From b2812665beaba5eb7484fd361d71d25c0ca2a4f5 Mon Sep 17 00:00:00 2001
From: Alexander Medvednikov <alexander@medvednikov.com>
Date: Wed, 15 Apr 2026 16:13:30 +0300
Subject: [PATCH] checker: fix invalid utf8 string warning in character literal
 (fixes #23741)

---
 vlib/v/checker/str.v                          | 59 ++++++++++++++++++-
 .../tests/invalid_utf8_escape_string.out      |  5 --
 .../tests/invalid_utf8_escape_string.vv       |  1 +
 3 files changed, 59 insertions(+), 6 deletions(-)

diff --git a/vlib/v/checker/str.v b/vlib/v/checker/str.v
index f55e19c45..0a4a43ec6 100644
--- a/vlib/v/checker/str.v
+++ b/vlib/v/checker/str.v
@@ -6,6 +6,7 @@ module checker
 import v.ast
 import v.token
 import encoding.utf8.validate
+import v.util
 
 fn (mut c Checker) get_default_fmt(ftyp ast.Type, typ ast.Type) u8 {
 	if ftyp.has_option_or_result() {
@@ -183,11 +184,67 @@ fn (mut c Checker) string_inter_lit(mut node ast.StringInterLiteral) ast.Type {
 
 const unicode_lit_overflow_message = 'unicode character exceeds max allowed value of 0x10ffff, consider using a unicode literal (\\u####)'
 
+fn is_source_char_escaped(source string, idx int) bool {
+	mut backslashes := 0
+	mut i := idx - 1
+	for i >= 0 && source[i] == `\\` {
+		backslashes++
+		i--
+	}
+	return (backslashes & 1) == 1
+}
+
+fn raw_string_literal_source(source string, approx_pos int) ?string {
+	if source.len == 0 {
+		return none
+	}
+	mut hint := approx_pos
+	if hint < 0 {
+		hint = 0
+	} else if hint >= source.len {
+		hint = source.len - 1
+	}
+	mut start := hint
+	for start >= 0 {
+		if source[start] in [`'`, `"`] && !is_source_char_escaped(source, start) {
+			quote := source[start]
+			is_raw := start > 0 && source[start - 1] == `r`
+			mut end := start + 1
+			for end < source.len {
+				if source[end] == quote && (is_raw || !is_source_char_escaped(source, end)) {
+					if end >= hint {
+						return source[start..end + 1]
+					}
+					break
+				}
+				end++
+			}
+		}
+		start--
+	}
+	return none
+}
+
+fn (c &Checker) source_string_literal_is_valid_utf8(node ast.StringLiteral) bool {
+	if node.pos.file_idx < 0 || node.pos.file_idx >= c.table.filelist.len {
+		return validate.utf8_string(node.val)
+	}
+	source := util.read_file(c.table.filelist[node.pos.file_idx]) or {
+		return validate.utf8_string(node.val)
+	}
+	raw_source := raw_string_literal_source(source, node.pos.pos) or {
+		return validate.utf8_string(node.val)
+	}
+	return validate.utf8_string(raw_source)
+}
+
 // unicode character literals are limited to a maximum value of 0x10ffff
 // https://stackoverflow.com/questions/52203351/why-unicode-is-restricted-to-0x10ffff
 @[direct_array_access]
 fn (mut c Checker) string_lit(mut node ast.StringLiteral) ast.Type {
-	valid_utf8 := validate.utf8_string(node.val)
+	// Validate the bytes that came from the source file, not the decoded string value.
+	// `\x..` escapes are allowed to produce arbitrary bytes intentionally.
+	valid_utf8 := c.source_string_literal_is_valid_utf8(node)
 	if !valid_utf8 {
 		c.note('invalid utf8 byte sequence in string literal', node.pos)
 	}
diff --git a/vlib/v/checker/tests/invalid_utf8_escape_string.out b/vlib/v/checker/tests/invalid_utf8_escape_string.out
index 965e7ae61..e69de29bb 100644
--- a/vlib/v/checker/tests/invalid_utf8_escape_string.out
+++ b/vlib/v/checker/tests/invalid_utf8_escape_string.out
@@ -1,5 +0,0 @@
-vlib/v/checker/tests/invalid_utf8_escape_string.vv:2:22: notice: invalid utf8 byte sequence in string literal
-    1 | fn main() {
-    2 |     println('\xF0\x81\xA0\x80')
-      |                         ~~~~~~
-    3 | }
diff --git a/vlib/v/checker/tests/invalid_utf8_escape_string.vv b/vlib/v/checker/tests/invalid_utf8_escape_string.vv
index 16deb72c1..793c4a030 100644
--- a/vlib/v/checker/tests/invalid_utf8_escape_string.vv
+++ b/vlib/v/checker/tests/invalid_utf8_escape_string.vv
@@ -1,3 +1,4 @@
 fn main() {
+	println('\xc0')
 	println('\xF0\x81\xA0\x80')
 }
-- 
2.39.5