align "the alignment of the destination if different from the default, 0 assumes the default"
+ * @param $is_volatile "True if this copy should be treated as volatile, i.e. it can't be optimized away."
+ *
+ * @require $len == 0 || dst + $len <= src || src + $len <= dst : "Ranges may not overlap"
+ **/
+macro void copy_inline(void* dst, void* src, usz $len, usz $dst_align = 0, usz $src_align = 0, bool $is_volatile = false)
+{
+	$$memcpy_inline(dst, src, $len, $is_volatile, $dst_align, $src_align);
+}
+
+/**
+ * Copy memory from src to dst but correctly handle the possibility of overlapping ranges.
+ *
+ * @param [&out] dst "The destination to copy to"
+ * @param [&in] src "The source to copy from"
+ * @param len "The number of bytes to copy"
+ * @param $dst_align "the alignment of the destination if different from the default, 0 assumes the default"
+ * @param $src_align "the alignment of the destination if different from the default, 0 assumes the default"
+ * @param $is_volatile "True if this copy should be treated as volatile, i.e. it can't be optimized away."
+ **/
+macro void move(void* dst, void* src, usz len, usz $dst_align = 0, usz $src_align = 0, bool $is_volatile = false)
+{
+	$$memmove(dst, src, len, $is_volatile, $dst_align, $src_align);
+}
+
+/**
+ * Sets all memory in a region to that of the provided byte.
+ *
+ * @param [&out] dst "The destination to copy to"
+ * @param val "The value to copy into memory"
+ * @param len "The number of bytes to copy"
+ * @param $dst_align "the alignment of the destination if different from the default, 0 assumes the default"
+ * @param $is_volatile "True if this copy should be treated as volatile, i.e. it can't be optimized away."
+ *
+ * @ensure !len || (dst[0] == val && dst[len - 1] == val)
+ **/
+macro void set(void* dst, char val, usz len, usz $dst_align = 0, bool $is_volatile = false)
+{
+	$$memset(dst, val, len, $is_volatile, $dst_align);
+}
+
+/**
+ * Sets all memory in a region to that of the provided byte. Never calls OS memset.
+ *
+ * @param [&out] dst "The destination to copy to"
+ * @param val "The value to copy into memory"
+ * @param $len "The number of bytes to copy"
+ * @param $dst_align "the alignment of the destination if different from the default, 0 assumes the default"
+ * @param $is_volatile "True if this copy should be treated as volatile, i.e. it can't be optimized away."
+ *
+ * @ensure !$len || (dst[0] == val && dst[$len - 1] == val)
+ **/
+macro void set_inline(void* dst, char val, usz $len, usz $dst_align = 0, bool $is_volatile = false)
+{
+	$$memset_inline(dst, val, $len, $is_volatile, $dst_align);
+}
+/**
+ * @require values::@inner_kind(a) == TypeKind.SLICE || values::@inner_kind(a) == TypeKind.POINTER
+ * @require values::@inner_kind(b) == TypeKind.SLICE || values::@inner_kind(b) == TypeKind.POINTER
+ * @require values::@inner_kind(a) != TypeKind.SLICE || len == -1
+ * @require values::@inner_kind(a) != TypeKind.POINTER || len > -1
+ * @require values::@assign_to(a, b) && values::@assign_to(b, a)
+ **/
+macro bool equals(a, b, isz len = -1, usz $align = 0)
+{
+	$if !$align:
+		$align = $typeof(a[0]).alignof;
+	$endif
+	void* x @noinit;
+	void* y @noinit;
+	$if values::@inner_kind(a) == TypeKind.SLICE:
+		len = a.len;
+		if (len != b.len) return false;
+		x = a.ptr;
+		y = b.ptr;
+	$else
+		x = a;
+		y = b;
+		assert(len >= 0, "A zero or positive length must be given when comparing pointers.");
+	$endif
+
+	if (!len) return true;
+	var $Type;
+	$switch ($align)
+		$case 1:
+			$Type = char;
+		$case 2:
+			$Type = ushort;
+		$case 4:
+			$Type = uint;
+		$case 8:
+		$default:
+			$Type = ulong;
+	$endswitch
+	var $step = $Type.sizeof;
+	usz end = len / $step;
+	for (usz i = 0; i < end; i++)
+	{
+		if ((($Type*)x)[i] != (($Type*)y)[i]) return false;
+	}
+	usz last = len % $align;
+	for (usz i = len - last; i < len; i++)
+	{
+		if (((char*)x)[i] != ((char*)y)[i]) return false;
+	}
+	return true;
+}
+
+
+macro type_alloc_must_be_aligned($Type)
+{
+	return $Type.alignof > DEFAULT_MEM_ALIGNMENT;
+}
+
+/**
+ * Run with a specific allocator inside of the macro body.
+ **/
+macro void @scoped(Allocator allocator; @body())
+{
+	Allocator old_allocator = allocator::thread_allocator;
+	allocator::thread_allocator = allocator;
+	defer allocator::thread_allocator = old_allocator;
+	@body();
+}
+
+macro void @report_heap_allocs_in_scope(;@body())
+{
+	TrackingAllocator tracker;
+	tracker.init(allocator::thread_allocator);
+	Allocator old_allocator = allocator::thread_allocator;
+	allocator::thread_allocator = &tracker;
+	defer
+	{
+		allocator::thread_allocator = old_allocator;
+		tracker.print_report();
+		tracker.free();
+	}
+	@body();
+}
+
+macro void @stack_mem(usz $size; @body(Allocator mem)) @builtin
+{
+	char[$size] buffer;
+	OnStackAllocator allocator;
+	allocator.init(&buffer, allocator::heap());
+	defer allocator.free();
+	@body(&allocator);
+}
+
+macro void @stack_pool(usz $size; @body) @builtin
+{
+	char[$size] buffer;
+	OnStackAllocator allocator;
+	allocator.init(&buffer, allocator::heap());
+	defer allocator.free();
+	mem::@scoped(&allocator)
+	{
+		@body();
+	};
+}
+
+struct TempState
+{
+	TempAllocator* old;
+	TempAllocator* current;
+	usz mark;
+}
+
+/**
+ * Push the current temp allocator. A push must always be balanced with a pop using the current state.
+ **/
+fn TempState temp_push(TempAllocator* other = null)
+{
+	TempAllocator* current = allocator::temp();
+	TempAllocator* old = current;
+	if (other == current)
+	{
+		current = allocator::temp_allocator_next();
+	}
+	return { old, current, current.used };
+}
+
+/**
+ * Pop the current temp allocator. A pop must always be balanced with a push.
+ **/
+fn void temp_pop(TempState old_state)
+{
+	assert(allocator::thread_temp_allocator == old_state.current, "Tried to pop temp allocators out of order.");
+	assert(old_state.current.used >= old_state.mark, "Tried to pop temp allocators out of order.");
+	old_state.current.reset(old_state.mark);
+	allocator::thread_temp_allocator = old_state.old;
+}
+
+macro void @pool(TempAllocator* #other_temp = null; @body) @builtin
+{
+	TempAllocator* current = allocator::temp();
+	var $has_arg = !$is_const(#other_temp);
+	$if $has_arg:
+		TempAllocator* original = current;
+		if (current == (void*)#other_temp) current = allocator::temp_allocator_next();
+	$endif
+	usz mark = current.used;
+	defer
+	{
+		current.reset(mark);
+		$if $has_arg:
+			allocator::thread_temp_allocator = original;
+		$endif;
+	}
+	@body();
+}
+
+import libc;
+
+
+module std::core::mem @if(WASM_NOLIBC);
+import std::core::mem::allocator @public;
+SimpleHeapAllocator wasm_allocator @private;
+extern int __heap_base;
+
+fn void initialize_wasm_mem() @init(1) @private
+{
+	allocator::wasm_memory.allocate_block(mem::DEFAULT_MEM_ALIGNMENT)!!; // Give us a valid null.
+	// Check if we need to move the heap.
+	uptr start = (uptr)&__heap_base;
+	if (start > mem::DEFAULT_MEM_ALIGNMENT) allocator::wasm_memory.use = start;
+	wasm_allocator.init(fn (x) => allocator::wasm_memory.allocate_block(x));
+	allocator::thread_allocator = &wasm_allocator;
+	allocator::init_default_temp_allocators();
+}
+
+module std::core::mem;
+
+
+macro TrackingEnv* get_tracking_env()
+{
+	$if env::TRACK_MEMORY:
+		return &&TrackingEnv { $$FILE, $$FUNC, $$LINE };
+	$else
+		return null;
+	$endif
+}
+
+macro @clone(value) @builtin @nodiscard
+{
+	return allocator::clone(allocator::heap(), value);
+}
+
+macro @tclone(value) @builtin @nodiscard
+{
+	return temp_new($typeof(value), value);
+}
+
+fn void* malloc(usz size) @builtin @inline @nodiscard
+{
+	return allocator::malloc(allocator::heap(), size);
+}
+
+/**
+ * Allocate using an aligned allocation. This is necessary for types with a default memory alignment
+ * exceeding DEFAULT_MEM_ALIGNMENT. IMPORTANT! It must be freed using free_aligned.
+ **/
+fn void* malloc_aligned(usz size, usz alignment) @builtin @inline @nodiscard
+{
+	return allocator::malloc_aligned(allocator::heap(), size, alignment)!!;
+}
+
+fn void* tmalloc(usz size, usz alignment = 0) @builtin @inline @nodiscard
+{
+	if (!size) return null;
+	return allocator::temp().acquire(size, NO_ZERO, alignment)!!;
+}
+
+/**
+ * @require $vacount < 2 : "Too many arguments."
+ * @require $vacount == 0 ||| $assignable($vaexpr[0], $Type) : "The second argument must be an initializer for the type"
+ * @require $Type.alignof <= DEFAULT_MEM_ALIGNMENT : "Types with alignment exceeding the default must use 'alloc_aligned' instead"
+ **/
+macro new($Type, ...) @nodiscard
+{
+	$if $vacount == 0:
+		return ($Type*)calloc($Type.sizeof);
+	$else
+		$Type* val = malloc($Type.sizeof);
+		*val = $vaexpr[0];
+		return val;
+	$endif
+}
+
+/**
+ * Allocate using an aligned allocation. This is necessary for types with a default memory alignment
+ * exceeding DEFAULT_MEM_ALIGNMENT. IMPORTANT! It must be freed using free_aligned.
+ * @require $vacount < 2 : "Too many arguments."
+ * @require $vacount == 0 ||| $assignable($vaexpr[0], $Type) : "The second argument must be an initializer for the type"
+ **/
+macro new_aligned($Type, ...) @nodiscard
+{
+	$if $vacount == 0:
+		return ($Type*)calloc_aligned($Type.sizeof, $Type.alignof);
+	$else
+		$Type* val = malloc_aligned($Type.sizeof, $Type.alignof);
+		*val = $vaexpr[0];
+		return val;
+	$endif
+}
+
+/**
+ * @require $Type.alignof <= DEFAULT_MEM_ALIGNMENT : "Types with alignment exceeding the default must use 'alloc_aligned' instead"
+ **/
+macro alloc($Type) @nodiscard
+{
+	return ($Type*)malloc($Type.sizeof);
+}
+
+/**
+ * Allocate using an aligned allocation. This is necessary for types with a default memory alignment
+ * exceeding DEFAULT_MEM_ALIGNMENT. IMPORTANT! It must be freed using free_aligned.
+ **/
+macro alloc_aligned($Type) @nodiscard
+{
+	return ($Type*)malloc_aligned($Type.sizeof, $Type.alignof);
+}
+
+/**
+ * @require $vacount < 2 : "Too many arguments."
+ * @require $vacount == 0 ||| $assignable($vaexpr[0], $Type) : "The second argument must be an initializer for the type"
+ **/
+macro temp_new($Type, ...) @nodiscard
+{
+	$if $vacount == 0:
+		return ($Type*)tcalloc($Type.sizeof) @inline;
+	$else
+		$Type* val = tmalloc($Type.sizeof) @inline;
+		*val = $vaexpr[0];
+		return val;
+	$endif
+}
+
+macro temp_alloc($Type) @nodiscard
+{
+	return tmalloc($Type.sizeof);
+}
+
+
+/**
+ * @require $Type.alignof <= DEFAULT_MEM_ALIGNMENT : "Types with alignment exceeding the default must use 'new_array_aligned' instead"
+ **/
+macro new_array($Type, usz elements) @nodiscard
+{
+	return allocator::new_array(allocator::heap(), $Type, elements);
+}
+
+/**
+ * Allocate using an aligned allocation. This is necessary for types with a default memory alignment
+ * exceeding DEFAULT_MEM_ALIGNMENT. IMPORTANT! It must be freed using free_aligned.
+ **/
+macro new_array_aligned($Type, usz elements) @nodiscard
+{
+	return allocator::new_array_aligned(allocator::heap(), $Type, elements);
+}
+
+/**
+ * @require $Type.alignof <= DEFAULT_MEM_ALIGNMENT : "Types with alignment exceeding the default must use 'alloc_array_aligned' instead"
+ **/
+macro alloc_array($Type, usz elements) @nodiscard
+{
+	return allocator::alloc_array(allocator::heap(), $Type, elements);
+}
+
+/**
+ * Allocate using an aligned allocation. This is necessary for types with a default memory alignment
+ * exceeding DEFAULT_MEM_ALIGNMENT. IMPORTANT! It must be freed using free_aligned.
+ **/
+macro alloc_array_aligned($Type, usz elements) @nodiscard
+{
+	return allocator::alloc_array_aligned(allocator::heap(), $Type, elements);
+}
+
+macro temp_alloc_array($Type, usz elements) @nodiscard
+{
+	return (($Type*)tmalloc($Type.sizeof * elements, $Type.alignof))[:elements];
+}
+
+macro temp_new_array($Type, usz elements) @nodiscard
+{
+	return (($Type*)tcalloc($Type.sizeof * elements, $Type.alignof))[:elements];
+}
+
+fn void* calloc(usz size) @builtin @inline @nodiscard
+{
+	return allocator::calloc(allocator::heap(), size);
+}
+
+/**
+ * Allocate using an aligned allocation. This is necessary for types with a default memory alignment
+ * exceeding DEFAULT_MEM_ALIGNMENT. IMPORTANT! It must be freed using free_aligned.
+ **/
+fn void* calloc_aligned(usz size, usz alignment) @builtin @inline @nodiscard
+{
+	return allocator::calloc_aligned(allocator::heap(), size, alignment)!!;
+}
+
+fn void* tcalloc(usz size, usz alignment = 0) @builtin @inline @nodiscard
+{
+	if (!size) return null;
+	return allocator::temp().acquire(size, ZERO, alignment)!!;
+}
+
+fn void* realloc(void *ptr, usz new_size) @builtin @inline @nodiscard
+{
+	return allocator::realloc(allocator::heap(), ptr, new_size);
+}
+
+fn void* realloc_aligned(void *ptr, usz new_size, usz alignment) @builtin @inline @nodiscard
+{
+	return allocator::realloc_aligned(allocator::heap(), ptr, new_size, alignment)!!;
+}
+
+fn void free(void* ptr) @builtin @inline
+{
+	return allocator::free(allocator::heap(), ptr);
+}
+
+fn void free_aligned(void* ptr) @builtin @inline
+{
+	return allocator::free_aligned(allocator::heap(), ptr);
+}
+
+fn void* trealloc(void* ptr, usz size, usz alignment = mem::DEFAULT_MEM_ALIGNMENT) @builtin @inline @nodiscard
+{
+	if (!size) return null;
+    if (!ptr) return tmalloc(size, alignment);
+	return allocator::temp().resize(ptr, size, alignment)!!;
+}
+

lib/std/core/mem_allocator.c3

@@ -0,0 +1,424 @@
+module std::core::mem::allocator;
+
+const DEFAULT_SIZE_PREFIX = usz.sizeof;
+const DEFAULT_SIZE_PREFIX_ALIGNMENT = usz.alignof;
+
+struct TrackingEnv
+{
+	String file;
+	String function;
+	uint line;
+}
+
+enum AllocInitType
+{
+	NO_ZERO,
+	ZERO
+}
+
+interface Allocator
+{
+	fn void reset(usz mark) @optional;
+	fn usz mark() @optional;
+	/**
+     * @require !alignment || math::is_power_of_2(alignment)
+     * @require alignment <= mem::MAX_MEMORY_ALIGNMENT `alignment too big`
+     * @require size > 0
+     **/
+	fn void*! acquire(usz size, AllocInitType init_type, usz alignment = 0);
+	/**
+     * @require !alignment || math::is_power_of_2(alignment)
+     * @require alignment <= mem::MAX_MEMORY_ALIGNMENT `alignment too big`
+     * @require ptr != null
+     * @require new_size > 0
+     **/
+	fn void*! resize(void* ptr, usz new_size, usz alignment = 0);
+	/**
+     * @require ptr != null
+     **/
+	fn void release(void* ptr, bool aligned);
+}
+
+def MemoryAllocFn = fn char[]!(usz);
+
+fault AllocationFailure
+{
+	OUT_OF_MEMORY,
+	CHUNK_TOO_LARGE,
+}
+
+fn usz alignment_for_allocation(usz alignment) @inline @private
+{
+	return alignment < mem::DEFAULT_MEM_ALIGNMENT ? mem::DEFAULT_MEM_ALIGNMENT : alignment;
+}
+
+macro void* malloc(Allocator allocator, usz size) @nodiscard
+{
+	return malloc_try(allocator, size)!!;
+}
+
+macro void*! malloc_try(Allocator allocator, usz size) @nodiscard
+{
+	if (!size) return null;
+	$if env::TESTING:
+		char* data = allocator.acquire(size, NO_ZERO)!;
+		mem::set(data, 0xAA, size, mem::DEFAULT_MEM_ALIGNMENT);
+		return data;
+	$else
+		return allocator.acquire(size, NO_ZERO);
+	$endif
+}
+
+macro void* calloc(Allocator allocator, usz size) @nodiscard
+{
+	return calloc_try(allocator, size)!!;
+}
+
+macro void*! calloc_try(Allocator allocator, usz size) @nodiscard
+{
+	if (!size) return null;
+	return allocator.acquire(size, ZERO);
+}
+
+macro void* realloc(Allocator allocator, void* ptr, usz new_size) @nodiscard
+{
+	return realloc_try(allocator, ptr, new_size)!!;
+}
+
+macro void*! realloc_try(Allocator allocator, void* ptr, usz new_size) @nodiscard
+{
+	if (!new_size)
+	{
+		free(allocator, ptr);
+		return null;
+	}
+	if (!ptr) return allocator.acquire(new_size, NO_ZERO);
+	return allocator.resize(ptr, new_size);
+}
+
+macro void free(Allocator allocator, void* ptr)
+{
+	if (!ptr) return;
+	$if env::TESTING:
+		((char*)ptr)[0] = 0xBA;
+	$endif
+	allocator.release(ptr, false);
+}
+
+macro void*! malloc_aligned(Allocator allocator, usz size, usz alignment) @nodiscard
+{
+	if (!size) return null;
+	$if env::TESTING:
+		char* data = allocator.acquire(size, NO_ZERO, alignment)!;
+		mem::set(data, 0xAA, size, mem::DEFAULT_MEM_ALIGNMENT);
+		return data;
+	$else
+		return allocator.acquire(size, NO_ZERO, alignment);
+	$endif
+}
+
+macro void*! calloc_aligned(Allocator allocator, usz size, usz alignment) @nodiscard
+{
+	if (!size) return null;
+	return allocator.acquire(size, ZERO, alignment);
+}
+
+macro void*! realloc_aligned(Allocator allocator, void* ptr, usz new_size, usz alignment) @nodiscard
+{
+	if (!new_size)
+	{
+		free_aligned(allocator, ptr);
+		return null;
+	}
+	if (!ptr)
+	{
+		return malloc_aligned(allocator, new_size, alignment);
+	}
+	return allocator.resize(ptr, new_size, alignment);
+}
+
+macro void free_aligned(Allocator allocator, void* ptr)
+{
+	if (!ptr) return;
+	$if env::TESTING:
+	((char*)ptr)[0] = 0xBA;
+	$endif
+	allocator.release(ptr, .aligned = true);
+}
+
+/**
+ * @require $Type.alignof <= mem::DEFAULT_MEM_ALIGNMENT : "Types with alignment exceeding the default must use 'new_aligned' instead"
+ * @require $vacount < 2 : "Too many arguments."
+ * @require $vacount == 0 ||| $assignable($vaexpr[0], $Type) : "The second argument must be an initializer for the type"
+ **/
+macro new(Allocator allocator, $Type, ...) @nodiscard
+{
+	$if $vacount == 0:
+		return ($Type*)calloc(allocator, $Type.sizeof);
+	$else
+		$Type* val = malloc(allocator, $Type.sizeof);
+		*val = $vaexpr[0];
+		return val;
+	$endif
+}
+
+/**
+ * @require $Type.alignof <= mem::DEFAULT_MEM_ALIGNMENT : "Types with alignment exceeding the default must use 'new_aligned' instead"
+ * @require $vacount < 2 : "Too many arguments."
+ * @require $vacount == 0 ||| $assignable($vaexpr[0], $Type) : "The second argument must be an initializer for the type"
+ **/
+macro new_try(Allocator allocator, $Type, ...) @nodiscard
+{
+	$if $vacount == 0:
+		return ($Type*)calloc_try(allocator, $Type.sizeof);
+	$else
+		$Type* val = malloc_try(allocator, $Type.sizeof)!;
+		*val = $vaexpr[0];
+		return val;
+	$endif
+}
+
+/**
+ * Allocate using an aligned allocation. This is necessary for types with a default memory alignment
+ * exceeding DEFAULT_MEM_ALIGNMENT. IMPORTANT! It must be freed using free_aligned.
+ * @require $vacount < 2 : "Too many arguments."
+ * @require $vacount == 0 ||| $assignable($vaexpr[0], $Type) : "The second argument must be an initializer for the type"
+ **/
+macro new_aligned($Type, ...) @nodiscard
+{
+	$if $vacount == 0:
+		return ($Type*)calloc_aligned(allocator, $Type.sizeof, $Type.alignof);
+	$else
+		$Type* val = malloc_aligned(allocator, $Type.sizeof, $Type.alignof);
+		*val = $vaexpr[0];
+		return val;
+	$endif
+}
+
+/**
+ * @require $Type.alignof <= mem::DEFAULT_MEM_ALIGNMENT
+ **/
+macro new_with_padding(Allocator allocator, $Type, usz padding) @nodiscard
+{
+	return ($Type*)calloc_try(allocator, $Type.sizeof + padding);
+}
+
+/**
+ * @require $Type.alignof <= mem::DEFAULT_MEM_ALIGNMENT : "Types with alignment exceeding the default must use 'alloc_aligned' instead"
+ **/
+macro alloc(Allocator allocator, $Type) @nodiscard
+{
+	return ($Type*)malloc(allocator, $Type.sizeof);
+}
+
+/**
+ * @require $Type.alignof <= mem::DEFAULT_MEM_ALIGNMENT : "Types with alignment exceeding the default must use 'alloc_aligned' instead"
+ **/
+macro alloc_try(Allocator allocator, $Type) @nodiscard
+{
+	return ($Type*)malloc_try(allocator, $Type.sizeof);
+}
+
+/**
+ * Allocate using an aligned allocation. This is necessary for types with a default memory alignment
+ * exceeding DEFAULT_MEM_ALIGNMENT. IMPORTANT! It must be freed using free_aligned.
+ **/
+macro alloc_aligned(Allocator allocator, $Type) @nodiscard
+{
+	return ($Type*)malloc_aligned(allocator, $Type.sizeof, $Type.alignof);
+}
+
+/**
+ * @require $Type.alignof <= mem::DEFAULT_MEM_ALIGNMENT
+ **/
+macro alloc_with_padding(Allocator allocator, $Type, usz padding) @nodiscard
+{
+	return ($Type*)malloc_try(allocator, $Type.sizeof + padding);
+}
+
+/**
+ * @require $Type.alignof <= mem::DEFAULT_MEM_ALIGNMENT : "Types with alignment exceeding the default must use 'new_array_aligned' instead"
+ **/
+macro new_array(Allocator allocator, $Type, usz elements) @nodiscard
+{
+	return new_array_try(allocator, $Type, elements)!!;
+}
+
+/**
+ * @require $Type.alignof <= mem::DEFAULT_MEM_ALIGNMENT : "Types with alignment exceeding the default must use 'new_array_aligned' instead"
+ **/
+macro new_array_try(Allocator allocator, $Type, usz elements) @nodiscard
+{
+	return (($Type*)calloc_try(allocator, $Type.sizeof * elements))[:elements];
+}
+
+/**
+ * Allocate using an aligned allocation. This is necessary for types with a default memory alignment
+ * exceeding DEFAULT_MEM_ALIGNMENT. IMPORTANT! It must be freed using free_aligned.
+ **/
+macro new_array_aligned(Allocator allocator, $Type, usz elements) @nodiscard
+{
+	return (($Type*)calloc_aligned(allocator, $Type.sizeof * elements, $Type.alignof))[:elements]!!;
+}
+
+/**
+ * @require $Type.alignof <= mem::DEFAULT_MEM_ALIGNMENT : "Types with alignment exceeding the default must use 'alloc_array_aligned' instead"
+ **/
+macro alloc_array(Allocator allocator, $Type, usz elements) @nodiscard
+{
+	return alloc_array_try(allocator, $Type, elements)!!;
+}
+
+/**
+ * Allocate using an aligned allocation. This is necessary for types with a default memory alignment
+ * exceeding DEFAULT_MEM_ALIGNMENT. IMPORTANT! It must be freed using free_aligned.
+ **/
+macro alloc_array_aligned(Allocator allocator, $Type, usz elements) @nodiscard
+{
+	return (($Type*)malloc_aligned(allocator, $Type.sizeof * elements, $Type.alignof))[:elements]!!;
+}
+
+/**
+ * @require $Type.alignof <= mem::DEFAULT_MEM_ALIGNMENT : "Types with alignment exceeding the default must use 'alloc_array_aligned' instead"
+ **/
+macro alloc_array_try(Allocator allocator, $Type, usz elements) @nodiscard
+{
+	return (($Type*)malloc_try(allocator, $Type.sizeof * elements))[:elements];
+}
+
+macro clone(Allocator allocator, value) @nodiscard
+{
+	return new(allocator, $typeof(value), value);
+}
+
+fn any clone_any(Allocator allocator, any value) @nodiscard
+{
+	usz size = value.type.sizeof;
+	void* data = malloc(allocator, size);
+	mem::copy(data, value.ptr, size);
+	return any_make(data, value.type);
+}
+
+
+/**
+ * @require bytes > 0
+ * @require alignment > 0
+ * @require bytes <= isz.max
+ **/
+macro void*! @aligned_alloc(#alloc_fn, usz bytes, usz alignment)
+{
+	if (alignment < void*.alignof) alignment = void*.alignof;
+	usz header = AlignedBlock.sizeof + alignment;
+	usz alignsize = bytes + header;
+	$if @typekind(#alloc_fn(bytes)) == OPTIONAL:
+		void* data = #alloc_fn(alignsize)!;
+	$else
+		void* data = #alloc_fn(alignsize);
+	$endif
+	void* mem = mem::aligned_pointer(data + AlignedBlock.sizeof, alignment);
+	AlignedBlock* desc = (AlignedBlock*)mem - 1;
+	assert(mem > data);
+	*desc = { bytes, data };
+	return mem;
+}
+
+struct AlignedBlock
+{
+	usz len;
+	void* start;
+}
+
+macro void! @aligned_free(#free_fn, void* old_pointer)
+{
+	AlignedBlock* desc = (AlignedBlock*)old_pointer - 1;
+	$if @typekind(#free_fn(desc.start)) == OPTIONAL:
+	#free_fn(desc.start)!;
+	$else
+	#free_fn(desc.start);
+	$endif
+}
+
+/**
+ * @require bytes > 0
+ * @require alignment > 0
+ **/
+macro void*! @aligned_realloc(#calloc_fn, #free_fn, void* old_pointer, usz bytes, usz alignment)
+{
+	AlignedBlock* desc = (AlignedBlock*)old_pointer - 1;
+	void* data_start = desc.start;
+	void* new_data = @aligned_alloc(#calloc_fn, bytes, alignment)!;
+	mem::copy(new_data, old_pointer, desc.len < bytes ? desc.len : bytes, 1, 1);
+	$if @typekind(#free_fn(data_start)) == OPTIONAL:
+	#free_fn(data_start)!;
+	$else
+	#free_fn(data_start);
+	$endif
+	return new_data;
+}
+
+
+// All allocators
+tlocal Allocator thread_allocator @private = &allocator::LIBC_ALLOCATOR;
+tlocal TempAllocator* thread_temp_allocator @private = null;
+tlocal TempAllocator*[2] temp_allocator_pair @private;
+Allocator temp_base_allocator @private = &allocator::LIBC_ALLOCATOR;
+
+macro TempAllocator* create_default_sized_temp_allocator(Allocator allocator) @local
+{
+	$switch (env::MEMORY_ENV)
+	$case NORMAL:
+		return new_temp_allocator(1024 * 256, allocator)!!;
+	$case SMALL:
+		return new_temp_allocator(1024 * 16, allocator)!!;
+	$case TINY:
+		return new_temp_allocator(1024 * 2, allocator)!!;
+	$case NONE:
+		unreachable("Temp allocator must explicitly created when memory-env is set to 'none'.");
+	$endswitch
+}
+
+macro Allocator heap() => thread_allocator;
+
+macro TempAllocator* temp()
+{
+	if (!thread_temp_allocator)
+	{
+		init_default_temp_allocators();
+	}
+	return thread_temp_allocator;
+}
+
+fn void init_default_temp_allocators() @private
+{
+	temp_allocator_pair[0] = create_default_sized_temp_allocator(temp_base_allocator);
+	temp_allocator_pair[1] = create_default_sized_temp_allocator(temp_base_allocator);
+	thread_temp_allocator = temp_allocator_pair[0];
+}
+
+fn void destroy_temp_allocators_after_exit() @finalizer(65535) @local @if(env::LIBC)
+{
+	destroy_temp_allocators();
+}
+
+/**
+ * Call this to destroy any memory used by the temp allocators. This will invalidate all temp memory.
+ **/
+fn void destroy_temp_allocators()
+{
+	if (!thread_temp_allocator) return;
+	temp_allocator_pair[0].destroy();
+	temp_allocator_pair[1].destroy();
+	temp_allocator_pair[..] = null;
+	thread_temp_allocator = null;
+}
+
+fn TempAllocator* temp_allocator_next() @private
+{
+	if (!thread_temp_allocator)
+	{
+		init_default_temp_allocators();
+		return thread_temp_allocator;
+	}
+	usz index = thread_temp_allocator == temp_allocator_pair[0] ? 1 : 0;
+	return thread_temp_allocator = temp_allocator_pair[index];
+}

lib/std/core/os/wasm_memory.c3

@@ -0,0 +1,32 @@
+module std::core::mem::allocator;
+
+
+const usz WASM_BLOCK_SIZE = 65536;
+
+WasmMemory wasm_memory;
+
+struct WasmMemory
+{
+	usz allocation;
+	uptr use;
+}
+
+fn char[]! WasmMemory.allocate_block(&self, usz bytes)
+{
+	if (!self.allocation)
+	{
+		self.allocation = $$wasm_memory_size(0) * WASM_BLOCK_SIZE;
+	}
+	isz bytes_required = bytes + self.use - self.allocation;
+	if (bytes_required <= 0)
+	{
+		defer self.use += bytes;
+		return ((char*)self.use)[:bytes];
+	}
+
+	usz blocks_required = (bytes_required + WASM_BLOCK_SIZE + 1) / WASM_BLOCK_SIZE;
+	if ($$wasm_memory_grow(0, blocks_required) == -1) return AllocationFailure.OUT_OF_MEMORY?;
+	self.allocation = $$wasm_memory_size(0) * WASM_BLOCK_SIZE;
+	defer self.use += bytes;
+	return ((char*)self.use)[:bytes];
+}

lib/std/core/private/cpu_detect.c3

@@ -0,0 +1,258 @@
+module std::core::cpudetect @if(env::X86 || env::X86_64);
+
+struct CpuId
+{
+	uint eax, ebx, ecx, edx;
+}
+fn CpuId x86_cpuid(uint eax, uint ecx = 0)
+{
+	int edx;
+	int ebx;
+	asm
+	{
+		movl $eax, eax;
+		movl $ecx, ecx;
+		cpuid;
+		movl eax, $eax;
+		movl ebx, $ebx;
+		movl ecx, $ecx;
+		movl edx, $edx;
+	}
+	return { eax, ebx, ecx, edx };
+}
+
+enum X86Feature
+{
+	ADX,
+	AES,
+	AMX_BF16,
+	AMX_COMPLEX,
+	AMX_FP16,
+	AMX_INT8,
+	AMX_TILE,
+	AVX,
+	AVX10_1_256,
+	AVX10_1_512,
+	AVX2,
+	AVX5124FMAPS,
+	AVX5124VNNIW,
+	AVX512BF16,
+	AVX512BITALG,
+	AVX512BW,
+	AVX512CD,
+	AVX512DQ,
+	AVX512ER,
+	AVX512F,
+	AVX512FP16,
+	AVX512IFMA,
+	AVX512PF,
+	AVX512VBMI,
+	AVX512VBMI2,
+	AVX512VL,
+	AVX512VNNI,
+	AVX512VP2INTERSECT,
+	AVX512VPOPCNTDQ,
+	AVXIFMA,
+	AVXNECONVERT,
+	AVXVNNI,
+	AVXVNNIINT16,
+	AVXVNNIINT8,
+	BMI,
+	BMI2,
+	CLDEMOTE,
+	CLFLUSHOPT,
+	CLWB,
+	CLZERO,
+	CMOV,
+	CMPCCXADD,
+	CMPXCHG16B,
+	CX8,
+	ENQCMD,
+	F16C,
+	FMA,
+	FMA4,
+	FSGSBASE,
+	FXSR,
+	GFNI,
+	HRESET,
+	INVPCID,
+	KL,
+	LWP,
+	LZCNT,
+	MMX,
+	MOVBE,
+	MOVDIR64B,
+	MOVDIRI,
+	MWAITX,
+	PCLMUL,
+	PCONFIG,
+	PKU,
+	POPCNT,
+	PREFETCHI,
+	PREFETCHWT1,
+	PRFCHW,
+	PTWRITE,
+	RAOINT,
+	RDPID,
+	RDPRU,
+	RDRND,
+	RDSEED,
+	RTM,
+	SAHF,
+	SERIALIZE,
+	SGX,
+	SHA,
+	SHA512,
+	SHSTK,
+	SM3,
+	SM4,
+	SSE,
+	SSE2,
+	SSE3,
+	SSE4_1,
+	SSE4_2,
+	SSE4_A,
+	SSSE3,
+	TBM,
+	TSXLDTRK,
+	UINTR,
+	USERMSR,
+	VAES,
+	VPCLMULQDQ,
+	WAITPKG,
+	WBNOINVD,
+	WIDEKL,
+	X87,
+	XOP,
+	XSAVE,
+	XSAVEC,
+	XSAVEOPT,
+	XSAVES,
+}
+
+uint128 x86_features;
+
+fn void add_feature_if_bit(X86Feature feature, uint register, int bit)
+{
+	if (register & 1U << bit) x86_features |= 1u128 << feature.ordinal;
+}
+
+fn void x86_initialize_cpu_features()
+{
+	uint max_level = x86_cpuid(0).eax;
+	CpuId feat = x86_cpuid(1);
+	CpuId leaf7 = max_level >= 8 ? x86_cpuid(7) : CpuId {};
+	CpuId leaf7s1 = leaf7.eax >= 1 ? x86_cpuid(7, 1) : CpuId {};
+	CpuId ext1 = x86_cpuid(0x80000000).eax >= 0x80000001 ? x86_cpuid(0x80000001) : CpuId {};
+	CpuId ext8 = x86_cpuid(0x80000000).eax >= 0x80000008 ? x86_cpuid(0x80000008) : CpuId {};
+	CpuId leaf_d = max_level >= 0xd ? x86_cpuid(0xd, 0x1) : CpuId {};
+	CpuId leaf_14 = max_level >= 0x14 ? x86_cpuid(0x14) : CpuId {};
+	CpuId leaf_19 = max_level >= 0x19 ? x86_cpuid(0x19) : CpuId {};
+	CpuId leaf_24 = max_level >= 0x24 ? x86_cpuid(0x24) : CpuId {};
+	add_feature_if_bit(ADX, leaf7.ebx, 19);
+	add_feature_if_bit(AES, feat.ecx, 25);
+	add_feature_if_bit(AMX_BF16, leaf7.edx, 22);
+	add_feature_if_bit(AMX_COMPLEX, leaf7s1.edx, 8);
+	add_feature_if_bit(AMX_FP16, leaf7s1.eax, 21);
+	add_feature_if_bit(AMX_INT8, leaf7.edx, 25);
+	add_feature_if_bit(AMX_TILE, leaf7.edx, 24);
+	add_feature_if_bit(AVX, feat.ecx, 28);
+	add_feature_if_bit(AVX10_1_256, leaf7s1.edx, 19);
+	add_feature_if_bit(AVX10_1_512, leaf_24.ebx, 18);
+	add_feature_if_bit(AVX2, leaf7.ebx, 5);
+	add_feature_if_bit(AVX5124FMAPS, leaf7.edx, 3);
+	add_feature_if_bit(AVX5124VNNIW, leaf7.edx, 2);
+	add_feature_if_bit(AVX512BF16, leaf7s1.eax, 5);
+	add_feature_if_bit(AVX512BITALG, leaf7.ecx, 12);
+	add_feature_if_bit(AVX512BW, leaf7.ebx, 30);
+	add_feature_if_bit(AVX512CD, leaf7.ebx, 28);
+	add_feature_if_bit(AVX512DQ, leaf7.ebx, 17);
+	add_feature_if_bit(AVX512ER, leaf7.ebx, 27);
+	add_feature_if_bit(AVX512F, leaf7.ebx, 16);
+	add_feature_if_bit(AVX512FP16, leaf7.edx, 23);
+	add_feature_if_bit(AVX512IFMA, leaf7.ebx, 21);
+	add_feature_if_bit(AVX512PF, leaf7.ebx, 26);
+	add_feature_if_bit(AVX512VBMI, leaf7.ecx, 1);
+	add_feature_if_bit(AVX512VBMI2, leaf7.ecx, 6);
+	add_feature_if_bit(AVX512VL, leaf7.ebx, 31);
+	add_feature_if_bit(AVX512VNNI, leaf7.ecx, 11);
+	add_feature_if_bit(AVX512VP2INTERSECT, leaf7.edx, 8);
+	add_feature_if_bit(AVX512VPOPCNTDQ, leaf7.ecx, 14);
+	add_feature_if_bit(AVXIFMA, leaf7s1.eax, 23);
+	add_feature_if_bit(AVXNECONVERT, leaf7s1.edx, 5);
+	add_feature_if_bit(AVXVNNI, leaf7s1.eax, 4);
+	add_feature_if_bit(AVXVNNIINT16, leaf7s1.edx, 10);
+	add_feature_if_bit(AVXVNNIINT8, leaf7s1.edx, 4);
+	add_feature_if_bit(BMI, leaf7.ebx, 3);
+	add_feature_if_bit(BMI2, leaf7.ebx, 8);
+	add_feature_if_bit(CLDEMOTE, leaf7.ecx, 25);
+	add_feature_if_bit(CLFLUSHOPT, leaf7.ebx, 23);
+	add_feature_if_bit(CLWB, leaf7.ebx, 24);
+	add_feature_if_bit(CLZERO, ext8.ecx, 0);
+	add_feature_if_bit(CMOV, feat.edx, 15);
+	add_feature_if_bit(CMPCCXADD, leaf7s1.eax, 7);
+	add_feature_if_bit(CMPXCHG16B, feat.ecx, 12);
+	add_feature_if_bit(CX8, feat.edx, 8);
+	add_feature_if_bit(ENQCMD, leaf7.ecx, 29);
+	add_feature_if_bit(F16C, feat.ecx, 29);
+	add_feature_if_bit(FMA, feat.ecx, 12);
+	add_feature_if_bit(FMA4, ext1.ecx, 16);
+	add_feature_if_bit(FSGSBASE, leaf7.ebx, 0);
+	add_feature_if_bit(FXSR, feat.edx, 24);
+	add_feature_if_bit(GFNI, leaf7.ecx, 8);
+	add_feature_if_bit(HRESET, leaf7s1.eax, 22);
+	add_feature_if_bit(INVPCID, leaf7.ebx, 10);
+	add_feature_if_bit(KL, leaf7.ecx, 23);
+	add_feature_if_bit(LWP, ext1.ecx, 15);
+	add_feature_if_bit(LZCNT, ext1.ecx, 5);
+	add_feature_if_bit(MMX, feat.edx, 23);
+	add_feature_if_bit(MOVBE, feat.ecx, 22);
+	add_feature_if_bit(MOVDIR64B, leaf7.ecx, 28);
+	add_feature_if_bit(MOVDIRI, leaf7.ecx, 27);
+	add_feature_if_bit(MWAITX, ext1.ecx, 29);
+	add_feature_if_bit(PCLMUL, feat.ecx, 1);
+	add_feature_if_bit(PCONFIG, leaf7.edx, 18);
+	add_feature_if_bit(PKU, leaf7.ecx, 4);
+	add_feature_if_bit(POPCNT, feat.ecx, 23);
+	add_feature_if_bit(PREFETCHI, leaf7s1.edx, 14);
+	add_feature_if_bit(PREFETCHWT1, leaf7.ecx, 0);
+	add_feature_if_bit(PRFCHW, ext1.ecx, 8);
+	add_feature_if_bit(PTWRITE, leaf_14.ebx, 4);
+	add_feature_if_bit(RAOINT, leaf7s1.eax, 3);
+	add_feature_if_bit(RDPID, leaf7.ecx, 22);
+	add_feature_if_bit(RDPRU, ext8.ecx, 4);
+	add_feature_if_bit(RDRND, feat.ecx, 30);
+	add_feature_if_bit(RDSEED, leaf7.ebx, 18);
+	add_feature_if_bit(RTM, leaf7.ebx, 11);
+	add_feature_if_bit(SAHF, ext1.ecx, 0);
+	add_feature_if_bit(SERIALIZE, leaf7.edx, 14);
+	add_feature_if_bit(SGX, leaf7.ebx, 2);
+	add_feature_if_bit(SHA, leaf7.ebx, 29);
+	add_feature_if_bit(SHA512, leaf7s1.eax, 0);
+	add_feature_if_bit(SHSTK, leaf7.ecx, 7);
+	add_feature_if_bit(SM3, leaf7s1.eax, 1);
+	add_feature_if_bit(SM4, leaf7s1.eax, 2);
+	add_feature_if_bit(SSE, feat.edx, 25);
+	add_feature_if_bit(SSE2, feat.edx, 26);
+	add_feature_if_bit(SSE3, feat.ecx, 0);
+	add_feature_if_bit(SSE4_1, feat.ecx, 19);
+	add_feature_if_bit(SSE4_2, feat.ecx, 20);
+	add_feature_if_bit(SSE4_A, ext1.ecx, 6);
+	add_feature_if_bit(SSSE3, feat.ecx, 9);
+	add_feature_if_bit(TBM, ext1.ecx, 21);
+	add_feature_if_bit(TSXLDTRK, leaf7.edx, 16);
+	add_feature_if_bit(UINTR, leaf7.edx, 5);
+	add_feature_if_bit(USERMSR, leaf7s1.edx, 15);
+	add_feature_if_bit(VAES, leaf7.ecx, 9);
+	add_feature_if_bit(VPCLMULQDQ, leaf7.ecx, 10);
+	add_feature_if_bit(WAITPKG, leaf7.ecx, 5);
+	add_feature_if_bit(WBNOINVD, ext8.ecx, 9);
+	add_feature_if_bit(WIDEKL, leaf_19.ebx, 2);
+	add_feature_if_bit(X87, feat.edx, 0);
+	add_feature_if_bit(XOP, ext1.ecx, 11);
+	add_feature_if_bit(XSAVE, feat.ecx, 26);
+	add_feature_if_bit(XSAVEC, leaf_d.eax, 1);
+	add_feature_if_bit(XSAVEOPT, leaf_d.eax, 0);
+	add_feature_if_bit(XSAVES, leaf_d.eax, 3);
+
+}

