From f34eb7d9f3b32b04a7a4e4da36ff48abcbc6da66 Mon Sep 17 00:00:00 2001
From: Aaron <2858172+lith-x@users.noreply.github.com>
Date: Mon, 23 Dec 2024 14:23:46 -0600
Subject: [PATCH] Prevent trailing _ for all numbers (#1706)

* moved _ check to scan_number_suffix
* Skipping underscore on list-operators command
* Disallow # from operator, update release notes
---
 releasenotes.md         |  2 ++
 src/compiler/compiler.c |  2 +-
 src/compiler/lexer.c    | 16 +++++-----------
 3 files changed, 8 insertions(+), 12 deletions(-)

diff --git a/releasenotes.md b/releasenotes.md
index a867b1ac1..f9f302d6f 100644
--- a/releasenotes.md
+++ b/releasenotes.md
@@ -17,6 +17,8 @@
 - Fix bug outputting exported functions without predefined extname.
 - Fix problem where crt1 was linked for dynamic libraries on Linux and BSD. #1710
 - Fix CRT detection on Arch Linux.
+- Fix lexer allowing a trailing underscore (_) with hex and binary literals.
+- Fix `--list-operators` CLI command printing underscore (_) and hash (#).
 
 ### Stdlib changes
 - Increase BitWriter.write_bits limit up to 32 bits.
diff --git a/src/compiler/compiler.c b/src/compiler/compiler.c
index ec532bfcf..0170235b0 100644
--- a/src/compiler/compiler.c
+++ b/src/compiler/compiler.c
@@ -939,7 +939,7 @@ void print_syntax(BuildOptions *options)
 			if (i == TOKEN_DOCS_START || i == TOKEN_DOCS_END) continue;
 			const char *name = token_type_to_string((TokenType)i);
 			char first_char = name[0];
-			if (first_char == '$' || first_char == '@'
+			if (first_char == '$' || first_char == '@' || first_char == '_' || first_char == '#'
 				|| (first_char >= 'a' && first_char <= 'z')
 				|| (first_char >= 'A' && first_char <= 'Z'))
 			{
diff --git a/src/compiler/lexer.c b/src/compiler/lexer.c
index 3fb0e36fe..e1659f6da 100644
--- a/src/compiler/lexer.c
+++ b/src/compiler/lexer.c
@@ -370,6 +370,11 @@ static inline bool scan_ident(Lexer *lexer, TokenType normal, TokenType const_to
  */
 static bool scan_number_suffix(Lexer *lexer, bool *is_float)
 {
+	if (prev(lexer) == '_')
+	{
+		backtrack(lexer);
+		return add_error_token_at_current(lexer, "The number ended with '_', which isn't allowed, please remove it.");
+	}
 	char c = peek(lexer);
 	if (!char_is_alphanum_(c)) return true;
 	switch (c | 32)
@@ -535,11 +540,6 @@ static inline bool scan_hex(Lexer *lexer)
 		is_float = true;
 		if (!scan_exponent(lexer)) return false;
 	}
-	if (prev(lexer) == '_')
-	{
-		backtrack(lexer);
-		return add_error_token_at_current(lexer, "The number ended with '_', which isn't allowed, please remove it.");
-	}
 	if (!scan_number_suffix(lexer, &is_float)) return false;
 	return new_token(lexer, is_float ? TOKEN_REAL : TOKEN_INTEGER, lexer->lexing_start);
 }
@@ -581,12 +581,6 @@ static inline bool scan_dec(Lexer *lexer)
 		is_float = true;
 		if (!scan_exponent(lexer)) return false;
 	}
-
-	if (prev(lexer) == '_')
-	{
-		backtrack(lexer);
-		return add_error_token_at_current(lexer, "The number ended with '_', which isn't allowed, please remove it.");
-	}
 	if (!scan_number_suffix(lexer, &is_float)) return false;
 	return new_token(lexer, is_float ? TOKEN_REAL : TOKEN_INTEGER, lexer->lexing_start);
 }