lib/std/core/private/macho_runtime.c3

@@ -0,0 +1,254 @@
+module std::core::machoruntime @if(env::DARWIN) @private;
+
+struct SegmentCommand64
+{
+	uint cmd;
+	uint cmdsize;
+	char[16] segname;
+	ulong vmaddr;
+	ulong vmsize;
+	ulong fileoff;
+	ulong filesize;
+	uint maxprot;
+	uint initprot;
+	uint nsects;
+	uint flags;
+}
+
+struct LoadCommand
+{
+	uint cmd;
+	uint cmdsize;
+}
+
+struct Section64
+{
+	char[16] sectname;
+	char[16] segname;
+	ulong addr;
+	ulong size;
+	uint offset;
+	uint align;
+	uint reloff;
+	uint nreloc;
+	uint flags;
+	uint reserved1;
+	uint reserved2;
+	uint reserved3;
+}
+
+struct MachHeader
+{
+	uint magic;
+	uint cputype;
+	uint cpusubtype;
+	uint filetype;
+	uint ncmds;
+	uint sizeofcmds;
+	uint flags;
+}
+
+struct MachHeader64
+{
+	inline MachHeader header;
+	uint reserved;
+}
+
+const LC_SEGMENT_64 = 0x19;
+
+fault MachoSearch
+{
+	NOT_FOUND
+}
+fn bool name_cmp(char* a, char[16]* b)
+{
+	for (usz i = 0; i < 16; i++)
+	{
+		if (a[i] != (*b)[i]) return false;
+		if (a[i] == '\0') return true;
+	}
+	return false;
+}
+
+fn SegmentCommand64*! find_segment(MachHeader* header, char* segname)
+{
+	LoadCommand* command = (void*)header + MachHeader64.sizeof;
+	for (uint i = 0; i < header.ncmds; i++)
+	{
+		if (command.cmd == LC_SEGMENT_64)
+		{
+			SegmentCommand64* segment = (SegmentCommand64*)command;
+			if (name_cmp(segname, &segment.segname)) return segment;
+		}
+		command = (void*)command + command.cmdsize;
+	}
+	return MachoSearch.NOT_FOUND?;
+}
+fn Section64*! find_section(SegmentCommand64* command, char* sectname)
+{
+	Section64* section = (void*)command + SegmentCommand64.sizeof;
+	for (uint i = 0; i < command.nsects; i++)
+	{
+		if (name_cmp(sectname, &section.sectname)) return section;
+		section++;
+	}
+	return MachoSearch.NOT_FOUND?;
+}
+
+macro find_segment_section_body(MachHeader* header, char* segname, char* sectname, $Type)
+{
+
+	Section64*! section = find_section(find_segment(header, segname), sectname);
+	if (catch section)
+	{
+		return $Type[] {};
+	}
+	$Type* ptr = (void*)header + section.offset;
+	return ptr[:section.size / $Type.sizeof];
+}
+
+def DyldCallback = fn void (MachHeader* mh, isz vmaddr_slide);
+
+extern fn void _dyld_register_func_for_add_image(DyldCallback);
+
+
+struct DlInfo
+{
+	char* dli_fname;
+	void* dli_fbase;
+	char* dli_sname;
+	void* dli_saddr;
+}
+
+extern fn void printf(char*, ...);
+extern fn int dladdr(MachHeader* mh, DlInfo* dlinfo);
+extern fn void* realloc(void* ptr, usz size);
+extern fn void* malloc(usz size);
+extern fn void free(void* ptr);
+
+def CallbackFn = fn void();
+struct Callback
+{
+	uint priority;
+	CallbackFn xtor;
+	Callback* next;
+}
+struct DynamicMethod
+{
+	void* fn_ptr;
+	char* sel;
+	union
+	{
+		DynamicMethod* next;
+		TypeId* type;
+	}
+}
+
+enum StartupState
+{
+	NOT_STARTED,
+	INIT,
+	RUN_CTORS,
+	READ_DYLIB,
+	RUN_DYLIB_CTORS,
+	RUN_DTORS,
+	SHUTDOWN
+}
+
+StartupState runtime_state = NOT_STARTED;
+
+Callback* ctor_first;
+Callback* dtor_first;
+
+fn void runtime_startup() @public @export("__c3_runtime_startup")
+{
+	if (runtime_state != NOT_STARTED) return;
+	runtime_state = INIT;
+	_dyld_register_func_for_add_image(&dl_reg_callback);
+	assert(runtime_state == INIT);
+	runtime_state = RUN_CTORS;
+	Callback* ctor = ctor_first;
+	while (ctor)
+	{
+		ctor.xtor();
+		ctor = ctor.next;
+	}
+	assert(runtime_state == RUN_CTORS);
+	runtime_state = READ_DYLIB;
+	ctor_first = null;
+}
+
+fn void runtime_finalize() @public @export("__c3_runtime_finalize")
+{
+	if (runtime_state != READ_DYLIB) return;
+	runtime_state = RUN_DTORS;
+	Callback* dtor = dtor_first;
+	while (dtor)
+	{
+		dtor.xtor();
+		dtor = dtor.next;
+	}
+	assert(runtime_state == RUN_DTORS);
+	runtime_state = SHUTDOWN;
+}
+
+fn void append_xxlizer(Callback** ref, Callback* cb)
+{
+	while (Callback* current = *ref, current)
+	{
+		if (current.priority > cb.priority)
+		{
+			cb.next = current;
+			break;
+		}
+		ref = &current.next;
+	}
+	*ref = cb;
+}
+
+struct TypeId
+{
+	char type;
+	TypeId* parentof;
+	DynamicMethod* dtable;
+	usz sizeof;
+	TypeId* inner;
+	usz len;
+	typeid[*] additional;
+}
+
+fn void dl_reg_callback(MachHeader* mh, isz vmaddr_slide)
+{
+	usz size = 0;
+	assert(runtime_state == INIT || runtime_state == READ_DYLIB, "State was %s", runtime_state);
+	foreach (&dm : find_segment_section_body(mh, "__DATA", "__c3_dynamic", DynamicMethod))
+	{
+		TypeId* type = dm.type;
+		dm.next = type.dtable;
+		type.dtable = dm;
+		DynamicMethod* m = dm;
+		while (m)
+		{
+			m = m.next;
+		}
+	}
+	foreach (&cb : find_segment_section_body(mh, "__DATA", "__c3dtor", Callback))
+	{
+		append_xxlizer(&dtor_first, cb);
+	}
+	foreach (&cb : find_segment_section_body(mh, "__DATA", "__c3ctor", Callback))
+	{
+		append_xxlizer(&ctor_first, cb);
+	}
+	if (runtime_state != READ_DYLIB) return;
+	runtime_state = RUN_DYLIB_CTORS;
+	Callback* ctor = ctor_first;
+	ctor_first = null;
+	while (ctor)
+	{
+		ctor.xtor();
+		ctor = ctor.next;
+	}
+	assert(runtime_state == RUN_DYLIB_CTORS);
+	runtime_state = READ_DYLIB;
+}

lib/std/core/private/main_stub.c3

@@ -0,0 +1,166 @@
+module std::core::main_stub;
+
+macro usz _strlen(ptr) @private
+{
+	usz len = 0;
+	while (ptr[len]) len++;
+	return len;
+}
+
+macro int @main_to_err_main(#m, int, char**)
+{
+	if (catch #m()) return 1;
+	return 0;
+}
+macro int @main_to_int_main(#m, int, char**) => #m();
+macro int @main_to_void_main(#m, int, char**)
+{
+	#m();
+	return 0;
+}
+
+macro String[] args_to_strings(int argc, char** argv) @private
+{
+	String[] list = mem::alloc_array(String, argc);
+	for (int i = 0; i < argc; i++)
+	{
+		char* arg = argv[i];
+		usz len = 0;
+		list[i] = (String)arg[:_strlen(arg)];
+	}
+	return list;
+}
+
+
+macro int @main_to_err_main_args(#m, int argc, char** argv)
+{
+	String[] list = args_to_strings(argc, argv);
+	defer free(list.ptr);
+	if (catch #m(list)) return 1;
+	return 0;
+}
+
+macro int @main_to_int_main_args(#m, int argc, char** argv)
+{
+	String[] list = args_to_strings(argc, argv);
+	defer free(list.ptr);
+	return #m(list);
+}
+
+macro int @main_to_void_main_args(#m, int argc, char** argv)
+{
+	String[] list = args_to_strings(argc, argv);
+	defer free(list.ptr);
+	#m(list);
+	return 0;
+}
+
+module std::core::main_stub @if(env::WIN32);
+
+extern fn Char16** _win_command_line_to_argv_w(ushort* cmd_line, int* argc_ptr) @extern("CommandLineToArgvW");
+
+macro String[] win_command_line_to_strings(ushort* cmd_line) @private
+{
+	int argc;
+	Char16** argv = _win_command_line_to_argv_w(cmd_line, &argc);
+	return wargs_strings(argc, argv);
+}
+
+macro String[] wargs_strings(int argc, Char16** argv) @private
+{
+	String[] list = mem::alloc_array(String, argc);
+	for (int i = 0; i < argc; i++)
+	{
+		Char16* arg = argv[i];
+		Char16[] argstring = arg[:_strlen(arg)];
+		list[i] = string::new_from_utf16(argstring) ?? "?".copy();
+	}
+	return list[:argc];
+}
+
+macro void release_wargs(String[] list) @private
+{
+	foreach (s : list) free(s.ptr);
+	free(list.ptr);
+}
+
+macro int @win_to_err_main_noargs(#m, void* handle, void* prev_handle, Char16* cmd_line, int show_cmd)
+{
+	if (catch #m()) return 1;
+	return 0;
+}
+macro int @win_to_int_main_noargs(#m, void* handle, void* prev_handle, Char16* cmd_line, int show_cmd) => #m();
+macro int @win_to_void_main_noargs(#m, void* handle, void* prev_handle, Char16* cmd_line, int show_cmd)
+{
+	#m();
+	return 0;
+}
+
+macro int @win_to_err_main_args(#m, void* handle, void* prev_handle, Char16* cmd_line, int show_cmd)
+{
+	String[] args = win_command_line_to_strings(cmd_line);
+	defer release_wargs(args);
+	if (catch #m(args)) return 1;
+	return 0;
+}
+
+macro int @win_to_int_main_args(#m, void* handle, void* prev_handle, Char16* cmd_line, int show_cmd)
+{
+	String[] args = win_command_line_to_strings(cmd_line);
+	defer release_wargs(args);
+	return #m(args);
+}
+
+macro int @win_to_void_main_args(#m, void* handle, void* prev_handle, Char16* cmd_line, int show_cmd)
+{
+	String[] args = win_command_line_to_strings(cmd_line);
+	defer release_wargs(args);
+	#m(args);
+	return 0;
+}
+
+macro int @win_to_err_main(#m, void* handle, void* prev_handle, Char16* cmd_line, int show_cmd)
+{
+	String[] args = win_command_line_to_strings(cmd_line);
+	defer release_wargs(args);
+	if (catch #m(handle, prev_handle, args, show_cmd)) return 1;
+	return 0;
+}
+
+macro int @win_to_int_main(#m, void* handle, void* prev_handle, Char16* cmd_line, int show_cmd)
+{
+	String[] args = win_command_line_to_strings(cmd_line);
+	defer release_wargs(args);
+	return #m(handle, prev_handle, args, show_cmd);
+}
+
+macro int @win_to_void_main(#m, void* handle, void* prev_handle, Char16* cmd_line, int show_cmd)
+{
+	String[] args = win_command_line_to_strings(cmd_line);
+	defer release_wargs(args);
+	#m(handle, prev_handle, args, show_cmd);
+	return 0;
+}
+
+macro int @wmain_to_err_main_args(#m, int argc, Char16** argv)
+{
+	String[] args = wargs_strings(argc, argv);
+	defer release_wargs(args);
+	if (catch #m(args)) return 1;
+	return 1;
+}
+
+macro int @wmain_to_int_main_args(#m, int argc, Char16** argv)
+{
+	String[] args = wargs_strings(argc, argv);
+	defer release_wargs(args);
+	return #m(args);
+}
+
+macro int @wmain_to_void_main_args(#m, int argc, Char16** argv)
+{
+	String[] args = wargs_strings(argc, argv);
+	defer release_wargs(args);
+	#m(args);
+	return 0;
+}

lib/std/core/runtime.c3

@@ -0,0 +1,250 @@
+// Copyright (c) 2021 Christoffer Lerno. All rights reserved.
+// Use of this source code is governed by the MIT license
+// a copy of which can be found in the LICENSE_STDLIB file.
+module std::core::runtime;
+import libc, std::time, std::io, std::sort;
+
+struct AnyRaw
+{
+	void* ptr;
+	typeid type;
+}
+
+struct SliceRaw
+{
+	void* ptr;
+	usz len;
+}
+
+def BenchmarkFn = fn void!();
+
+struct BenchmarkUnit
+{
+	String name;
+	BenchmarkFn func;
+}
+
+fn BenchmarkUnit[] benchmark_collection_create(Allocator allocator = allocator::heap())
+{
+	BenchmarkFn[] fns = $$BENCHMARK_FNS;
+	String[] names = $$BENCHMARK_NAMES;
+	BenchmarkUnit[] benchmarks = allocator::alloc_array(allocator, BenchmarkUnit, names.len);
+	foreach (i, benchmark : fns)
+	{
+		benchmarks[i] = { names[i], fns[i] };
+	}
+	return benchmarks;
+}
+
+const DEFAULT_BENCHMARK_WARMUP_ITERATIONS = 3;
+const DEFAULT_BENCHMARK_MAX_ITERATIONS = 10000;
+
+uint benchmark_warmup_iterations @private = DEFAULT_BENCHMARK_WARMUP_ITERATIONS;
+uint benchmark_max_iterations @private = DEFAULT_BENCHMARK_MAX_ITERATIONS;
+
+fn void set_benchmark_warmup_iterations(uint value) @builtin
+{
+    benchmark_warmup_iterations = value;
+}
+
+fn void set_benchmark_max_iterations(uint value) @builtin
+{
+    assert(value > 0);
+    benchmark_max_iterations = value;
+}
+
+fn bool run_benchmarks(BenchmarkUnit[] benchmarks)
+{
+	int benchmarks_passed = 0;
+	int benchmark_count = benchmarks.len;
+	usz max_name;
+
+	foreach (&unit : benchmarks)
+	{
+		if (max_name < unit.name.len) max_name = unit.name.len;
+	}
+
+	usz len = max_name + 9;
+
+	DString name = dstring::temp_with_capacity(64);
+	name.append_repeat('-', len / 2);
+	name.append(" BENCHMARKS ");
+	name.append_repeat('-', len - len / 2);
+
+	io::printn(name);
+
+	name.clear();
+
+	long sys_clock_started;
+	long sys_clock_finished;
+	long sys_clocks;
+	Clock clock;
+	anyfault err;
+
+	foreach(unit : benchmarks)
+	{
+		defer name.clear();
+		name.appendf("Benchmarking %s ", unit.name);
+		name.append_repeat('.', max_name - unit.name.len + 2);
+		io::printf("%s ", name.str_view());
+
+		for (uint i = 0; i < benchmark_warmup_iterations; i++)
+		{
+			err = @catch(unit.func()) @inline;
+			@volatile_load(err);
+		}
+
+		clock = std::time::clock::now();
+		sys_clock_started = $$sysclock();
+
+		for (uint i = 0; i < benchmark_max_iterations; i++)
+		{
+			err = @catch(unit.func()) @inline;
+			@volatile_load(err);
+		}
+
+		sys_clock_finished = $$sysclock();
+		NanoDuration nano_seconds = clock.mark();
+		sys_clocks = sys_clock_finished - sys_clock_started;
+
+		if (err)
+		{
+			io::printfn("[failed] Failed due to: %s", err);
+			continue;
+		}
+
+		io::printfn("[ok] %.2f ns, %.2f CPU's clocks", (float)nano_seconds / benchmark_max_iterations, (float)sys_clocks / benchmark_max_iterations);
+		benchmarks_passed++;
+	}
+
+	io::printfn("\n%d benchmark%s run.\n", benchmark_count, benchmark_count > 1 ? "s" : "");
+	io::printfn("Benchmarks Result: %s. %d passed, %d failed.",
+		benchmarks_passed < benchmark_count ? "FAILED" : "ok",
+		benchmarks_passed,
+		benchmark_count - benchmarks_passed);
+
+	return benchmark_count == benchmarks_passed;
+}
+
+fn bool default_benchmark_runner()
+{
+	@pool()
+	{
+		return run_benchmarks(benchmark_collection_create(allocator::temp()));
+	};
+}
+
+def TestFn = fn void!();
+
+struct TestUnit
+{
+	String name;
+	TestFn func;
+}
+
+fn TestUnit[] test_collection_create(Allocator allocator = allocator::heap())
+{
+	TestFn[] fns = $$TEST_FNS;
+	String[] names = $$TEST_NAMES;
+	TestUnit[] tests = allocator::alloc_array(allocator, TestUnit, names.len);
+	foreach (i, test : fns)
+	{
+		tests[i] = { names[i], fns[i] };
+	}
+	return tests;
+}
+
+struct TestContext
+{
+	JmpBuf buf;
+}
+
+// Sort the tests by their name in ascending order.
+fn int cmp_test_unit(TestUnit a, TestUnit b)
+{
+	usz an = a.name.len;
+	usz bn = b.name.len;
+	if (an > bn) @swap(a, b);
+	foreach (i, ac : a.name)
+	{
+		char bc = b.name[i];
+		if (ac != bc) return an > bn ? bc - ac : ac - bc;
+	}
+	return (int)(an - bn);
+}
+
+TestContext* test_context @private;
+
+fn void test_panic(String message, String file, String function, uint line)
+{
+	io::printn("[error]");
+	io::print("\n Error: ");
+	io::print(message);
+	io::printn();
+	io::printfn("  - in %s %s:%s.\n", function, file, line);
+	libc::longjmp(&test_context.buf, 1);
+}
+
+fn bool run_tests(TestUnit[] tests)
+{
+	usz max_name;
+	foreach (&unit : tests)
+	{
+		if (max_name < unit.name.len) max_name = unit.name.len;
+	}
+	quicksort(tests, &cmp_test_unit);
+
+	TestContext context;
+	test_context = &context;
+
+	PanicFn old_panic = builtin::panic;
+	defer builtin::panic = old_panic;
+	builtin::panic = &test_panic;
+	int tests_passed = 0;
+	int test_count = tests.len;
+	DString name = dstring::temp_with_capacity(64);
+	usz len = max_name + 9;
+	name.append_repeat('-', len / 2);
+	name.append(" TESTS ");
+	name.append_repeat('-', len - len / 2);
+	io::printn(name);
+	name.clear();
+	foreach(unit : tests)
+	{
+		defer name.clear();
+		name.appendf("Testing %s ", unit.name);
+		name.append_repeat('.', max_name - unit.name.len + 2);
+		io::printf("%s ", name.str_view());
+		if (libc::setjmp(&context.buf) == 0)
+		{
+			if (catch err = unit.func())
+			{
+				io::printfn("[failed] Failed due to: %s", err);
+				continue;
+			}
+			io::printn("[ok]");
+			tests_passed++;
+		}
+	}
+	io::printfn("\n%d test%s run.\n", test_count, test_count > 1 ? "s" : "");
+	io::printfn("Test Result: %s. %d passed, %d failed.",
+		tests_passed < test_count ? "FAILED" : "ok", tests_passed, test_count - tests_passed);
+	return test_count == tests_passed;
+}
+
+fn bool default_test_runner()
+{
+	@pool()
+	{
+		return run_tests(test_collection_create(allocator::temp()));
+	};
+}
+
+module std::core::runtime @if(WASM_NOLIBC);
+
+extern fn void __wasm_call_ctors();
+fn void wasm_initialize() @extern("_initialize") @wasm
+{
+	// The linker synthesizes this to call constructors.
+	__wasm_call_ctors();
+}

lib/std/core/sanitizer/asan.c3

@@ -0,0 +1,127 @@
+// Add this to your code to suppress leak detection or set other default options
+// fn ZString __asan_default_options() @export("__asan_default_options") @if(env::ADDRESS_SANITIZER)
+// {
+//   return "detect_leaks=0";
+// }
+
+// Add this to break on error
+// asan::set_error_report_callback(fn (ZString err)
+// {
+//   breakpoint();
+// });
+
+module std::core::sanitizer::asan;
+
+def ErrorCallback = fn void (ZString);
+
+/**
+ * Marks a memory region ([addr, addr+size)) as unaddressable.
+ *
+ * This memory must be previously allocated by your program. Instrumented
+ * code is forbidden from accessing addresses in this region until it is
+ * unpoisoned. This function is not guaranteed to poison the entire region -
+ * it could poison only a subregion of [addr, addr+size) due to ASan
+ * alignment restrictions.
+ *
+ * NOTE This function is not thread-safe because no two threads can poison or
+ * unpoison memory in the same memory region simultaneously.
+ *
+ * @param addr "Start of memory region."
+ * @param size "Size of memory region."
+ **/
+macro poison_memory_region(void* addr, usz size)
+{
+	$if env::ADDRESS_SANITIZER: 
+		__asan_poison_memory_region(addr, size);
+	$endif 
+}
+
+/**
+ * Marks a memory region ([addr, addr+size)) as addressable.
+ *
+ * This memory must be previously allocated by your program. Accessing
+ * addresses in this region is allowed until this region is poisoned again.
+ * This function could unpoison a super-region of [addr, addr+size) due
+ * to ASan alignment restrictions.
+ *
+ * NOTE This function is not thread-safe because no two threads can
+ * poison or unpoison memory in the same memory region simultaneously.
+ *
+ * @param addr "Start of memory region."
+ * @param size "Size of memory region."
+ **/
+macro unpoison_memory_region(void* addr, usz size)
+{
+	$if env::ADDRESS_SANITIZER: 
+		__asan_unpoison_memory_region(addr, size);
+	$endif 
+}
+
+/**
+ * Checks if an address is poisoned.
+ * @return "True if 'addr' is poisoned (that is, 1-byte read/write access to this address would result in an error report from ASan). Otherwise returns false."
+ * @param addr "Address to check."
+ **/
+macro bool address_is_poisoned(void* addr)
+{
+	$if env::ADDRESS_SANITIZER:
+		return (bool)__asan_address_is_poisoned(addr);
+	$else
+		return false;
+	$endif
+}
+
+/**
+ * Checks if a region is poisoned.
+ *
+ * If at least one byte in [beg, beg+size) is poisoned, returns the
+ * address of the first such byte. Otherwise returns 0.
+ *
+ * @param beg "Start of memory region."
+ * @param size "Start of memory region."
+ * @return "Address of first poisoned byte."
+ **/
+macro void* region_is_poisoned(void* beg, usz size)
+{
+	$if env::ADDRESS_SANITIZER:
+		return __asan_region_is_poisoned(addr);
+	$else
+		return null;
+	$endif
+}
+
+/**
+ * Sets the callback function to be called during ASan error reporting.
+ **/
+fn void set_error_report_callback(ErrorCallback callback)
+{
+	$if env::ADDRESS_SANITIZER:
+		__asan_set_error_report_callback(callback);
+	$endif
+}
+
+module std::core::sanitizer::asan @if(env::ADDRESS_SANITIZER);
+
+extern fn void __asan_poison_memory_region(void* addr, usz size);
+extern fn void __asan_unpoison_memory_region(void* addr, usz size);
+extern fn CInt __asan_address_is_poisoned(void* addr);
+extern fn void* __asan_region_is_poisoned(void* beg, usz size);
+extern fn void __asan_describe_address(void* addr);
+extern fn CInt __asan_report_present();
+extern fn void* __asan_get_report_pc();
+extern fn void* __asan_get_report_bp();
+extern fn void* __asan_get_report_sp();
+extern fn void* __asan_get_report_address();
+extern fn CInt __asan_get_report_access_type();
+extern fn usz __asan_get_report_access_size();
+extern fn ZString __asan_get_report_description();
+extern fn ZString __asan_locate_address(void* addr, char* name, usz name_size, void** region_address, usz* region_size);
+extern fn usz __asan_get_alloc_stack(void* addr, void** trace, usz size, CInt* thread_id);
+extern fn usz __asan_get_free_stack(void* addr, void** trace, usz size, CInt* thread_id);
+extern fn void __asan_get_shadow_mapping(usz* shadow_scale, usz* shadow_offset);
+extern fn void __asan_set_error_report_callback(ErrorCallback callback);
+extern fn void __asan_print_accumulated_stats();
+extern fn void* __asan_get_current_fake_stack();
+extern fn void* __asan_addr_is_in_fake_stack(void* fake_stack, void* addr, void** beg, void** end);
+extern fn void __asan_handle_no_return();
+extern fn CInt __asan_update_allocation_context(void* addr);

lib/std/core/sanitizer/sanitizer.c3

@@ -0,0 +1,80 @@
+module std::core::sanitizer;
+
+macro void annotate_contiguous_container(void* beg, void* end, void* old_mid, void* new_mid)
+{
+	$if env::ADDRESS_SANITIZER:
+		__sanitizer_annotate_contiguous_container(beg, end, old_mid, new_mid);
+	$endif
+}
+
+macro void annotate_double_ended_contiguous_container(void* storage_beg, void* storage_end, void* old_container_beg, void* old_container_end, void* new_container_beg, void* new_container_end)
+{
+	$if env::ADDRESS_SANITIZER:
+		__sanitizer_annotate_double_ended_contiguous_container(storage_beg, storage_end, old_container_beg, old_container_end, new_container_beg, new_container_end);
+	$endif
+}
+
+macro void print_stack_trace()
+{
+	$if env::ADDRESS_SANITIZER:
+		__sanitizer_print_stack_trace();
+	$endif
+}
+
+fn void set_death_callback(VoidFn callback)
+{
+	$if env::ANY_SANITIZER:
+		__sanitizer_set_death_callback(callback);
+	$endif
+}
+
+module std::core::sanitizer @if (env::ANY_SANITIZER);
+
+struct __Sanitizer_sandbox_arguments
+{
+	CInt coverage_sandboxed;
+	iptr coverage_fd;
+	CUInt coverage_max_block_size;
+}
+
+extern fn void __sanitizer_set_report_path(ZString path);
+extern fn void __sanitizer_set_report_fd(void* fd);
+extern fn ZString __sanitizer_get_report_path();
+extern fn void __sanitizer_sandbox_on_notify(__Sanitizer_sandbox_arguments* args);
+extern fn void __sanitizer_report_error_summary(ZString error_summary);
+extern fn ushort __sanitizer_unaligned_load16(void* p);
+extern fn uint __sanitizer_unaligned_load32(void* p);
+extern fn ulong __sanitizer_unaligned_load64(void* p);
+extern fn void __sanitizer_unaligned_store16(void* p, ushort x);
+extern fn void __sanitizer_unaligned_store32(void* p, uint x);
+extern fn void __sanitizer_unaligned_store64(void* p, ulong x);
+extern fn CInt __sanitizer_acquire_crash_state();
+extern fn void __sanitizer_annotate_contiguous_container(void* beg, void* end, void* old_mid, void* new_mid);
+extern fn void __sanitizer_annotate_double_ended_contiguous_container(void* storage_beg, void* storage_end,
+	void* old_container_beg, void* old_container_end,
+	void* new_container_beg, void* new_container_end);
+extern fn CInt __sanitizer_verify_contiguous_container(void* beg, void* mid, void* end);
+extern fn CInt __sanitizer_verify_double_ended_contiguous_container(
+	void* storage_beg, void* container_beg,
+	void* container_end, void* storage_end);
+extern fn void* __sanitizer_contiguous_container_find_bad_address(void* beg, void* mid, void* end);
+extern fn void* __sanitizer_double_ended_contiguous_container_find_bad_address(
+	void* storage_beg, void* container_beg,
+	void* container_end, void* storage_end);
+
+extern fn void __sanitizer_print_stack_trace();
+extern fn void __sanitizer_symbolize_pc(void* pc, ZString fmt, char* out_buf, usz out_buf_size);
+extern fn void __sanitizer_symbolize_global(void* data_ptr, ZString fmt, char* out_buf, usz out_buf_size);
+extern fn void __sanitizer_set_death_callback(VoidFn callback);
+extern fn void __sanitizer_weak_hook_memcmp(void* called_pc, void* s1, void* s2, usz n, CInt result);
+extern fn void __sanitizer_weak_hook_strncmp(void* called_pc, ZString s1, ZString s2, usz n, CInt result);
+extern fn void __sanitizer_weak_hook_strncasecmp(void* called_pc, ZString s1, ZString s2, usz n, CInt result);
+extern fn void __sanitizer_weak_hook_strcmp(void* called_pc, ZString s1, ZString s2, CInt result);
+extern fn void __sanitizer_weak_hook_strcasecmp(void* called_pc, ZString s1, ZString s2, CInt result);
+extern fn void __sanitizer_weak_hook_strstr(void* called_pc, ZString s1, ZString s2, char* result);
+extern fn void __sanitizer_weak_hook_strcasestr(void* called_pc, ZString s1, ZString s2, char* result);
+extern fn void __sanitizer_weak_hook_memmem(void* called_pc, void* s1, usz len1, void* s2, usz len2, void* result);
+extern fn void __sanitizer_print_memory_profile(usz top_percent, usz max_number_of_contexts);
+extern fn void __sanitizer_start_switch_fiber(void** fake_stack_save, void* bottom, usz size);
+extern fn void __sanitizer_finish_switch_fiber(void* fake_stack_save, void** bottom_old, usz* size_old);
+extern fn CInt __sanitizer_get_module_and_offset_for_pc(void* pc, char* module_path, usz module_path_len, void** pc_offset);

lib/std/core/sanitizer/tsan.c3

@@ -0,0 +1,39 @@
+module std::core::sanitizer::tsan;
+
+distinct MutexFlags = inline CUInt;
+
+const MutexFlags MUTEX_LINKER_INIT      = 1 << 0;
+const MutexFlags MUTEX_WRITE_REENTRANT  = 1 << 1;
+const MutexFlags MUTEX_READ_REENTRANT   = 1 << 2;
+const MutexFlags MUTEX_NOT_STATIC       = 1 << 8;
+const MutexFlags MUTEX_READ_LOCK        = 1 << 3;
+const MutexFlags MUTEX_TRY_LOCK         = 1 << 4;
+const MutexFlags MUTEX_TRY_LOCK_FAILED  = 1 << 5;
+const MutexFlags MUTEX_RECURSIVE_LOCK   = 1 << 6;
+const MutexFlags MUTEX_RECURSIVE_UNLOCK = 1 << 7;
+const MutexFlags MUTEX_TRY_READ_LOCK = MUTEX_READ_LOCK | MUTEX_TRY_LOCK;
+const MutexFlags MUTEX_TRY_READ_LOCK_FAILED = MUTEX_TRY_READ_LOCK | MUTEX_TRY_LOCK_FAILED;
+
+macro void mutex_create(void* addr, MutexFlags flags) { $if env::THREAD_SANITIZER: __tsan_mutex_create(addr, flags); $endif }
+macro void mutex_destroy(void* addr, MutexFlags flags) { $if env::THREAD_SANITIZER: __tsan_mutex_destroy(addr, flags); $endif }
+macro void mutex_pre_lock(void* addr, MutexFlags flags) { $if env::THREAD_SANITIZER: __tsan_mutex_pre_lock(addr, flags); $endif }
+macro void mutex_post_lock(void* addr, MutexFlags flags, CInt recursion) { $if env::THREAD_SANITIZER: __tsan_mutex_post_lock(addr, flags, recursion); $endif }
+macro CInt mutex_pre_unlock(void* addr, MutexFlags flags) { $if env::THREAD_SANITIZER: return __tsan_mutex_pre_unlock(addr, flags); $else return 0; $endif }
+macro void mutex_post_unlock(void* addr, MutexFlags flags) { $if env::THREAD_SANITIZER: __tsan_mutex_post_unlock(addr, flags); $endif }
+macro void mutex_pre_signal(void* addr, MutexFlags flags) { $if env::THREAD_SANITIZER: __tsan_mutex_pre_signal(addr, flags); $endif }
+macro void mutex_post_signal(void* addr, MutexFlags flags) { $if env::THREAD_SANITIZER: __tsan_mutex_post_signal(addr, flags); $endif }
+macro void mutex_pre_divert(void* addr, MutexFlags flags) { $if env::THREAD_SANITIZER: __tsan_mutex_pre_divert(addr, flags); $endif }
+macro void mutex_post_divert(void* addr, MutexFlags flags) { $if env::THREAD_SANITIZER: __tsan_mutex_post_divert(addr, flags); $endif }
+
+module std::core::sanitizer::tsan @if(env::THREAD_SANITIZER) @private;
+
+extern fn void __tsan_mutex_create(void* addr, CUInt flags);
+extern fn void __tsan_mutex_destroy(void* addr, CUInt flags);
+extern fn void __tsan_mutex_pre_lock(void* addr, CUInt flags);
+extern fn void __tsan_mutex_post_lock(void* addr, CUInt flags, CInt recursion);
+extern fn CInt __tsan_mutex_pre_unlock(void* addr, CUInt flags);
+extern fn void __tsan_mutex_post_unlock(void* addr, CUInt flags);
+extern fn void __tsan_mutex_pre_signal(void* addr, CUInt flags);
+extern fn void __tsan_mutex_post_signal(void* addr, CUInt flags);
+extern fn void __tsan_mutex_pre_divert(void* addr, CUInt flags);
+extern fn void __tsan_mutex_post_divert(void* addr, CUInt flags);

lib/std/core/string.c3

@@ -0,0 +1,740 @@
+module std::core::string;
+import std::ascii;
+
+distinct String @if(!$defined(String)) = inline char[];
+distinct ZString = inline char*;
+distinct WString = inline Char16*;
+def Char32 = uint;
+def Char16 = ushort;
+
+fault UnicodeResult
+{
+	INVALID_UTF8,
+	INVALID_UTF16,
+	CONVERSION_FAILED,
+}
+
+const uint SURROGATE_OFFSET @private = 0x10000;
+const uint SURROGATE_GENERIC_MASK @private = 0xF800;
+const uint SURROGATE_MASK @private = 0xFC00;
+const uint SURROGATE_CODEPOINT_MASK @private = 0x03FF;
+const uint SURROGATE_BITS @private = 10;
+const uint SURROGATE_LOW_VALUE @private = 0xDC00;
+const uint SURROGATE_HIGH_VALUE @private = 0xD800;
+
+fault NumberConversion
+{
+	EMPTY_STRING,
+	NEGATIVE_VALUE,
+	MALFORMED_INTEGER,
+	INTEGER_OVERFLOW,
+	MALFORMED_FLOAT,
+	FLOAT_OUT_OF_RANGE,
+}
+
+/**
+ * Return a temporary String created using the formatting function.
+ *
+ * @param [in] fmt `The formatting string`
+ **/
+macro String tformat(String fmt, ...)
+{
+	DString str = dstring::temp_with_capacity(fmt.len + $vacount * 8);
+	str.appendf(fmt, $vasplat);
+	return str.str_view();
+}
+
+/**
+ * Return a temporary ZString created using the formatting function.
+ *
+ * @param [in] fmt `The formatting string`
+ **/
+macro ZString tformat_zstr(String fmt, ...)
+{
+	DString str = dstring::temp_with_capacity(fmt.len + $vacount * 8);
+	str.appendf(fmt, $vasplat);
+	return str.zstr_view();
+}
+
+/**
+ * Return a new String created using the formatting function.
+ *
+ * @param [in] fmt `The formatting string`
+ * @param [inout] allocator `The allocator to use`
+ **/
+macro String new_format(String fmt, ..., Allocator allocator = allocator::heap())
+{
+	@pool(allocator)
+	{
+		DString str = dstring::temp_with_capacity(fmt.len + $vacount * 8);
+		str.appendf(fmt, $vasplat);
+		return str.copy_str(allocator);
+	};
+}
+
+/**
+ * Return a new ZString created using the formatting function.
+ *
+ * @param [in] fmt `The formatting string`
+ * @param [inout] allocator `The allocator to use`
+ **/
+macro ZString new_format_zstr(String fmt, ..., Allocator allocator = allocator::heap())
+{
+	@pool(allocator)
+	{
+		DString str = dstring::temp_with_capacity(fmt.len + $vacount * 8);
+		str.appendf(fmt, $vasplat);
+		return str.copy_zstr(allocator);
+	};
+}
+
+/**
+ * Check if a character is in a set.
+ *
+ * @param c `the character to check`
+ * @param [in] set `The formatting string`
+ * @pure
+ * @return `True if a character is in the set`
+ **/
+macro bool char_in_set(char c, String set)
+{
+	foreach (ch : set) if (ch == c) return true;
+	return false;
+}
+
+fn String join_new(String[] s, String joiner, Allocator allocator = allocator::heap())
+{
+	if (!s)
+	{
+		return (String)allocator::new_array(allocator, char, 2)[:0];
+	}
+
+	usz total_size = joiner.len * s.len;
+	foreach (String* &str : s)
+	{
+		total_size += str.len;
+	}
+	@pool(allocator)
+	{
+		DString res = dstring::temp_with_capacity(total_size);
+		res.append(s[0]);
+		foreach (String* &str : s[1..])
+		{
+			res.append(joiner);
+			res.append(*str);
+		}
+		return res.copy_str(allocator);
+	};
+}
+
+/**
+ * Remove characters from the front and end of a string.
+ *
+ * @param [in] string `The string to trim`
+ * @param [in] to_trim `The set of characters to trim, defaults to whitespace`
+ * @pure
+ * @return `a substring of the string passed in`
+ **/
+fn String String.trim(string, String to_trim = "\t\n\r ")
+{
+	usz start = 0;
+	usz len = string.len;
+	while (start < len && char_in_set(string[start], to_trim)) start++;
+	if (start == len) return string[:0];
+	usz end = len - 1;
+	while (end > start && char_in_set(string[end], to_trim)) end--;
+	return string[start..end];
+}
+
+/**
+ * Check if the String starts with the needle.
+ *
+ * @param [in] string
+ * @param [in] needle
+ * @pure
+ * @return `'true' if the string starts with the needle`
+ **/
+fn bool String.starts_with(string, String needle)
+{
+	if (needle.len > string.len) return false;
+	if (!needle.len) return true;
+	return string[:needle.len] == needle;
+}
+
+/**
+ * Check if the String ends with the needle.
+ *
+ * @param [in] string
+ * @param [in] needle
+ * @pure
+ * @return `'true' if the string ends with the needle`
+ **/
+fn bool String.ends_with(string, String needle)
+{
+	if (needle.len > string.len) return false;
+	if (!needle.len) return true;
+	return string[^needle.len..] == needle;
+}
+
+/**
+ * Strip the front of the string if the prefix exists.
+ *
+ * @param [in] string
+ * @param [in] needle
+ * @pure
+ * @return `the substring with the prefix removed`
+ **/
+fn String String.strip(string, String needle)
+{
+	if (!needle.len || !string.starts_with(needle)) return string;
+	return string[needle.len..];
+}
+
+/**
+ * Strip the end of the string if the suffix exists.
+ *
+ * @param [in] string
+ * @param [in] needle
+ * @pure
+ * @return `the substring with the suffix removed`
+ **/
+fn String String.strip_end(string, String needle)
+{
+	if (!needle.len || !string.ends_with(needle)) return string;
+	// Note that this is the safe way if we want to support zero length.
+	return string[:(string.len - needle.len)];
+}
+
+
+/**
+ * Split a string into parts, e.g "a|b|c" split with "|" yields { "a", "b", "c" }
+ *
+ * @param [in] s
+ * @param [in] needle
+ * @param [&inout] allocator "The allocator, defaults to the heap allocator"
+ * @param max "Max number of elements, 0 means no limit, defaults to 0"
+ * @require needle.len > 0 "The needle must be at least 1 character long"
+ * @ensure return.len > 0
+ **/
+fn String[] String.split(s, String needle, usz max = 0, Allocator allocator = allocator::heap())
+{
+	usz capacity = 16;
+	usz i = 0;
+	String* holder = allocator::alloc_array(allocator, String, capacity);
+	bool no_more = false;
+	while (!no_more)
+	{
+		usz! index = i == max - 1 ? SearchResult.MISSING? : s.index_of(needle);
+		String res @noinit;
+		if (try index)
+		{
+			res = s[:index];
+			s = s[index + needle.len..];
+		}
+		else
+		{
+			res = s;
+			no_more = true;
+		}
+		if (i == capacity)
+		{
+			capacity *= 2;
+			holder = allocator::realloc(allocator, holder, String.sizeof * capacity);
+		}
+		holder[i++] = res;
+	}
+	return holder[:i];
+}
+
+/**
+ * This function is identical to String.split, but implicitly uses the
+ * temporary allocator.
+ *
+ * @param [in] s
+ * @param [in] needle
+ * @param max "Max number of elements, 0 means no limit, defaults to 0"
+ **/
+fn String[] String.tsplit(s, String needle, usz max = 0)
+{
+	return s.split(needle, max, allocator::temp()) @inline;
+}
+
+/**
+ * Check if a substring is found in the string.
+
+ * @param [in] s
+ * @param [in] needle "The string to look for."
+ * @pure
+ * @return "true if the string contains the substring, false otherwise"
+ **/
+fn bool String.contains(s, String needle)
+{
+	return @ok(s.index_of(needle));
+}
+
+/**
+ * Find the index of the first incidence of a string.
+ *
+ * @param [in] s
+ * @param needle "The character to look for"
+ * @pure
+ * @ensure return < s.len
+ * @return "the index of the needle"
+ * @return! SearchResult.MISSING "if the needle cannot be found"
+ **/
+fn usz! String.index_of_char(s, char needle)
+{
+	foreach (i, c : s)
+	{
+		if (c == needle) return i;
+	}
+	return SearchResult.MISSING?;
+}
+
+/**
+ * Find the index of the first incidence of a character.
+ *
+ * @param [in] s
+ * @param needle "The character to look for"
+ * @param start_index "The index to start with, may exceed max index."
+ * @pure
+ * @ensure return < s.len
+ * @return "the index of the needle"
+ * @return! SearchResult.MISSING "if the needle cannot be found starting from the start_index"
+ **/
+fn usz! String.index_of_char_from(s, char needle, usz start_index)
+{
+	usz len = s.len;
+	if (len <= start_index) return SearchResult.MISSING?;
+	for (usz i = start_index; i < len; i++)
+	{
+		if (s[i] == needle) return i;
+	}
+	return SearchResult.MISSING?;
+}
+
+/**
+ * Find the index of the first incidence of a character starting from the end.
+ *
+ * @param [in] s
+ * @param needle "the character to find"
+ * @pure
+ * @ensure return < s.len
+ * @return "the index of the needle"
+ * @return! SearchResult.MISSING "if the needle cannot be found"
+ **/
+fn usz! String.rindex_of_char(s, char needle)
+{
+	foreach_r (i, c : s)
+	{
+		if (c == needle) return i;
+	}
+	return SearchResult.MISSING?;
+}
+
+/**
+ * Find the index of the first incidence of a string.
+ *
+ * @param [in] s
+ * @param [in] needle
+ * @pure
+ * @ensure return < s.len
+ * @require needle.len > 0 : "The needle must be len 1 or more"
+ * @return "the index of the needle"
+ * @return! SearchResult.MISSING "if the needle cannot be found"
+ **/
+fn usz! String.index_of(s, String needle)
+{
+	usz needed = needle.len;
+	if (needed > 0 && s.len >= needed)
+	{
+		char first = needle[0];
+		foreach (i, c: s[..^needed])
+		{
+			if (c == first && s[i:needed] == needle) return i;
+		}
+	}
+	return SearchResult.MISSING?;
+}
+
+/**
+ * Find the index of the last incidence of a string.
+ *
+ * @param [in] s
+ * @param [in] needle
+ * @pure
+ * @ensure return < s.len
+ * @require needle.len > 0 "The needle must be len 1 or more"
+ * @return "the index of the needle"
+ * @return! SearchResult.MISSING "if the needle cannot be found"
+ **/
+fn usz! String.rindex_of(s, String needle)
+{
+	usz needed = needle.len;
+	if (needed > 0 && s.len >= needed)
+	{
+		char first = needle[0];
+		foreach_r (i, c: s[..^needed])
+		{
+			if (c == first && s[i:needed] == needle) return i;
+		}
+	}
+	return SearchResult.MISSING?;
+}
+
+fn String ZString.str_view(str)
+{
+	return (String)(str[:str.len()]);
+}
+
+fn usz ZString.char_len(str)
+{
+	usz len = 0;
+	char* ptr = (char*)str;
+	while (char c = ptr++[0])
+	{
+		if (c & 0xC0 != 0x80) len++;
+	}
+	return len;
+}
+
+fn usz ZString.len(str)
+{
+	usz len = 0;
+	char* ptr = (char*)str;
+	while (char c = ptr++[0]) len++;
+	return len;
+}
+
+
+fn ZString String.zstr_copy(s, Allocator allocator = allocator::heap())
+{
+	usz len = s.len;
+	char* str = allocator::malloc(allocator, len + 1);
+	mem::copy(str, s.ptr, len);
+	str[len] = 0;
+	return (ZString)str;
+}
+
+fn String String.concat(s1, String s2, Allocator allocator = allocator::heap())
+{
+	usz full_len = s1.len + s2.len;
+	char* str = allocator::malloc(allocator, full_len + 1);
+	usz s1_len = s1.len;
+	mem::copy(str, s1.ptr, s1_len);
+	mem::copy(str + s1_len, s2.ptr, s2.len);
+	str[full_len] = 0;
+	return (String)str[:full_len];
+}
+
+fn String String.tconcat(s1, String s2) => s1.concat(s2, allocator::temp());
+
+
+fn ZString String.zstr_tcopy(s) => s.zstr_copy(allocator::temp()) @inline;
+
+fn String String.copy(s, Allocator allocator = allocator::heap())
+{
+	usz len = s.len;
+	char* str = allocator::malloc(allocator, len + 1);
+	mem::copy(str, s.ptr, len);
+	str[len] = 0;
+	return (String)str[:len];
+}
+
+fn void String.free(&s, Allocator allocator = allocator::heap())
+{
+	if (!s.len) return;
+	allocator::free(allocator, s.ptr);
+	*s = "";
+}
+
+fn String String.tcopy(s) => s.copy(allocator::temp()) @inline;
+
+fn String ZString.copy(z, Allocator allocator = allocator::temp())
+{
+	return z.str_view().copy(allocator) @inline;
+}
+
+fn String ZString.tcopy(z)
+{
+	return z.str_view().copy(allocator::temp()) @inline;
+}
+
+/**
+ * Convert an UTF-8 string to UTF-16
+ * @return "The UTF-16 string as a slice, allocated using the given allocator"
+ * @return! UnicodeResult.INVALID_UTF8 "If the string contained an invalid UTF-8 sequence"
+ * @return! AllocationFailure "If allocation of the string fails"
+ **/
+fn Char16[]! String.to_new_utf16(s, Allocator allocator = allocator::heap())
+{
+	usz len16 = conv::utf16len_for_utf8(s);
+	Char16* data = allocator::alloc_array_try(allocator, Char16, len16 + 1)!;
+	conv::utf8to16_unsafe(s, data)!;
+	data[len16] = 0;
+	return data[:len16];
+}
+
+/**
+ * Convert an UTF-8 string to UTF-16
+ * @return "The UTF-16 string as a slice, allocated using the given allocator"
+ * @return! UnicodeResult.INVALID_UTF8 "If the string contained an invalid UTF-8 sequence"
+ * @return! AllocationFailure "If allocation of the string fails"
+ **/
+fn Char16[]! String.to_temp_utf16(s)
+{
+	return s.to_new_utf16(allocator::temp());
+}
+
+fn WString! String.to_new_wstring(s, Allocator allocator = allocator::heap())
+{
+	return (WString)s.to_new_utf16(allocator).ptr;
+}
+
+fn WString! String.to_temp_wstring(s)
+{
+	return (WString)s.to_temp_utf16().ptr;
+}
+
+fn Char32[]! String.to_new_utf32(s, Allocator allocator = allocator::heap())
+{
+	usz codepoints = conv::utf8_codepoints(s);
+	Char32* data = allocator::alloc_array_try(allocator, Char32, codepoints + 1)!;
+	conv::utf8to32_unsafe(s, data)!;
+	data[codepoints] = 0;
+	return data[:codepoints];
+}
+
+fn Char32[]! String.to_temp_utf32(s)
+{
+	return s.to_new_utf32(allocator::temp());
+}
+
+/**
+ * Convert a string to ASCII lower case.
+ *
+ * @param [inout] s
+ * @pure
+ **/
+fn void String.convert_ascii_to_lower(s)
+{
+	foreach (&c : s) if (c.is_upper() @pure) *c += 'a' - 'A';
+}
+
+fn String String.new_ascii_to_lower(s, Allocator allocator = allocator::heap())
+{
+	String copy = s.copy(allocator);
+	copy.convert_ascii_to_lower();
+	return copy;
+}
+
+fn String String.temp_ascii_to_lower(s, Allocator allocator = allocator::heap())
+{
+	return s.new_ascii_to_lower(allocator::temp());
+}
+
+/**
+ * Convert a string to ASCII upper case.
+ *
+ * @param [inout] s
+ * @pure
+ **/
+fn void String.convert_ascii_to_upper(s)
+{
+	foreach (&c : s) if (c.is_lower() @pure) *c -= 'a' - 'A';
+}
+
+/**
+ * Returns a string converted to ASCII upper case.
+ *
+ * @param [in] s
+ * @param [inout] allocator
+ *
+ * @return `a new String converted to ASCII upper case.`
+ **/
+fn String String.new_ascii_to_upper(s, Allocator allocator = allocator::heap())
+{
+	String copy = s.copy(allocator);
+	copy.convert_ascii_to_upper();
+	return copy;
+}
+
+fn StringIterator String.iterator(s)
+{
+	return { s, 0 };
+}
+
+/**
+ * @param [in] s
+ * @return `a temporary String converted to ASCII upper case.`
+ **/
+fn String String.temp_ascii_to_upper(s)
+{
+	return s.new_ascii_to_upper(allocator::temp());
+}
+
+fn String! new_from_utf32(Char32[] utf32, Allocator allocator = allocator::heap())
+{
+	usz len = conv::utf8len_for_utf32(utf32);
+	char* data = allocator::malloc_try(allocator, len + 1)!;
+	defer catch allocator::free(allocator, data);
+	conv::utf32to8_unsafe(utf32, data);
+	data[len] = 0;
+	return (String)data[:len];
+}
+
+fn String! new_from_utf16(Char16[] utf16, Allocator allocator = allocator::heap())
+{
+	usz len = conv::utf8len_for_utf16(utf16);
+	char* data = allocator::malloc_try(allocator, len + 1)!;
+	defer catch allocator::free(allocator, data);
+	conv::utf16to8_unsafe(utf16, data)!;
+	data[len] = 0;
+	return (String)data[:len];
+}
+
+fn String! new_from_wstring(WString wstring, Allocator allocator = allocator::heap())
+{
+	usz utf16_len;
+	while (wstring[utf16_len] != 0) utf16_len++;
+	Char16[] utf16 = wstring[:utf16_len];
+	return new_from_utf16(utf16, allocator);
+}
+
+fn String! temp_from_wstring(WString wstring) => new_from_wstring(wstring, allocator::temp()) @inline;
+fn String! temp_from_utf16(Char16[] utf16) => new_from_utf16(utf16, allocator::temp()) @inline;
+
+fn usz String.utf8_codepoints(s)
+{
+	usz len = 0;
+	foreach (char c : s)
+	{
+		if (c & 0xC0 != 0x80) len++;
+	}
+	return len;
+}
+
+
+/**
+ * @require (base <= 10 && base > 1) || base == 16 : "Unsupported base"
+ **/
+macro String.to_integer(string, $Type, int base = 10)
+{
+	usz len = string.len;
+	usz index = 0;
+	char* ptr = string.ptr;
+	while (index < len && ascii::is_blank_m(ptr[index])) index++;
+	if (len == index) return NumberConversion.EMPTY_STRING?;
+	bool is_negative;
+	switch (string[index])
+	{
+		case '-':
+			if ($Type.min == 0) return NumberConversion.NEGATIVE_VALUE?;
+			is_negative = true;
+			index++;
+		case '+':
+			index++;
+		default:
+			break;
+	}
+	if (len == index) return NumberConversion.MALFORMED_INTEGER?;
+	$Type base_used = ($Type)base;
+	if (string[index] == '0' && base == 10)
+	{
+		index++;
+		if (index == len) return ($Type)0;
+		switch (string[index])
+		{
+			case 'x':
+			case 'X':
+				base_used = 16;
+				index++;
+			case 'b':
+			case 'B':
+				base_used = 2;
+				index++;
+			case 'o':
+			case 'O':
+				base_used = 8;
+				index++;
+			default:
+				break;
+		}
+		if (len == index) return NumberConversion.MALFORMED_INTEGER?;
+	}
+	$Type value = 0;
+	while (index != len)
+	{
+		char c = {|
+			char ch = string[index++];
+			if (base_used != 16 || ch < 'A') return (char)(ch - '0');
+			if (ch <= 'F') return (char)(ch - 'A' + 10);
+			if (ch < 'a') return NumberConversion.MALFORMED_INTEGER?;
+			if (ch > 'f') return NumberConversion.MALFORMED_INTEGER?;
+			return (char)(ch - 'a' + 10);
+		|}!;
+		if (c >= base_used) return NumberConversion.MALFORMED_INTEGER?;
+		value = {|
+			if (is_negative)
+			{
+				$Type new_value = value * base_used - c;
+				if (new_value > value) return NumberConversion.INTEGER_OVERFLOW?;
+				return new_value;
+			}
+			$Type new_value = value * base_used + c;
+			if (new_value < value) return NumberConversion.INTEGER_OVERFLOW?;
+			return new_value;
+		|}!;
+	}
+	return value;
+}
+
+fn int128! String.to_int128(s, int base = 10) => s.to_integer(int128, base);
+fn long! String.to_long(s, int base = 10) => s.to_integer(long, base);
+fn int! String.to_int(s, int base = 10) => s.to_integer(int, base);
+fn short! String.to_short(s, int base = 10) => s.to_integer(short, base);
+fn ichar! String.to_ichar(s, int base = 10) => s.to_integer(ichar, base);
+
+fn uint128! String.to_uint128(s, int base = 10) => s.to_integer(uint128, base);
+fn ulong! String.to_ulong(s, int base = 10) => s.to_integer(ulong, base);
+fn uint! String.to_uint(s, int base = 10) => s.to_integer(uint, base);
+fn ushort! String.to_ushort(s, int base = 10) => s.to_integer(ushort, base);
+fn char! String.to_uchar(s, int base = 10) => s.to_integer(char, base);
+
+fn double! String.to_double(s) => s.to_real(double);
+fn float! String.to_float(s) => s.to_real(float);
+
+fn Splitter String.splitter(self, String split)
+{
+	return Splitter { self, split, 0 };
+}
+
+struct Splitter
+{
+	String string;
+	String split;
+	usz current;
+}
+
+fn void Splitter.reset(&self)
+{
+	self.current = 0;
+}
+
+fn String! Splitter.next(&self)
+{
+	usz len = self.string.len;
+	usz current = self.current;
+	if (current >= len) return IteratorResult.NO_MORE_ELEMENT?;
+	String remaining = self.string[current..];
+	usz! next = remaining.index_of(self.split);
+	if (try next)
+	{
+		defer self.current = current + next + self.split.len;
+		return remaining[:next];
+	}
+	self.current = len;
+	return remaining;
+}
+
+

lib/std/core/string_iterator.c3

@@ -0,0 +1,49 @@
+module std::core::string::iterator;
+
+struct StringIterator
+{
+	String utf8;
+	usz current;
+}
+
+fn void StringIterator.reset(&self)
+{
+	self.current = 0;
+}
+
+fn Char32! StringIterator.next(&self)
+{
+    usz len = self.utf8.len;
+    usz current = self.current;
+    if (current >= len) return IteratorResult.NO_MORE_ELEMENT?;
+    usz read = (len - current < 4 ? len - current : 4);
+    Char32 res = conv::utf8_to_char32(&self.utf8[current], &read)!;
+    self.current += read;
+    return res;
+}
+
+fn Char32! StringIterator.peek(&self)
+{
+    usz len = self.utf8.len;
+    usz current = self.current;
+    if (current >= len) return IteratorResult.NO_MORE_ELEMENT?;
+    usz read = (len - current < 4 ? len - current : 4);
+    Char32 res = conv::utf8_to_char32(&self.utf8[current], &read)!;
+    return res;
+}
+
+fn bool StringIterator.has_next(&self)
+{
+    return self.current < self.utf8.len;
+}
+
+fn Char32! StringIterator.get(&self)
+{
+    usz len = self.utf8.len;
+    usz current = self.current;
+    usz read = (len - current < 4 ? len - current : 4);
+    usz index = current > read ? current - read : 0;
+    if (index >= len) return IteratorResult.NO_MORE_ELEMENT?;
+    Char32 res = conv::utf8_to_char32(&self.utf8[index], &read)!;
+    return res;
+}

lib/std/core/string_to_real.c3

@@ -0,0 +1,487 @@
+module std::core::string;
+import std::math;
+
+// Float parsing based on code in Musl floatscan.c by Rich Felker.
+// Musl uses the MIT license, copied below:
+// ----------------------------------------------------------------------
+//   Copyright © 2005-2014 Rich Felker, et al.
+//
+//   Permission is hereby granted, free of charge, to any person obtaining
+//   a copy of this software and associated documentation files (the
+//   "Software"), to deal in the Software without restriction, including
+//   without limitation the rights to use, copy, modify, merge, publish,
+//   distribute, sublicense, and/or sell copies of the Software, and to
+//   permit persons to whom the Software is furnished to do so, subject to
+//   the following conditions:
+//
+//   The above copyright notice and this permission notice shall be
+//   included in all copies or substantial portions of the Software.
+//
+//   THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+//   EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+//   MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+//   IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY
+//   CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+//   TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+//   SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+//   ----------------------------------------------------------------------
+
+const KMAX = 128;
+const MASK = KMAX - 1;
+const B1B_DIG = 2;
+const uint[2] B1B_MAX = { 9007199, 254740991 };
+
+/**
+ * @require chars.len > 0
+ **/
+macro double! decfloat(char[] chars, int $bits, int $emin, int sign)
+{
+	uint[KMAX] x;
+	const uint[2] TH = B1B_MAX;
+	int emax = - $emin - $bits + 3;
+
+	const int[*] P10S = { 10, 100, 1000, 10000, 100000, 1000000, 10000000, 100000000 };
+	usz index;
+	bool got_digit = chars[0] == '0';
+	bool got_rad;
+	long lrp, dc;
+	int k, j, lnz;
+	usz len = chars.len;
+	usz last_char = len - 1;
+
+	assert(len);
+
+	char c @noinit;
+	// Skip past first characters
+	while ((c = chars[index]) == '0')
+	{
+		if (index == last_char) return sign * 0.0;
+		index++;
+	}
+
+	if (c == '.')
+	{
+		got_rad = true;
+		if (index == last_char)
+		{
+			if (!got_digit) return NumberConversion.MALFORMED_FLOAT?;
+			return sign * 0.0;
+		}
+		if (index != last_char && (c = chars[++index]) == '0')
+		{
+			lrp--;
+			got_digit = true;
+			while (last_char != index && (c = chars[++index]) == '0')
+			{
+				lrp--;
+			}
+		}
+	}
+
+	while (c - '0' < 10u || c == '.')
+	{
+		switch
+		{
+			case c == '.':
+				if (got_rad) return NumberConversion.MALFORMED_FLOAT?;
+				got_rad = true;
+				lrp = dc;
+			case k < KMAX - 3:
+				dc++;
+				if (c != '0') lnz = (int)dc;
+				if (j)
+				{
+					x[k] = x[k] * 10 + c - '0';
+				}
+				else
+				{
+					x[k] = c - '0';
+				}
+				if (++j == 9)
+				{
+					k++;
+					j = 0;
+				}
+				got_digit = true;
+			default:
+				dc++;
+				if (c != '0') x[KMAX - 4] |= 1;
+
+		}
+		if (index == last_char) break;
+		assert(index < last_char);
+		c = chars[++index];
+	}
+	if (!got_rad) lrp = dc;
+	if (!got_digit) return NumberConversion.MALFORMED_FLOAT?;
+	if ((c | 32) == 'e')
+	{
+		if (last_char == index) return NumberConversion.MALFORMED_FLOAT?;
+		long e10 = String.to_long((String)chars[index + 1..]) ?? NumberConversion.MALFORMED_FLOAT?!;
+		lrp += e10;
+	}
+	else if (index != last_char)
+	{
+		return NumberConversion.MALFORMED_FLOAT?;
+	}
+	// Handle zero specially to avoid nasty special cases later
+	if (!x[0]) return sign * 0.0;
+
+	// Optimize small integers (w/no exponent) and over/under-flow
+	if (lrp == dc && dc < 10 && ($bits > 30 || (ulong)x[0] >> $bits == 0)) return sign * (double)x[0];
+	if (lrp > - $emin / 2) return NumberConversion.FLOAT_OUT_OF_RANGE?;
+	if (lrp < $emin - 2 * math::DOUBLE_MANT_DIG) return NumberConversion.FLOAT_OUT_OF_RANGE?;
+
+	// Align incomplete final B1B digit
+	if (j)
+	{
+		for (; j < 9; j++) x[k] *= 10;
+		k++;
+		j = 0;
+	}
+
+	int a;
+	int z = k;
+	int e2;
+	long rp = lrp;
+
+	// Optimize small to mid-size integers (even in exp. notation)
+	if (lnz < 9 && lnz <= rp && rp < 18)
+	{
+		if (rp == 9) return sign * (double)x[0];
+		if (rp < 9) return sign * (double)x[0] / P10S[8 - rp];
+		int bitlim = $bits - 3 * (int)(rp - 9);
+		if (bitlim > 30 || x[0] >> bitlim == 0) return sign * (double)x[0] * P10S[rp - 10];
+	}
+
+	// Align radix point to B1B digit boundary
+	if (rp % 9)
+	{
+		long rpm9 = rp >= 0 ? rp % 9 : rp % 9 + 9;
+		int p10 = P10S[8 - rpm9];
+		uint carry = 0;
+		for (k = a; k != z; k++)
+		{
+			uint tmp = x[k] % p10;
+			x[k] = x[k] / p10 + carry;
+			carry = 1000000000 / p10 * tmp;
+			if (k == a && !x[k])
+			{
+				a = (a + 1) & MASK;
+				rp -= 9;
+			}
+		}
+		if (carry) x[z++] = carry;
+		rp += 9 - rpm9;
+	}
+
+	// Upscale until desired number of bits are left of radix point
+	while (rp < 9 * B1B_DIG || (rp == 9 * B1B_DIG && x[a] < TH[0]))
+	{
+		uint carry = 0;
+		e2 -= 29;
+		for (k = (z - 1) & MASK; ; k = (k - 1) & MASK)
+		{
+			ulong tmp = (ulong)x[k] << 29 + carry;
+			if (tmp > 1000000000)
+			{
+				carry = (uint)(tmp / 1000000000);
+				x[k] = (uint)(tmp % 1000000000);
+			}
+			else
+			{
+				carry = 0;
+				x[k] = (uint)tmp;
+			}
+			if (k == (z - 1) & MASK && k != a && !x[k]) z = k;
+			if (k == a) break;
+		}
+		if (carry)
+		{
+			rp += 9;
+			a = (a - 1) & MASK;
+			if (a == z)
+			{
+				z = (z - 1) & MASK;
+				x[(z - 1) & MASK] |= x[z];
+			}
+			x[a] = carry;
+		}
+	}
+
+	// Downscale until exactly number of bits are left of radix point
+	while (true)
+	{
+		uint carry = 0;
+		int sh = 1;
+		int i;
+		for (i = 0; i < B1B_DIG; i++)
+		{
+			k = (a + i) & MASK;
+			if (k == z || x[k] < TH[i])
+			{
+				i = B1B_DIG;
+				break;
+			}
+			if (x[(a + i) & MASK] > TH[i]) break;
+		}
+		if (i == B1B_DIG && rp == 9 * B1B_DIG) break;
+		if (rp > 9 + 9 * B1B_DIG) sh = 9;
+		e2 += sh;
+		for (k = a; k != z; k = (k+1) & MASK)
+		{
+			uint tmp = x[k] & (1 << sh - 1);
+			x[k] = x[k] >> sh + carry;
+			carry = (1000000000 >> sh) * tmp;
+			if (k == a && !x[k])
+			{
+				a = (a + 1) & MASK;
+				i--;
+				rp -= 9;
+			}
+		}
+		if (carry)
+		{
+			if ((z + 1) & MASK != a)
+			{
+				x[z] = carry;
+				z = (z + 1) & MASK;
+			}
+			else
+			{
+				x[(z - 1) & MASK] |= 1;
+			}
+		}
+	}
+
+	// Assemble desired bits into floating point variable
+	double y;
+	int i;
+	for (i = 0; i < B1B_DIG; i++)
+	{
+		if ((a + i) & MASK == z) x[(z = (z + 1) & MASK) - 1] = 0;
+		y = 1000000000.0 * y + x[(a + i) & MASK];
+	}
+
+	y *= sign;
+
+	bool denormal;
+	/* Limit precision for denormal results */
+	uint bits = $bits;
+	if (bits > math::DOUBLE_MANT_DIG + e2 - $emin)
+	{
+		bits = math::DOUBLE_MANT_DIG + e2 - $emin;
+		if (bits < 0) bits = 0;
+		denormal = true;
+	}
+
+	// Calculate bias term to force rounding, move out lower bits
+	double bias;
+	double frac;
+	if (bits < math::DOUBLE_MANT_DIG)
+	{
+		bias = math::copysign(math::scalbn(1, 2 * math::DOUBLE_MANT_DIG - bits - 1), y);
+		frac = y % math::scalbn(1, math::DOUBLE_MANT_DIG - bits);
+		y -= frac;
+		y += bias;
+	}
+
+	// Process tail of decimal input so it can affect rounding
+	if ((a + i) & MASK != z)
+	{
+		uint t = x[(a + i) & MASK];
+		switch
+		{
+			case t < 500000000 && (t || (a + i + 1) & MASK != z):
+				frac += 0.25 * sign;
+			case t > 500000000:
+				frac += 0.75 * sign;
+			case t == 500000000:
+				if ((a + i + 1) & MASK == z)
+				{
+					frac += 0.5 * sign;
+				}
+				else
+				{
+					frac += 0.75 * sign;
+				}
+		}
+		if (math::DOUBLE_MANT_DIG - bits >= 2 && !(frac % 1)) frac++;
+	}
+
+	y += frac;
+	y -= bias;
+
+	if (((e2 + math::DOUBLE_MANT_DIG) & int.max) > emax - 5)
+	{
+		if (math::abs(y) >= 0x1p53)
+		{
+			if (denormal && bits == math::DOUBLE_MANT_DIG + e2 - $emin) denormal = false;
+			y *= 0.5;
+			e2++;
+		}
+		if (e2 + math::DOUBLE_MANT_DIG > emax || (denormal && frac)) return NumberConversion.MALFORMED_FLOAT?;
+	}
+	return math::scalbn(y, e2);
+}
+
+macro double! hexfloat(char[] chars, int $bits, int $emin, int sign)
+{
+	double scale = 1;
+	uint x;
+	long rp;
+	long dc;
+	char c @noinit;
+	bool got_rad;
+	bool got_digit;
+	bool got_tail;
+	usz len = chars.len;
+	usz last_char = len - 1;
+	usz index;
+	double y;
+
+	// Skip past first characters
+	while ((c = chars[index]) == '0')
+	{
+		if (index == last_char) return 0.0;
+		index++;
+	}
+	if (c == '.')
+	{
+		got_rad = true;
+		if (index == last_char)
+		{
+			if (!got_digit) return NumberConversion.MALFORMED_FLOAT?;
+			return sign * 0.0;
+		}
+		if (index != last_char && (c = chars[++index]) == '0')
+		{
+			rp--;
+			got_digit = true;
+			while (last_char != index && (c = chars[++index]) == '0')
+			{
+				rp--;
+			}
+		}
+	}
+
+	while ((c - '0') < 10u || ((c | 32) - 'a') < 6u || c == '.')
+	{
+		if (c == '.')
+		{
+			if (got_rad) return NumberConversion.MALFORMED_FLOAT?;
+			got_rad = true;
+			rp = dc;
+		}
+		else
+		{
+			got_digit = true;
+			int d = {|
+				if (c > '9') return (c | 32) + 10 - 'a';
+				return c - '0';
+			|};
+			switch
+			{
+				case dc < 8:
+					x = x * 16 + d;
+				case dc < math::DOUBLE_MANT_DIG / 4 + 1:
+					y += d * (scale /= 16);
+					got_tail = true;
+				case d && !got_tail:
+					y += 0.5 * scale;
+					got_tail = true;
+			}
+			dc++;
+		}
+		if (index == last_char) break;
+		c = chars[++index];
+	}
+	if (!got_digit) return NumberConversion.MALFORMED_FLOAT?;
+	if (!got_rad) rp = dc;
+	for (; dc < 8; dc++) x *= 16;
+
+	long e2;
+	if ((c | 32) == 'p')
+	{
+		long e2val = String.to_long((String)chars[index + 1..]) ?? (NumberConversion.MALFORMED_FLOAT?)!;
+		e2 = e2val;
+	}
+	e2 += 4 * rp - 32;
+	if (!x) return sign * 0.0;
+	if (e2 > -$emin) return NumberConversion.FLOAT_OUT_OF_RANGE?;
+	if (e2 < $emin - 2 * math::DOUBLE_MANT_DIG) return NumberConversion.FLOAT_OUT_OF_RANGE?;
+
+	while (x < 0x80000000)
+	{
+		if (y >= 0.5)
+		{
+			x += x + 1;
+			y += y - 1;
+		}
+		else
+		{
+			x += x;
+			y += y;
+		}
+		e2--;
+	}
+	int bits = $bits;
+	if ($bits > 32 + e2 - $emin)
+	{
+		bits = (int)(32 + e2 - $emin);
+		if (bits < 0) bits = 0;
+	}
+	double bias;
+	if (bits < math::DOUBLE_MANT_DIG)
+	{
+		bias = math::copysign(math::scalbn(1, 32 + math::DOUBLE_MANT_DIG - bits - 1), (double)sign);
+	}
+
+	if (bits < 32 && y && !(x & 1))
+	{
+		x++;
+		y = 0;
+	}
+	y = bias + sign * (double)x + sign * y;
+	y -= bias;
+	if (!y) return NumberConversion.FLOAT_OUT_OF_RANGE?;
+
+	return math::scalbn(y, (int)e2);
+}
+
+macro String.to_real(chars, $Type) @private
+{
+	int sign = 1;
+	$switch ($Type)
+	$case float:
+		const int BITS = math::FLOAT_MANT_DIG;
+		const int EMIN = math::FLOAT_MIN_EXP - BITS;
+	$case double:
+		const int BITS = math::DOUBLE_MANT_DIG;
+		const int EMIN = math::DOUBLE_MIN_EXP - BITS;
+	$case float128:
+		$error "Not yet supported";
+	$default:
+		$error "Unexpected type";
+	$endswitch
+
+	while (chars.len && chars[0] == ' ') chars = chars[1..];
+	if (!chars.len) return NumberConversion.MALFORMED_FLOAT?;
+	switch (chars[0])
+	{
+		case '-':
+			sign = -1;
+			nextcase;
+		case '+':
+			chars = chars[1..];
+	}
+	if (chars == "infinity" || chars == "INFINITY") return sign * $Type.inf;
+	if (chars == "NAN" || chars == "nan") return $Type.nan;
+
+	if (chars.len > 2 && chars[0] == '0' && (chars[1] | 32) == 'x')
+	{
+		return ($Type)hexfloat((char[])chars[2..], BITS, EMIN, sign);
+	}
+	return ($Type)decfloat((char[])chars, BITS, EMIN, sign);
+}
+

lib/std/core/types.c3

@@ -0,0 +1,337 @@
+
+module std::core::types;
+import libc;
+
+
+fault ConversionResult
+{
+	VALUE_OUT_OF_RANGE,
+	VALUE_OUT_OF_UNSIGNED_RANGE,
+}
+/**
+ * @require $Type.kindof.is_int() || $Type.kindof == TypeKind.ENUM "Argument was not an integer"
+ **/
+macro any_to_int(any v, $Type)
+{
+	typeid any_type = v.type;
+	TypeKind kind = any_type.kindof;
+	if (kind == TypeKind.ENUM)
+	{
+		any_type = any_type.inner;
+		kind = any_type.kindof;
+	}
+	bool is_mixed_signed = $Type.kindof != any_type.kindof;
+	$Type max = $Type.max;
+	$Type min = $Type.min;
+	switch (any_type)
+	{
+		case ichar:
+			ichar c = *(char*)v.ptr;
+			if (is_mixed_signed && c < 0) return ConversionResult.VALUE_OUT_OF_UNSIGNED_RANGE?;
+			return ($Type)c;
+		case short:
+			short s = *(short*)v.ptr;
+			if (is_mixed_signed && s < 0) return ConversionResult.VALUE_OUT_OF_UNSIGNED_RANGE?;
+			if (s > max || s < min) return ConversionResult.VALUE_OUT_OF_RANGE?;
+			return ($Type)s;
+		case int:
+			int i = *(int*)v.ptr;
+			if (is_mixed_signed && i < 0) return ConversionResult.VALUE_OUT_OF_UNSIGNED_RANGE?;
+			if (i > max || i < min) return ConversionResult.VALUE_OUT_OF_RANGE?;
+			return ($Type)i;
+		case long:
+			long l = *(long*)v.ptr;
+			if (is_mixed_signed && l < 0) return ConversionResult.VALUE_OUT_OF_UNSIGNED_RANGE?;
+			if (l > max || l < min) return ConversionResult.VALUE_OUT_OF_RANGE?;
+			return ($Type)l;
+		case int128:
+			int128 i = *(int128*)v.ptr;
+			if (is_mixed_signed && i < 0) return ConversionResult.VALUE_OUT_OF_UNSIGNED_RANGE?;
+			if (i > max || i < min) return ConversionResult.VALUE_OUT_OF_RANGE?;
+			return ($Type)i;
+		case char:
+			char c = *(char*)v.ptr;
+			if (c > max) return ConversionResult.VALUE_OUT_OF_RANGE?;
+			return ($Type)c;
+		case ushort:
+			ushort s = *(ushort*)v.ptr;
+			if (s > max || s < min) return ConversionResult.VALUE_OUT_OF_RANGE?;
+			return ($Type)s;
+		case uint:
+			uint i = *(uint*)v.ptr;
+			if (i > max || i < min) return ConversionResult.VALUE_OUT_OF_RANGE?;
+			return ($Type)i;
+		case ulong:
+			ulong l = *(ulong*)v.ptr;
+			if (l > max || l < min) return ConversionResult.VALUE_OUT_OF_RANGE?;
+			return ($Type)l;
+		case uint128:
+			uint128 i = *(uint128*)v.ptr;
+			if (i > max || i < min) return ConversionResult.VALUE_OUT_OF_RANGE?;
+			return ($Type)i;
+		default:
+			unreachable();
+	}
+}
+
+fn bool typeid.is_subtype_of(self, typeid other)
+{
+	while (self != void.typeid)
+	{
+		if (self == other) return true;
+		self = self.parentof;
+	}
+	return false;
+}
+
+macro bool is_subtype_of($Type, $OtherType)
+{
+	var $typeid = $Type.typeid;
+	$switch ($Type)
+		$case $OtherType: return true;
+		$default: return false;
+	$endswitch
+}
+macro bool is_numerical($Type)
+{
+	var $kind = $Type.kindof;
+	$if $kind == TypeKind.DISTINCT:
+		return is_numerical($typefrom($Type.inner));
+	$else
+		return $kind == TypeKind.SIGNED_INT || $kind == TypeKind.UNSIGNED_INT || $kind == TypeKind.FLOAT
+			|| $kind == TypeKind.VECTOR;
+	$endif
+}
+
+fn bool TypeKind.is_int(kind) @inline
+{
+	return kind == TypeKind.SIGNED_INT || kind == TypeKind.UNSIGNED_INT;
+}
+
+macro bool is_slice_convertable($Type)
+{
+	$switch ($Type.kindof)
+		$case SLICE:
+			return true;
+		$case POINTER:
+			return $Type.inner.kindof == TypeKind.ARRAY;
+		$default:
+			return false;
+	$endswitch
+}
+
+macro bool is_bool($Type) @const => $Type.kindof == TypeKind.BOOL;
+macro bool is_int($Type) @const  => $Type.kindof == TypeKind.SIGNED_INT || $Type.kindof == TypeKind.UNSIGNED_INT;
+
+macro bool is_indexable($Type) @const
+{
+	return $defined($Type{}[0]);
+}
+
+macro bool is_ref_indexable($Type) @const
+{
+	return $defined(&$Type{}[0]);
+}
+
+macro bool is_intlike($Type) @const
+{
+	$switch ($Type.kindof)
+		$case SIGNED_INT:
+		$case UNSIGNED_INT:
+			return true;
+		$case VECTOR:
+			return $Type.inner.kindof == TypeKind.SIGNED_INT || $Type.inner.kindof == TypeKind.UNSIGNED_INT;
+		$default:
+			return false;
+	$endswitch
+}
+
+macro bool is_underlying_int($Type) @const
+{
+	$switch ($Type.kindof)
+		$case SIGNED_INT:
+		$case UNSIGNED_INT:
+			return true;
+		$case DISTINCT:
+			return is_underlying_int($typefrom($Type.inner));
+		$default:
+			return false;
+	$endswitch
+}
+
+macro bool is_float($Type) @const  => $Type.kindof == TypeKind.FLOAT;
+
+macro bool is_floatlike($Type) @const
+{
+	$switch ($Type.kindof)
+		$case FLOAT:
+			return true;
+		$case VECTOR:
+			return $Type.inner.kindof == TypeKind.FLOAT;
+		$default:
+			return false;
+	$endswitch
+}
+
+macro bool is_vector($Type) @const
+{
+	return $Type.kindof == TypeKind.VECTOR;
+}
+
+macro TypeKind inner_kind($Type) @const
+{
+	$if $Type.kindof == TypeKind.DISTINCT:
+		return inner_kind($typefrom($Type.inner));
+	$else
+		return $Type.kindof;
+	$endif
+}
+
+macro bool is_same($TypeA, $TypeB) @const
+{
+	return $TypeA.typeid == $TypeB.typeid;
+}
+
+macro bool @has_same(#a, #b, ...) @const
+{
+	var $type_a = @typeid(#a);
+	$if $type_a != @typeid(#b):
+		return false;
+	$endif
+	$for (var $i = 0; $i < $vacount; $i++)
+		$if @typeid($vaexpr[$i]) != $type_a:
+			return false;
+		$endif
+	$endfor
+	return true;
+}
+
+macro bool may_load_atomic($Type) @const
+{
+	$switch ($Type.kindof)
+	$case SIGNED_INT:
+	$case UNSIGNED_INT:
+	$case POINTER:
+	$case FLOAT:
+		return true;
+	$case DISTINCT:
+		return may_load_atomic($Type.inner);
+	$default:
+		return false;
+	$endswitch
+}
+
+macro lower_to_atomic_compatible_type($Type) @const
+{
+	$switch ($Type.kindof)
+    $case SIGNED_INT:
+	  $case UNSIGNED_INT:
+      return $Type.typeid;
+		$case DISTINCT:
+			return lower_to_atomic_compatible_type($Type.inner);
+		$case FLOAT:
+			$switch ($Type)
+				$case float16: 
+					return ushort.typeid;
+				$case float: 
+					return uint.typeid;
+				$case double:
+					return ulong.typeid;
+				$case float128:
+					return uint128.typeid;
+				$default: 
+					return void.typeid;
+			$endswitch
+		$default:
+			return void.typeid;
+  $endswitch
+}
+
+macro bool is_promotable_to_floatlike($Type) @const  => types::is_floatlike($Type) || types::is_int($Type);
+macro bool is_promotable_to_float($Type) @const  => types::is_float($Type) || types::is_int($Type);
+
+macro bool is_same_vector_type($Type1, $Type2) @const
+{
+	$if $Type1.kindof != TypeKind.VECTOR:
+		return $Type2.kindof != TypeKind.VECTOR;
+	$else
+		return $Type1.inner == $Type2.inner && $Type1.len == $Type2.len;
+	$endif
+}
+
+macro bool is_equatable_type($Type) @const
+{
+	$if $defined($Type.less) || $defined($Type.compare_to) || $defined($Type.equals):
+		return true;
+	$else
+		return $Type.is_eq;
+	$endif
+}
+
+/**
+ * Checks if a type implements the copy protocol.
+ **/
+macro bool implements_copy($Type) @const
+{
+	return $defined($Type.copy) && $defined($Type.free);
+}
+
+macro bool is_equatable_value(value) @deprecated
+{
+	return is_equatable_type($typeof(value));
+}
+
+macro bool @equatable_value(#value) @const
+{
+	return is_equatable_type($typeof(#value));
+}
+
+macro bool @comparable_value(#value) @const
+{
+	$if $defined(#value.less) || $defined(#value.compare_to):
+		return true;
+	$else
+		return $typeof(#value).is_ordered;
+	$endif
+}
+
+macro bool is_comparable_value(value) @deprecated
+{
+	$if $defined(value.less) || $defined(value.compare_to):
+		return true;
+	$else
+		return $typeof(value).is_ordered;
+	$endif
+}
+
+enum TypeKind : char
+{
+	VOID,
+	BOOL,
+	SIGNED_INT,
+	UNSIGNED_INT,
+	FLOAT,
+	TYPEID,
+	ANYFAULT,
+	ANY,
+	ENUM,
+	FAULT,
+	STRUCT,
+	UNION,
+	BITSTRUCT,
+	FUNC,
+	OPTIONAL,
+	ARRAY,
+	SLICE,
+	VECTOR,
+	DISTINCT,
+	POINTER,
+	INTERFACE,
+}
+
+struct TypeEnum
+{
+	TypeKind type;
+	usz elements;
+}
+

lib/std/core/values.c3

@@ -0,0 +1,46 @@
+module std::core::values;
+
+macro typeid @typeid(#value) @const @builtin => $typeof(#value).typeid;
+macro TypeKind @typekind(#value) @const @builtin => $typeof(#value).kindof;
+macro bool @typeis(#value, $Type) @const @builtin => $typeof(#value).typeid == $Type.typeid;
+/**
+ * Return true if two values have the same type before any conversions.
+ **/
+macro bool @is_same_type(#value1, #value2) @const => $typeof(#value1).typeid == $typeof(#value2).typeid;
+macro bool @is_bool(#value) @const => types::is_bool($typeof(#value));
+macro bool @is_int(#value) @const => types::is_int($typeof(#value));
+macro bool @is_floatlike(#value) @const => types::is_floatlike($typeof(#value));
+macro bool @is_float(#value) @const => types::is_float($typeof(#value));
+macro bool @is_promotable_to_floatlike(#value) @const => types::is_promotable_to_floatlike($typeof(#value));
+macro bool @is_promotable_to_float(#value) @const => types::is_promotable_to_float($typeof(#value));
+macro bool @is_vector(#value) @const => types::is_vector($typeof(#value));
+macro bool @is_same_vector_type(#value1, #value2) @const => types::is_same_vector_type($typeof(#value1), $typeof(#value2));
+macro bool @assign_to(#value1, #value2) @const => $assignable(#value1, $typeof(#value2));
+
+macro promote_int(x)
+{
+	$if @is_int(x):
+		return (double)x;
+	$else
+		return x;
+	$endif
+}
+
+macro promote_int_same(x, y)
+{
+	$if @is_int(x):
+		$switch
+		$case @is_vector(y) &&& $typeof(y).inner == float.typeid:
+			return (float)x;
+		$case $typeof(y).typeid == float.typeid:
+			return (float)x;
+		$default:
+			return (double)x;
+		$endswitch
+	$else
+		return x;
+	$endif
+}
+
+macro TypeKind @inner_kind(#value) @const => types::inner_kind($typeof(#value));
+

lib/std/crypto/crypto.c3

@@ -0,0 +1,2 @@
+module std::crypto;
+

lib/std/crypto/rc4.c3

@@ -0,0 +1,75 @@
+module std::crypto::rc4;
+// Copyright (c) 2021 Christoffer Lerno. All rights reserved.
+// Use of this source code is governed by the MIT license
+// a copy of which can be found in the LICENSE_STDLIB file.
+
+struct Rc4
+{
+	uint i, j;
+	char[256] state;
+}
+
+/**
+ * Initialize the RC4 state.
+ *
+ * @param [in] key "The key to use"
+ * @require key.len > 0 "The key must be at least 1 byte long"
+ **/
+fn void Rc4.init(&self, char[] key)
+{
+	// Init the state matrix
+	foreach (char i, &c : self.state) *c = i;
+	for (int i = 0, int j = 0; i < 256; i++)
+	{
+		j = (j + self.state[i] + key[i % key.len]) & 0xFF;
+		@swap(self.state[i], self.state[j]);
+	}
+	self.i = 0;
+	self.j = 0;
+}
+
+/**
+ * Run a single pass of en/decryption using a particular key.
+ * @param [in] key
+ * @param [inout] data
+ **/
+fn void crypt(char[] key, char[] data)
+{
+	Rc4 rc4;
+	rc4.init(key);
+	rc4.crypt(data, data);
+}
+
+/**
+ * Encrypt or decrypt a sequence of bytes.
+ *
+ * @param [in] in "The input"
+ * @param [out] out "The output"
+ * @require in.len <= out.len "Output would overflow"
+ **/
+fn void Rc4.crypt(&self, char[] in, char[] out)
+{
+	uint i = self.i;
+	uint j = self.j;
+	char* state = &self.state;
+	isz len = in.len;
+	foreach (idx, c : in)
+	{
+		i = (i + 1) & 0xFF;
+		j = (j + state[i]) & 0xFF;
+		@swap(state[i], state[j]);
+		out[idx] = in[idx] ^ state[(state[i] + state[j]) & 0xFF];
+	}
+	self.i = i;
+	self.j = j;
+}
+
+/**
+ * Clear the rc4 state.
+ *
+ * @param [&out] self "The RC4 State"
+ **/
+fn void Rc4.destroy(&self)
+{
+	*self = {};
+}

lib/std/encoding/base64.c3

@@ -0,0 +1,288 @@
+module std::encoding::base64;
+import std::core::bitorder;
+
+// The implementation is based on https://www.rfc-editor.org/rfc/rfc4648
+// Specifically this section:
+// https://www.rfc-editor.org/rfc/rfc4648#section-4
+
+const STD_ALPHABET = "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/";
+const URL_ALPHABET = "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789-_";
+
+const MASK @private = 0b111111;
+
+struct Base64Encoder
+{
+	int padding;
+	String alphabet;
+}
+
+fault Base64Error
+{
+	DUPLICATE_IN_ALPHABET,
+	PADDING_IN_ALPHABET,
+	DESTINATION_TOO_SMALL,
+	INVALID_PADDING,
+	INVALID_CHARACTER,
+}
+
+/**
+ * @param alphabet "The alphabet used for encoding."
+ * @param padding "Set to a negative value to disable padding."
+ * @require alphabet.len == 64
+ * @require padding < 256
+ * @return! Base64Error.DUPLICATE_IN_ALPHABET, Base64Error.PADDING_IN_ALPHABET
+ **/
+fn void! Base64Encoder.init(&self, String alphabet, int padding = '=')
+{
+	check_alphabet(alphabet, padding)!;
+	*self = { .padding = padding, .alphabet = alphabet };
+}
+
+/**
+ * Calculate the size of the encoded data.
+ * @param n "Size of the input to be encoded."
+ * @return "The size of the input once encoded."
+ **/
+fn usz Base64Encoder.encode_len(&self, usz n)
+{
+	if (self.padding >= 0) return (n + 2) / 3 * 4;
+	usz trailing = n % 3;
+	return n / 3 * 4 + (trailing * 4 + 2) / 3;
+}
+
+/**
+ * Encode the content of src into dst, which must be properly sized.
+ * @param src "The input to be encoded."
+ * @param dst "The encoded input."
+ * @return "The encoded size."
+ * @return! Base64Error.DESTINATION_TOO_SMALL
+ **/
+fn usz! Base64Encoder.encode(&self, char[] src, char[] dst)
+{
+	if (src.len == 0) return 0;
+	usz dn = self.encode_len(src.len);
+	if (dst.len < dn) return Base64Error.DESTINATION_TOO_SMALL?;
+	usz trailing = src.len % 3;
+	char[] src3 = src[:^trailing];
+
+	while (src3.len > 0)
+	{
+		uint group = (uint)src3[0] << 16 | (uint)src3[1] << 8 | (uint)src3[2];
+		dst[0] = self.alphabet[group >> 18 & MASK];
+		dst[1] = self.alphabet[group >> 12 & MASK];
+		dst[2] = self.alphabet[group >> 6 & MASK];
+		dst[3] = self.alphabet[group & MASK];
+		dst = dst[4..];
+		src3 = src3[3..];
+	}
+
+	// Encode the remaining bytes according to:
+	// https://www.rfc-editor.org/rfc/rfc4648#section-3.5
+	switch (trailing)
+	{
+		case 1:
+			uint group = (uint)src[^1] << 16;
+			dst[0] = self.alphabet[group >> 18 & MASK];
+			dst[1] = self.alphabet[group >> 12 & MASK];
+			if (self.padding >= 0)
+			{
+				char pad = (char)self.padding;
+				dst[2] = pad;
+				dst[3] = pad;
+			}
+		case 2:
+			uint group = (uint)src[^2] << 16 | (uint)src[^1] << 8;
+			dst[0] = self.alphabet[group >> 18 & MASK];
+			dst[1] = self.alphabet[group >> 12 & MASK];
+			dst[2] = self.alphabet[group >> 6 & MASK];
+			if (self.padding >= 0)
+			{
+				char pad = (char)self.padding;
+				dst[3] = pad;
+			}
+	}
+	return dn;
+}
+
+struct Base64Decoder
+{
+	int padding;
+	String alphabet;
+	char[256] reverse;
+	char invalid;
+}
+
+/**
+ * @param alphabet "The alphabet used for encoding."
+ * @param padding "Set to a negative value to disable padding."
+ * @require alphabet.len == 64
+ * @require padding < 256
+ * @return! Base64Error.DUPLICATE_IN_ALPHABET, Base64Error.PADDING_IN_ALPHABET
+ **/
+fn void! Base64Decoder.init(&self, String alphabet, int padding = '=')
+{
+	check_alphabet(alphabet, padding)!;
+	*self = { .padding = padding, .alphabet = alphabet };
+
+	bool[256] checked;
+	foreach (i, c : alphabet)
+	{
+		checked[c] = true;
+		self.reverse[c] = (char)i;
+	}
+	if (padding < 0)
+	{
+		self.invalid = 255;
+		return;
+	}
+	// Find a character for invalid neither in the alphabet nor equal to the padding.
+	char pad = (char)padding;
+	foreach (i, ok : checked)
+	{
+		if (!ok && (char)i != pad)
+		{
+			self.invalid = (char)i;
+			break;
+		}
+	}
+}
+
+/**
+ * Calculate the size of the decoded data.
+ * @param n "Size of the input to be decoded."
+ * @return "The size of the input once decoded."
+ * @return! Base64Error.INVALID_PADDING
+ **/
+fn usz! Base64Decoder.decode_len(&self, usz n)
+{
+	usz dn = n / 4 * 3;
+	usz trailing = n % 4;
+	if (self.padding >= 0)
+	{
+		if (trailing != 0) return Base64Error.INVALID_PADDING?;
+		// source size is multiple of 4
+	}
+	else
+	{
+		if (trailing == 1) return Base64Error.INVALID_PADDING?;
+		dn += trailing * 3 / 4;
+	}
+	return dn;
+}
+
+/**
+ * Decode the content of src into dst, which must be properly sized.
+ * @param src "The input to be decoded."
+ * @param dst "The decoded input."
+ * @return "The decoded size."
+ * @return! Base64Error.DESTINATION_TOO_SMALL, Base64Error.INVALID_PADDING, Base64Error.INVALID_CHARACTER
+ **/
+fn usz! Base64Decoder.decode(&self, char[] src, char[] dst)
+{
+	if (src.len == 0) return 0;
+	usz dn = self.decode_len(src.len)!;
+	if (dst.len < dn) return Base64Error.DESTINATION_TOO_SMALL?;
+
+	usz trailing = src.len % 4;
+	char[] src4 = src;
+	switch
+	{
+		case self.padding < 0:
+			src4 = src[:^trailing];
+		default:
+			// If there is padding, keep the last 4 bytes for later.
+			// NB. src.len >= 4 as decode_len passed
+			trailing = 4;
+			char pad = (char)self.padding;
+			if (src[^1] == pad) src4 = src[:^4];
+	}
+	while (src4.len > 0)
+	{
+		char c0 = self.reverse[src4[0]];
+		char c1 = self.reverse[src4[1]];
+		char c2 = self.reverse[src4[2]];
+		char c3 = self.reverse[src4[3]];
+		switch (self.invalid)
+		{
+			case c0:
+			case c1:
+			case c2:
+			case c3:
+				return Base64Error.INVALID_CHARACTER?;
+		}
+		uint group = (uint)c0 << 18 | (uint)c1 << 12 | (uint)c2 << 6 | (uint)c3;
+		dst[0] = (char)(group >> 16);
+		dst[1] = (char)(group >> 8);
+		dst[2] = (char)group;
+		dst = dst[3..];
+		src4 = src4[4..];
+	}
+
+	if (trailing == 0) return dn;
+
+	src = src[^trailing..];
+	char c0 = self.reverse[src[0]];
+	char c1 = self.reverse[src[1]];
+	if (c0 == self.invalid || c1 == self.invalid) return Base64Error.INVALID_PADDING?;
+	if (self.padding < 0)
+	{
+		switch (src.len)
+		{
+			case 2:
+				uint group = (uint)c0 << 18 | (uint)c1 << 12;
+				dst[0] = (char)(group >> 16);
+			case 3:
+				char c2 = self.reverse[src[2]];
+				if (c2 == self.invalid) return Base64Error.INVALID_CHARACTER?;
+				uint group = (uint)c0 << 18 | (uint)c1 << 12 | (uint)c2 << 6;
+				dst[0] = (char)(group >> 16);
+				dst[1] = (char)(group >> 8);
+		}
+	}
+	else
+	{
+		// Valid paddings are:
+		// 2: xx==
+		// 1: xxx=
+		char pad = (char)self.padding;
+		switch (pad)
+		{
+			case src[2]:
+				if (src[3] != pad) return Base64Error.INVALID_PADDING?;
+				uint group = (uint)c0 << 18 | (uint)c1 << 12;
+				dst[0] = (char)(group >> 16);
+				dn -= 2;
+			case src[3]:
+				char c2 = self.reverse[src[2]];
+				if (c2 == self.invalid) return Base64Error.INVALID_CHARACTER?;
+				uint group = (uint)c0 << 18 | (uint)c1 << 12 | (uint)c2 << 6;
+				dst[0] = (char)(group >> 16);
+				dst[1] = (char)(group >> 8);
+				dn -= 1;
+		}
+	}
+	return dn;
+}
+
+// Make sure that all bytes in the alphabet are unique and
+// the padding is not present in the alphabet.
+fn void! check_alphabet(String alphabet, int padding) @local
+{
+	bool[256] checked;
+	if (padding < 0)
+	{
+		foreach (c : alphabet)
+		{
+			if (checked[c]) return Base64Error.DUPLICATE_IN_ALPHABET?;
+			checked[c] = true;
+		}
+		return;
+	}
+	char pad = (char)padding;
+	foreach (c : alphabet)
+	{
+		if (c == pad) return Base64Error.PADDING_IN_ALPHABET?;
+		if (checked[c]) return Base64Error.DUPLICATE_IN_ALPHABET?;
+		checked[c] = true;
+	}
+}

lib/std/encoding/csv.c3

@@ -0,0 +1,64 @@
+module std::encoding::csv;
+import std::io;
+
+struct CsvReader
+{
+	InStream stream;
+	String separator;
+}
+
+fn void CsvReader.init(&self, InStream stream, String separator = ",")
+{
+	self.stream = stream;
+	self.separator = separator;
+}
+
+fn String[]! CsvReader.read_new_row(self, Allocator allocator = allocator::heap())
+{
+	return self.read_new_row_with_allocator(allocator::temp()) @inline;
+}
+
+fn String[]! CsvReader.read_new_row_with_allocator(self, Allocator allocator = allocator::heap())
+{
+	@pool(allocator)
+	{
+		return io::treadline(self.stream).split(self.separator, .allocator = allocator);
+	};
+}
+
+fn String[]! CsvReader.read_temp_row(self)
+{
+	return self.read_new_row_with_allocator(allocator::temp()) @inline;
+}
+
+fn void! CsvReader.skip_row(self) @maydiscard
+{
+	@pool()
+	{
+		(void)io::treadline(self.stream);
+	};
+}
+
+macro CsvReader.@each_row(self, int rows = int.max; @body(String[] row))
+{
+	InputStream* stream = self.stream;
+	String sep = self.separator;
+	while (rows--)
+	{
+		@stack_mem(512; Allocator mem)
+		{
+			String[] parts;
+			@pool()
+			{
+				String! s = stream.treadline();
+				if (catch err = s)
+				{
+					if (err == IoError.EOF) return;
+					return err?;
+				}
+				parts = s.split(sep, .allocator = mem);
+			};
+			@body(parts);
+		};
+	}
+}

lib/std/encoding/json.c3

@@ -0,0 +1,387 @@
+// Copyright (c) 2023 Christoffer Lerno. All rights reserved.
+// Use of this source code is governed by the MIT license
+// a copy of which can be found in the LICENSE_STDLIB file.
+module std::encoding::json;
+import std::io;
+import std::ascii;
+import std::collections::object;
+
+fault JsonParsingError
+{
+	EOF,
+	UNEXPECTED_CHARACTER,
+	INVALID_ESCAPE_SEQUENCE,
+	DUPLICATE_MEMBERS,
+	INVALID_NUMBER,
+}
+
+fn Object*! parse_string(String s, Allocator allocator = allocator::heap())
+{
+	return parse(ByteReader{}.init(s), allocator);
+}
+
+fn Object*! temp_parse_string(String s)
+{
+	return parse(ByteReader{}.init(s), allocator::temp());
+}
+
+fn Object*! parse(InStream s, Allocator allocator = allocator::heap())
+{
+	JsonContext context = { .last_string = dstring::new_with_capacity(64, allocator), .stream = s, .allocator = allocator };
+	defer context.last_string.free();
+	@pool(allocator)
+	{
+		return parse_any(&context);
+	};
+}
+
+fn Object*! temp_parse(InStream s)
+{
+	return parse(s, allocator::temp());
+}
+
+// -- Implementation follows --
+
+enum JsonTokenType @local
+{
+	NO_TOKEN,
+	LBRACE,
+	LBRACKET,
+	COMMA,
+	COLON,
+	RBRACE,
+	RBRACKET,
+	STRING,
+	NUMBER,
+	TRUE,
+	FALSE,
+	NULL,
+	EOF,
+}
+
+struct JsonContext @local
+{
+	uint line;
+	InStream stream;
+	Allocator allocator;
+	JsonTokenType token;
+	DString last_string;
+	double last_number;
+	char current;
+	bitstruct : char {
+		bool skip_comments;
+		bool reached_end;
+		bool pushed_back;
+	}
+}
+
+
+fn Object*! parse_from_token(JsonContext* context, JsonTokenType token) @local
+{
+	switch (token)
+	{
+		case NO_TOKEN: unreachable();
+		case LBRACE: return parse_map(context);
+		case LBRACKET: return parse_array(context);
+		case COMMA:
+		case RBRACE:
+		case RBRACKET:
+		case COLON: return JsonParsingError.UNEXPECTED_CHARACTER?;
+		case STRING: return object::new_string(context.last_string.str_view(), context.allocator);
+		case NUMBER: return object::new_float(context.last_number, context.allocator);
+		case TRUE: return object::new_bool(true);
+		case FALSE: return object::new_bool(false);
+		case NULL: return object::new_null();
+		case EOF: return JsonParsingError.EOF?;
+	}
+}
+fn Object*! parse_any(JsonContext* context) @local
+{
+	return parse_from_token(context, advance(context));
+}
+
+fn JsonTokenType! lex_number(JsonContext *context, char c) @local
+{
+	@pool()
+	{
+		DString t = dstring::temp_with_capacity(32);
+		bool negate = c == '-';
+		if (negate)
+		{
+			t.append(c);
+			c = read_next(context)!;
+		}
+		while (c.is_digit())
+		{
+			t.append(c);
+			c = read_next(context)!;
+		}
+		if (c == '.')
+		{
+			t.append(c);
+			while (c = read_next(context)!, c.is_digit())
+			{
+				t.append(c);
+			}
+		}
+		if ((c | 32) == 'e')
+		{
+			t.append(c);
+			c = read_next(context)!;
+			switch (c)
+			{
+				case '-':
+				case '+':
+					t.append(c);
+					c = read_next(context)!;
+			}
+			if (!c.is_digit()) return JsonParsingError.INVALID_NUMBER?;
+			while (c.is_digit())
+			{
+				t.append(c);
+				c = read_next(context)!;
+			}
+		}
+		pushback(context, c);
+		double! d = t.str_view().to_double() ?? JsonParsingError.INVALID_NUMBER?;
+		context.last_number = d!;
+		return NUMBER;
+	};
+}
+
+fn Object*! parse_map(JsonContext* context) @local
+{
+	Object* map = object::new_obj(context.allocator);
+	JsonTokenType token = advance(context)!;
+	defer catch map.free();
+
+	DString temp_key = dstring::new_with_capacity(32, context.allocator);
+	defer temp_key.free();
+	while (token != JsonTokenType.RBRACE)
+	{
+		if (token != JsonTokenType.STRING) return JsonParsingError.UNEXPECTED_CHARACTER?;
+		DString string = context.last_string;
+		if (map.has_key(string.str_view())) return JsonParsingError.DUPLICATE_MEMBERS?;
+		// Copy the key to our temp holder. We do this to work around the issue
+		// if the temp allocator should be used as the default allocator.
+		temp_key.clear();
+		temp_key.append(string);
+		parse_expected(context, COLON)!;
+		Object* element = parse_any(context)!;
+		map.set(temp_key.str_view(), element);
+		token = advance(context)!;
+		if (token == JsonTokenType.COMMA)
+		{
+			token = advance(context)!;
+			continue;
+		}
+		if (token != JsonTokenType.RBRACE) return JsonParsingError.UNEXPECTED_CHARACTER?;
+	}
+	return map;
+}
+
+fn Object*! parse_array(JsonContext* context) @local
+{
+	Object* list = object::new_obj(context.allocator);
+	defer catch list.free();
+	JsonTokenType token = advance(context)!;
+	while (token != JsonTokenType.RBRACKET)
+	{
+		Object* element = parse_from_token(context, token)!;
+		list.push(element);
+		token = advance(context)!;
+		if (token == JsonTokenType.COMMA)
+		{
+			token = advance(context)!;
+			continue;
+		}
+		if (token != JsonTokenType.RBRACKET) return JsonParsingError.UNEXPECTED_CHARACTER?;
+	}
+	return list;
+}
+
+fn void pushback(JsonContext* context, char c) @local
+{
+	if (!context.reached_end)
+	{
+		assert(!context.pushed_back);
+		context.pushed_back = true;
+		context.current = c;
+	}
+}
+
+fn char! read_next(JsonContext* context) @local
+{
+	if (context.reached_end) return '\0';
+	if (context.pushed_back)
+	{
+		context.pushed_back = false;
+		return context.current;
+	}
+	char! c = context.stream.read_byte();
+	if (catch err = c)
+	{
+		case IoError.EOF:
+			context.reached_end = true;
+			return '\0';
+		default:
+			return err?;
+	}
+	if (c == 0)
+	{
+		context.reached_end = true;
+	}
+	return c;
+}
+
+fn JsonTokenType! advance(JsonContext* context) @local
+{
+	char c;
+	// Skip whitespace
+	while WS: (c = read_next(context)!)
+	{
+		switch (c)
+		{
+			case '\n':
+				context.line++;
+				nextcase;
+			case ' ':
+			case '\t':
+			case '\r':
+			case '\v':
+				continue;
+			case '/':
+				if (!context.skip_comments) break;
+				c = read_next(context)!;
+				if (c != '*')
+				{
+					pushback(context, c);
+					break WS;
+				}
+				while COMMENT: (true)
+				{
+					// Skip to */
+					while (c = read_next(context)!)
+					{
+						if (c == '\n') context.line++;
+						if (c != '*') continue;
+						// Skip through all the '*'
+						while (c = read_next(context)!)
+						{
+							if (c == '\n') context.line++;
+							if (c != '*') break;
+						}
+						if (c == '/') break COMMENT;
+					}
+				}
+				continue;
+			default:
+				break WS;
+		}
+	}
+	switch (c)
+	{
+		case '\0':
+			return IoError.EOF?;
+		case '{':
+			return LBRACE;
+		case '}':
+			return RBRACE;
+		case '[':
+			return LBRACKET;
+		case ']':
+			return RBRACKET;
+		case ':':
+			return COLON;
+		case ',':
+			return COMMA;
+		case '"':
+			return lex_string(context);
+		case '-':
+		case '0'..'9':
+			return lex_number(context, c);
+		case 't':
+			match(context, "rue")!;
+			return TRUE;
+		case 'f':
+			match(context, "alse")!;
+			return FALSE;
+		case 'n':
+			match(context, "ull")!;
+			return NULL;
+		default:
+			return JsonParsingError.UNEXPECTED_CHARACTER?;
+	}
+}
+
+fn void! match(JsonContext* context, String str) @local
+{
+	foreach (c : str)
+	{
+		char l = read_next(context)!;
+		if (l != c) return JsonParsingError.UNEXPECTED_CHARACTER?;
+	}
+}
+
+fn void! parse_expected(JsonContext* context, JsonTokenType token) @local
+{
+	if (advance(context)! != token) return JsonParsingError.UNEXPECTED_CHARACTER?;
+}
+
+fn JsonTokenType! lex_string(JsonContext* context)
+{
+	context.last_string.clear();
+	while LOOP: (true)
+	{
+		char c = read_next(context)!;
+		switch (c)
+		{
+			case '\0':
+				return JsonParsingError.EOF?;
+			case 1..31:
+				return JsonParsingError.UNEXPECTED_CHARACTER?;
+			case '"':
+				break LOOP;
+			case '\\':
+				break;
+			default:
+				context.last_string.append(c);
+				continue;
+		}
+		c = read_next(context)!;
+		switch (c)
+		{
+			case '\0':
+				return JsonParsingError.EOF?;
+			case 1..31:
+				return JsonParsingError.UNEXPECTED_CHARACTER?;
+			case '"':
+			case '\\':
+			case '/':
+				break;
+			case 'b':
+				c = '\b';
+			case 'f':
+				c = '\f';
+			case 'n':
+				c = '\n';
+			case 'r':
+				c = '\r';
+			case 't':
+				c = '\t';
+			case 'u':
+				uint val;
+				for (int i = 0; i < 4; i++)
+				{
+					c = read_next(context)!;
+					if (!c.is_xdigit()) return JsonParsingError.INVALID_ESCAPE_SEQUENCE?;
+					val = val << 4 + (c > '9' ? (c | 32) - 'a' + 10 : c - '0');
+				}
+				context.last_string.append_char32(val);
+				continue;
+			default:
+				 return JsonParsingError.INVALID_ESCAPE_SEQUENCE?;
+		}
+	}
+	return STRING;
+}

lib/std/hash/adler32.c3

@@ -0,0 +1,53 @@
+// Copyright (c) 2021 Christoffer Lerno. All rights reserved.
+// Use of this source code is governed by the MIT license
+// a copy of which can be found in the LICENSE_STDLIB file.
+
+module std::hash::adler32;
+
+const uint ADLER_CONST @private = 65521;
+
+struct Adler32
+{
+	uint a;
+	uint b;
+}
+
+fn void Adler32.init(&self)
+{
+	*self = { 1, 0 };
+}
+
+fn void Adler32.updatec(&self, char c)
+{
+	self.a = (self.a + c) % ADLER_CONST;
+	self.b = (self.b + self.a) % ADLER_CONST;
+}
+
+fn void Adler32.update(&self, char[] data)
+{
+	uint a = self.a;
+	uint b = self.b;
+	foreach (char x : data)
+	{
+		a = (a + x) % ADLER_CONST;
+		b = (b + a) % ADLER_CONST;
+	}
+	*self = { a, b };
+}
+
+fn uint Adler32.final(&self)
+{
+	return (self.b << 16) | self.a;
+}
+
+fn uint encode(char[] data)
+{
+	uint a = 1;
+	uint b = 0;
+	foreach (char x : data)
+	{
+		a = (a + x) % ADLER_CONST;
+		b = (b + a) % ADLER_CONST;
+	}
+	return (b << 16) | a;
+}

lib/std/hash/crc32.c3

@@ -0,0 +1,111 @@
+// Copyright (c) 2021 Christoffer Lerno. All rights reserved.
+// Use of this source code is governed by the MIT license
+// a copy of which can be found in the LICENSE_STDLIB file.
+module std::hash::crc32;
+
+struct Crc32
+{
+	uint result;
+}
+
+fn void Crc32.init(&self, uint seed = 0)
+{
+	self.result = ~seed;
+}
+
+fn void Crc32.updatec(&self, char c)
+{
+	self.result = (self.result >> 8) ^ CRC32_TABLE[(self.result ^ c) & 0xFF];
+}
+
+fn void Crc32.update(&self, char[] data)
+{
+	uint result = self.result;
+	foreach (char x : data)
+	{
+		result = (result >> 8) ^ CRC32_TABLE[(result ^ x) & 0xFF];
+	}
+	self.result = result;
+}
+
+fn uint Crc32.final(&self)
+{
+	return ~self.result;
+}
+
+fn uint encode(char[] data)
+{
+	uint result = ~(uint)(0);
+	foreach (char x : data)
+	{
+		result = (result >> 8) ^ CRC32_TABLE[(result ^ x) & 0xFF];
+	}
+	return ~result;
+}
+
+const uint[256] CRC32_TABLE @private = {
+	0x00000000, 0x77073096, 0xee0e612c, 0x990951ba,
+	0x076dc419, 0x706af48f, 0xe963a535, 0x9e6495a3,
+	0x0edb8832, 0x79dcb8a4, 0xe0d5e91e, 0x97d2d988,
+	0x09b64c2b, 0x7eb17cbd, 0xe7b82d07, 0x90bf1d91,
+	0x1db71064, 0x6ab020f2, 0xf3b97148, 0x84be41de,
+	0x1adad47d, 0x6ddde4eb, 0xf4d4b551, 0x83d385c7,
+	0x136c9856, 0x646ba8c0, 0xfd62f97a, 0x8a65c9ec,
+	0x14015c4f, 0x63066cd9, 0xfa0f3d63, 0x8d080df5,
+	0x3b6e20c8, 0x4c69105e, 0xd56041e4, 0xa2677172,
+	0x3c03e4d1, 0x4b04d447, 0xd20d85fd, 0xa50ab56b,
+	0x35b5a8fa, 0x42b2986c, 0xdbbbc9d6, 0xacbcf940,
+	0x32d86ce3, 0x45df5c75, 0xdcd60dcf, 0xabd13d59,
+	0x26d930ac, 0x51de003a, 0xc8d75180, 0xbfd06116,
+	0x21b4f4b5, 0x56b3c423, 0xcfba9599, 0xb8bda50f,
+	0x2802b89e, 0x5f058808, 0xc60cd9b2, 0xb10be924,
+	0x2f6f7c87, 0x58684c11, 0xc1611dab, 0xb6662d3d,
+	0x76dc4190, 0x01db7106, 0x98d220bc, 0xefd5102a,
+	0x71b18589, 0x06b6b51f, 0x9fbfe4a5, 0xe8b8d433,
+	0x7807c9a2, 0x0f00f934, 0x9609a88e, 0xe10e9818,
+	0x7f6a0dbb, 0x086d3d2d, 0x91646c97, 0xe6635c01,
+	0x6b6b51f4, 0x1c6c6162, 0x856530d8, 0xf262004e,
+	0x6c0695ed, 0x1b01a57b, 0x8208f4c1, 0xf50fc457,
+	0x65b0d9c6, 0x12b7e950, 0x8bbeb8ea, 0xfcb9887c,
+	0x62dd1ddf, 0x15da2d49, 0x8cd37cf3, 0xfbd44c65,
+	0x4db26158, 0x3ab551ce, 0xa3bc0074, 0xd4bb30e2,
+	0x4adfa541, 0x3dd895d7, 0xa4d1c46d, 0xd3d6f4fb,
+	0x4369e96a, 0x346ed9fc, 0xad678846, 0xda60b8d0,
+	0x44042d73, 0x33031de5, 0xaa0a4c5f, 0xdd0d7cc9,
+	0x5005713c, 0x270241aa, 0xbe0b1010, 0xc90c2086,
+	0x5768b525, 0x206f85b3, 0xb966d409, 0xce61e49f,
+	0x5edef90e, 0x29d9c998, 0xb0d09822, 0xc7d7a8b4,
+	0x59b33d17, 0x2eb40d81, 0xb7bd5c3b, 0xc0ba6cad,
+	0xedb88320, 0x9abfb3b6, 0x03b6e20c, 0x74b1d29a,
+	0xead54739, 0x9dd277af, 0x04db2615, 0x73dc1683,
+	0xe3630b12, 0x94643b84, 0x0d6d6a3e, 0x7a6a5aa8,
+	0xe40ecf0b, 0x9309ff9d, 0x0a00ae27, 0x7d079eb1,
+	0xf00f9344, 0x8708a3d2, 0x1e01f268, 0x6906c2fe,
+	0xf762575d, 0x806567cb, 0x196c3671, 0x6e6b06e7,
+	0xfed41b76, 0x89d32be0, 0x10da7a5a, 0x67dd4acc,
+	0xf9b9df6f, 0x8ebeeff9, 0x17b7be43, 0x60b08ed5,
+	0xd6d6a3e8, 0xa1d1937e, 0x38d8c2c4, 0x4fdff252,
+	0xd1bb67f1, 0xa6bc5767, 0x3fb506dd, 0x48b2364b,
+	0xd80d2bda, 0xaf0a1b4c, 0x36034af6, 0x41047a60,
+	0xdf60efc3, 0xa867df55, 0x316e8eef, 0x4669be79,
+	0xcb61b38c, 0xbc66831a, 0x256fd2a0, 0x5268e236,
+	0xcc0c7795, 0xbb0b4703, 0x220216b9, 0x5505262f,
+	0xc5ba3bbe, 0xb2bd0b28, 0x2bb45a92, 0x5cb36a04,
+	0xc2d7ffa7, 0xb5d0cf31, 0x2cd99e8b, 0x5bdeae1d,
+	0x9b64c2b0, 0xec63f226, 0x756aa39c, 0x026d930a,
+	0x9c0906a9, 0xeb0e363f, 0x72076785, 0x05005713,
+	0x95bf4a82, 0xe2b87a14, 0x7bb12bae, 0x0cb61b38,
+	0x92d28e9b, 0xe5d5be0d, 0x7cdcefb7, 0x0bdbdf21,
+	0x86d3d2d4, 0xf1d4e242, 0x68ddb3f8, 0x1fda836e,
+	0x81be16cd, 0xf6b9265b, 0x6fb077e1, 0x18b74777,
+	0x88085ae6, 0xff0f6a70, 0x66063bca, 0x11010b5c,
+	0x8f659eff, 0xf862ae69, 0x616bffd3, 0x166ccf45,
+	0xa00ae278, 0xd70dd2ee, 0x4e048354, 0x3903b3c2,
+	0xa7672661, 0xd06016f7, 0x4969474d, 0x3e6e77db,
+	0xaed16a4a, 0xd9d65adc, 0x40df0b66, 0x37d83bf0,
+	0xa9bcae53, 0xdebb9ec5, 0x47b2cf7f, 0x30b5ffe9,
+	0xbdbdf21c, 0xcabac28a, 0x53b39330, 0x24b4a3a6,
+	0xbad03605, 0xcdd70693, 0x54de5729, 0x23d967bf,
+	0xb3667a2e, 0xc4614ab8, 0x5d681b02, 0x2a6f2b94,
+	0xb40bbe37, 0xc30c8ea1, 0x5a05df1b, 0x2d02ef8d,
+};

lib/std/hash/crc64.c3

@@ -0,0 +1,111 @@
+// Copyright (c) 2021 Christoffer Lerno. All rights reserved.
+// Use of this source code is governed by the MIT license
+// a copy of which can be found in the LICENSE_STDLIB file.
+module std::hash::crc64;
+
+struct Crc64
+{
+	ulong result;
+}
+
+fn void Crc64.init(&self, uint seed = 0)
+{
+	self.result = seed;
+}
+
+fn void Crc64.updatec(&self, char c)
+{
+	self.result = (self.result << 8) ^ CRC64_TABLE[(char)((self.result >> 56) ^ c)];
+}
+
+fn void Crc64.update(&self, char[] data)
+{
+	ulong result = self.result;
+	foreach (char x : data)
+	{
+		result = (result << 8) ^ CRC64_TABLE[(char)((result >> 56) ^ x)];
+	}
+	self.result = result;
+}
+
+fn ulong Crc64.final(&self)
+{
+	return self.result;
+}
+
+fn ulong encode(char[] data)
+{
+	ulong result = (ulong)(0);
+	foreach (char x : data)
+	{
+		result = (result << 8) ^ CRC64_TABLE[(char)((result >> 56) ^ x)];
+	}
+	return result;
+}
+
+const ulong[256] CRC64_TABLE @private = {
+	0x0000000000000000, 0x42f0e1eba9ea3693, 0x85e1c3d753d46d26, 0xc711223cfa3e5bb5,
+	0x493366450e42ecdf, 0x0bc387aea7a8da4c, 0xccd2a5925d9681f9, 0x8e224479f47cb76a,
+	0x9266cc8a1c85d9be, 0xd0962d61b56fef2d, 0x17870f5d4f51b498, 0x5577eeb6e6bb820b,
+	0xdb55aacf12c73561, 0x99a54b24bb2d03f2, 0x5eb4691841135847, 0x1c4488f3e8f96ed4,
+	0x663d78ff90e185ef, 0x24cd9914390bb37c, 0xe3dcbb28c335e8c9, 0xa12c5ac36adfde5a,
+	0x2f0e1eba9ea36930, 0x6dfeff5137495fa3, 0xaaefdd6dcd770416, 0xe81f3c86649d3285,
+	0xf45bb4758c645c51, 0xb6ab559e258e6ac2, 0x71ba77a2dfb03177, 0x334a9649765a07e4,
+	0xbd68d2308226b08e, 0xff9833db2bcc861d, 0x388911e7d1f2dda8, 0x7a79f00c7818eb3b,
+	0xcc7af1ff21c30bde, 0x8e8a101488293d4d, 0x499b3228721766f8, 0x0b6bd3c3dbfd506b,
+	0x854997ba2f81e701, 0xc7b97651866bd192, 0x00a8546d7c558a27, 0x4258b586d5bfbcb4,
+	0x5e1c3d753d46d260, 0x1cecdc9e94ace4f3, 0xdbfdfea26e92bf46, 0x990d1f49c77889d5,
+	0x172f5b3033043ebf, 0x55dfbadb9aee082c, 0x92ce98e760d05399, 0xd03e790cc93a650a,
+	0xaa478900b1228e31, 0xe8b768eb18c8b8a2, 0x2fa64ad7e2f6e317, 0x6d56ab3c4b1cd584,
+	0xe374ef45bf6062ee, 0xa1840eae168a547d, 0x66952c92ecb40fc8, 0x2465cd79455e395b,
+	0x3821458aada7578f, 0x7ad1a461044d611c, 0xbdc0865dfe733aa9, 0xff3067b657990c3a,
+	0x711223cfa3e5bb50, 0x33e2c2240a0f8dc3, 0xf4f3e018f031d676, 0xb60301f359dbe0e5,
+	0xda050215ea6c212f, 0x98f5e3fe438617bc, 0x5fe4c1c2b9b84c09, 0x1d14202910527a9a,
+	0x93366450e42ecdf0, 0xd1c685bb4dc4fb63, 0x16d7a787b7faa0d6, 0x5427466c1e109645,
+	0x4863ce9ff6e9f891, 0x0a932f745f03ce02, 0xcd820d48a53d95b7, 0x8f72eca30cd7a324,
+	0x0150a8daf8ab144e, 0x43a04931514122dd, 0x84b16b0dab7f7968, 0xc6418ae602954ffb,
+	0xbc387aea7a8da4c0, 0xfec89b01d3679253, 0x39d9b93d2959c9e6, 0x7b2958d680b3ff75,
+	0xf50b1caf74cf481f, 0xb7fbfd44dd257e8c, 0x70eadf78271b2539, 0x321a3e938ef113aa,
+	0x2e5eb66066087d7e, 0x6cae578bcfe24bed, 0xabbf75b735dc1058, 0xe94f945c9c3626cb,
+	0x676dd025684a91a1, 0x259d31cec1a0a732, 0xe28c13f23b9efc87, 0xa07cf2199274ca14,
+	0x167ff3eacbaf2af1, 0x548f120162451c62, 0x939e303d987b47d7, 0xd16ed1d631917144,
+	0x5f4c95afc5edc62e, 0x1dbc74446c07f0bd, 0xdaad56789639ab08, 0x985db7933fd39d9b,
+	0x84193f60d72af34f, 0xc6e9de8b7ec0c5dc, 0x01f8fcb784fe9e69, 0x43081d5c2d14a8fa,
+	0xcd2a5925d9681f90, 0x8fdab8ce70822903, 0x48cb9af28abc72b6, 0x0a3b7b1923564425,
+	0x70428b155b4eaf1e, 0x32b26afef2a4998d, 0xf5a348c2089ac238, 0xb753a929a170f4ab,
+	0x3971ed50550c43c1, 0x7b810cbbfce67552, 0xbc902e8706d82ee7, 0xfe60cf6caf321874,
+	0xe224479f47cb76a0, 0xa0d4a674ee214033, 0x67c58448141f1b86, 0x253565a3bdf52d15,
+	0xab1721da49899a7f, 0xe9e7c031e063acec, 0x2ef6e20d1a5df759, 0x6c0603e6b3b7c1ca,
+	0xf6fae5c07d3274cd, 0xb40a042bd4d8425e, 0x731b26172ee619eb, 0x31ebc7fc870c2f78,
+	0xbfc9838573709812, 0xfd39626eda9aae81, 0x3a28405220a4f534, 0x78d8a1b9894ec3a7,
+	0x649c294a61b7ad73, 0x266cc8a1c85d9be0, 0xe17dea9d3263c055, 0xa38d0b769b89f6c6,
+	0x2daf4f0f6ff541ac, 0x6f5faee4c61f773f, 0xa84e8cd83c212c8a, 0xeabe6d3395cb1a19,
+	0x90c79d3fedd3f122, 0xd2377cd44439c7b1, 0x15265ee8be079c04, 0x57d6bf0317edaa97,
+	0xd9f4fb7ae3911dfd, 0x9b041a914a7b2b6e, 0x5c1538adb04570db, 0x1ee5d94619af4648,
+	0x02a151b5f156289c, 0x4051b05e58bc1e0f, 0x87409262a28245ba, 0xc5b073890b687329,
+	0x4b9237f0ff14c443, 0x0962d61b56fef2d0, 0xce73f427acc0a965, 0x8c8315cc052a9ff6,
+	0x3a80143f5cf17f13, 0x7870f5d4f51b4980, 0xbf61d7e80f251235, 0xfd913603a6cf24a6,
+	0x73b3727a52b393cc, 0x31439391fb59a55f, 0xf652b1ad0167feea, 0xb4a25046a88dc879,
+	0xa8e6d8b54074a6ad, 0xea16395ee99e903e, 0x2d071b6213a0cb8b, 0x6ff7fa89ba4afd18,
+	0xe1d5bef04e364a72, 0xa3255f1be7dc7ce1, 0x64347d271de22754, 0x26c49cccb40811c7,
+	0x5cbd6cc0cc10fafc, 0x1e4d8d2b65facc6f, 0xd95caf179fc497da, 0x9bac4efc362ea149,
+	0x158e0a85c2521623, 0x577eeb6e6bb820b0, 0x906fc95291867b05, 0xd29f28b9386c4d96,
+	0xcedba04ad0952342, 0x8c2b41a1797f15d1, 0x4b3a639d83414e64, 0x09ca82762aab78f7,
+	0x87e8c60fded7cf9d, 0xc51827e4773df90e, 0x020905d88d03a2bb, 0x40f9e43324e99428,
+	0x2cffe7d5975e55e2, 0x6e0f063e3eb46371, 0xa91e2402c48a38c4, 0xebeec5e96d600e57,
+	0x65cc8190991cb93d, 0x273c607b30f68fae, 0xe02d4247cac8d41b, 0xa2dda3ac6322e288,
+	0xbe992b5f8bdb8c5c, 0xfc69cab42231bacf, 0x3b78e888d80fe17a, 0x7988096371e5d7e9,
+	0xf7aa4d1a85996083, 0xb55aacf12c735610, 0x724b8ecdd64d0da5, 0x30bb6f267fa73b36,
+	0x4ac29f2a07bfd00d, 0x08327ec1ae55e69e, 0xcf235cfd546bbd2b, 0x8dd3bd16fd818bb8,
+	0x03f1f96f09fd3cd2, 0x41011884a0170a41, 0x86103ab85a2951f4, 0xc4e0db53f3c36767,
+	0xd8a453a01b3a09b3, 0x9a54b24bb2d03f20, 0x5d45907748ee6495, 0x1fb5719ce1045206,
+	0x919735e51578e56c, 0xd367d40ebc92d3ff, 0x1476f63246ac884a, 0x568617d9ef46bed9,
+	0xe085162ab69d5e3c, 0xa275f7c11f7768af, 0x6564d5fde549331a, 0x279434164ca30589,
+	0xa9b6706fb8dfb2e3, 0xeb46918411358470, 0x2c57b3b8eb0bdfc5, 0x6ea7525342e1e956,
+	0x72e3daa0aa188782, 0x30133b4b03f2b111, 0xf7021977f9cceaa4, 0xb5f2f89c5026dc37,
+	0x3bd0bce5a45a6b5d, 0x79205d0e0db05dce, 0xbe317f32f78e067b, 0xfcc19ed95e6430e8,
+	0x86b86ed5267cdbd3, 0xc4488f3e8f96ed40, 0x0359ad0275a8b6f5, 0x41a94ce9dc428066,
+	0xcf8b0890283e370c, 0x8d7be97b81d4019f, 0x4a6acb477bea5a2a, 0x089a2aacd2006cb9,
+	0x14dea25f3af9026d, 0x562e43b4931334fe, 0x913f6188692d6f4b, 0xd3cf8063c0c759d8,
+	0x5dedc41a34bbeeb2, 0x1f1d25f19d51d821, 0xd80c07cd676f8394, 0x9afce626ce85b507,
+};

lib/std/hash/fnv32a.c3

@@ -0,0 +1,41 @@
+// Copyright (c) 2021 Christoffer Lerno. All rights reserved.
+// Use of this source code is governed by the MIT license
+// a copy of which can be found in the LICENSE_STDLIB file.
+module std::hash::fnv32a;
+
+distinct Fnv32a = uint;
+
+const FNV32A_START @private = 0x811c9dc5;
+const FNV32A_MUL @private = 0x01000193;
+
+macro void @update(uint* &h, char x) @private => *h = (*h * FNV32A_MUL) ^ x;
+
+fn void Fnv32a.init(&self)
+{
+	*self = FNV32A_START;
+}
+
+fn void Fnv32a.update(&self, char[] data)
+{
+	uint h = (uint)*self;
+	foreach (char x : data)
+	{
+		@update(h, x);
+	}
+	*self = (Fnv32a)h;
+}
+
+macro void Fnv32a.update_char(&self, char c)
+{
+	@update(*self, x);
+}
+
+fn uint encode(char[] data)
+{
+	uint h = FNV32A_START;
+	foreach (char x : data)
+	{
+		@update(h, x);
+	}
+	return h;
+}

lib/std/hash/fnv64a.c3

@@ -0,0 +1,41 @@
+// Copyright (c) 2021 Christoffer Lerno. All rights reserved.
+// Use of this source code is governed by the MIT license
+// a copy of which can be found in the LICENSE_STDLIB file.
+module std::hash::fnv64a;
+
+distinct Fnv64a = ulong;
+
+const FNV64A_START @private = 0xcbf29ce484222325;
+const FNV64A_MUL @private = 0x00000100000001b3;
+
+macro void @update(ulong* &h, char x) @private => *h = (*h * FNV64A_MUL) ^ x;
+
+fn void Fnv64a.init(&self)
+{
+	*self = FNV64A_START;
+}
+
+fn void Fnv64a.update(&self, char[] data)
+{
+	ulong h = (ulong)*self;
+	foreach (char x : data)
+	{
+		@update(h, x);
+	}
+	*self = (Fnv64a)h;
+}
+
+macro void Fnv64a.update_char(&self, char c)
+{
+	@update(*self, x);
+}
+
+fn ulong encode(char[] data)
+{
+	ulong h = FNV64A_START;
+	foreach (char x : data)
+	{
+		@update(h, x);
+	}
+	return h;
+}

lib/std/hash/sha1.c3

@@ -0,0 +1,241 @@
+// Copyright (c) 2021 Christoffer Lerno. All rights reserved.
+// Use of this source code is governed by the MIT license
+// a copy of which can be found in the LICENSE_STDLIB file.
+//
+// Implementation was off Steve Reid's SHA-1 C implementation
+
+module std::hash::sha1;
+import std::bits;
+
+struct Sha1
+{
+	uint[5] state;
+	uint[2] count;
+	char[64] buffer;
+}
+
+fn void Sha1.init(&self)
+{
+	// SHA1 initialization constants
+	*self = {
+		.state = {
+			0x67452301,
+			0xEFCDAB89,
+			0x98BADCFE,
+			0x10325476,
+			0xC3D2E1F0
+		}
+	};
+}
+
+/**
+ * @param [in] data
+ * @require data.len <= uint.max
+ **/
+fn void Sha1.update(&self, char[] data)
+{
+	uint j = self.count[0];
+	uint len = data.len;
+	if ((self.count[0] += len << 3) < j) self.count[1]++;
+	self.count[1] += len >> 29;
+	j = (j >> 3) & 63;
+	uint i;
+	if (j + len > 63)
+	{
+		i = 64 - j;
+		self.buffer[j..] = data[:i];
+		sha1_transform(&self.state, &self.buffer);
+		for (; i + 63 < len; i += 64)
+		{
+			sha1_transform(&self.state, &data[i]);
+		}
+		j = 0;
+	}
+	self.buffer[j:len - i] = data[i..];
+}
+
+
+fn char[20] Sha1.final(&self)
+{
+	char[8] finalcount;
+	for (uint i = 0; i < 8; i++)
+	{
+		finalcount[i] = (char)((self.count[(i >= 4 ? 0 : 1)] >> ((3 - (i & 3)) * 8)) & 0xFF);
+	}
+	self.update(char[] { 0o200 });
+	while ((self.count[0] & 504) != 448)
+	{
+		self.update(char[] { 0 });
+	}
+
+	self.update(&finalcount);
+	char[20] digest;
+	for (uint i = 0; i < 20; i++)
+	{
+		digest[i] = (char)((self.state[i >> 2] >> ((3 - (i & 3)) * 8)) & 0xFF);
+	}
+
+	// Clear mem
+	mem::clear(self, Sha1.sizeof);
+	finalcount = {};
+	return digest;
+}
+
+union Long16 @local
+{
+	char[64] c;
+	uint[16] l;
+}
+
+macro @blk(&block, i) @local
+{
+	return (block.l[i & 15] = (block.l[(i + 13) & 15] ^ block.l[(i + 8) & 15]
+		^ block.l[(i + 2) & 15] ^ block.l[i & 15]).rotl(1));
+}
+
+macro @blk0(&block, i) @local
+{
+	$if env::BIG_ENDIAN:
+		return block.l[i];
+	$else
+		return block.l[i] = (block.l[i].rotl(24) & 0xFF00FF00)
+			| (block.l[i].rotl(8) & 0x00FF00FF);
+	$endif
+}
+
+macro @r0(&block, v, &wref, x, y, &z, i) @local
+{
+	var w = *wref;
+	*z += ((w & (x ^ y)) ^ y) + @blk0(*block, i) + 0x5A827999 + v.rotl(5);
+	*wref = w.rotl(30);
+}
+
+macro @r1(&block, v, &wref, x, y, &z, i) @local
+{
+	var w = *wref;
+	*z += ((w & (x ^ y)) ^ y) + @blk(*block, i) + 0x5A827999 + v.rotl(5);
+	*wref = w.rotl(30);
+}
+
+macro @r2(&block, v, &wref, x, y, &z, i) @local
+{
+	var w = *wref;
+	*z += (w ^ x ^ y) + @blk(*block, i) + 0x6ED9EBA1 + v.rotl(5);
+	*wref = w.rotl(30);
+}
+
+macro @r3(&block, v, &wref, x, y, &z, i) @local
+{
+	var w = *wref;
+	*z += (((w | x) & y) | (w & x)) + @blk(*block, i) + 0x8F1BBCDC + v.rotl(5);
+	*wref = w.rotl(30);
+}
+
+macro @r4(&block, v, &wref, x, y, &z, i) @local
+{
+	var w = *wref;
+	*z += (w ^ x ^ y) + @blk(*block, i) + 0xCA62C1D6 + v.rotl(5);
+	*wref = w.rotl(30);
+}
+
+/**
+ * @param [&inout] state
+ * @param [&in] buffer
+ **/
+fn void sha1_transform(uint* state, char* buffer) @local
+{
+	Long16 block;
+	block.c[..] = buffer[:64];
+	uint a = state[0];
+	uint b = state[1];
+	uint c = state[2];
+	uint d = state[3];
+	uint e = state[4];
+	@r0(block, a, b, c, d, e, 0);
+	@r0(block, e, a, b, c, d, 1);
+	@r0(block, d, e, a, b, c, 2);
+	@r0(block, c, d, e, a, b, 3);
+	@r0(block, b, c, d, e, a, 4);
+	@r0(block, a, b, c, d, e, 5);
+	@r0(block, e, a, b, c, d, 6);
+	@r0(block, d, e, a, b, c, 7);
+	@r0(block, c, d, e, a, b, 8);
+	@r0(block, b, c, d, e, a, 9);
+	@r0(block, a, b, c, d, e, 10);
+	@r0(block, e, a, b, c, d, 11);
+	@r0(block, d, e, a, b, c, 12);
+	@r0(block, c, d, e, a, b, 13);
+	@r0(block, b, c, d, e, a, 14);
+	@r0(block, a, b, c, d, e, 15);
+	@r1(block, e, a, b, c, d, 16);
+	@r1(block, d, e, a, b, c, 17);
+	@r1(block, c, d, e, a, b, 18);
+	@r1(block, b, c, d, e, a, 19);
+	@r2(block, a, b, c, d, e, 20);
+	@r2(block, e, a, b, c, d, 21);
+	@r2(block, d, e, a, b, c, 22);
+	@r2(block, c, d, e, a, b, 23);
+	@r2(block, b, c, d, e, a, 24);
+	@r2(block, a, b, c, d, e, 25);
+	@r2(block, e, a, b, c, d, 26);
+	@r2(block, d, e, a, b, c, 27);
+	@r2(block, c, d, e, a, b, 28);
+	@r2(block, b, c, d, e, a, 29);
+	@r2(block, a, b, c, d, e, 30);
+	@r2(block, e, a, b, c, d, 31);
+	@r2(block, d, e, a, b, c, 32);
+	@r2(block, c, d, e, a, b, 33);
+	@r2(block, b, c, d, e, a, 34);
+	@r2(block, a, b, c, d, e, 35);
+	@r2(block, e, a, b, c, d, 36);
+	@r2(block, d, e, a, b, c, 37);
+	@r2(block, c, d, e, a, b, 38);
+	@r2(block, b, c, d, e, a, 39);
+	@r3(block, a, b, c, d, e, 40);
+	@r3(block, e, a, b, c, d, 41);
+	@r3(block, d, e, a, b, c, 42);
+	@r3(block, c, d, e, a, b, 43);
+	@r3(block, b, c, d, e, a, 44);
+	@r3(block, a, b, c, d, e, 45);
+	@r3(block, e, a, b, c, d, 46);
+	@r3(block, d, e, a, b, c, 47);
+	@r3(block, c, d, e, a, b, 48);
+	@r3(block, b, c, d, e, a, 49);
+	@r3(block, a, b, c, d, e, 50);
+	@r3(block, e, a, b, c, d, 51);
+	@r3(block, d, e, a, b, c, 52);
+	@r3(block, c, d, e, a, b, 53);
+	@r3(block, b, c, d, e, a, 54);
+	@r3(block, a, b, c, d, e, 55);
+	@r3(block, e, a, b, c, d, 56);
+	@r3(block, d, e, a, b, c, 57);
+	@r3(block, c, d, e, a, b, 58);
+	@r3(block, b, c, d, e, a, 59);
+	@r4(block, a, b, c, d, e, 60);
+	@r4(block, e, a, b, c, d, 61);
+	@r4(block, d, e, a, b, c, 62);
+	@r4(block, c, d, e, a, b, 63);
+	@r4(block, b, c, d, e, a, 64);
+	@r4(block, a, b, c, d, e, 65);
+	@r4(block, e, a, b, c, d, 66);
+	@r4(block, d, e, a, b, c, 67);
+	@r4(block, c, d, e, a, b, 68);
+	@r4(block, b, c, d, e, a, 69);
+	@r4(block, a, b, c, d, e, 70);
+	@r4(block, e, a, b, c, d, 71);
+	@r4(block, d, e, a, b, c, 72);
+	@r4(block, c, d, e, a, b, 73);
+	@r4(block, b, c, d, e, a, 74);
+	@r4(block, a, b, c, d, e, 75);
+	@r4(block, e, a, b, c, d, 76);
+	@r4(block, d, e, a, b, c, 77);
+	@r4(block, c, d, e, a, b, 78);
+	@r4(block, b, c, d, e, a, 79);
+	state[0] += a;
+	state[1] += b;
+	state[2] += c;
+	state[3] += d;
+	state[4] += e;
+	a = b = c = d = e = 0;
+	buffer[:64] = 0;
+}

lib/std/io/bits.c3

@@ -0,0 +1,85 @@
+module std::io;
+
+struct BitReader
+{
+	InStream reader;
+	uint bits;
+	uint len;
+}
+
+fn void BitReader.init(&self, InStream byte_reader)
+{
+	*self = { .reader = byte_reader };
+}
+
+fn void BitReader.clear(&self) @inline
+{
+	self.len = 0;
+}
+
+/**
+ * @require nbits <= 8
+ * @require self.len + nbits <= uint.sizeof * 8
+ **/
+fn char! BitReader.read_bits(&self, uint nbits)
+{
+	uint bits = self.bits;
+	if (self.len < nbits)
+	{
+		// New bits are pushed right.
+		char c = self.reader.read_byte()!;
+		bits <<= 8;
+		bits |= c;
+		self.bits = bits;
+		self.len += 8;
+	}
+	self.len -= nbits;
+	uint mask = (1 << nbits) - 1;
+	return (char)((bits >> self.len) & mask);
+}
+
+struct BitWriter
+{
+	OutStream writer;
+	uint bits;
+	uint len;
+}
+
+fn void BitWriter.init(&self, OutStream byte_writer)
+{
+	*self = { .writer = byte_writer };
+}
+
+fn void! BitWriter.flush(&self)
+{
+	if (self.len == 0) return;
+	uint bits = self.bits << (32 - self.len);
+	uint n = (self.len + 7) / 8;
+	char[4] buffer;
+	bitorder::write(bits, &buffer, UIntBE);
+	io::write_all(self.writer, buffer[:n])!;
+	self.len = 0;
+}
+
+/**
+ * @require nbits <= 8
+ **/
+fn void! BitWriter.write_bits(&self, uint bits, uint nbits)
+{
+	if (nbits == 0) return;
+	uint n = self.len + nbits;
+	uint to_write = n / 8;
+	uint left = n % 8;
+	if (to_write > 0)
+	{
+		ulong lbits;
+		if (self.len > 0) lbits = (ulong)self.bits << (64 - self.len);
+		lbits |= (ulong)(bits >> left) << (64 - (n - left));
+		char[8] buffer;
+		bitorder::write(lbits, &buffer, ULongBE);
+		io::write_all(self.writer, buffer[:to_write])!;
+	}
+	self.bits <<= left;
+	self.bits |= bits & ((1 << left) - 1);
+	self.len = left;
+}

lib/std/io/file.c3

@@ -0,0 +1,191 @@
+module std::io;
+import libc;
+
+struct File (InStream, OutStream)
+{
+	CFile file;
+}
+
+module std::io::file;
+import libc, std::io::path, std::io::os;
+
+fn File! open(String filename, String mode)
+{
+	return from_handle(os::native_fopen(filename, mode));
+}
+
+fn File! open_path(Path path, String mode)
+{
+	return from_handle(os::native_fopen(path.str_view(), mode));
+}
+
+fn File from_handle(CFile file)
+{
+	return { .file = file };
+}
+
+fn bool is_file(String path)
+{
+	return os::native_is_file(path);
+}
+
+fn usz! get_size(String path)
+{
+	return os::native_file_size(path);
+}
+
+fn void! delete(String filename) => os::native_remove(filename) @inline;
+
+
+/**
+ * @require self.file != null
+ **/
+fn void! File.reopen(&self, String filename, String mode)
+{
+	self.file = os::native_freopen(self.file, filename, mode)!;
+}
+
+/**
+ * @require self.file != null
+ **/
+fn usz! File.seek(&self, isz offset, Seek seek_mode = Seek.SET) @dynamic
+{
+	os::native_fseek(self.file, offset, seek_mode)!;
+	return os::native_ftell(self.file);
+}
+
+
+/*
+Implement later
+/**
+ * @require self.file == null
+ **/
+fn void! File.memopen(File* file, char[] data, String mode)
+{
+	@pool()
+	{
+		file.file = libc::memopen(data.ptr, data.len, mode.zstr_tcopy(), file.file);
+		// TODO errors
+	};
+}
+*/
+
+
+/**
+ * @require self.file != null
+ */
+fn void! File.write_byte(&self, char c) @dynamic
+{
+	if (!libc::fputc(c, self.file)) return IoError.EOF?;
+}
+
+/**
+ * @param [&inout] self
+ */
+fn void! File.close(&self) @inline @dynamic
+{
+	if (self.file && libc::fclose(self.file))
+	{
+		switch (libc::errno())
+		{
+			case errno::ECONNRESET:
+			case errno::EBADF: return IoError.FILE_NOT_VALID?;
+			case errno::EINTR: return IoError.INTERRUPTED?;
+			case errno::EDQUOT:
+			case errno::EFAULT:
+			case errno::EAGAIN:
+			case errno::EFBIG:
+			case errno::ENETDOWN:
+			case errno::ENETUNREACH:
+			case errno::ENOSPC:
+			case errno::EIO: return IoError.INCOMPLETE_WRITE?;
+			default: return IoError.UNKNOWN_ERROR?;
+		}
+	}
+	self.file = null;
+}
+
+/**
+ * @require self.file
+ */
+fn bool File.eof(&self) @inline
+{
+	return libc::feof(self.file) != 0;
+}
+
+/**
+ * @param [in] buffer
+ */
+fn usz! File.read(&self, char[] buffer) @dynamic
+{
+	return os::native_fread(self.file, buffer);
+}
+
+/**
+ * @param [out] buffer
+ * @require self.file `File must be initialized`
+ */
+fn usz! File.write(&self, char[] buffer) @dynamic
+{
+	return os::native_fwrite(self.file, buffer);
+}
+
+
+fn char! File.read_byte(&self) @dynamic
+{
+	int c = libc::fgetc(self.file);
+	if (c == -1) return IoError.EOF?;
+	return (char)c;
+}
+
+/**
+ * Load up to buffer.len characters. Returns IoError.OVERFLOW if the file is longer
+ * than the buffer.
+ *
+ * @param filename "The path to the file to read"
+ * @param [in] buffer "The buffer to read to"
+ **/
+fn char[]! load_buffer(String filename, char[] buffer)
+{
+	File file = open(filename, "rb")!;
+	defer (void)file.close();
+	usz len = file.seek(0, END)!;
+	if (len > buffer.len) return IoError.OVERFLOW?;
+	file.seek(0, SET)!;
+	usz read = 0;
+	while (read < len)
+	{
+		read += file.read(buffer[read:len - read])!;
+	}
+	return buffer[:len];
+
+}
+
+fn char[]! load_new(String filename, Allocator allocator = allocator::heap())
+{
+	File file = open(filename, "rb")!;
+	defer (void)file.close();
+	usz len = file.seek(0, END)!;
+	file.seek(0, SET)!;
+	char* data = allocator::malloc_try(allocator, len)!;
+	defer catch allocator::free(allocator, data);
+	usz read = 0;
+	while (read < len)
+	{
+		read += file.read(data[read:len - read])!;
+	}
+	return data[:len];
+}
+
+fn char[]! load_temp(String filename)
+{
+	return load_new(filename, allocator::temp());
+}
+
+/**
+ * @require self.file `File must be initialized`
+ */
+fn void! File.flush(&self) @dynamic
+{
+	libc::fflush(self.file);
+}

lib/std/io/formatter.c3

@@ -0,0 +1,463 @@
+module std::io;
+import std::collections::map;
+import libc;
+
+const int PRINTF_NTOA_BUFFER_SIZE = 256;
+
+interface Printable
+{
+	fn String to_new_string(Allocator allocator) @optional;
+	fn usz! to_format(Formatter* formatter) @optional;
+}
+
+fault PrintFault
+{
+	BUFFER_EXCEEDED,
+	INTERNAL_BUFFER_EXCEEDED,
+	INVALID_FORMAT_STRING,
+	MISSING_ARG,
+	INVALID_ARGUMENT_TYPE,
+}
+
+fault FormattingFault
+{
+	UNTERMINATED_FORMAT,
+	MISSING_ARG,
+	INVALID_WIDTH_ARG,
+	INVALID_FORMAT_TYPE,
+}
+
+def OutputFn = fn void!(void* buffer, char c);
+def FloatType = double;
+
+
+
+fn usz! Formatter.printf(&self, String format, args...)
+{
+	return self.vprintf(format, args) @inline;
+}
+
+struct Formatter
+{
+	void *data;
+	OutputFn out_fn;
+	struct
+	{
+		PrintFlags flags;
+		uint width;
+		uint prec;
+		usz idx;
+	}
+}
+
+bitstruct PrintFlags : uint
+{
+	bool zeropad;
+	bool left;
+	bool plus;
+	bool space;
+	bool hash;
+	bool uppercase;
+	bool precision;
+}
+
+fn void Formatter.init(&self, OutputFn out_fn, void* data = null)
+{
+	*self = { .data = data, .out_fn = out_fn};
+}
+
+fn usz! Formatter.out(&self, char c) @private
+{
+	self.out_fn(self.data, c)!;
+	return 1;
+}
+
+fn usz! Formatter.print_with_function(&self, Printable arg)
+{
+	if (&arg.to_format)
+	{
+		PrintFlags old = self.flags;
+		uint old_width = self.width;
+		uint old_prec = self.prec;
+		defer
+		{
+			self.flags = old;
+			self.width = old_width;
+			self.prec = old_prec;
+		}
+		if (!arg) return self.out_substr("(null)");
+		return arg.to_format(self);
+	}
+	if (&arg.to_new_string)
+	{
+		PrintFlags old = self.flags;
+		uint old_width = self.width;
+		uint old_prec = self.prec;
+		defer
+		{
+			self.flags = old;
+			self.width = old_width;
+			self.prec = old_prec;
+		}
+		if (!arg) return self.out_substr("(null)");
+		@stack_mem(1024; Allocator mem)
+		{
+			return self.out_substr(arg.to_new_string(mem));
+		};
+	}
+	return SearchResult.MISSING?;
+}
+
+
+fn usz! Formatter.out_str(&self, any arg) @private
+{
+	switch (arg.type.kindof)
+	{
+		case TYPEID:
+			return self.out_substr("typeid");
+		case VOID:
+			return self.out_substr("void");
+		case ANYFAULT:
+		case FAULT:
+			return self.out_substr((*(anyfault*)arg.ptr).nameof);
+		case ANY:
+			return self.out_str(*(any*)arg);
+		case OPTIONAL:
+			unreachable();
+		case SIGNED_INT:
+		case UNSIGNED_INT:
+			PrintFlags flags = self.flags;
+			uint width = self.width;
+			defer
+			{
+				self.flags = flags;
+				self.width = width;
+			}
+			self.flags = {};
+			self.width = 0;
+			return self.ntoa_any(arg, 10);
+		case FLOAT:
+			PrintFlags flags = self.flags;
+			uint width = self.width;
+			defer
+			{
+				self.flags = flags;
+				self.width = width;
+			}
+			self.flags = {};
+			self.width = 0;
+			return self.ftoa(float_from_any(arg)!!);
+		case BOOL:
+			return self.out_substr(*(bool*)arg.ptr ? "true" : "false");
+		default:
+	}
+	usz! n = self.print_with_function((Printable)arg);
+	if (try n) return n;
+	if (@catch(n) != SearchResult.MISSING) n!;
+	switch (arg.type.kindof)
+	{
+		case ENUM:
+			usz i = types::any_to_int(arg, usz)!!;
+			assert(i < arg.type.names.len, "Illegal enum value found, numerical value was %d.", i);
+			return self.out_substr(arg.type.names[i]);
+		case STRUCT:
+			return self.out_substr("<struct>");
+		case UNION:
+			return self.out_substr("<union>");
+		case BITSTRUCT:
+			return self.out_substr("<bitstruct>");
+		case FUNC:
+			return self.out_substr("<function>");
+		case DISTINCT:
+			if (arg.type == ZString.typeid)
+			{
+				return self.out_substr(*(ZString*)arg ? ((ZString*)arg).str_view() : "(null)");
+			}
+			if (arg.type == DString.typeid)
+			{
+				return self.out_substr(*(DString*)arg ? ((DString*)arg).str_view() : "(null)");
+			}
+			return self.out_str(arg.as_inner());
+		case POINTER:
+			typeid inner = arg.type.inner;
+			void** pointer = arg.ptr;
+			if (arg.type.inner != void.typeid)
+			{
+				any deref = any_make(*pointer, inner);
+				n = self.print_with_function((Printable)deref);
+				if (try n) return n;
+				if (@catch(n) != SearchResult.MISSING) n!;
+			}
+			PrintFlags flags = self.flags;
+			uint width = self.width;
+			defer
+			{
+				self.flags = flags;
+				self.width = width;
+			}
+			self.width = 0;
+			self.out_substr("0x")!;
+			return self.ntoa_any(arg, 16);
+		case ARRAY:
+			// this is SomeType[*] so grab the "SomeType"
+			PrintFlags flags = self.flags;
+			uint width = self.width;
+			defer
+			{
+				self.flags = flags;
+				self.width = width;
+			}
+			self.flags = {};
+			self.width = 0;
+			typeid inner = arg.type.inner;
+			usz size = inner.sizeof;
+			usz alen = arg.type.len;
+			// Pretend this is a String
+			void* ptr = (void*)arg.ptr;
+			usz len = self.out('[')!;
+			for (usz i = 0; i < alen; i++)
+			{
+				if (i != 0) len += self.out_substr(", ")!;
+				len += self.out_str(any_make(ptr, inner))!;
+				ptr += size;
+			}
+			len += self.out(']')!;
+			return len;
+		case VECTOR:
+			PrintFlags flags = self.flags;
+			uint width = self.width;
+			defer
+			{
+				self.flags = flags;
+				self.width = width;
+			}
+			self.flags = {};
+			self.width = 0;
+			// this is SomeType[*] so grab the "SomeType"
+			typeid inner = arg.type.inner;
+			usz size = inner.sizeof;
+			usz vlen = arg.type.len;
+			// Pretend this is a String
+			void* ptr = (void*)arg.ptr;
+			usz len = self.out_substr("[<")!;
+			for (usz i = 0; i < vlen; i++)
+			{
+				if (i != 0) len += self.out_substr(", ")!;
+				len += self.out_str(any_make(ptr, inner))!;
+				ptr += size;
+			}
+			len += self.out_substr(">]")!;
+			return len;
+		case SLICE:
+			// this is SomeType[] so grab the "SomeType"
+			typeid inner = arg.type.inner;
+			if (inner == char.typeid)
+			{
+				return self.out_substr(*(String*)arg);
+			}
+			if (inner == void.typeid) inner = char.typeid;
+			PrintFlags flags = self.flags;
+			uint width = self.width;
+			defer
+			{
+				self.flags = flags;
+				self.width = width;
+			}
+			self.flags = {};
+			self.width = 0;
+			usz size = inner.sizeof;
+			// Pretend this is a String
+			String* temp = (void*)arg.ptr;
+			void* ptr = (void*)temp.ptr;
+			usz slen = temp.len;
+			usz len = self.out('[')!;
+			for (usz i = 0; i < slen; i++)
+			{
+				if (i != 0) len += self.out_substr(", ")!;
+				len += self.out_str(any_make(ptr, inner))!;
+				ptr += size;
+			}
+			len += self.out(']')!;
+			return len;
+		default:
+	}
+	return self.out_substr("Invalid type");
+}
+
+
+
+
+fn void! out_null_fn(void* data @unused, char c @unused) @private
+{
+}
+
+
+
+fn usz! Formatter.vprintf(&self, String format, any[] anys)
+{
+	if (!self.out_fn)
+	{
+		// use null output function
+		self.out_fn = &out_null_fn;
+	}
+	usz total_len;
+	usz format_len = format.len;
+	usz variant_index = 0;
+	for (usz i = 0; i < format_len; i++)
+	{
+		// format specifier?  %[flags][width][.precision][length]
+		char c = format[i];
+		if (c != '%')
+		{
+			// no
+			total_len += self.out(c)!;
+			continue;
+		}
+		i++;
+		if (i >= format_len) return PrintFault.INVALID_FORMAT_STRING?;
+		c = format[i];
+		if (c == '%')
+		{
+			total_len += self.out(c)!;
+			continue;
+		}
+		// evaluate flags
+		self.flags = {};
+		while FLAG_EVAL: (true)
+		{
+			switch (c)
+			{
+				case '0': self.flags.zeropad = true;
+				case '-': self.flags.left = true;
+				case '+': self.flags.plus = true;
+				case ' ': self.flags.space = true;
+				case '#': self.flags.hash = true;
+				default: break FLAG_EVAL;
+			}
+			if (++i >= format_len) return PrintFault.INVALID_FORMAT_STRING?;
+			c = format[i];
+		}
+		// evaluate width field
+		int w = printf_parse_format_field(anys.ptr, anys.len, &variant_index, format.ptr, format.len, &i)!;
+		c = format[i];
+		if (w < 0)
+		{
+			self.flags.left = true;
+			w = -w;
+		}
+		self.width = w;
+		// evaluate precision field
+		self.prec = 0;
+		if (c == '.')
+		{
+			self.flags.precision = true;
+			if (++i >= format_len) return PrintFault.INVALID_FORMAT_STRING?;
+			int prec = printf_parse_format_field(anys.ptr, anys.len, &variant_index, format.ptr, format.len, &i)!;
+			self.prec = prec < 0 ? 0 : prec;
+			c = format[i];
+		}
+
+		// evaluate specifier
+		uint base = 0;
+		if (variant_index >= anys.len) return PrintFault.MISSING_ARG?;
+		any current = anys[variant_index++];
+		switch (c)
+		{
+			case 'd':
+				base = 10;
+				self.flags.hash = false;
+			case 'X' :
+				self.flags.uppercase = true;
+				nextcase;
+			case 'x' :
+				base = 16;
+			case 'O':
+				self.flags.uppercase = true;
+				nextcase;
+			case 'o' :
+				base = 8;
+			case 'B':
+				self.flags.uppercase = true;
+				nextcase;
+			case 'b' :
+				base = 2;
+			case 'A':
+				self.flags.uppercase = true;
+				nextcase;
+			case 'a':
+				total_len += self.atoa(float_from_any(current)!!)!;
+				continue;
+			case 'F' :
+				self.flags.uppercase = true;
+				nextcase;
+			case 'f':
+				total_len += self.ftoa(float_from_any(current)!!)!;
+				continue;
+			case 'E':
+				self.flags.uppercase = true;
+				nextcase;
+			case 'e':
+				total_len += self.etoa(float_from_any(current)!!)!;
+				continue;
+			case 'G':
+				self.flags.uppercase = true;
+				nextcase;
+			case 'g':
+				total_len += self.gtoa(float_from_any(current)!!)!;
+				continue;
+			case 'c':
+				total_len += self.out_char(current)!;
+				continue;
+			case 's':
+				if (self.flags.left)
+				{
+					usz len = self.out_str(current)!;
+					total_len += len;
+					total_len += self.pad(' ', self.width, len)!;
+					continue;
+				}
+				OutputFn out_fn = self.out_fn;
+				self.out_fn = (OutputFn)&out_null_fn;
+				usz len = self.out_str(current)!;
+				self.out_fn = out_fn;
+				total_len += self.pad(' ', self.width, len)!;
+				total_len += self.out_str(current)!;
+				continue;
+			case 'p':
+				self.flags.zeropad = true;
+				self.flags.hash = true;
+				base = 16;
+			default:
+				return PrintFault.INVALID_FORMAT_STRING?;
+		}
+		if (base != 10)
+		{
+			self.flags.plus = false;
+			self.flags.space = false;
+		}
+		// ignore '0' flag when precision is given
+		if (self.flags.precision) self.flags.zeropad = false;
+
+		bool is_neg;
+		uint128 v = int_from_any(current, &is_neg)!!;
+
+		total_len += self.ntoa(v, is_neg, base)!;
+	}
+	// termination
+	//	out((char)0, buffer, idx < maxlen ? idx : maxlen - 1U, maxlen);
+
+	// return written chars without terminating \0
+	return total_len;
+}
+
+
+fn usz! Formatter.print(&self, String str)
+{
+	if (!self.out_fn)
+	{
+		// use null output function
+		self.out_fn = &out_null_fn;
+	}
+	foreach (c : str) self.out(c)!;
+	return self.idx;
+}

lib/std/io/formatter_private.c3

@@ -0,0 +1,667 @@
+module std::io;
+import std::math;
+
+const char[16] XDIGITS_H = "0123456789ABCDEF";
+const char[16] XDIGITS_L = "0123456789abcdef";
+
+fn usz! Formatter.adjust(&self, usz len) @local
+{
+	if (!self.flags.left) return 0;
+	return self.pad(' ', self.width, len);
+}
+
+fn uint128! int_from_any(any arg, bool *is_neg) @private
+{
+	switch (arg.type.kindof)
+	{
+		case TypeKind.POINTER:
+			*is_neg = false;
+			return (uint128)(uptr)*(void**)arg.ptr;
+		case TypeKind.DISTINCT:
+		case TypeKind.ENUM:
+			return int_from_any(arg.as_inner(), is_neg);
+		default:
+	}
+	*is_neg = false;
+	switch (arg)
+	{
+		case bool:
+			return (uint128)*arg;
+		case ichar:
+			int val = *arg;
+			return (*is_neg = val < 0) ? (~(uint128)val) + 1 : (uint128)val;
+		case short:
+			int val = *arg;
+			return (*is_neg = val < 0) ? (~(uint128)val) + 1 : (uint128)val;
+		case int:
+			int val = *arg;
+			return (*is_neg = val < 0) ? (~(uint128)val) + 1 : (uint128)val;
+		case long:
+			long val = *arg;
+			return (*is_neg = val < 0) ? (~(uint128)val) + 1 : (uint128)val;
+		case int128:
+			int128 val = *arg;
+			return (*is_neg = val < 0) ? (~(uint128)val) + 1 : (uint128)val;
+		case char:
+			return *arg;
+		case ushort:
+			return *arg;
+		case uint:
+			return *arg;
+		case ulong:
+			return *arg;
+		case uint128:
+			return *arg;
+		case float:
+			float f = *arg;
+			return (uint128)((*is_neg = f < 0) ? -f : f);
+		case double:
+			double d = *arg;
+			return (uint128)((*is_neg = d < 0) ? -d : d);
+		default:
+			return PrintFault.INVALID_ARGUMENT_TYPE?;
+	}
+}
+
+fn FloatType! float_from_any(any arg) @private
+{
+	$if env::F128_SUPPORT:
+		if (arg.type == float128.typeid) return (FloatType)*((float128*)arg.ptr);
+	$endif
+	if (arg.type.kindof == TypeKind.DISTINCT)
+	{
+		return float_from_any(arg.as_inner());
+	}
+	switch (arg)
+	{
+		case bool:
+			return (FloatType)*arg;
+		case ichar:
+			return *arg;
+		case short:
+			return *arg;
+		case int:
+			return *arg;
+		case long:
+			return *arg;
+		case int128:
+			return *arg;
+		case char:
+			return *arg;
+		case ushort:
+			return *arg;
+		case uint:
+			return *arg;
+		case ulong:
+			return *arg;
+		case uint128:
+			return *arg;
+		case float:
+			return (FloatType)*arg;
+		case double:
+			return (FloatType)*arg;
+		default:
+			return PrintFault.INVALID_ARGUMENT_TYPE?;
+	}
+}
+
+
+/**
+ * Read a simple integer value, typically for formatting.
+ *
+ * @param [inout] len_ptr "the length remaining."
+ * @param [in] buf "the buf to read from."
+ * @param maxlen "the maximum len that can be read."
+ * @return "The result of the atoi."
+ **/
+fn uint simple_atoi(char* buf, usz maxlen, usz* len_ptr) @inline @private
+{
+	uint i = 0;
+	usz len = *len_ptr;
+	while (len < maxlen)
+	{
+		char c = buf[len];
+		if (!c.is_digit()) break;
+		i = i * 10 + c - '0';
+		len++;
+	}
+	*len_ptr = len;
+	return i;
+}
+
+fn usz! Formatter.out_substr(&self, String str) @private
+{
+	usz l = conv::utf8_codepoints(str);
+	uint prec = self.prec;
+	if (self.flags.precision && l < prec) l = prec;
+	usz index = 0;
+	usz chars = str.len;
+	char* ptr = str.ptr;
+	while (index < chars)
+	{
+		char c = ptr[index];
+		// Break if we have precision set and we ran out...
+		if (c & 0xC0 != 0x80 && self.flags.precision && !prec--) break;
+		self.out(c)!;
+		index++;
+	}
+	return index;
+}
+
+fn usz! Formatter.pad(&self, char c, isz width, isz len) @inline
+{
+	isz delta = width - len;
+	for (isz i = 0; i < delta; i++) self.out(c)!;
+	return max(0, delta);
+}
+
+fn char* fmt_u(uint128 x, char* s)
+{
+	for (; x > ulong.max; x /= 10) *--s = '0' + (char)(x % 10);
+	for (ulong y = (ulong)x; y; y /= 10) *--s = '0' + (char)(y % 10);
+	return s;
+}
+
+fn usz! Formatter.out_chars(&self, char[] s)
+{
+	foreach (c : s) self.out(c)!;
+	return s.len;
+}
+
+enum FloatFormatting
+{
+	FLOAT,
+	EXPONENTIAL,
+	ADAPTIVE,
+	HEX
+}
+
+fn usz! Formatter.etoa(&self, double y) => self.floatformat(EXPONENTIAL, y);
+fn usz! Formatter.ftoa(&self, double y) => self.floatformat(FLOAT, y);
+fn usz! Formatter.gtoa(&self, double y) => self.floatformat(ADAPTIVE, y);
+fn usz! Formatter.atoa(&self, double y) => self.floatformat(HEX, y);
+
+fn usz! Formatter.floatformat(&self, FloatFormatting formatting, double y) @private
+{
+	// This code is heavily based on musl's printf code
+	const BUF_SIZE = (math::DOUBLE_MANT_DIG + 28) / 29 + 1
+		+ (math::DOUBLE_MAX_EXP + math::DOUBLE_MANT_DIG + 28 + 8) / 9;
+	uint[BUF_SIZE] big;
+	bool is_neg = false;
+	if (math::signbit(y))
+	{
+		is_neg = true;
+		y = -y;
+	}
+	isz pl = is_neg || self.flags.plus ? 1 : 0;
+	// Print inf/nan
+	if (!math::is_finite(y))
+	{
+		usz len;
+		// Add padding
+		if (!self.flags.left) len += self.pad(' ', self.width, 3 + pl)!;
+		String s = self.flags.uppercase ? "INF" : "inf";
+		if (y != y) s = self.flags.uppercase ? "NAN" : "nan";
+		len += s.len;
+		if (pl) len += self.out(is_neg ? '-' : '+')!;
+		len += self.out_chars(s)!;
+		if (self.flags.left) len += self.pad(' ', self.width, 3 + pl)!;
+		return len;
+	}
+	// Rescale
+	int e2;
+
+	y = math::frexp(y, &e2) * 2;
+	if (y) e2--;
+	char[12] ebuf0;
+	char* ebuf = 12 + (char*)&ebuf0;
+	char[9 + math::DOUBLE_MANT_DIG / 4] buf_array;
+	char* buf = &buf_array;
+	isz p = self.flags.precision ? self.prec : -1;
+	if (formatting == HEX)
+	{
+		double round = 8.0;
+		// 0x / 0X
+		pl += 2;
+		if (p > 0 && p < math::DOUBLE_MANT_DIG / 4 - 1)
+		{
+			int re = math::DOUBLE_MANT_DIG / 4 - 1 - (int)p;
+			round *= 1 << (math::DOUBLE_MANT_DIG % 4);
+			while (re--) round *= 16;
+			if (is_neg)
+			{
+				y = -y;
+				y -= round;
+				y += round;
+				y = -y;
+			}
+			else
+			{
+				y += round;
+				y -= round;
+			}
+		}
+		// Reverse print
+		char* estr = fmt_u(e2 < 0 ? (int128)-e2 : (int128)e2, ebuf);
+		if (estr == ebuf) *--estr = '0';
+		*--estr = (e2 < 0 ? '-' : '+');
+		*--estr = self.flags.uppercase ? 'P' : 'p';
+		char* s = buf;
+		char* xdigits = self.flags.uppercase ? &XDIGITS_H : &XDIGITS_L;
+		do
+		{
+			int x = (int)y;
+			*s++ = xdigits[x];
+			y = 16 * (y - x);
+			if (s - buf == 1 && (y || p > 0 || self.flags.hash)) *s++ = '.';
+		} while (y);
+		isz outlen = s - buf;
+		isz explen = ebuf - estr;
+		if (p > int.max - 2 - explen - pl) return PrintFault.INTERNAL_BUFFER_EXCEEDED?;
+		usz len;
+		usz l = p && outlen - 2 < p
+				? p + 2 + explen
+				: outlen + explen;
+		if (!self.flags.left && !self.flags.zeropad) len += self.pad(' ', self.width, pl + l)!;
+		if (is_neg || self.flags.plus) len += self.out(is_neg ? '-' : '+')!;
+		len += self.out_chars(self.flags.uppercase ? "0X" : "0x")!;
+		if (self.flags.zeropad) len += self.pad('0', self.width, pl + l)!;
+		len += self.out_chars(buf[:outlen])!;
+		len += self.pad('0', l - outlen - explen, 0)!;
+		len += self.out_chars(estr[:explen])!;
+		if (self.flags.left) len += self.pad(' ', self.width, pl + l)!;
+		return len;
+	}
+	if (p < 0) p = 6;
+	if (y)
+	{
+		y *= 0x1p28;
+		e2 -= 28;
+	}
+
+	uint* a, z, r;
+	if (e2 < 0)
+	{
+		a = r = z = &big;
+	}
+	else
+	{
+		a = r = z = (uint*)&big + big.len - math::DOUBLE_MANT_DIG - 1;
+	}
+
+	do
+	{
+		uint v = z++[0] = (uint)y;
+		y = 1000000000 * (y - v);
+	} while (y);
+
+	while (e2 > 0)
+	{
+		uint carry = 0;
+		int sh = math::min(29, e2);
+		for (uint* d = z - 1; d >= a; d--)
+		{
+			ulong x = (ulong)*d << sh + carry;
+			*d = (uint)(x % 1000000000);
+			carry = (uint)(x / 1000000000);
+		}
+		if (carry) *--a = carry;
+		while (z > a && !z[-1]) z--;
+		e2 -= sh;
+	}
+
+	while (e2 < 0)
+	{
+		uint carry = 0;
+		uint* b;
+		int sh = math::min(9, -e2);
+		int need = (int)(1 + (p + math::DOUBLE_MANT_DIG / 3u + 8) / 9);
+		for (uint* d = a; d < z; d++)
+		{
+			// CHECK THIS
+			uint rm = *d & ((1 << sh) - 1);
+			*d = (*d >> sh) + carry;
+			carry = (1000000000 >> sh) * rm;
+		}
+		if (!a[0]) a++;
+		if (carry) z++[0] = carry;
+		// Avoid (slow!) computation past requested precision
+		b = formatting == FLOAT ? r : a;
+		if (z - b > need) z = b + need;
+		e2 += sh;
+	}
+
+	int e;
+	if (a < z)
+	{
+		for (int i = 10, e = (int)(9 * (r - a)); *a >= i; i *= 10, e++);
+	}
+
+	// Perform rounding: j is precision after the radix (possibly neg)
+	int j = (int)(p - (isz)(formatting == FLOAT ? 0 : e - (int)(formatting == ADAPTIVE && p)));
+	if (j < 9 * (z - r - 1))
+	{
+		uint x;
+		// We avoid C's broken division of negative numbers
+		uint* d = r + 1 + ((j + 9 * math::DOUBLE_MAX_EXP) / 9 - math::DOUBLE_MAX_EXP);
+		j += 9 * math::DOUBLE_MAX_EXP;
+		j %= 9;
+		int i;
+		for (i = 10, j++; j < 9; i *= 10, j++);
+		x = *d % i;
+		// Are there any significant digits past j?
+		if (x || (d + 1) != z)
+		{
+			double round = 2 / math::DOUBLE_EPSILON;
+			double small;
+			if (((*d / i) & 1) || (i == 1000000000 && d > a && (d[-1] & 1)))
+			{
+				round += 2;
+			}
+			switch
+			{
+				case x < i / 2:
+					small = 0x0.8p0;
+				case x == i / 2 && d + 1 == z:
+					small = 0x1.0p0;
+				default:
+					small = 0x1.8p0;
+			}
+			if (pl && is_neg)
+			{
+				round *= -1;
+				small *= -1;
+			}
+			*d -= x;
+			// Decide whether to round by probing round+small
+			if (round + small != round)
+			{
+				*d = *d + i;
+				while (*d > 999999999)
+				{
+					*d-- = 0;
+					if (d < a) *--a = 0;
+					(*d)++;
+				}
+				for (i = 10, e = (int)(9 * (r - a)); *a >= i; i *= 10, e++);
+			}
+		}
+		if (z > d + 1) z = d + 1;
+	}
+	for (; z>a && !z[-1]; z--);
+
+	if (formatting == ADAPTIVE)
+	{
+		if (!p) p++;
+		if (p > e && e >= -4)
+		{
+			formatting = FLOAT;
+			p -= (isz)e + 1;
+		}
+		else
+		{
+			formatting = EXPONENTIAL;
+			p--;
+		}
+		if (!self.flags.hash)
+		{
+			// Count trailing zeros in last place
+			if (z > a && z[-1])
+			{
+				for (int i = 10, j = 0; z[-1] % i == 0; i *= 10, j++);
+			}
+			else
+			{
+				j = 9;
+			}
+			if (formatting == FLOAT)
+			{
+				p = math::min(p, math::max((isz)0, 9 * (z - r - 1) - j));
+			}
+			else
+			{
+				p = math::min(p, math::max((isz)0, 9 * (z - r - 1) + e - j));
+
+			}
+		}
+	}
+	if (p > int.max - 1 - (isz)(p || self.flags.hash)) return PrintFault.INTERNAL_BUFFER_EXCEEDED?;
+	int l = (int)(1 + p + (isz)(p || self.flags.hash));
+	char* estr @noinit;
+	if (formatting == FLOAT)
+	{
+		if (e > int.max - l) return PrintFault.INTERNAL_BUFFER_EXCEEDED?;
+		if (e > 0) l += e;
+	}
+	else
+	{
+		estr = fmt_u((uint128)(e < 0 ? -e : e), ebuf);
+		while (ebuf - estr < 2) (--estr)[0] = '0';
+		*--estr = (e < 0 ? '-' : '+');
+		*--estr = self.flags.uppercase ? 'E' : 'e';
+		if (ebuf - estr > (isz)int.max - l) return PrintFault.INTERNAL_BUFFER_EXCEEDED?;
+		l += (int)(ebuf - estr);
+	}
+	if (l > int.max - pl) return PrintFault.INTERNAL_BUFFER_EXCEEDED?;
+	usz len;
+	if (!self.flags.left && !self.flags.zeropad) len += self.pad(' ', self.width, pl + l)!;
+	if (is_neg || self.flags.plus) len += self.out(is_neg ? '-' : '+')!;
+	if (self.flags.zeropad) len += self.pad('0', self.width, pl + l)!;
+	if (formatting == FLOAT)
+	{
+		if (a > r) a = r;
+		uint* d = a;
+		for (; d <= r; d++)
+		{
+			char* s = fmt_u(*d, buf + 9);
+			switch
+			{
+				case d != a:
+					while (s > buf) (--s)[0] = '0';
+				case s == buf + 9:
+					*--s = '0';
+			}
+			len += self.out_chars(s[:buf + 9 - s])!;
+		}
+		if (p || self.flags.hash) len += self.out('.')!;
+		for (; d < z && p > 0; d++, p -= 9)
+		{
+			char* s = fmt_u(*d, buf + 9);
+			while (s > buf) *--s = '0';
+			len += self.out_chars(s[:math::min((isz)9, p)])!;
+		}
+		len += self.pad('0', p + 9, 9)!;
+	}
+	else
+	{
+		if (z <= a) z = a + 1;
+		for (uint* d = a; d < z && p >= 0; d++)
+		{
+			char* s = fmt_u(*d, buf + 9);
+			if (s == buf + 9) (--s)[0] = '0';
+			if (d != a)
+			{
+				while (s > buf) (--s)[0] = '0';
+			}
+			else
+			{
+				len += self.out(s++[0])!;
+				if (p > 0 || self.flags.hash) len += self.out('.')!;
+			}
+			len += self.out_chars(s[:math::min(buf + 9 - s, p)])!;
+			p -= buf + 9 - s;
+		}
+		len += self.pad('0', p + 18, 18)!;
+		len += self.out_chars(estr[:ebuf - estr])!;
+	}
+
+	if (self.flags.left) len += self.pad(' ', self.width, pl + l)!;
+
+	return len;
+}
+
+fn usz! Formatter.ntoa(&self, uint128 value, bool negative, uint base) @private
+{
+	char[PRINTF_NTOA_BUFFER_SIZE] buf @noinit;
+	usz len;
+
+	// no hash for 0 values
+	if (!value) self.flags.hash = false;
+
+	// write if precision != 0 or value is != 0
+	if (!self.flags.precision || value)
+	{
+		char past_10 = (self.flags.uppercase ? 'A' : 'a') - 10;
+		do
+		{
+			if (len >= PRINTF_NTOA_BUFFER_SIZE) return PrintFault.INTERNAL_BUFFER_EXCEEDED?;
+			char digit = (char)(value % base);
+			buf[len++] = digit + (digit < 10 ? '0' : past_10);
+			value /= base;
+		}
+		while (value);
+	}
+	return self.ntoa_format((String)buf[:PRINTF_NTOA_BUFFER_SIZE], len, negative, base);
+}
+
+fn usz! Formatter.ntoa_format(&self, String buf, usz len, bool negative, uint base) @private
+{
+	// pad leading zeros
+	if (!self.flags.left)
+	{
+		if (self.width && self.flags.zeropad && (negative || self.flags.plus || self.flags.space)) self.width--;
+		while (len < self.prec)
+		{
+			if (len >= buf.len) return PrintFault.INTERNAL_BUFFER_EXCEEDED?;
+			buf[len++] = '0';
+		}
+		while (self.flags.zeropad && len < self.width)
+		{
+			if (len >= buf.len) return PrintFault.INTERNAL_BUFFER_EXCEEDED?;
+			buf[len++] = '0';
+		}
+	}
+
+	// handle hash
+	if (self.flags.hash && base != 10)
+	{
+		if (!self.flags.precision && len && len == self.prec && len == self.width)
+		{
+			len--;
+			if (len) len--;
+		}
+		if (base != 10)
+		{
+			if (len + 1 >= buf.len) return PrintFault.INTERNAL_BUFFER_EXCEEDED?;
+			switch (base)
+			{
+				case 16:
+					buf[len++] = self.flags.uppercase ? 'X' : 'x';
+				case 8:
+					buf[len++] = self.flags.uppercase ? 'O' : 'o';
+				case 2:
+					buf[len++] = self.flags.uppercase ? 'B' : 'b';
+				default:
+					unreachable();
+			}
+			buf[len++] = '0';
+		}
+	}
+
+	switch (true)
+	{
+		case negative:
+			if (len >= buf.len) return PrintFault.INTERNAL_BUFFER_EXCEEDED?;
+			buf[len++] = '-';
+		case self.flags.plus:
+			if (len >= buf.len) return PrintFault.INTERNAL_BUFFER_EXCEEDED?;
+			buf[len++] = '+';
+		case self.flags.space:
+			if (len >= buf.len) return PrintFault.INTERNAL_BUFFER_EXCEEDED?;
+			buf[len++] = ' ';
+	}
+	if (len) self.out_reverse(buf[:len])!;
+	return len;
+}
+
+
+fn usz! Formatter.ntoa_any(&self, any arg, uint base) @private
+{
+	bool is_neg;
+	uint128 val = int_from_any(arg, &is_neg)!!;
+	return self.ntoa(val, is_neg, base) @inline;
+}
+
+fn usz! Formatter.out_char(&self, any arg) @private
+{
+	usz len = 1;
+	uint l = 1;
+	// pre padding
+	len += self.adjust(l)!;
+	// char output
+	Char32 c = types::any_to_int(arg, uint) ?? 0xFFFD;
+	switch (true)
+	{
+		case c < 0x7f:
+			self.out((char)c)!;
+		case c < 0x7ff:
+			self.out((char)(0xC0 | c >> 6))!;
+			self.out((char)(0x80 | (c & 0x3F)))!;
+		case c < 0xffff:
+			self.out((char)(0xE0 | c >> 12))!;
+			self.out((char)(0x80 | (c >> 6 & 0x3F)))!;
+			self.out((char)(0x80 | (c & 0x3F)))!;
+		default:
+			self.out((char)(0xF0 | c >> 18))!;
+			self.out((char)(0x80 | (c >> 12 & 0x3F)))!;
+			self.out((char)(0x80 | (c >> 6 & 0x3F)))!;
+			self.out((char)(0x80 | (c & 0x3F)))!;
+	}
+	len += self.adjust(l)!;
+	return len;
+}
+
+
+fn usz! Formatter.out_reverse(&self, char[] buf) @private
+{
+	usz n;
+	usz buffer_start_idx = self.idx;
+	usz len = buf.len;
+	// pad spaces up to given width
+	if (!self.flags.zeropad && !self.flags.left)
+	{
+    	n += self.pad(' ', self.width, len)!;
+	}
+	// reverse string
+	while (len) n += self.out(buf[--len])!;
+
+	// append pad spaces up to given width
+	n += self.adjust(n)!;
+	return n;
+}
+
+fn void! printf_advance_format(usz format_len, usz *index_ptr) @inline @private
+{
+	usz val = ++(*index_ptr);
+	if (val >= format_len) return FormattingFault.UNTERMINATED_FORMAT?;
+}
+
+fn any! next_any(any* args_ptr, usz args_len, usz* arg_index_ptr) @inline @private
+{
+	if (*arg_index_ptr >= args_len) return FormattingFault.MISSING_ARG?;
+	return args_ptr[(*arg_index_ptr)++];
+}
+
+fn int! printf_parse_format_field(
+	any* args_ptr, usz args_len, usz* args_index_ptr,
+	char* format_ptr, usz format_len, usz* index_ptr) @inline @private
+{
+	char c = format_ptr[*index_ptr];
+	if (c.is_digit()) return simple_atoi(format_ptr, format_len, index_ptr);
+	if (c != '*') return 0;
+	printf_advance_format(format_len, index_ptr)!;
+	any val = next_any(args_ptr, args_len, args_index_ptr)!;
+	if (!val.type.kindof.is_int()) return FormattingFault.INVALID_WIDTH_ARG?;
+	uint! intval = types::any_to_int(val, int);
+	return intval ?? FormattingFault.INVALID_WIDTH_ARG?;
+}

lib/std/io/io.c3

@@ -0,0 +1,403 @@
+// Copyright (c) 2021-2022 Christoffer Lerno. All rights reserved.
+// Use of this source code is governed by the MIT license
+// a copy of which can be found in the LICENSE_STDLIB file.
+module std::io;
+import libc;
+
+enum Seek
+{
+	SET,
+	CURSOR,
+	END
+}
+
+fault IoError
+{
+	ALREADY_EXISTS,
+	BUSY,
+	CANNOT_READ_DIR,
+	DIR_NOT_EMPTY,
+	EOF,
+	FILE_CANNOT_DELETE,
+	FILE_IS_DIR,
+	FILE_IS_PIPE,
+	FILE_NOT_DIR,
+	FILE_NOT_FOUND,
+	FILE_NOT_VALID,
+	GENERAL_ERROR,
+	ILLEGAL_ARGUMENT,
+	INCOMPLETE_WRITE,
+	INTERRUPTED,
+	INVALID_POSITION,
+	INVALID_PUSHBACK,
+	NAME_TOO_LONG,
+	NOT_SEEKABLE,
+	NO_PERMISSION,
+	OUT_OF_SPACE,
+	OVERFLOW,
+	READ_ONLY,
+	SYMLINK_FAILED,
+	TOO_MANY_DESCRIPTORS,
+	UNEXPECTED_EOF,
+	UNKNOWN_ERROR,
+	UNSUPPORTED_OPERATION,
+	WOULD_BLOCK,
+}
+
+
+/**
+ * Read from a stream (default is stdin) to the next "\n"
+ * or to the end of the stream, whatever comes first.
+ * "\r" will be filtered from the String.
+ *
+ * @param stream `The stream to read from.`
+ * @require @is_instream(stream) `The stream must implement InStream.`
+ * @param [inout] allocator `the allocator to use.`
+ * @return `The string containing the data read.`
+ **/
+macro String! readline(stream = io::stdin(), Allocator allocator = allocator::heap())
+{
+	bool $is_stream = @typeis(stream, InStream);
+	$if $is_stream:
+		$typeof(&stream.read_byte) func = &stream.read_byte;
+		char val = func((void*)stream)!;
+	$else
+		char val = stream.read_byte()!;
+	$endif
+
+    if (val == '\n') return "";
+    @pool(allocator)
+    {
+        DString str = dstring::temp_with_capacity(256);
+    	if (val != '\r') str.append(val);
+    	while (1)
+    	{
+    	    $if $is_stream:
+    	        char! c = func((void*)stream);
+    	    $else
+    	        char! c = stream.read_byte();
+    	    $endif
+    		if (catch err = c)
+    		{
+    			if (err == IoError.EOF) break;
+    			return err?;
+    		}
+    		if (c == '\r') continue;
+    		if (c == '\n') break;
+    		str.append_char(c);
+    	}
+    	return str.copy_str(allocator);
+    };
+}
+
+/**
+ * Reads a string, see `readline`, except the it is allocated
+ * on the temporary allocator and does not need to be freed.
+ *
+ * @param stream `The stream to read from.`
+ * @require @is_instream(stream) `The stream must implement InStream.`
+ * @return `The temporary string containing the data read.`
+ **/
+macro String! treadline(stream = io::stdin())
+{
+	return readline(stream, allocator::temp()) @inline;
+}
+
+/**
+ * Print a value to a stream.
+ *
+ * @param out `the stream to print to`
+ * @param x `the value to print`
+ * @require @is_outstream(out) `The output must implement OutStream.`
+ * @return `the number of bytes printed.`
+ */
+macro usz! fprint(out, x)
+{
+	var $Type = $typeof(x);
+	$switch ($Type)
+		$case String: return out.write(x);
+		$case ZString: return out.write(x.str_view());
+		$case DString: return out.write(x.str_view());
+		$default:
+			$if $assignable(x, String):
+				return out.write((String)x);
+			$else
+				return fprintf(out, "%s", x);
+			$endif
+	$endswitch
+}
+
+/**
+ * Prints using a 'printf'-style formatting string.
+ * See `printf` for details on formatting.
+ *
+ * @param [inout] out `The OutStream to print to`
+ * @param [in] format `The printf-style format string`
+ * @return `the number of characters printed`
+ **/
+fn usz! fprintf(OutStream out, String format, args...)
+{
+	Formatter formatter;
+	formatter.init(&out_putstream_fn, &out);
+	return formatter.vprintf(format, args);
+}
+
+/**
+ * Prints using a 'printf'-style formatting string,
+ * appending '\n' at the end. See `printf`.
+ *
+ * @param [inout] out `The OutStream to print to`
+ * @param [in] format `The printf-style format string`
+ * @return `the number of characters printed`
+ **/
+fn usz! fprintfn(OutStream out, String format, args...) @maydiscard
+{
+	Formatter formatter;
+	formatter.init(&out_putstream_fn, &out);
+	usz len = formatter.vprintf(format, args)!;
+	out.write_byte('\n')!;
+	if (&out.flush) out.flush()!;
+	return len + 1;
+}
+
+/**
+ * @require @is_outstream(out) "The output must implement OutStream"
+ */
+macro usz! fprintn(out, x = "")
+{
+	usz len = fprint(out, x)!;
+	out.write_byte('\n')!;
+	$switch
+		$case @typeid(out) == OutStream.typeid:
+			if (&out.flush) out.flush()!;
+		$case $defined(out.flush):
+			out.flush()!;
+	$endswitch
+	return len + 1;
+}
+
+/**
+ * Print any value to stdout.
+ **/
+macro void print(x)
+{
+	(void)fprint(io::stdout(), x);
+}
+
+/**
+ * Print any value to stdout, appending an '\n’ after.
+ *
+ * @param x "The value to print"
+ **/
+macro void printn(x = "")
+{
+	(void)fprintn(io::stdout(), x);
+}
+
+/**
+ * Print any value to stderr.
+ **/
+macro void eprint(x)
+{
+	(void)fprint(io::stderr(), x);
+}
+
+/**
+ * Print any value to stderr, appending an '\n’ after.
+ *
+ * @param x "The value to print"
+ **/
+macro void eprintn(x)
+{
+	(void)fprintn(io::stderr(), x);
+}
+
+
+fn void! out_putstream_fn(void* data, char c) @private
+{
+	OutStream* stream = data;
+	return (*stream).write_byte(c);
+}
+
+fn void! out_putchar_fn(void* data @unused, char c) @private
+{
+	libc::putchar(c);
+}
+
+/**
+ * Prints using a 'printf'-style formatting string.
+ * To print integer numbers, use "%d" or "%x"/"%X,
+ * the latter gives the hexadecimal representation.
+ *
+ * All types can be printed using "%s" which gives
+ * the default representation of the value.
+ *
+ * To create a custom output for a type, implement
+ * the Printable interface.
+ *
+ * @param [in] format `The printf-style format string`
+ * @return `the number of characters printed`
+ **/
+fn usz! printf(String format, args...) @maydiscard
+{
+	Formatter formatter;
+	formatter.init(&out_putchar_fn);
+	return formatter.vprintf(format, args);
+}
+
+/**
+ * Prints using a 'printf'-style formatting string,
+ * appending '\n' at the end. See `printf`.
+ *
+ * @param [in] format `The printf-style format string`
+ * @return `the number of characters printed`
+ **/
+fn usz! printfn(String format, args...) @maydiscard
+{
+	Formatter formatter;
+	formatter.init(&out_putchar_fn);
+	usz len = formatter.vprintf(format, args)!;
+	putchar('\n');
+	io::stdout().flush()!;
+	return len + 1;
+}
+
+/**
+ * Prints using a 'printf'-style formatting string
+ * to stderr.
+ *
+ * @param [in] format `The printf-style format string`
+ * @return `the number of characters printed`
+ **/
+fn usz! eprintf(String format, args...) @maydiscard
+{
+	Formatter formatter;
+	OutStream stream = stderr();
+	formatter.init(&out_putstream_fn, &stream);
+	return formatter.vprintf(format, args);
+}
+
+
+/**
+ * Prints using a 'printf'-style formatting string,
+ * to stderr appending '\n' at the end. See `printf`.
+ *
+ * @param [in] format `The printf-style format string`
+ * @return `the number of characters printed`
+ **/
+fn usz! eprintfn(String format, args...) @maydiscard
+{
+	Formatter formatter;
+	OutStream stream = stderr();
+	formatter.init(&out_putstream_fn, &stream);
+	usz len = formatter.vprintf(format, args)! + 1;
+	stderr().write_byte('\n')!;
+	stderr().flush()!;
+	return len;
+}
+
+/**
+ * Prints using a 'printf'-style formatting string,
+ * to a string buffer. See `printf`.
+ *
+ * @param [inout] buffer `The buffer to print to`
+ * @param [in] format `The printf-style format string`
+ * @return `a slice formed from the "buffer" with the resulting length.`
+ **/
+fn char[]! bprintf(char[] buffer, String format, args...) @maydiscard
+{
+	Formatter formatter;
+	BufferData data = { .buffer = buffer };
+	formatter.init(&out_buffer_fn, &data);
+	usz size = formatter.vprintf(format, args)!;
+	return buffer[:data.written];
+}
+
+// Used to print to a buffer.
+fn void! out_buffer_fn(void *data, char c) @private
+{
+	BufferData *buffer_data = data;
+	if (buffer_data.written >= buffer_data.buffer.len) return PrintFault.BUFFER_EXCEEDED?;
+	buffer_data.buffer[buffer_data.written++] = c;
+}
+
+// Used for buffer printing
+struct BufferData @private
+{
+	char[] buffer;
+	usz written;
+}
+
+// Only available with LIBC
+module std::io @if (env::LIBC);
+import libc;
+
+/**
+ * Libc `putchar`, prints a single character to stdout.
+ **/
+fn void putchar(char c) @inline
+{
+	libc::putchar(c);
+}
+
+/**
+ * Get standard out.
+ *
+ * @return `stdout as a File`
+ **/
+fn File* stdout()
+{
+	static File file;
+	if (!file.file) file = file::from_handle(libc::stdout());
+	return &file;
+}
+
+/**
+ * Get standard err.
+ *
+ * @return `stderr as a File`
+ **/
+fn File* stderr()
+{
+	static File file;
+	if (!file.file) file = file::from_handle(libc::stderr());
+	return &file;
+}
+
+/**
+ * Get standard in.
+ *
+ * @return `stdin as a File`
+ **/
+fn File* stdin()
+{
+	static File file;
+	if (!file.file) file = file::from_handle(libc::stdin());
+	return &file;
+}
+
+module std::io @if(!env::LIBC);
+
+File stdin_file;
+File stdout_file;
+File stderr_file;
+
+fn void putchar(char c) @inline
+{
+	(void)stdout_file.write_byte(c);
+}
+
+fn File* stdout()
+{
+	return &stdout_file;
+}
+
+fn File* stderr()
+{
+	return &stderr_file;
+}
+
+fn File* stdin()
+{
+	return &stdin_file;
+}

lib/std/io/os/chdir.c3

@@ -0,0 +1,30 @@
+module std::io::os;
+import std::io::path, libc, std::os;
+
+macro void! native_chdir(Path path)
+{
+	$switch
+		$case env::POSIX:
+			if (posix::chdir(path.as_zstr()))
+			{
+				switch (libc::errno())
+				{
+					case errno::EACCES: return IoError.NO_PERMISSION?;
+					case errno::ENAMETOOLONG: return IoError.NAME_TOO_LONG?;
+					case errno::ENOTDIR: return IoError.FILE_NOT_DIR?;
+					case errno::ENOENT: return IoError.FILE_NOT_FOUND?;
+					case errno::ELOOP: return IoError.SYMLINK_FAILED?;
+					default: return IoError.GENERAL_ERROR?;
+				}
+			}
+		$case env::WIN32:
+			@pool()
+			{
+				// TODO improve with better error handling.
+				if (win32::setCurrentDirectoryW(path.str_view().to_temp_utf16()!!)) return;
+			};
+			return IoError.GENERAL_ERROR?;
+		$default:
+			return IoError.UNSUPPORTED_OPERATION?;
+	$endswitch
+}

lib/std/io/os/file_libc.c3

@@ -0,0 +1,129 @@
+module std::io::os @if(env::LIBC);
+import libc;
+
+/**
+ * @require mode.len > 0
+ * @require filename.len > 0
+ **/
+fn void*! native_fopen(String filename, String mode) @inline
+{
+	@pool()
+	{
+		$if env::WIN32:
+			void* file = libc::_wfopen(filename.to_temp_wstring(), mode.to_temp_wstring())!;
+		$else
+			void* file = libc::fopen(filename.zstr_tcopy(), mode.zstr_tcopy());
+		$endif
+			return file ?: file_open_errno()?;
+	};
+}
+
+fn void! native_remove(String filename)
+{
+	@pool()
+	{
+		$if env::WIN32:
+			CInt result = libc::_wremove(filename.to_temp_wstring())!;
+		$else
+			CInt result = libc::remove(filename.zstr_tcopy());
+		$endif
+		if (result)
+		{
+			switch (libc::errno())
+			{
+				case errno::ENOENT:
+					return IoError.FILE_NOT_FOUND?;
+				case errno::EACCES:
+				default:
+					return IoError.FILE_CANNOT_DELETE?;
+			}
+		}
+	};
+}
+
+/**
+ * @require mode.len > 0
+ * @require filename.len > 0
+ **/
+fn void*! native_freopen(void* file, String filename, String mode) @inline
+{
+	@pool()
+	{
+		$if env::WIN32:
+			file = libc::_wfreopen(filename.to_temp_wstring(), mode.to_temp_wstring(), file)!;
+		$else
+			file = libc::freopen(filename.zstr_tcopy(), mode.zstr_tcopy(), file);
+		$endif
+			return file ?: file_open_errno()?;
+	};
+}
+
+fn void! native_fseek(void* file, isz offset, Seek seek_mode) @inline
+{
+	if (libc::fseek(file, (SeekIndex)offset, (CInt)seek_mode)) return file_seek_errno()?;
+}
+
+
+fn usz! native_ftell(CFile file) @inline
+{
+	long index = libc::ftell(file);
+	return index >= 0 ? (usz)index : file_seek_errno()?;
+}
+
+fn usz! native_fwrite(CFile file, char[] buffer) @inline
+{
+	return libc::fwrite(buffer.ptr, 1, buffer.len, file);
+}
+
+fn usz! native_fread(CFile file, char[] buffer) @inline
+{
+	return libc::fread(buffer.ptr, 1, buffer.len, file);
+}
+
+macro anyfault file_open_errno() @local
+{
+	switch (libc::errno())
+	{
+		case errno::EACCES: return IoError.NO_PERMISSION;
+		case errno::EDQUOT: return IoError.OUT_OF_SPACE;
+		case errno::EBADF: return IoError.FILE_NOT_VALID;
+		case errno::EEXIST: return IoError.ALREADY_EXISTS;
+		case errno::EINTR: return IoError.INTERRUPTED;
+		case errno::EFAULT: return IoError.GENERAL_ERROR;
+		case errno::EISDIR: return IoError.FILE_IS_DIR;
+		case errno::ELOOP: return IoError.SYMLINK_FAILED;
+		case errno::EMFILE: return IoError.TOO_MANY_DESCRIPTORS;
+		case errno::ENAMETOOLONG: return IoError.NAME_TOO_LONG;
+		case errno::ENFILE: return IoError.OUT_OF_SPACE;
+		case errno::ENOTDIR: return IoError.FILE_NOT_DIR;
+		case errno::ENOENT: return IoError.FILE_NOT_FOUND;
+		case errno::ENOSPC: return IoError.OUT_OF_SPACE;
+		case errno::ENXIO: return IoError.FILE_NOT_FOUND;
+		case errno::EOVERFLOW: return IoError.OVERFLOW;
+		case errno::EROFS: return IoError.READ_ONLY;
+		case errno::EOPNOTSUPP: return IoError.UNSUPPORTED_OPERATION;
+		case errno::EIO: return IoError.INCOMPLETE_WRITE;
+		case errno::EWOULDBLOCK: return IoError.WOULD_BLOCK;
+		default: return IoError.UNKNOWN_ERROR;
+	}
+}
+
+macro anyfault file_seek_errno() @local
+{
+	switch (libc::errno())
+	{
+		case errno::ESPIPE: return IoError.FILE_IS_PIPE;
+		case errno::EPIPE: return IoError.FILE_IS_PIPE;
+		case errno::EOVERFLOW: return IoError.OVERFLOW;
+		case errno::ENXIO: return IoError.FILE_NOT_FOUND;
+		case errno::ENOSPC: return IoError.OUT_OF_SPACE;
+		case errno::EIO: return IoError.INCOMPLETE_WRITE;
+		case errno::EINVAL: return IoError.INVALID_POSITION;
+		case errno::EINTR: return IoError.INTERRUPTED;
+		case errno::EFBIG: return IoError.OUT_OF_SPACE;
+		case errno::EBADF: return IoError.FILE_NOT_VALID;
+		case errno::EAGAIN: return IoError.WOULD_BLOCK;
+		default: return IoError.UNKNOWN_ERROR;
+	}
+}
+

lib/std/io/os/file_nolibc.c3

@@ -0,0 +1,75 @@
+module std::io::os @if(env::NO_LIBC);
+import libc;
+
+def FopenFn = fn void*!(String, String);
+def FreopenFn = fn void*!(void*, String, String);
+def FcloseFn = fn void!(void*);
+def FseekFn = fn void!(void*, isz, Seek);
+def FtellFn = fn usz!(void*);
+def FwriteFn = fn usz!(void*, char[] buffer);
+def FreadFn = fn usz!(void*, char[] buffer);
+def RemoveFn = fn void!(String);
+
+FopenFn native_fopen_fn @weak @if(!$defined(native_fopen_fn));
+FcloseFn native_fclose_fn @weak @if(!$defined(native_fclose_fn));
+FreopenFn native_freopen_fn @weak @if(!$defined(native_freopen_fn));
+FseekFn native_fseek_fn @weak @if(!$defined(native_fseek_fn));
+FtellFn native_ftell_fn @weak @if(!$defined(native_ftell_fn));
+FwriteFn native_fwrite_fn @weak @if(!$defined(native_fwrite_fn));
+FreadFn native_fread_fn @weak @if(!$defined(native_fread_fn));
+RemoveFn native_remove_fn @weak @if(!$defined(native_remove_fn));
+
+/**
+ * @require mode.len > 0
+ * @require filename.len > 0
+ **/
+fn void*! native_fopen(String filename, String mode) @inline
+{
+	if (native_fopen_fn) return native_fopen_fn(filename, mode);
+	return IoError.UNSUPPORTED_OPERATION?;
+}
+
+/**
+ * Delete a file.
+ *
+ * @require filename.len > 0
+ **/
+fn void! native_remove(String filename) @inline
+{
+	if (native_remove_fn) return native_remove_fn(filename);
+	return IoError.UNSUPPORTED_OPERATION?;
+}
+
+/**
+ * @require mode.len > 0
+ * @require filename.len > 0
+ **/
+fn void*! native_freopen(void* file, String filename, String mode) @inline
+{
+	if (native_freopen_fn) return native_freopen_fn(file, filename, mode);
+	return IoError.UNSUPPORTED_OPERATION?;
+}
+
+fn void! native_fseek(void* file, isz offset, Seek seek_mode) @inline
+{
+	if (native_fseek_fn) return native_fseek_fn(file, offset, seek_mode);
+	return IoError.UNSUPPORTED_OPERATION?;
+}
+
+fn usz! native_ftell(CFile file) @inline
+{
+	if (native_ftell_fn) return native_ftell_fn(file);
+	return IoError.UNSUPPORTED_OPERATION?;
+}
+
+fn usz! native_fwrite(CFile file, char[] buffer) @inline
+{
+	if (native_fwrite_fn) return native_fwrite_fn(file, buffer);
+	return IoError.UNSUPPORTED_OPERATION?;
+}
+
+fn usz! native_fread(CFile file, char[] buffer) @inline
+{
+	if (native_fread_fn) return native_fread_fn(file, buffer);
+	return IoError.UNSUPPORTED_OPERATION?;
+}

lib/std/io/os/fileinfo.c3

@@ -0,0 +1,116 @@
+module std::io::os;
+import libc, std::os, std::io;
+
+fn void! native_stat(Stat* stat, String path) @if(env::DARWIN || env::LINUX)
+{
+	@pool()
+	{
+		$if env::DARWIN || env::LINUX:
+			int res = libc::stat(path.zstr_tcopy(), stat);
+		$else
+			unreachable("Stat unimplemented");
+			int res = 0;
+		$endif
+		if (res != 0)
+		{
+			switch (libc::errno())
+			{
+				case errno::EBADF:
+					return IoError.FILE_NOT_VALID?;
+				case errno::EFAULT:
+					unreachable("Invalid stat");
+				case errno::EIO:
+					return IoError.GENERAL_ERROR?;
+				case errno::EACCES:
+					return IoError.NO_PERMISSION?;
+				case errno::ELOOP:
+					return IoError.NO_PERMISSION?;
+				case errno::ENAMETOOLONG:
+					return IoError.NAME_TOO_LONG?;
+				case errno::ENOENT:
+					return IoError.FILE_NOT_FOUND?;
+				case errno::ENOTDIR:
+					return IoError.FILE_NOT_DIR?;
+				case errno::EOVERFLOW:
+					return IoError.GENERAL_ERROR?;
+				default:
+					return IoError.UNKNOWN_ERROR?;
+			}
+		}
+	};
+}
+
+fn usz! native_file_size(String path) @if(env::WIN32)
+{
+	@pool()
+	{
+		Win32_FILE_ATTRIBUTE_DATA data;
+		win32::getFileAttributesExW(path.to_temp_wstring()!, Win32_GET_FILEEX_INFO_LEVELS.STANDARD, &data);
+		Win32_LARGE_INTEGER size;
+		size.lowPart = data.nFileSizeLow;
+		size.highPart = data.nFileSizeHigh;
+		return (usz)size.quadPart;
+	};
+}
+
+fn usz! native_file_size(String path) @if(!env::WIN32 && !env::DARWIN)
+{
+	File f = file::open(path, "r")!;
+	defer (void)f.close();
+	return f.seek(0, Seek.END)!;
+}
+
+fn usz! native_file_size(String path) @if(env::DARWIN)
+{
+	Stat stat;
+	native_stat(&stat, path)!;
+	return stat.st_size;
+}
+
+fn bool native_file_or_dir_exists(String path)
+{
+	$switch
+		$case env::DARWIN:
+		$case env::LINUX:
+			Stat stat;
+			return @ok(native_stat(&stat, path));
+		$case env::WIN32:
+			@pool()
+			{
+				return (bool)win32::pathFileExistsW(path.to_temp_utf16()) ?? false;
+			};
+		$case env::POSIX:
+			@pool()
+			{
+				return posix::access(path.zstr_tcopy(), 0 /* F_OK */) != -1;
+			};
+		$default:
+			unreachable("Not supported");
+	$endswitch
+}
+
+fn bool native_is_file(String path)
+{
+	$switch
+		$case env::DARWIN:
+		$case env::LINUX:
+			Stat stat;
+			return @ok(native_stat(&stat, path)) && libc_S_ISTYPE(stat.st_mode, libc::S_IFREG);
+		$default:
+			File! f = file::open(path, "r");
+			defer (void)f.close();
+			return @ok(f);
+	$endswitch
+}
+
+fn bool native_is_dir(String path)
+{
+	$if env::DARWIN || env::LINUX:
+		Stat stat;
+		return @ok(native_stat(&stat, path)) && libc_S_ISTYPE(stat.st_mode, libc::S_IFDIR);
+	$else
+		return native_file_or_dir_exists(path) && !native_is_file(path);
+	$endif
+}
+
+

lib/std/io/os/getcwd.c3

@@ -0,0 +1,41 @@
+module std::io::os;
+import libc, std::os;
+
+macro String! getcwd(Allocator allocator = allocator::heap())
+{
+	$switch
+		$case env::WIN32:
+			const DEFAULT_BUFFER = 256;
+			Char16[DEFAULT_BUFFER] buffer;
+			WString res = win32::_wgetcwd(&buffer, DEFAULT_BUFFER);
+			bool free = false;
+			defer if (free) libc::free(res);
+			if (!res)
+			{
+				if (libc::errno() != errno::ERANGE) return IoError.GENERAL_ERROR?;
+				res = win32::_wgetcwd(null, 0);
+				free = true;
+			}
+			Char16[] str16 = res[:win32::wcslen(res)];
+			return string::new_from_utf16(str16, allocator);
+
+		$case env::POSIX:
+			const usz DEFAULT_BUFFER = 256;
+			char[DEFAULT_BUFFER] buffer;
+			ZString res = posix::getcwd(&buffer, DEFAULT_BUFFER);
+			bool free = false;
+			if (!res)
+			{
+				// Improve error
+				if (libc::errno() != errno::ERANGE) return IoError.GENERAL_ERROR?;
+				res = posix::getcwd(null, 0);
+				free = true;
+			}
+			defer if (free) libc::free((void*)res);
+			return res.copy(allocator);
+
+		$default:
+			return IoError.UNSUPPORTED_OPERATION?;
+	$endswitch
+}
+

lib/std/io/os/ls.c3

@@ -0,0 +1,51 @@
+module std::io::os @if(env::POSIX);
+import std::io, std::os;
+
+fn PathList! native_ls(Path dir, bool no_dirs, bool no_symlinks, String mask, Allocator allocator)
+{
+	PathList list;
+	list.new_init(.allocator = allocator);
+	DIRPtr directory = posix::opendir(dir.str_view() ? dir.as_zstr() : (ZString)".");
+	defer if (directory) posix::closedir(directory);
+	if (!directory) return (path::is_dir(dir) ? IoError.CANNOT_READ_DIR : IoError.FILE_NOT_DIR)?;
+	Posix_dirent* entry;
+	while ((entry = posix::readdir(directory)))
+	{
+		String name = ((ZString)&entry.name).str_view();
+		if (!name || name == "." || name == "..") continue;
+		if (entry.d_type == posix::DT_LNK && no_symlinks) continue;
+		if (entry.d_type == posix::DT_DIR && no_dirs) continue;
+		Path path = path::new(name, allocator)!!;
+		list.push(path);
+	}
+	return list;
+}
+
+module std::io::os @if(env::WIN32);
+import std::time, std::os, std::io;
+
+fn PathList! native_ls(Path dir, bool no_dirs, bool no_symlinks, String mask, Allocator allocator)
+{
+	PathList list;
+	list.new_init(.allocator = allocator);
+
+	@pool(allocator)
+	{
+		WString result = dir.str_view().tconcat(`\*`).to_temp_wstring()!!;
+		Win32_WIN32_FIND_DATAW find_data;
+		Win32_HANDLE find = win32::findFirstFileW(result, &find_data);
+		if (find == win32::INVALID_HANDLE_VALUE) return IoError.CANNOT_READ_DIR?;
+		defer win32::findClose(find);
+		do
+		{
+			if (no_dirs && (find_data.dwFileAttributes & win32::FILE_ATTRIBUTE_DIRECTORY)) continue;
+			@pool(allocator)
+			{
+				String filename = string::temp_from_wstring((WString)&find_data.cFileName)!;
+				if (filename == ".." || filename == ".") continue;
+				list.push(path::new(filename, allocator)!);
+			};
+		} while (win32::findNextFileW(find, &find_data));
+		return list;
+	};
+}

lib/std/io/os/mkdir.c3

@@ -0,0 +1,50 @@
+module std::io::os;
+import libc;
+import std::io::path;
+import std::os::win32;
+import std::os::posix;
+
+
+macro bool! native_mkdir(Path path, MkdirPermissions permissions)
+{
+	$switch
+		$case env::POSIX:
+			if (!posix::mkdir(path.as_zstr(), permissions == NORMAL ? 0o777 : 0o700)) return true;
+			switch (libc::errno())
+			{
+				case errno::EACCES:
+				case errno::EPERM:
+				case errno::EROFS:
+				case errno::EFAULT: return IoError.NO_PERMISSION?;
+				case errno::ENAMETOOLONG: return IoError.NAME_TOO_LONG?;
+				case errno::EDQUOT:
+				case errno::ENOSPC: return IoError.OUT_OF_SPACE?;
+				case errno::EISDIR:
+				case errno::EEXIST: return false;
+				case errno::ELOOP: return IoError.SYMLINK_FAILED?;
+				case errno::ENOTDIR: return IoError.FILE_NOT_FOUND?;
+				default: return IoError.GENERAL_ERROR?;
+			}
+		$case env::WIN32:
+			@pool()
+			{
+				// TODO security attributes
+				if (win32::createDirectoryW(path.str_view().to_temp_utf16()!!, null)) return true;
+				switch (win32::getLastError())
+				{
+					case win32::ERROR_ACCESS_DENIED:
+						return IoError.NO_PERMISSION?;
+					case win32::ERROR_DISK_FULL:
+						return IoError.OUT_OF_SPACE?;
+					case win32::ERROR_ALREADY_EXISTS:
+						return false;
+					case win32::ERROR_PATH_NOT_FOUND:
+						return IoError.FILE_NOT_FOUND?;
+					default:
+						return IoError.GENERAL_ERROR?;
+				}
+			};
+		$default:
+			return IoError.UNSUPPORTED_OPERATION?;
+	$endswitch
+}

lib/std/io/os/rmdir.c3

@@ -0,0 +1,48 @@
+module std::io::os;
+import libc;
+import std::io::path;
+import std::os::win32;
+import std::os::posix;
+
+macro bool! native_rmdir(Path path)
+{
+	$switch
+		$case env::POSIX:
+			if (!posix::rmdir(path.as_zstr())) return true;
+			switch (libc::errno())
+			{
+				case errno::EBUSY: return IoError.BUSY?;
+				case errno::EACCES:
+				case errno::EPERM:
+				case errno::EROFS:
+				case errno::EFAULT: return IoError.NO_PERMISSION?;
+				case errno::ENAMETOOLONG: return IoError.NAME_TOO_LONG?;
+				case errno::ENOTDIR:
+				case errno::ENOENT: return false;
+				case errno::ENOTEMPTY: return IoError.DIR_NOT_EMPTY?;
+				case errno::ELOOP: return IoError.SYMLINK_FAILED?;
+				default: return IoError.GENERAL_ERROR?;
+			}
+		$case env::WIN32:
+			@pool()
+			{
+				if (win32::removeDirectoryW(path.str_view().to_temp_utf16()!!)) return true;
+				switch (win32::getLastError())
+				{
+					case win32::ERROR_ACCESS_DENIED:
+						return IoError.NO_PERMISSION?;
+					case win32::ERROR_CURRENT_DIRECTORY:
+						return IoError.BUSY?;
+					case win32::ERROR_DIR_NOT_EMPTY:
+						return IoError.DIR_NOT_EMPTY?;
+					case win32::ERROR_DIRECTORY:
+					case win32::ERROR_PATH_NOT_FOUND:
+						return false;
+					default:
+						return IoError.GENERAL_ERROR?;
+				}
+			};
+		$default:
+			return IoError.UNSUPPORTED_OPERATION?;
+	$endswitch
+}

lib/std/io/os/rmtree.c3

@@ -0,0 +1,64 @@
+module std::io::os @if(env::POSIX);
+import std::io, std::os, libc;
+
+/**
+ * @require dir.str_view()
+ **/
+fn void! native_rmtree(Path dir)
+{
+	DIRPtr directory = posix::opendir(dir.as_zstr());
+	defer if (directory) posix::closedir(directory);
+	if (!directory) return path::is_dir(dir) ? IoError.CANNOT_READ_DIR? : IoError.FILE_NOT_DIR?;
+	Posix_dirent* entry;
+	while ((entry = posix::readdir(directory)))
+	{
+		@pool()
+		{
+			String name = ((ZString)&entry.name).str_view();
+			if (!name || name == "." || name == "..") continue;
+			Path new_path = dir.temp_append(name)!;
+			if (entry.d_type == posix::DT_DIR)
+			{
+				native_rmtree(new_path)!;
+				continue;
+			}
+			if (libc::remove(new_path.as_zstr()))
+			{
+				// TODO improve
+				return IoError.GENERAL_ERROR?;
+			}
+		};
+	}
+	os::native_rmdir(dir)!;
+}
+
+module std::io::os @if(env::WIN32);
+import std::io, std::time, std::os;
+
+fn void! native_rmtree(Path path)
+{
+	Win32_WIN32_FIND_DATAW find_data;
+	String s = path.str_view().tconcat("\\*");
+	Win32_HANDLE find = win32::findFirstFileW(s.to_temp_utf16(), &find_data)!;
+
+	if (find == win32::INVALID_HANDLE_VALUE) return IoError.CANNOT_READ_DIR?;
+	defer win32::findClose(find);
+	do
+	{
+		@pool()
+		{
+			String filename = string::new_from_wstring((WString)&find_data.cFileName, allocator::temp())!;
+			if (filename == "." || filename == "..") continue;
+			Path file_path = path.temp_append(filename)!;
+			if (find_data.dwFileAttributes & win32::FILE_ATTRIBUTE_DIRECTORY)
+			{
+				native_rmtree(file_path)!;
+			}
+			else
+			{
+				win32::deleteFileW(file_path.str_view().to_temp_wstring()!!);
+			}
+		};
+	} while (win32::findNextFileW(find, &find_data) != 0);
+	os::native_rmdir(path)!;
+}

lib/std/io/os/temp_directory.c3

@@ -0,0 +1,32 @@
+module std::io::os @if(env::LIBC);
+import std::io::path, std::os;
+
+fn Path! native_temp_directory(Allocator allocator = allocator::heap()) @if(!env::WIN32)
+{
+	foreach (String env : { "TMPDIR", "TMP", "TEMP", "TEMPDIR" })
+	{
+		String tmpdir = env::get_var(env) ?? "";
+		if (tmpdir) return path::new(tmpdir, allocator);
+	}
+	return path::new("/tmp", allocator);
+}
+
+fn Path! native_temp_directory(Allocator allocator = allocator::heap()) @if(env::WIN32)
+{
+	@pool(allocator)
+	{
+		Win32_DWORD len = win32::getTempPathW(0, null);
+		if (!len) return IoError.GENERAL_ERROR?;
+		Char16[] buff = mem::temp_alloc_array(Char16, len + (usz)1);
+		if (!win32::getTempPathW(len, buff)) return IoError.GENERAL_ERROR?;
+		return path::new(string::temp_from_utf16(buff[:len]), allocator);
+	};
+}
+
+module std::io::os @if(env::NO_LIBC);
+import std::io::path;
+
+macro Path! native_temp_directory(Allocator allocator = allocator::heap())
+{
+	return IoError.UNSUPPORTED_OPERATION?;
+}

lib/std/io/path.c3

@@ -0,0 +1,595 @@
+module std::io::path;
+import std::collections::list, std::io::os;
+import std::os::win32;
+
+const PathEnv DEFAULT_PATH_ENV = env::WIN32 ? PathEnv.WIN32 : PathEnv.POSIX;
+const char PREFERRED_SEPARATOR_WIN32 = '\\';
+const char PREFERRED_SEPARATOR_POSIX = '/';
+const char PREFERRED_SEPARATOR = env::WIN32 ? PREFERRED_SEPARATOR_WIN32 : PREFERRED_SEPARATOR_POSIX;
+
+def PathList = List(<Path>);
+
+fault PathResult
+{
+	INVALID_PATH,
+	NO_PARENT,
+}
+
+def Path = PathImp;
+
+struct PathImp (Printable)
+{
+	String path_string;
+	PathEnv env;
+}
+
+enum PathEnv
+{
+	WIN32,
+	POSIX
+}
+
+fn Path! new_cwd(Allocator allocator = allocator::heap())
+{
+	@pool(allocator)
+	{
+		return new(os::getcwd(allocator::temp()), allocator);
+	};
+}
+
+fn Path! getcwd(Allocator allocator = allocator::heap()) @deprecated("Use new_cwd()")
+{
+	@pool(allocator)
+	{
+		return new(os::getcwd(allocator::temp()), allocator);
+	};
+}
+
+fn bool is_dir(Path path) => os::native_is_dir(path.str_view());
+fn bool is_file(Path path) => os::native_is_file(path.str_view());
+fn usz! file_size(Path path) => os::native_file_size(path.str_view());
+fn bool exists(Path path) => os::native_file_or_dir_exists(path.str_view());
+fn Path! temp_cwd() => new_cwd(allocator::temp()) @inline;
+fn Path! tgetcwd() @deprecated("Use temp_cwd()") => new_cwd(allocator::temp()) @inline;
+fn void! chdir(Path path) => os::native_chdir(path) @inline;
+fn Path! temp_directory(Allocator allocator = allocator::heap()) => os::native_temp_directory(allocator);
+fn void! delete(Path path) => os::native_remove(path.str_view()) @inline;
+
+macro bool is_separator(char c, PathEnv path_env = DEFAULT_PATH_ENV)
+{
+	return c == '/' || (c == '\\' && path_env == PathEnv.WIN32);
+}
+
+macro bool is_posix_separator(char c)
+{
+	return c == '/' || c == '\\';
+}
+
+macro bool is_win32_separator(char c)
+{
+	return c == '/' || c == '\\';
+}
+
+fn PathList! ls(Path dir, bool no_dirs = false, bool no_symlinks = false, String mask = "", Allocator allocator = allocator::heap()) @deprecated("use new_ls")
+{
+	return new_ls(dir, no_dirs, no_symlinks, mask, allocator);
+
+}
+fn PathList! temp_ls(Path dir, bool no_dirs = false, bool no_symlinks = false, String mask = "")
+{
+	return new_ls(dir, no_dirs, no_symlinks, mask, allocator::temp()) @inline;
+}
+
+fn PathList! new_ls(Path dir, bool no_dirs = false, bool no_symlinks = false, String mask = "", Allocator allocator = allocator::heap())
+{
+	$if $defined(os::native_ls):
+		return os::native_ls(dir, no_dirs, no_symlinks, mask, allocator);
+	$else
+		return IoError.UNSUPPORTED_OPERATION?;
+	$endif
+}
+
+enum MkdirPermissions
+{
+	NORMAL,
+	USER_ONLY,
+	USER_AND_ADMIN
+}
+
+fn bool! mkdir(Path path, bool recursive = false, MkdirPermissions permissions = NORMAL)
+{
+	if (!path.path_string.len) return PathResult.INVALID_PATH?;
+	if (is_dir(path)) return false;
+	if (exists(path)) return IoError.FILE_NOT_DIR?;
+
+	if (recursive)
+	{
+		if (try parent = path.parent()) mkdir(parent, true, permissions)!;
+	}
+	if (!is_dir(path.parent()) ?? false) return IoError.CANNOT_READ_DIR?;
+
+	return os::native_mkdir(path, permissions);
+}
+
+fn bool! rmdir(Path path)
+{
+	if (!path.path_string.len) return PathResult.INVALID_PATH?;
+	return os::native_rmdir(path);
+}
+
+fn void! rmtree(Path path)
+{
+	if (!path.path_string.len) return PathResult.INVALID_PATH?;
+	$if $defined(os::native_rmtree):
+		return os::native_rmtree(path);
+	$else
+		return IoError.UNSUPPORTED_OPERATION?;
+	$endif
+}
+
+fn Path! new(String path, Allocator allocator = allocator::heap(), PathEnv path_env = DEFAULT_PATH_ENV)
+{
+	return { normalize(path.copy(allocator), path_env), path_env };
+}
+
+fn Path! temp_new(String path, PathEnv path_env = DEFAULT_PATH_ENV)
+{
+	return new(path, allocator::temp(), path_env);
+}
+
+fn Path! new_win32_wstring(WString path, Allocator allocator = allocator::heap())
+{
+	@pool(allocator)
+	{
+		return path::new(string::temp_from_wstring(path)!, .allocator = allocator);
+	};
+}
+
+fn Path! new_windows(String path, Allocator allocator = allocator::heap())
+{
+	return new(path, allocator, WIN32);
+}
+
+fn Path! new_posix(String path, Allocator allocator = allocator::heap())
+{
+	return new(path, allocator, POSIX);
+}
+
+fn bool Path.equals(self, Path p2)
+{
+	return self.env == p2.env && self.path_string == p2.path_string;
+}
+
+fn Path! Path.append(self, String filename, Allocator allocator = allocator::heap()) @deprecated("Use path.new_append(...)")
+{
+	return self.new_append(filename, allocator) @inline;
+}
+
+/**
+ * Append the string to the current path.
+ *
+ * @param [in] filename
+ **/
+fn Path! Path.new_append(self, String filename, Allocator allocator = allocator::heap())
+{
+	if (!self.path_string.len) return new(filename, allocator, self.env)!;
+	assert(!is_separator(self.path_string[^1], self.env));
+
+	@pool(allocator)
+	{
+		DString dstr = dstring::temp_with_capacity(self.path_string.len + 1 + filename.len);
+		dstr.append(self.path_string);
+		dstr.append(PREFERRED_SEPARATOR);
+		dstr.append(filename);
+		return { normalize(dstr.copy_str(allocator), self.env), self.env };
+	};
+}
+
+fn Path! Path.temp_append(self, String filename) => self.new_append(filename, allocator::temp());
+
+fn Path! Path.tappend(self, String filename) @deprecated("Use path.temp_append(...)") => self.new_append(filename, allocator::temp());
+
+fn usz Path.start_of_base_name(self) @local
+{
+	String path_str = self.path_string;
+	if (!path_str.len) return 0;
+	if (self.env == PathEnv.WIN32)
+	{
+		if (try index = path_str.rindex_of_char('\\'))
+		{
+			// c:\ style path, we're done!
+			if (path_str[0] != '\\') return index + 1;
+			// Handle \\server\foo
+			// Find the \ before "foo"
+			usz last_index = 2 + path_str[2..].index_of_char('\\')!!;
+			// If they don't match, we're done
+			assert(last_index <= index, "Invalid normalized, path %d vs %s in %s", last_index, index, path_str);
+			if (last_index != index) return index + 1;
+			// Otherwise just default to the volume length.
+		}
+		return volume_name_len(path_str, self.env)!!;
+	}
+	return path_str.rindex_of_char('/') + 1 ?? 0;
+}
+
+fn bool! Path.is_absolute(self)
+{
+	String path_str = self.str_view();
+	if (!path_str.len) return false;
+	usz path_start = volume_name_len(path_str, self.env)!;
+	if (path_start > 0 && path_str[0] == '\\') return true;
+	return path_start < path_str.len && is_separator(path_str[path_start], self.env);
+}
+
+fn Path! Path.absolute(self, Allocator allocator = allocator::heap()) @deprecated("Use path.new_absolute()")
+{
+	return self.new_absolute(allocator) @inline;
+}
+
+/**
+ * @require self.env == DEFAULT_PATH_ENV : "This method is only available on native paths"
+ **/
+fn Path! Path.new_absolute(self, Allocator allocator = allocator::heap())
+{
+	String path_str = self.str_view();
+	if (!path_str.len) return PathResult.INVALID_PATH?;
+	if (self.is_absolute()!) return new(path_str, allocator, self.env);
+	if (path_str == ".")
+	{
+		@pool(allocator)
+		{
+			String cwd = os::getcwd(allocator::temp())!;
+			return new(cwd, allocator, self.env);
+		};
+	}
+	$if DEFAULT_PATH_ENV == WIN32:
+		@pool(allocator)
+		{
+			const usz BUFFER_LEN = 4096;
+			WString buffer = (WString)mem::temp_alloc_array(Char16, BUFFER_LEN);
+			buffer = win32::_wfullpath(buffer, path_str.to_temp_wstring()!, BUFFER_LEN);
+			if (!buffer) return PathResult.INVALID_PATH?;
+			return { string::new_from_wstring(buffer, allocator), WIN32 };
+		};
+	$else
+		String cwd = os::getcwd(allocator::temp())!;
+		return Path { cwd, self.env }.new_append(path_str, allocator)!;
+	$endif
+}
+
+fn String Path.basename(self)
+{
+	usz basename_start = self.start_of_base_name();
+	String path_str = self.path_string;
+	if (basename_start == path_str.len) return "";
+	return path_str[basename_start..];
+}
+
+
+fn String Path.dirname(self)
+{
+	usz basename_start = self.start_of_base_name();
+	String path_str = self.path_string;
+	if (basename_start == 0) return ".";
+	usz start = volume_name_len(path_str, self.env)!!;
+	if (basename_start <= start + 1)
+	{
+		if (self.env == WIN32 && basename_start > start && path_str[0..1] == `\\`)
+		{
+			return path_str[:basename_start - 1];
+		}
+		return path_str[:basename_start];
+	}
+	return path_str[:basename_start - 1];
+}
+
+/**
+ * Test if the path has the given extension, so given the path /foo/bar.c3
+ * this would be true matching the extension "c3"
+ *
+ * @param [in] extension `The extension name (not including the leading '.')`
+ * @require extension.len > 0 : `The extension cannot be empty`
+ * @return `true if the extension matches`
+ **/
+fn bool Path.has_extension(self, String extension)
+{
+	String basename = self.basename();
+	if (basename.len <= extension.len) return false;
+	if (basename[^extension.len + 1] != '.') return false;
+	return basename[^extension.len..] == extension;
+}
+
+fn String! Path.extension(self)
+{
+	String basename = self.basename();
+	usz index = basename.rindex_of(".")!;
+	// Plain ".foo" does not have an
+	if (index == 0) return SearchResult.MISSING?;
+	if (index == basename.len) return "";
+	return basename[index + 1..];
+}
+
+fn String Path.volume_name(self)
+{
+	usz len = volume_name_len(self.str_view(), self.env)!!;
+	if (!len) return "";
+	return self.path_string[:len];
+}
+
+fn usz! volume_name_len(String path, PathEnv path_env) @local
+{
+	usz len = path.len;
+	if (len < 2 || path_env != PathEnv.WIN32) return 0;
+	switch (path[0])
+	{
+		case '\\':
+			// "\\" paths.. must be longer than 2
+			if (len == 2) return 0;
+			int count = 1;
+			while (count < len && path[count] == '\\') count++;
+			// Not 2 => folded paths
+			if (count != 2) return 0;
+			// Check that we have a name followed by '\'
+			isz base_found = 0;
+			for (usz i = 2; i < len; i++)
+			{
+				char c = path[i];
+				if (is_win32_separator(c))
+				{
+					if (base_found) return i;
+					base_found = i;
+					continue;
+				}
+				if (is_reserved_win32_path_char(c)) return PathResult.INVALID_PATH?;
+			}
+			if (base_found > 0 && base_found + 1 < len) return len;
+			return PathResult.INVALID_PATH?;
+		case 'A'..'Z':
+		case 'a'..'z':
+			return path[1] == ':' ? 2 : 0;
+		default:
+			return 0;
+	}
+}
+
+fn Path! Path.parent(self)
+{
+	if (self.path_string.len == 1 && is_separator(self.path_string[0], self.env)) return PathResult.NO_PARENT?;
+	foreach_r(i, c : self.path_string)
+	{
+		if (is_separator(c, self.env))
+		{
+			return { self.path_string[:i], self.env };
+		}
+	}
+	return PathResult.NO_PARENT?;
+}
+
+fn String! normalize(String path_str, PathEnv path_env = DEFAULT_PATH_ENV)
+{
+	if (!path_str.len) return "";
+	usz path_start = volume_name_len(path_str, path_env)!;
+	if (path_start > 0 && path_env == PathEnv.WIN32)
+	{
+		for (usz i = 0; i < path_start; i++) if (path_str[i] == '/') path_str[i] = '\\';
+	}
+	usz path_len = path_str.len;
+	if (path_start == path_len) return path_str;
+	char path_separator = path_env == PathEnv.WIN32 ? PREFERRED_SEPARATOR_WIN32 : PREFERRED_SEPARATOR_POSIX;
+	usz len = path_start;
+	bool has_root = is_separator(path_str[path_start], path_env);
+	if (has_root)
+	{
+		path_str[len++] = path_separator;
+		path_start++;
+	}
+	// It is safe to write it as true, since we already dealt with /foo.
+	// This allows us to avoid checking whether it is the start of the path.
+	bool previous_was_separator = true;
+
+	for (usz i = path_start; i < path_len; i++)
+	{
+		char c = path_str[i];
+		// Fold foo///bar into foo/bar
+		if (is_separator(c, path_env))
+		{
+			// Fold //
+			if (previous_was_separator) continue;
+
+			// New /, so mark and rewrite
+			path_str.ptr[len++] = path_separator;
+			previous_was_separator = true;
+			continue;
+		}
+
+		// The rest are names of the path elements, so check that the
+		// characters are valid.
+		if (is_reserved_path_char(c, path_env)) return PathResult.INVALID_PATH?;
+
+		// If we have '.' after a separator
+		if (c == '.' && previous_was_separator)
+		{
+			// Get the number of dots until next separator, expecting 1 or 2
+			bool is_last = i == path_len - 1;
+			int dots = 1;
+			if (!is_last)
+			{
+				char next = path_str[i + 1];
+				switch
+				{
+					case next == '.':
+						dots = 2;
+						is_last = i == path_len - 2;
+						if (!is_last && !is_separator(path_str[i + 2], path_env))
+						{
+							dots = 0;
+						}
+					case !is_separator(next, path_env):
+						dots = 0;
+				}
+			}
+			switch (dots)
+			{
+				case 1:
+					// /./abc -> skip to /./abc
+					//  ^                  ^
+					i++;
+					continue;
+				case 2:
+					// This is an error: /a/../..
+					if (len == path_start && has_root) return PathResult.INVALID_PATH?;
+
+					// If this .. at the start, or after ../? If so, we just copy ..
+					if (len == path_start ||
+						(len - path_start >= 3 && path_str[len - 1] == path_separator
+						&& path_str[len - 3] == '.' && path_str[len - 3] == '.' &&
+						(len - 3 == 0 || path_str[len - 4] == path_separator)))
+					{
+						if (i != len)
+						{
+							path_str[len] = '.';
+							path_str[len + 1] = '.';
+						}
+						len += 2;
+						if (len < path_len) path_str[len++] = path_separator;
+						i += 2;
+						continue;
+					}
+					// Step back, now looking at '/' abc/def/. -> abc/def/
+					len--;
+					// Step back until finding a separator or the start.
+					while (len > path_start && !is_separator(path_str[len - 1], path_env))
+					{
+						len--;
+					}
+					// Reading, we go from /../abc to /../abc
+					//                      ^            ^
+					i += 2;
+					continue;
+				default:
+					break;
+
+			}
+		}
+		if (i != len) path_str[len] = c;
+		previous_was_separator = false;
+		len++;
+	}
+	if (len > path_start + 1 && is_separator(path_str[len - 1], path_env)) len--;
+	if (path_str.len > len) path_str.ptr[len] = 0;
+	// Empty path after normalization -> "."
+	if (!len) return ".";
+	return path_str[:len];
+}
+
+fn ZString Path.as_zstr(self) => (ZString)self.path_string.ptr;
+
+fn String Path.root_directory(self)
+{
+	String path_str = self.str_view();
+	usz len = path_str.len;
+	if (!len) return "";
+	if (path_str == ".") return ".";
+	if (self.env == PathEnv.WIN32)
+	{
+		usz root_len = volume_name_len(path_str, self.env)!!;
+		if (root_len == len || !is_win32_separator(path_str[root_len])) return "";
+		return path_str[root_len..root_len];
+	}
+	if (!is_posix_separator(path_str[0])) return "";
+	for (usz i = 1; i < len; i++)
+	{
+		if (is_posix_separator(path_str[i]))
+		{
+			return path_str[:i];
+		}
+	}
+	return path_str;
+}
+
+def PathWalker = fn bool! (Path, bool is_dir, void*);
+
+/*
+ * Walk the path recursively. PathWalker is run on every file and
+ * directory found. Return true to abort the walk.
+ * @require self.env == DEFAULT_PATH_ENV : "This method is only available on native paths"
+ */
+fn bool! Path.walk(self, PathWalker w, void* data)
+{
+	const PATH_MAX = 512;
+	@stack_mem(PATH_MAX; Allocator allocator)
+	{
+		Path abs = self.new_absolute(allocator)!;
+		PathList files = new_ls(abs, .allocator = allocator)!;
+		foreach (f : files)
+		{
+			if (f.str_view() == "." || f.str_view() == "..") continue;
+			f = abs.new_append(f.str_view(), allocator)!;
+			bool is_directory = is_dir(f);
+			if (w(f, is_directory, data)!) return true;
+			if (is_directory && f.walk(w, data)!) return true;
+		}
+	};
+	return false;
+}
+
+fn String Path.str_view(self) @inline
+{
+	return self.path_string;
+}
+
+
+fn bool Path.has_suffix(self, String str)
+{
+	return self.str_view().ends_with(str);
+}
+
+fn void Path.free_with_allocator(self, Allocator allocator)
+{
+	allocator::free(allocator, self.path_string.ptr);
+}
+
+fn void Path.free(self)
+{
+	free(self.path_string.ptr);
+}
+
+
+fn usz! Path.to_format(&self, Formatter* formatter) @dynamic
+{
+	return formatter.print(self.str_view());
+}
+
+fn String Path.to_new_string(&self, Allocator allocator = allocator::heap()) @dynamic
+{
+	return self.str_view().copy(allocator);
+}
+
+const bool[256] RESERVED_PATH_CHAR_POSIX = {
+	[0] = true,
+	['/'] = true,
+};
+const bool[256] RESERVED_PATH_CHAR_WIN32 = {
+	[0..31] = true,
+	['>'] = true,
+	['<'] = true,
+	[':'] = true,
+	['\"'] = true,
+	['/'] = true,
+	['\\'] = true,
+	['|'] = true,
+	['?'] = true,
+	['*'] = true,
+};
+
+macro bool is_reserved_win32_path_char(char c)
+{
+	return RESERVED_PATH_CHAR_WIN32[c];
+}
+
+macro bool is_reserved_path_char(char c, PathEnv path_env = DEFAULT_PATH_ENV)
+{
+	return path_env == PathEnv.WIN32
+		? RESERVED_PATH_CHAR_WIN32[c]
+		: RESERVED_PATH_CHAR_POSIX[c];
+}

lib/std/io/stream.c3

@@ -0,0 +1,228 @@
+module std::io;
+import std::math;
+
+interface InStream
+{
+	fn void! close() @optional;
+	fn usz! seek(isz offset, Seek seek) @optional;
+	fn usz len() @optional;
+	fn usz! available() @optional;
+	fn usz! read(char[] buffer);
+	fn char! read_byte();
+	fn usz! write_to(OutStream out) @optional;
+	fn void! pushback_byte() @optional;
+}
+
+
+interface OutStream
+{
+	fn void! destroy() @optional;
+	fn void! close() @optional;
+	fn void! flush() @optional;
+	fn usz! write(char[] bytes);
+	fn void! write_byte(char c);
+	fn usz! read_to(InStream in) @optional;
+}
+
+fn usz! available(InStream s)
+{
+	if (&s.available) return s.available();
+	if (&s.seek)
+	{
+		usz curr = s.seek(0, Seek.CURSOR)!;
+		usz len = s.seek(0, Seek.END)!;
+		s.seek(curr, Seek.SET)!;
+		return len - curr;
+	}
+	return 0;
+}
+
+macro bool @is_instream(#expr)
+{
+	return $assignable(#expr, InStream);
+}
+
+macro bool @is_outstream(#expr)
+{
+	return $assignable(#expr, OutStream);
+}
+
+/**
+ * @param [&out] ref
+ * @require @is_instream(stream)
+ **/
+macro usz! read_any(stream, any ref)
+{
+	return read_all(stream, ((char*)ref)[:ref.type.sizeof]);
+}
+
+/**
+ * @param [&in] ref "the object to write."
+ * @require @is_outstream(stream)
+ * @ensure return == ref.type.sizeof
+ */
+macro usz! write_any(stream, any ref)
+{
+	return write_all(stream, ((char*)ref)[:ref.type.sizeof]);
+}
+
+/**
+ * @require @is_instream(stream)
+ */
+macro usz! read_all(stream, char[] buffer)
+{
+	if (buffer.len == 0) return 0;
+	usz n = stream.read(buffer)!;
+	if (n != buffer.len) return IoError.UNEXPECTED_EOF?;
+	return n;
+}
+
+/**
+ * @require @is_outstream(stream)
+ */
+macro usz! write_all(stream, char[] buffer)
+{
+	if (buffer.len == 0) return 0;
+	usz n = stream.write(buffer)!;
+	if (n != buffer.len) return IoError.INCOMPLETE_WRITE?;
+	return n;
+}
+
+macro usz! @read_using_read_byte(&s, char[] buffer)
+{
+	usz len = 0;
+	foreach (&cptr : buffer)
+	{
+		char! c = s.read_byte();
+		if (catch err = c)
+		{
+			case IoError.EOF: return len;
+			default: return err?;
+		}
+		*cptr = c;
+		len++;
+	}
+	return len;
+}
+
+macro void! @write_byte_using_write(&s, char c)
+{
+	char[1] buff = { c };
+	(*s).write(&buff)!;
+}
+
+
+macro char! @read_byte_using_read(&s)
+{
+	char[1] buffer;
+	usz read = (*s).read(&buffer)!;
+	if (read != 1) return IoError.EOF?;
+	return buffer[0];
+}
+
+def ReadByteFn = fn char!();
+
+
+macro usz! @write_using_write_byte(&s, char[] bytes)
+{
+	foreach (c : bytes) s.write_byte(self, c)!;
+	return bytes.len;
+}
+
+macro void! @pushback_using_seek(&s)
+{
+	s.seek(-1, CURSOR)!;
+}
+
+fn usz! copy_to(InStream in, OutStream dst, char[] buffer = {})
+{
+	if (buffer.len) return copy_through_buffer(in, dst, buffer);
+	if (&in.write_to) return in.write_to(dst);
+	if (&dst.read_to) return dst.read_to(in);
+	$switch (env::MEMORY_ENV)
+		$case NORMAL:
+			return copy_through_buffer(in, dst, &&char[4096]{});
+		$case SMALL:
+			return copy_through_buffer(in, dst, &&char[1024]{});
+		$case TINY:
+		$case NONE:
+			return copy_through_buffer(in, dst, &&(char[256]{}));
+	$endswitch
+}
+
+macro usz! copy_through_buffer(InStream in, OutStream dst, char[] buffer) @local
+{
+	usz total_copied;
+	while (true)
+	{
+		usz! len = in.read(buffer);
+		if (catch err = len)
+		{
+			case IoError.EOF: return total_copied;
+			default: return err?;
+		}
+		if (!len) return total_copied;
+		usz written = dst.write(buffer[:len])!;
+		total_copied += len;
+		if (written != len) return IoError.INCOMPLETE_WRITE?;
+	}
+}
+
+const char[*] MAX_VARS @private = { [2] = 3, [4] = 5, [8] = 10 };
+
+/**
+ * @require @is_instream(stream)
+ * @require @typekind(x_ptr) == POINTER && $typeof(x_ptr).inner.kindof.is_int()
+ **/
+macro usz! read_varint(stream, x_ptr)
+{
+	var $Type = $typefrom($typeof(x_ptr).inner);
+	const MAX = MAX_VARS[$Type.sizeof];
+	$Type x;
+	uint shift;
+	usz n;
+	for (usz i = 0; i < MAX; i++)
+	{
+		char! c = stream.read_byte();
+		if (catch err = c)
+		{
+			case IoError.EOF:
+				return IoError.UNEXPECTED_EOF?;
+			default:
+				return err?;
+		}
+		n++;
+		if (c & 0x80 == 0)
+		{
+			if (i + 1 == MAX && c > 1) break;
+			x |= c << shift;
+			$if $Type.kindof == SIGNED_INT:
+				x = x & 1 == 0 ? x >> 1 : ~(x >> 1);
+			$endif
+			*x_ptr = x;
+			return n;
+		}
+		x |= (c & 0x7F) << shift;
+		shift += 7;
+	}
+	return MathError.OVERFLOW?;
+}
+/**
+ * @require @is_outstream(stream)
+ * @require @typekind(x).is_int()
+ **/
+macro usz! write_varint(stream, x)
+{
+	var $Type = $typeof(x);
+	const MAX = MAX_VARS[$Type.sizeof];
+	char[MAX] buffer @noinit;
+	usz i;
+	while (x >= 0x80)
+	{
+		buffer[i] = (char)(x | 0x80);
+		x >>= 7;
+		i++;
+	}
+	buffer[i] = (char)x;
+	return write_all(stream, buffer[:i + 1]);
+}

lib/std/io/stream/buffer.c3

@@ -0,0 +1,133 @@
+module std::io;
+
+struct ReadBuffer (InStream)
+{
+	InStream wrapped_stream;
+	char[] bytes;
+	usz read_idx;
+	usz write_idx;
+}
+
+/**
+ * Buffer reads from a stream.
+ * @param [inout] self
+ * @require bytes.len > 0
+ * @require self.bytes.len == 0 "Init may not run on already initialized data"
+ **/
+fn ReadBuffer* ReadBuffer.init(&self, InStream wrapped_stream, char[] bytes)
+{
+	*self = { .wrapped_stream = wrapped_stream, .bytes = bytes };
+	return self;
+}
+fn String ReadBuffer.str_view(&self) @inline
+{
+	return (String)self.bytes[self.read_idx:self.write_idx - self.read_idx];
+}
+
+fn void! ReadBuffer.close(&self) @dynamic
+{
+	if (&self.wrapped_stream.close) self.wrapped_stream.close()!;
+}
+
+fn usz! ReadBuffer.read(&self, char[] bytes) @dynamic
+{
+	if (self.read_idx == self.write_idx)
+	{
+		if (self.read_idx == 0 && bytes.len >= self.bytes.len)
+		{
+			// Read directly into the input buffer.
+			return self.wrapped_stream.read(bytes)!;
+		}
+		self.refill()!;
+	}
+	usz n = min(self.write_idx - self.read_idx, bytes.len);
+	bytes[:n] = self.bytes[self.read_idx:n];
+	self.read_idx += n;
+	return n;
+}
+
+fn char! ReadBuffer.read_byte(&self) @dynamic
+{
+	if (self.read_idx == self.write_idx) self.refill()!;
+	if (self.read_idx == self.write_idx) return IoError.EOF?;
+	char c = self.bytes[self.read_idx];
+	self.read_idx++;
+	return c;
+}
+
+fn void! ReadBuffer.refill(&self) @local @inline
+{
+	self.read_idx = 0;
+	self.write_idx = self.wrapped_stream.read(self.bytes)!;
+}
+
+struct WriteBuffer (OutStream)
+{
+	OutStream wrapped_stream;
+	char[] bytes;
+	usz index;
+}
+
+/**
+ * Buffer writes to a stream. Call `flush` when done writing to the buffer.
+ * @param [inout] self
+ * @require bytes.len > 0 "Non-empty buffer required"
+ * @require self.bytes.len == 0 "Init may not run on already initialized data"
+ **/
+fn WriteBuffer* WriteBuffer.init(&self, OutStream wrapped_stream, char[] bytes)
+{
+	*self = { .wrapped_stream = wrapped_stream, .bytes = bytes };
+	return self;
+}
+
+fn String WriteBuffer.str_view(&self) @inline
+{
+	return (String)self.bytes[:self.index];
+}
+
+fn void! WriteBuffer.close(&self) @dynamic
+{
+	if (&self.wrapped_stream.close) return self.wrapped_stream.close();
+}
+
+fn void! WriteBuffer.flush(&self) @dynamic
+{
+	self.write_pending()!;
+	if (&self.wrapped_stream.flush) self.wrapped_stream.flush()!;
+}
+
+fn usz! WriteBuffer.write(&self, char[] bytes) @dynamic
+{
+	usz n = self.bytes.len - self.index;
+	if (bytes.len < n)
+	{
+		// Enough room in the buffer.
+		self.bytes[self.index:bytes.len] = bytes[..];
+		self.index += bytes.len;
+		return bytes.len;
+	}
+	self.write_pending()!;
+	if (bytes.len >= self.bytes.len)
+	{
+		// Write directly to the stream.
+		return self.wrapped_stream.write(bytes);
+	}
+	// Buffer the data.
+	self.bytes[:bytes.len] = bytes[..];
+	self.index = bytes.len;
+	return bytes.len;
+}
+
+fn void! WriteBuffer.write_byte(&self, char c) @dynamic
+{
+	usz n = self.bytes.len - self.index;
+	if (n == 0) self.write_pending()!;
+	self.bytes[0] = c;
+	self.index = 1;
+}
+
+fn void! WriteBuffer.write_pending(&self) @local
+{
+	self.index -= self.wrapped_stream.write(self.bytes[:self.index])!;
+	if (self.index != 0) return IoError.INCOMPLETE_WRITE?;
+}

lib/std/io/stream/bytebuffer.c3

@@ -0,0 +1,148 @@
+module std::io;
+import std::math;
+
+struct ByteBuffer (InStream, OutStream)
+{
+	Allocator allocator;
+	usz max_read;
+	char[] bytes;
+	usz read_idx;
+	usz write_idx;
+	bool has_last;
+}
+
+/**
+ * ByteBuffer provides a streamable read/write buffer.
+ * max_read defines how many bytes might be kept before its internal buffer is shrinked.
+ * @require self.bytes.len == 0 "Buffer already initialized."
+ **/
+fn ByteBuffer*! ByteBuffer.new_init(&self, usz max_read, usz initial_capacity = 16, Allocator allocator = allocator::heap())
+{
+	*self = { .allocator = allocator, .max_read = max_read };
+	initial_capacity = max(initial_capacity, 16);
+	self.grow(initial_capacity)!;
+	return self;
+}
+
+fn ByteBuffer*! ByteBuffer.temp_init(&self, usz max_read, usz initial_capacity = 16)
+{
+	return self.new_init(max_read, initial_capacity, allocator::temp());
+}
+
+/**
+ * @require buf.len > 0
+ * @require self.bytes.len == 0 "Buffer already initialized."
+ **/
+fn ByteBuffer*! ByteBuffer.init_with_buffer(&self, char[] buf)
+{
+	*self = { .max_read = buf.len, .bytes = buf };
+	return self;
+}
+
+fn void ByteBuffer.free(&self)
+{
+	if (self.allocator) allocator::free(self.allocator, self.bytes);
+	*self = {};
+}
+
+fn usz! ByteBuffer.write(&self, char[] bytes) @dynamic
+{
+	usz cap = self.bytes.len - self.write_idx;
+	if (cap < bytes.len) self.grow(bytes.len)!;
+	self.bytes[self.write_idx:bytes.len] = bytes[..];
+	self.write_idx += bytes.len;
+	return bytes.len;
+}
+
+fn void! ByteBuffer.write_byte(&self, char c) @dynamic
+{
+	usz cap = self.bytes.len - self.write_idx;
+	if (cap == 0) self.grow(1)!;
+	self.bytes[self.write_idx] = c;
+	self.write_idx++;
+}
+
+fn usz! ByteBuffer.read(&self, char[] bytes) @dynamic
+{
+	usz readable = self.write_idx - self.read_idx;
+	if (readable == 0)
+	{
+		self.has_last = false;
+		return IoError.EOF?;
+	}
+	usz n = min(readable, bytes.len);
+	bytes[:n] = self.bytes[self.read_idx:n];
+	self.read_idx += n;
+	self.has_last = n > 0;
+	self.shrink();
+	return n;
+}
+
+fn char! ByteBuffer.read_byte(&self) @dynamic
+{
+	usz readable = self.write_idx - self.read_idx;
+	if (readable == 0)
+	{
+		self.has_last = false;
+		return IoError.EOF?;
+	}
+	char c = self.bytes[self.read_idx];
+	self.read_idx++;
+	self.has_last = true;
+	self.shrink();
+	return c;
+}
+
+/*
+ * Only the last byte of a successful read can be pushed back.
+ */
+fn void! ByteBuffer.pushback_byte(&self) @dynamic
+{
+	if (!self.has_last) return IoError.EOF?;
+	assert(self.read_idx > 0);
+	self.read_idx--;
+	self.has_last = false;
+}
+
+fn usz! ByteBuffer.seek(&self, isz offset, Seek seek) @dynamic
+{
+	switch (seek)
+	{
+		case SET:
+			if (offset < 0 || offset > self.write_idx) return IoError.INVALID_POSITION?;
+			self.read_idx = offset;
+			return offset;
+		case CURSOR:
+			if ((offset < 0 && self.read_idx < -offset) ||
+				(offset > 0 && self.read_idx + offset > self.write_idx)) return IoError.INVALID_POSITION?;
+			self.read_idx += offset;
+		case END:
+			if (offset < 0 || offset > self.write_idx) return IoError.INVALID_POSITION?;
+			self.read_idx = self.write_idx - offset;
+	}
+	return self.read_idx;
+}
+
+fn usz! ByteBuffer.available(&self) @inline @dynamic
+{
+	return self.write_idx - self.read_idx;
+}
+
+fn void! ByteBuffer.grow(&self, usz n)
+{
+	n = math::next_power_of_2(n);
+	char* p = allocator::realloc_aligned(self.allocator, self.bytes, n, .alignment = char.alignof)!;
+	self.bytes = p[:n];
+}
+
+macro ByteBuffer.shrink(&self)
+{
+	if (self.read_idx >= self.max_read)
+	{
+		// Drop the read data besides the last byte (for pushback_byte).
+		usz readable = self.write_idx - self.read_idx;
+		self.bytes[:1 + readable] = self.bytes[self.read_idx - 1:1 + readable];
+		self.write_idx = 1 + readable;
+		self.read_idx = 1;
+	}
+}

lib/std/io/stream/bytereader.c3

@@ -0,0 +1,68 @@
+module std::io;
+
+struct ByteReader (InStream)
+{
+	char[] bytes;
+	usz index;
+}
+
+fn usz ByteReader.len(&self) @dynamic
+{
+	return self.bytes.len;
+}
+
+fn ByteReader* ByteReader.init(&self, char[] bytes)
+{
+	*self = { .bytes = bytes };
+	return self;
+}
+
+fn usz! ByteReader.read(&self, char[] bytes) @dynamic
+{
+	if (self.index >= self.bytes.len) return IoError.EOF?;
+	usz len = min(self.bytes.len - self.index, bytes.len);
+	if (len == 0) return 0;
+	mem::copy(bytes.ptr, &self.bytes[self.index], len);
+	self.index += len;
+	return len;
+}
+
+fn char! ByteReader.read_byte(&self) @dynamic
+{
+	if (self.index >= self.bytes.len) return IoError.EOF?;
+	return self.bytes[self.index++];
+}
+
+fn void! ByteReader.pushback_byte(&self) @dynamic
+{
+	if (!self.index) return IoError.INVALID_PUSHBACK?;
+	self.index--;
+}
+
+fn usz! ByteReader.seek(&self, isz offset, Seek seek) @dynamic
+{
+	isz new_index;
+	switch (seek)
+	{
+		case SET: new_index = offset;
+		case CURSOR: new_index = self.index + offset;
+		case END: new_index = self.bytes.len + offset;
+	}
+	if (new_index < 0) return IoError.INVALID_POSITION?;
+	self.index = new_index;
+	return new_index;
+}
+
+fn usz! ByteReader.write_to(&self, OutStream writer) @dynamic
+{
+	if (self.index >= self.bytes.len) return 0;
+	usz written = writer.write(self.bytes[self.index..])!;
+	self.index += written;
+	assert(self.index <= self.bytes.len);
+	return written;
+}
+
+fn usz! ByteReader.available(&self) @inline @dynamic
+{
+	return max(0, self.bytes.len - self.index);
+}

lib/std/io/stream/bytewriter.c3

@@ -0,0 +1,113 @@
+module std::io;
+import std::math;
+
+struct ByteWriter (OutStream)
+{
+	char[] bytes;
+	usz index;
+	Allocator allocator;
+}
+
+/**
+ * @param [&inout] self
+ * @param [&inout] allocator
+ * @require self.bytes.len == 0 "Init may not run on already initialized data"
+ * @ensure (bool)allocator, self.index == 0
+ **/
+fn ByteWriter* ByteWriter.new_init(&self, Allocator allocator = allocator::heap())
+{
+	*self = { .bytes = {}, .allocator = allocator };
+	return self;
+}
+
+/**
+ * @param [&inout] self
+ * @require self.bytes.len == 0 "Init may not run on already initialized data"
+ * @ensure self.index == 0
+ **/
+fn ByteWriter* ByteWriter.temp_init(&self)
+{
+	return self.new_init(allocator::temp()) @inline;
+}
+
+fn ByteWriter* ByteWriter.init_with_buffer(&self, char[] data)
+{
+	*self = { .bytes = data, .allocator = null };
+	return self;
+}
+
+fn void! ByteWriter.destroy(&self) @dynamic
+{
+	if (!self.allocator) return;
+	if (void* ptr = self.bytes.ptr) allocator::free(self.allocator, ptr);
+	*self = { };
+}
+
+fn String ByteWriter.str_view(&self) @inline
+{
+	return (String)self.bytes[:self.index];
+}
+
+fn void! ByteWriter.ensure_capacity(&self, usz len) @inline
+{
+	if (self.bytes.len > len) return;
+	if (!self.allocator) return IoError.OUT_OF_SPACE?;
+	if (len < 16) len = 16;
+	usz new_capacity = math::next_power_of_2(len);
+	char* new_ptr = allocator::realloc_try(self.allocator, self.bytes.ptr, new_capacity)!;
+	self.bytes = new_ptr[:new_capacity];
+}
+
+fn usz! ByteWriter.write(&self, char[] bytes) @dynamic
+{
+	self.ensure_capacity(self.index + bytes.len)!;
+	mem::copy(&self.bytes[self.index], bytes.ptr, bytes.len);
+	self.index += bytes.len;
+	return bytes.len;
+}
+
+fn void! ByteWriter.write_byte(&self, char c) @dynamic
+{
+	self.ensure_capacity(self.index + 1)!;
+	self.bytes[self.index++] = c;
+}
+
+/**
+ * @param [&inout] self
+ * @param reader
+ **/
+fn usz! ByteWriter.read_from(&self, InStream reader) @dynamic
+{
+	usz start_index = self.index;
+	if (&reader.available)
+	{
+		while (usz available = reader.available()!)
+		{
+			self.ensure_capacity(self.index + available)!;
+			usz read = reader.read(self.bytes[self.index..])!;
+			self.index += read;
+		}
+		return self.index - start_index;
+	}
+	if (self.bytes.len == 0)
+	{
+		self.ensure_capacity(16)!;
+	}
+	while (true)
+	{
+		// See how much we can read.
+		usz len_to_read = self.bytes.len - self.index;
+		// Less than 16 bytes? Double the capacity
+		if (len_to_read < 16)
+		{
+			self.ensure_capacity(self.bytes.len * 2)!;
+			len_to_read = self.bytes.len - self.index;
+		}
+		// Read into the rest of the buffer
+		usz read = reader.read(self.bytes[self.index..])!;
+		self.index += read;
+		// Ok, we reached the end.
+		if (read < len_to_read) return self.index - start_index;
+		// Otherwise go another round
+	}
+}

lib/std/io/stream/limitreader.c3

@@ -0,0 +1,44 @@
+module std::io;
+
+struct LimitReader (InStream)
+{
+	InStream wrapped_stream;
+	usz limit;
+}
+
+/**
+ * @param [&inout] wrapped_stream "The stream to read from"
+ * @param limit "The max limit to read"
+ **/
+fn LimitReader* LimitReader.init(&self, InStream wrapped_stream, usz limit)
+{
+	*self = { .wrapped_stream = wrapped_stream, .limit = limit };
+	return self;
+}
+
+fn void! LimitReader.close(&self) @dynamic
+{
+	if (&self.wrapped_stream.close) return self.wrapped_stream.close();
+}
+
+
+fn usz! LimitReader.read(&self, char[] bytes) @dynamic
+{
+	if (self.limit == 0) return IoError.EOF?;
+	usz m = min(bytes.len, self.limit);
+	usz n = self.wrapped_stream.read(bytes[:m])!;
+	self.limit -= n;
+	return n;
+}
+
+fn char! LimitReader.read_byte(&self) @dynamic
+{
+	if (self.limit == 0) return IoError.EOF?;
+	defer try self.limit--;
+	return self.wrapped_stream.read_byte();
+}
+
+fn usz! LimitReader.available(&self) @inline @dynamic
+{
+	return self.limit;
+}

lib/std/io/stream/scanner.c3

@@ -0,0 +1,124 @@
+module std::io;
+
+struct Scanner (InStream)
+{
+	InStream wrapped_stream;
+	char[] buf;
+	usz pattern_idx;
+	usz read_idx;
+}
+
+/**
+ * Scanner provides a way to read delimited data (with newlines as the default).
+ * The supplied buffer must be at least as large as the expected data length
+ * including its pattern.
+ *
+ * @param [&in] stream "The stream to read data from."
+ * @require buffer.len > 0 "Non-empty buffer required."
+ **/
+fn void Scanner.init(&self, InStream stream, char[] buffer)
+{
+	*self = { .wrapped_stream = stream, .buf = buffer };
+}
+
+
+/**
+ * Return and clear any remaining unscanned data.
+ **/
+fn char[] Scanner.flush(&self) @dynamic
+{
+	assert(self.read_idx >= self.pattern_idx);
+	usz n = self.read_idx - self.pattern_idx;
+	char[] buf = self.buf[self.pattern_idx:n];
+	self.pattern_idx = 0;
+	self.read_idx = 0;
+	return buf;
+}
+
+fn void! Scanner.close(&self) @dynamic
+{
+	if (&self.wrapped_stream.close) return self.wrapped_stream.close();
+}
+
+/**
+ * Scan the stream for the next split character and return data up to the match.
+ * @require pattern.len > 0 "Non-empty pattern required."
+ * @require self.buf.len > pattern.len "Pattern too large."
+ **/
+fn char[]! Scanner.scan(&self, String pattern = "\n")
+{
+	if (self.read_idx == 0)
+	{
+		// First read.
+		self.read_idx = self.refill(self.buf)!;
+		self.pattern_idx = 0;
+	}
+	assert(self.read_idx >= self.pattern_idx);
+	usz n = self.read_idx - self.pattern_idx;
+	char[] buf = self.buf[self.pattern_idx:n];
+	if (try i = self.find(buf, pattern))
+	{
+		self.pattern_idx += i + pattern.len;
+		return buf[:i];
+	}
+	if (self.pattern_idx == 0 || self.read_idx < self.buf.len)
+	{
+		// Split pattern not found with maximized search, abort.
+		// Split pattern not found and already read as much as possible.
+		return SearchResult.MISSING?;
+	}
+	// Split pattern not found: maximize the search and try one more time.
+	self.buf[:n] = buf[..];
+	self.pattern_idx = 0;
+
+	buf = self.buf[n..];
+	usz p = self.refill(buf)!;
+	self.read_idx = n + p;
+
+	buf = buf[:p];
+	usz i = self.find(buf, pattern)!;
+	self.pattern_idx = n + i + pattern.len;
+
+	return self.buf[:n + i];
+}
+
+macro usz! Scanner.find(&self, buf, pattern) @private
+{
+	return ((String)buf).index_of(pattern);
+}
+
+macro usz! Scanner.refill(&self, buf) @private
+{
+	usz! n = self.wrapped_stream.read(buf);
+	if (catch err = n)
+	{
+		case IoError.EOF:
+			return SearchResult.MISSING?;
+		default:
+			return err?;
+	}
+	return n;
+}
+
+fn usz! Scanner.read(&self, char[] bytes) @dynamic
+{
+	usz n;
+	if (self.pattern_idx < self.read_idx)
+	{
+		n = min(bytes.len, self.read_idx - self.pattern_idx);
+		bytes[:n] = self.buf[self.pattern_idx:n];
+		self.pattern_idx += n;
+		bytes = bytes[n..];
+	}
+	n += self.wrapped_stream.read(bytes)!;
+	return n;
+}
+
+fn char! Scanner.read_byte(&self) @dynamic
+{
+	if (self.pattern_idx < self.read_idx)
+	{
+		return self.buf[self.pattern_idx++];
+	}
+	return self.wrapped_stream.read_byte();
+}

lib/std/libc/libc.c3

@@ -0,0 +1,663 @@
+// Copyright (c) 2021 Christoffer Lerno. All rights reserved.
+// Use of this source code is governed by the MIT license
+// a copy of which can be found in the LICENSE_STDLIB file.
+module libc;
+
+
+// Constants need to be per os/arch
+const int EXIT_FAILURE = 1;
+const int EXIT_SUCCESS = 0;
+const int RAND_MAX = 0x7fffffff;
+
+struct DivResult
+{
+	CInt quot;
+	CInt rem;
+}
+
+struct LongDivResult
+{
+	CLong quot;
+	CLong rem;
+}
+
+fn Errno errno()
+{
+	return (Errno)os::errno();
+}
+
+fn void errno_set(Errno e)
+{
+	os::errno_set((int)e);
+}
+
+distinct Errno = inline CInt;
+def TerminateFunction = fn void();
+def CompareFunction = fn int(void*, void*);
+def JmpBuf = uptr[$$JMP_BUF_SIZE];
+def Fd = CInt;
+def Fpos_t = long; // TODO make sure fpos is correct on all targets.
+def SignalFunction = fn void(CInt);
+
+const CInt SIGHUP = 1;
+const CInt SIGINT = 2;
+const CInt SIGQUIT = 3;
+const CInt SIGILL = 4;
+const CInt SIGTRAP = 5;
+const CInt SIGABTR = 6;
+const CInt SIGBUS = BSD_FLAVOR_SIG ? 10 : 7; // Or Mips
+const CInt SIGFPE = 8;
+const CInt SIGKILL = 9;
+const CInt SIGSEGV = 11;
+const CInt SIGSYS = BSD_FLAVOR_SIG ? 12 : 31;
+const CInt SIGPIPE = 13;
+const CInt SIGALRM = 14;
+const CInt SIGTERM = 15;
+const CInt SIGURG = BSD_FLAVOR_SIG ? 16 : 23;
+const CInt SIGSTOP = BSD_FLAVOR_SIG ? 17 : 19;
+const CInt SIGTSTP = BSD_FLAVOR_SIG ? 18 : 20;
+const CInt SIGCONT = BSD_FLAVOR_SIG ? 19 : 18;
+const CInt SIGCHLD = BSD_FLAVOR_SIG ? 20 : 17;
+
+const bool BSD_FLAVOR_SIG @local = env::OPENBSD || env::DARWIN || env::FREEBSD || env::NETBSD;
+
+def Time_t = $typefrom(env::WIN32 ? long.typeid : CLong.typeid);
+def Off_t = $typefrom(env::WIN32 ? int.typeid : usz.typeid);
+
+struct Timespec
+{
+	Time_t tv_sec;
+	CLong tv_nsec;
+}
+
+module libc @if(env::LIBC);
+
+extern fn void abort();
+extern fn CInt abs(CInt n);
+extern fn ZString asctime(Tm* timeptr);
+extern fn ZString asctime_r(Tm* timeptr, char* buf);
+extern fn CInt atexit(TerminateFunction func);
+extern fn double atof(char* str);
+extern fn int atoi(char* str);
+extern fn CLongLong atoll(char* str);
+extern fn void bsearch(void* key, void *base, usz items, usz size, CompareFunction compare);
+extern fn void* calloc(usz count, usz size);
+extern fn void clearerr(CFile stream);
+extern fn Clock_t clock();
+extern fn CInt close(CInt fd) @if(!env::WIN32);
+extern fn double difftime(Time_t time1, Time_t time2) @if(!env::WIN32);
+extern fn DivResult div(CInt numer, CInt denom);
+extern fn void exit(CInt status);
+extern fn CInt fclose(CFile stream);
+extern fn CFile fdopen(CInt fd, ZString mode) @if(!env::WIN32);
+extern fn CInt feof(CFile stream);
+extern fn CInt ferror(CFile stream);
+extern fn CInt fflush(CFile stream);
+extern fn CInt fgetc(CFile stream);
+extern fn ZString fgets(char* string, CInt n, CFile stream);
+extern fn CInt fgetpos(CFile stream, Fpos_t* pos);
+extern fn Fd fileno(CFile stream) @if(!env::WIN32);
+extern fn CFile fopen(ZString filename, ZString mode);
+extern fn CInt fprintf(CFile stream, ZString format, ...);
+extern fn CInt fputc(CInt c, CFile stream);
+extern fn CInt fputs(ZString string, CFile stream);
+extern fn usz fread(void* ptr, usz size, usz nmemb, CFile stream);
+extern fn void* free(void*);
+extern fn CFile freopen(ZString filename, ZString mode, CFile stream);
+extern fn CInt fscanf(CFile stream, ZString format, ...);
+extern fn CInt fseek(CFile stream, SeekIndex offset, CInt whence) @if(!env::WIN32);
+extern fn CInt fsetpos(CFile stream, Fpos_t* pos);
+extern fn SeekIndex ftell(CFile stream) @if(!env::WIN32);
+extern fn usz fwrite(void* ptr, usz size, usz nmemb, CFile stream);
+extern fn CInt getc(CFile stream);
+extern fn CInt getchar();
+extern fn ZString getenv(ZString name);
+extern fn ZString gets(char* buffer);
+extern fn Tm* gmtime(Time_t* timer);
+extern fn Tm* gmtime_r(Time_t *timer, Tm* buf) @if(!env::WIN32);
+extern fn CLong labs(CLong x);
+extern fn LongDivResult ldiv(CLong number, CLong denom);
+extern fn Tm* localtime(Time_t* timer);
+extern fn Tm* localtime_r(Time_t* timer, Tm* result) @if(!env::WIN32);
+extern fn void longjmp(JmpBuf* buffer, CInt value);
+extern fn void* malloc(usz size);
+extern fn void* memchr(void* str, CInt c, usz n);
+extern fn CInt memcmp(void* buf1, void* buf2, usz count);
+extern fn void* memcpy(void* dest, void* src, usz n);
+extern fn void* memmove(void* dest, void* src, usz n);
+extern fn void* memset(void* dest, CInt value, usz n);
+extern fn Time_t* mktime(Tm* time) @if(!env::WIN32);
+extern fn void perror(ZString string);
+extern fn CInt printf(ZString format, ...);
+extern fn CInt putc(CInt c, CFile stream);
+extern fn CInt putchar(CInt c);
+extern fn CInt puts(ZString str);
+extern fn void qsort(void* base, usz items, usz size, CompareFunction compare);
+extern fn CInt raise(CInt signal);
+extern fn CInt rand();
+extern fn isz read(Fd fd, void* buf, usz nbyte) @if(!env::WIN32);
+extern fn void* realloc(void* ptr, usz size);
+extern fn CInt remove(ZString filename);
+extern fn CInt rename(ZString old_name, ZString new_name);
+extern fn void rewind(CFile stream);
+extern fn CInt scanf(ZString format, ...);
+extern fn void setbuf(CFile stream, char* buffer);
+extern fn int setenv(ZString name, ZString value, CInt overwrite);
+extern fn CInt setjmp(JmpBuf* buffer) @if(!env::WIN32);
+extern fn void setvbuf(CFile stream, char* buf, CInt type, usz size);
+extern fn SignalFunction signal(CInt sig, SignalFunction function);
+extern fn CInt snprintf(char* buffer, usz size, ZString format, ...);
+extern fn CInt sprintf(char* buffer, ZString format, ...);
+extern fn void srand(uint seed);
+extern fn CInt sscanf(char* buffer, ZString format, ...);
+extern fn ZString strcat(ZString dest, ZString src);
+extern fn char* strchr(char* str, CInt c);
+extern fn CInt strcmp(ZString str1, ZString str2);
+extern fn CInt strcoll(ZString str1, ZString str2);
+extern fn usz strcspn(ZString str1, ZString str2);
+extern fn ZString strcpy(ZString dst, ZString src);
+extern fn ZString strerror(CInt errn);
+extern fn usz strftime(char* dest, usz maxsize, ZString format, Tm* timeptr);
+extern fn usz strlen(ZString str);
+extern fn ZString strncat(char* dest, char* src, usz n);
+extern fn CInt strncmp(char* str1, char* str2, usz n);
+extern fn char* strncpy(char* dst, char* src, usz n);
+extern fn CULong stroul(char* str, char** endptr, int base);
+extern fn char* strpbrk(ZString str1, ZString str2);
+extern fn usz strspn(ZString str1, ZString str2);
+extern fn ZString strptime(char* buf, ZString format, Tm* tm);
+extern fn char* strrchr(ZString str, CInt c);
+extern fn char* strstr(ZString haystack, ZString needle);
+extern fn double strtod(char* str, char** endptr);
+extern fn float strtof(char* str, char** endptr);
+extern fn ZString strtok(ZString str, ZString delim);
+extern fn CLong strtol(char* str, char** endptr, CInt base);
+extern fn CULong strtul(char* str, char** endptr, CInt base);
+extern fn usz strxfrm(char* dest, ZString src, usz n);
+extern fn CInt system(ZString str);
+extern fn Time_t timegm(Tm *timeptr) @if(!env::WIN32);
+extern fn ZString tmpnam(ZString str);
+extern fn CInt ungetc(CInt c, CFile stream);
+extern fn CInt unsetenv(ZString name);
+extern fn isz write(Fd fd, void* buffer, usz count) @if(!env::WIN32);
+
+extern fn CFile fmemopen(void* ptr, usz size, ZString mode);
+extern fn isz getline(char** linep, usz* linecapp, CFile stream);
+extern fn CInt timespec_get(TimeSpec* ts, CInt base);
+extern fn CInt nanosleep(TimeSpec* req, TimeSpec* remaining);
+extern fn ZString ctime(Time_t *timer);
+extern fn Time_t time(Time_t *timer);
+
+const CInt STDIN_FD = 0;
+const CInt STDOUT_FD = 1;
+const CInt STDERR_FD = 2;
+
+module libc @if(env::LINUX);
+extern CFile __stdin @extern("stdin");
+extern CFile __stdout @extern("stdout");
+extern CFile __stderr @extern("stderr");
+extern fn usz malloc_usable_size(void* ptr);
+macro usz malloc_size(void* ptr) => malloc_usable_size(ptr);
+extern fn void* aligned_alloc(usz align, usz size);
+macro CFile stdin() => __stdin;
+macro CFile stdout() => __stdout;
+macro CFile stderr() => __stderr;
+
+module libc @if(env::DARWIN);
+extern CFile __stdinp;
+extern CFile __stdoutp;
+extern CFile __stderrp;
+extern fn usz malloc_size(void* ptr);
+extern fn void* aligned_alloc(usz align, usz size);
+macro CFile stdin() => __stdinp;
+macro CFile stdout() => __stdoutp;
+macro CFile stderr() => __stderrp;
+
+module libc @if(env::WIN32);
+macro usz malloc_size(void* ptr) => _msize(ptr);
+macro CFile stdin() => __acrt_iob_func(STDIN_FD);
+macro CFile stdout() => __acrt_iob_func(STDOUT_FD);
+macro CFile stderr() => __acrt_iob_func(STDERR_FD);
+
+module libc @if(env::LIBC && !env::WIN32 && !env::LINUX && !env::DARWIN);
+macro CFile stdin() { return (CFile*)(uptr)STDIN_FD; }
+macro CFile stdout() { return (CFile*)(uptr)STDOUT_FD; }
+macro CFile stderr() { return (CFile*)(uptr)STDERR_FD; }
+
+module libc @if(!env::LIBC);
+
+fn void longjmp(JmpBuf* buffer, CInt value) @weak @extern("longjmp") @nostrip
+{
+	unreachable("longjmp unavailable");
+}
+
+fn CInt setjmp(JmpBuf* buffer) @weak @extern("setjmp") @nostrip
+{
+	unreachable("setjmp unavailable");
+}
+
+fn void* malloc(usz size) @weak @extern("malloc") @nostrip
+{
+	unreachable("malloc unavailable");
+}
+fn void* calloc(usz count, usz size) @weak @extern("calloc") @nostrip
+{
+	unreachable("calloc unavailable");
+}
+fn void* free(void*) @weak @extern("free")
+{
+	unreachable("free unavailable");
+}
+
+fn void* realloc(void* ptr, usz size) @weak @extern("realloc") @nostrip
+{
+	unreachable("realloc unavailable");
+}
+
+fn void* memcpy(void* dest, void* src, usz n) @weak @extern("memcpy") @nostrip
+{
+	for (usz i = 0; i < n; i++) ((char*)dest)[i] = ((char*)src)[i];
+	return dest;
+}
+
+fn void* memmove(void* dest, void* src, usz n) @weak @extern("memmove") @nostrip
+{
+	return memcpy(dest, src, n) @inline;
+}
+
+fn void* memset(void* dest, CInt value, usz n) @weak @extern("memset") @nostrip
+{
+	for (usz i = 0; i < n; i++) ((char*)dest)[i] = (char)value;
+	return dest;
+}
+
+fn int fseek(CFile stream, SeekIndex offset, int whence) @weak @extern("fseek") @nostrip
+{
+	unreachable("'fseek' not available.");
+}
+fn CFile fopen(ZString filename, ZString mode) @weak @extern("fopen") @nostrip
+{
+	unreachable("'fopen' not available.");
+}
+
+fn CFile freopen(ZString filename, ZString mode, CFile stream) @weak @extern("fopen") @nostrip
+{
+	unreachable("'freopen' not available.");
+}
+
+fn usz fwrite(void* ptr, usz size, usz nmemb, CFile stream) @weak @extern("fwrite") @nostrip
+{
+	unreachable("'fwrite' not available.");
+}
+
+fn usz fread(void* ptr, usz size, usz nmemb, CFile stream) @weak @extern("fread") @nostrip
+{
+	unreachable("'fread' not available.");
+}
+
+fn CFile fclose(CFile) @weak @extern("fclose") @nostrip
+{
+	unreachable("'fclose' not available.");
+}
+
+fn int fflush(CFile stream) @weak @extern("fflush") @nostrip
+{
+	unreachable("'fflush' not available.");
+}
+
+fn int fputc(int c, CFile stream) @weak @extern("fputc") @nostrip
+{
+	unreachable("'fputc' not available.");
+}
+
+fn char* fgets(ZString str, int n, CFile stream) @weak @extern("fgets") @nostrip
+{
+	unreachable("'fgets' not available.");
+}
+
+fn int fgetc(CFile stream) @weak @extern("fgetc") @nostrip
+{
+	unreachable("'fgetc' not available.");
+}
+
+fn int feof(CFile stream) @weak @extern("feof") @nostrip
+{
+	unreachable("'feof' not available.");
+}
+
+fn int putc(int c, CFile stream) @weak @extern("putc") @nostrip
+{
+	unreachable("'putc' not available.");
+}
+fn int putchar(int c) @weak @extern("putchar") @nostrip
+{
+	unreachable("'putchar' not available.");
+}
+fn int puts(ZString str) @weak @extern("puts") @nostrip
+{
+	unreachable("'puts' not available.");
+}
+
+module libc;
+
+// stdio
+
+def CFile = void*;
+
+
+const HAS_MALLOC_SIZE = env::LINUX || env::WIN32 || env::DARWIN;
+
+// The following needs to be set per arch+os
+// For now I have simply pulled the defaults from MacOS
+const int SEEK_SET = 0;
+const int SEEK_CUR = 1;
+const int SEEK_END = 2;
+const int _IOFBF = 0; // Fully buffered
+const int _IOLBF = 1; // Line buffered
+const int _IONBF = 2; // Unbuffered
+const int BUFSIZ = 1024;
+const int EOF = -1;
+const int FOPEN_MAX = 20;
+const int FILENAME_MAX = 1024;
+
+macro bool libc_S_ISTYPE(value, mask) @builtin => (value & S_IFMT) == mask;
+const S_IFMT   = 0o170000; // type of file mask
+const S_IFIFO  = 0o010000; // named pipe (fifo)
+const S_IFCHR  = 0o020000; // character special
+const S_IFDIR  = 0o040000; // directory
+const S_IFBLK  = 0o060000; // block special
+const S_IFREG  = 0o100000; // regular
+const S_IFLNK  = 0o120000; // symbolic link
+const S_IFSOCK = 0o140000; // socket
+const S_ISUID  = 0o004000; // Set user id on execution
+const S_ISGID  = 0o002000; // Set group id on execution
+const S_ISVTX  = 0o001000; // Save swapped text even after use
+const S_IRUSR  = 0o000400; // Read permission, owner
+const S_IWUSR  = 0o000200; // Write permission, owner
+const S_IXUSR  = 0o000100; // Execute/search permission, owner
+
+def SeekIndex = CLong;
+
+// vsprintf vprintf not supported
+
+// time.h
+
+struct Tm
+{
+	CInt	tm_sec;		// seconds after the minute [0-60]
+	CInt	tm_min;		// minutes after the hour [0-59]
+	CInt	tm_hour;	// hours since midnight [0-23]
+	CInt	tm_mday;	// day of the month [1-31]
+	CInt	tm_mon;		// months since January [0-11]
+	CInt	tm_year;	// years since 1900
+	CInt	tm_wday;	// days since Sunday [0-6]
+	CInt	tm_yday;	// days since January 1 [0-365]
+	CInt	tm_isdst;	// Daylight Savings Time flag
+	TimeOffset tm_gmtoff @if(!env::WIN32);	/* offset from UTC in seconds */
+	char *tm_zone @if(!env::WIN32);	/* timezone abbreviation */
+	CInt tm_nsec @if(env::WASI);
+}
+
+struct TimeSpec
+{
+	Time_t s;
+	ulong ns @if(env::WIN32);
+	CLong ns @if(!env::WIN32);
+}
+
+
+def Clock_t = int @if(env::WIN32);
+def Clock_t = CULong @if(!env::WIN32);
+
+def TimeOffset = int @if(env::WASI) ;
+def TimeOffset = CLong @if(!env::WASI) ;
+
+const int TIME_UTC = 1;
+
+
+// Likely wrong, must be per platform.
+const CLOCKS_PER_SEC = 1000000;
+
+module libc::errno;
+
+const Errno OK                       = 0;
+const Errno EPERM                    = 1; // Operation not permitted
+const Errno ENOENT                   = 2; // No such file or directory
+const Errno ESRCH                    = 3; // No such process
+const Errno EINTR                    = 4; // Interrupted system call
+const Errno EIO                      = 5; // I/O error
+const Errno ENXIO                    = 6; // No such device or address
+const Errno E2BIG                    = 7; // Argument list too long
+const Errno ENOEXEC                  = 8; // Exec format error
+const Errno EBADF                    = 9; // Bad file number
+const Errno ECHILD                   = 10; // No child processes
+
+
+const Errno EAGAIN @if(env::DARWIN)  = 35; // Try again Macos
+const Errno EAGAIN @if(!env::DARWIN) = 11; // Try again
+
+const Errno ENOMEM                   = 12; // Out of memory
+const Errno EACCES                   = 13; // Permission denied
+const Errno EFAULT                   = 14; // Bad address
+const Errno ENOTBLK                  = 15; // Block device required, not on Win32
+const Errno EBUSY                    = 16; // Device or resource busy
+const Errno EEXIST                   = 17; // File exists
+const Errno EXDEV                    = 18; // Cross-device link
+const Errno ENODEV                   = 19; // No such device
+const Errno ENOTDIR                  = 20; // Not a directory
+const Errno EISDIR                   = 21; // Is a directory
+const Errno EINVAL                   = 22; // Invalid argument
+const Errno ENFILE                   = 23; // File table overflow
+const Errno EMFILE                   = 24; // Too many open files
+const Errno ENOTTY                   = 25; // Not a typewriter
+const Errno ETXTBSY                  = 26; // Text file busy, not on Win32
+const Errno EFBIG                    = 27; // File too large
+const Errno ENOSPC                   = 28; // No space left on device
+const Errno ESPIPE                   = 29; // Illegal seek
+const Errno EROFS                    = 30; // Read-only file system
+const Errno EMLINK                   = 31; // Too many links
+const Errno EPIPE                    = 32; // Broken pipe
+const Errno EDOM                     = 33; // Math argument out of domain of func
+const Errno ERANGE                   = 34; // Math result not representable
+
+// https://developer.apple.com/library/archive/documentation/System/Conceptual/ManPages_iPhoneOS/man2/intro.2.html
+module libc::errno @if(env::DARWIN);
+const Errno EWOULDBLOCK              = EAGAIN; // Operation would block
+const Errno EDEADLK                  = 11; // Resource deadlock would occur
+const Errno EINPROGRESS              = 36; // Operation now in progress
+const Errno EALREADY                 = 37; // Operation already in progress
+const Errno ENOTSOCK                 = 38; // Socket operation on non-socket
+const Errno EDESTADDRREQ             = 39; // Destination address required
+const Errno EMSGSIZE                 = 40; // Message too long
+const Errno EPROTOTYPE               = 41; // Protocol wrong type for socket
+const Errno ENOPROTOOPT              = 42; // Protocol not available
+const Errno EPROTONOSUPPORT          = 43; // Protocol not supported
+const Errno ESOCKTNOSUPPORT          = 44; // Socket type not supported
+const Errno ENOTSUP                  = 45; // Not supported
+const Errno EPFNOSUPPORT             = 46; // Protocol family not supported
+const Errno EAFNOSUPPORT             = 47; // Address family not supported by protocol family
+const Errno EADDRINUSE               = 48; // Address already in use
+const Errno EADDRNOTAVAIL            = 49; // Cannot assign requested address
+const Errno ENETDOWN                 = 50; // Network is down
+const Errno ENETUNREACH              = 51; // Network is unreachable
+const Errno ENETRESET                = 52; // Network dropped connection on reset
+const Errno ECONNABORTED             = 53; // Software caused connection abort
+const Errno ECONNRESET               = 54; // Connection reset by peer
+const Errno ENOBUFS                  = 55; // No buffer space available
+const Errno EISCONN                  = 56; // Socket is already connected
+const Errno ENOTCONN                 = 57; // Socket is not connected
+const Errno ESHUTDOWN                = 58; // Cannot send after socket shutdown
+const Errno ETIMEDOUT                = 60; // Operation timed out
+const Errno ECONNREFUSED             = 61; // Connection refused
+const Errno ELOOP                    = 62; // Too many levels of symbolic links
+const Errno ENAMETOOLONG             = 63; // File name too long
+const Errno EHOSTDOWN                = 64; // Host is down
+const Errno EHOSTUNREACH             = 65; // No route to host
+const Errno ENOTEMPTY                = 66; // Directory not empty
+const Errno EPROCLIM                 = 67; // Too many processes
+const Errno EUSERS                   = 68; // Too many users
+const Errno EDQUOT                   = 69; // Disc quota exceeded
+const Errno ESTALE                   = 70; // Stale NFS file handle
+const Errno EBADRPC                  = 72; // RPC struct is bad
+const Errno ERPCMISMATCH             = 73; // RPC version wrong
+const Errno EPROGUNAVAIL             = 74; // RPC prog. not avail
+const Errno EPROGMISMATCH            = 75; // Program version wrong
+const Errno EPROCUNAVAIL             = 76; // Bad procedure for program
+const Errno ENOLCK                   = 77; // No locks available
+const Errno ENOSYS                   = 78; // Function not implemented
+const Errno EFTYPE                   = 79; // Inappropriate file type or format
+const Errno EAUTH                    = 80; // Authentication error
+const Errno ENEEDAUTH                = 81; // Need authenticator
+const Errno EPWROFF                  = 82; // Device power is off
+const Errno EDEVERR                  = 83; // Device error
+const Errno EOVERFLOW                = 84; // Value too large to be stored in data type
+const Errno EBADEXEC                 = 85; // Bad executable (or shared library)
+const Errno EBADARCH                 = 86; // Bad CPU type in executable
+const Errno ESHLIBVERS               = 87; // Shared library version mismatch
+const Errno EBADMACHO                = 88; // Malformed Mach-o file
+const Errno ECANCELED                = 89; // Operation canceled
+const Errno EIDRM                    = 90; // Identifier removed
+const Errno ENOMSG                   = 91; // No message of desired type
+const Errno EILSEQ                   = 92; // Illegal byte sequence
+const Errno ENOATTR                  = 93; // Attribute not found
+const Errno EBADMSG                  = 94; // Bad message
+const Errno EMULTIHOP                = 95; // Reserved
+const Errno ENODATA                  = 96; // No message available
+const Errno ENOLINK                  = 97; // Reserved
+const Errno ENOSR                    = 98; // No STREAM resources
+const Errno ENOSTR                   = 99; // Not a STREAM
+const Errno EPROTO                   = 100; // Protocol error
+const Errno ETIME                    = 101; // STREAM ioctl() timeout
+const Errno EOPNOTSUPP               = 102; // Operation not supported on socket
+
+module libc::errno @if(env::WIN32);
+const Errno EDEADLK                  = 36; // Resource deadlock would occur Win32
+const Errno ENAMETOOLONG             = 38; // File name too long Win32
+const Errno ENOTEMPTY                = 41; // Directory not empty
+const Errno ELOOP                    = 114; // Too many symbolic links encountered
+const Errno EOVERFLOW                = 132; // Value too large for defined data type
+const Errno ENETDOWN                 = 116; // Network is down
+const Errno ECONNRESET               = 108; // Connection reset by peer
+const Errno ENETUNREACH              = 118; // Network is unreachable
+const Errno ENETRESET                = 117; // Network dropped connection because of reset
+const Errno EOPNOTSUPP               = 130; // Operation not supported on transport endpoint
+const Errno ETIMEDOUT                = 138; // Connection timed out
+const Errno EALREADY                 = 103; // Operation already in progress
+const Errno EINPROGRESS              = 112; // Operation now in progress Win32
+const Errno EDQUOT                   = -122; // Quota exceeded, not in Win32
+const Errno EWOULDBLOCK              = 140; // Operation would block
+
+module libc::errno @if(!env::WIN32 && !env::DARWIN);
+const Errno EDEADLK                  = 35; // Resource deadlock would occur Linux (others?)
+const Errno ENAMETOOLONG             = 36; // File name too long Linux (others?)
+const Errno ENOTEMPTY                = 39; // Directory not empty
+const Errno ELOOP                    = 40; // Too many symbolic links encountered
+const Errno EWOULDBLOCK              = EAGAIN; // Operation would block
+const Errno EOVERFLOW                = 75; // Value too large for defined data type
+const Errno ENOTSOCK                 = 88; // Socket operation on non-socket
+const Errno EOPNOTSUPP               = 95; // Operation not supported on transport endpoint
+const Errno EADDRINUSE               = 98; // Address already in use
+const Errno EADDRNOTAVAIL            = 99; // Cannot assign requested address
+const Errno ENETDOWN                 = 100; // Network is down
+const Errno ENETUNREACH              = 101; // Network is unreachable
+const Errno ENETRESET                = 102; // Network dropped connection because of reset
+const Errno ECONNRESET               = 104; // Connection reset by peer
+const Errno EISCONN                  = 106; // Socket is already connected
+const Errno ETIMEDOUT                = 110; // Connection timed out
+const Errno ECONNREFUSED             = 111; // Connection refused
+const Errno EALREADY                 = 114; // Operation already in progress
+const Errno EINPROGRESS              = 115; // Operation now in progress
+const Errno EDQUOT                   = 122; // Quota exceeded
+
+
+/*
+const Errno ENOLCK = 37;  /* No record locks available */
+const Errno ENOSYS = 38;  /* Function not implemented */
+
+const Errno ENOMSG = 42;  /* No message of desired type */
+const Errno EIDRM = 43;  /* Identifier removed */
+const Errno ECHRNG = 44;  /* Channel number out of range */
+const Errno EL2NSYNC = 45;  /* Level 2 not synchronized */
+const Errno EL3HLT = 46;  /* Level 3 halted */
+const Errno EL3RST = 47;  /* Level 3 reset */
+const Errno ELNRNG = 48;  /* Link number out of range */
+const Errno EUNATCH = 49;  /* Protocol driver not attached */
+const Errno ENOCSI = 50;  /* No CSI structure available */
+const Errno EL2HLT = 51;  /* Level 2 halted */
+const Errno EBADE = 52;  /* Invalid exchange */
+const Errno EBADR = 53;  /* Invalid request descriptor */
+const Errno EXFULL = 54;  /* Exchange full */
+const Errno ENOANO = 55;  /* No anode */
+const Errno EBADRQC = 56;  /* Invalid request code */
+const Errno EBADSLT = 57;  /* Invalid slot */
+
+const Errno EBFONT = 59;  /* Bad font file format */
+const Errno ENOSTR = 60;  /* Device not a stream */
+const Errno ENODATA = 61;  /* No data available */
+const Errno ETIME = 62;  /* Timer expired */
+const Errno ENOSR = 63;  /* Out of streams resources */
+const Errno ENONET = 64;  /* Machine is not on the network */
+const Errno ENOPKG = 65;  /* Package not installed */
+const Errno EREMOTE = 66;  /* Object is remote */
+const Errno ENOLINK = 67;  /* Link has been severed */
+const Errno EADV = 68;  /* Advertise error */
+const Errno ESRMNT = 69;  /* Srmount error */
+const Errno ECOMM = 70;  /* Communication error on send */
+const Errno EPROTO = 71;  /* Protocol error */
+const Errno EMULTIHOP = 72;  /* Multihop attempted */
+const Errno EDOTDOT = 73;  /* RFS specific error */
+const Errno EBADMSG = 74;  /* Not a data message */
+const Errno ENOTUNIQ = 76;  /* Name not unique on network */
+const Errno EBADFD = 77;  /* File descriptor in bad state */
+const Errno EREMCHG = 78;  /* Remote address changed */
+const Errno ELIBACC = 79;  /* Can not access a needed shared library */
+const Errno ELIBBAD = 80;  /* Accessing a corrupted shared library */
+const Errno ELIBSCN = 81;  /* .lib section in a.out corrupted */
+const Errno ELIBMAX = 82;  /* Attempting to link in too many shared libraries */
+const Errno ELIBEXEC = 83;  /* Cannot exec a shared library directly */
+const Errno EILSEQ = 84;  /* Illegal byte sequence */
+const Errno ERESTART = 85;  /* Interrupted system call should be restarted */
+const Errno ESTRPIPE = 86;  /* Streams pipe error */
+const Errno EUSERS = 87;  /* Too many users */
+const Errno ENOTSOCK = 88;  /* Socket operation on non-socket */
+const Errno EDESTADDRREQ = 89;  /* Destination address required */
+const Errno EMSGSIZE = 90;  /* Message too long */
+const Errno EPROTOTYPE = 91;  /* Protocol wrong type for socket */
+const Errno ENOPROTOOPT = 92;  /* Protocol not available */
+const Errno EPROTONOSUPPORT = 93;  /* Protocol not supported */
+const Errno ESOCKTNOSUPPORT = 94;  /* Socket type not supported */
+const Errno EPFNOSUPPORT = 96;  /* Protocol family not supported */
+const Errno EAFNOSUPPORT = 97;  /* Address family not supported by protocol */
+const Errno EADDRINUSE = 98;  /* Address already in use */
+const Errno ECONNABORTED = 103; /* Software caused connection abort */
+const Errno ENOBUFS = 105; /* No buffer space available */
+const Errno EISCONN = 106; /* Transport endpoint is already connected */
+const Errno ENOTCONN = 107; /* Transport endpoint is not connected */
+const Errno ESHUTDOWN = 108; /* Cannot send after transport endpoint shutdown */
+const Errno ETOOMANYREFS = 109; /* Too many references: cannot splice */
+const Errno ECONNREFUSED = 111; /* Connection refused */
+const Errno EHOSTDOWN = 112; /* Host is down */
+const Errno EHOSTUNREACH = 113; /* No route to host */
+*/
+
+
+/*
+const Errno ESTALE = 116; /* Stale NFS file handle */
+const Errno EUCLEAN = 117; /* Structure needs cleaning */
+const Errno ENOTNAM = 118; /* Not a XENIX named type file */
+const Errno ENAVAIL = 119; /* No XENIX semaphores available */
+const Errno EISNAM = 120; /* Is a named type file */
+const Errno EREMOTEIO = 121; /* Remote I/O error */
+
+const Errno ENOMEDIUM = 123; /* No medium found */
+const Errno EMEDIUMTYPE = 124; /* Wrong medium type */
+const Errno ECANCELED = 125; /* Operation Canceled */
+const Errno ENOKEY = 126; /* Required key not available */
+const Errno EKEYEXPIRED = 127; /* Key has expired */
+const Errno EKEYREVOKED = 128; /* Key has been revoked */
+const Errno EKEYREJECTED = 129; /* Key was rejected by service */
+
+const Errno EOWNERDEAD = 130; /* Owner died */
+const Errno ENOTRECOVERABLE = 131; /* State not recoverable */
+*/
+

lib/std/libc/libc_extra.c3

@@ -0,0 +1,22 @@
+module libc;
+import std::time;
+
+/**
+ * @require self >= 0
+ **/
+fn TimeSpec NanoDuration.to_timespec(self) @inline
+{
+	CLong ns = (CLong)(self % 1000_000_000);
+	Time_t sec = (Time_t)(self / 1000_000_000);
+	return { .s = sec, .ns = ns };
+}
+
+/**
+ * @require self >= 0
+ **/
+fn TimeSpec Duration.to_timespec(self) @inline
+{
+	CLong ns = (CLong)(1000 * (self % time::SEC));
+	Time_t sec = (Time_t)(self / time::SEC);
+	return { .s = sec, .ns = ns };
+}