From 634b579b63d8c3d098c47005fa4138c0e0e9f117 Mon Sep 17 00:00:00 2001
From: Abdul Wahab <abdulwahabone@gmail.com>
Date: Thu, 10 Sep 2026 11:16:24 +0500
Subject: [PATCH] Sampled contrast: url() is a token, not a suffix of an
 identifier

The sheet url rewrite and its comment scan both matched `url(` inside a
longer identifier, so `myurl(a /* url(x.png) */)` read as an unquoted
url and the comment inside it as url content. Both now require a
non-identifier byte before `url(`, the way the tokenizer does; the
regex re-emits that byte unchanged. A test pins a custom function
wrapping a comment beside a real url on the same rule.

AI assistance: Claude Code (Claude Fable 5.1), on maintainer instruction.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
---
 crates/html/src/cascade/build.rs   | 16 +++++++++++++---
 crates/html/tests/cascade_units.rs |  8 ++++++++
 2 files changed, 21 insertions(+), 3 deletions(-)

diff --git a/crates/html/src/cascade/build.rs b/crates/html/src/cascade/build.rs
index e9ba85e3c..235776e35 100644
--- a/crates/html/src/cascade/build.rs
+++ b/crates/html/src/cascade/build.rs
@@ -145,8 +145,10 @@ pub fn collect_static_css_text(
 // whitespace, quotes, parens, braces, or semicolons (CSS syntax), so a
 // stray `url(` can never pair with a `)` in a later rule.
 static CSS_URL_RE: Lazy<Regex> = Lazy::new(|| {
+    // The leading group keeps `url(` from matching inside a longer
+    // identifier such as `myurl(`; the closure emits it unchanged.
     Regex::new(
-        r#"(?i)url\(\s*(?:"((?:[^"\\\n\r]|\\.)*)"|'((?:[^'\\\n\r]|\\.)*)'|([^)"'(\s{};]*))\s*\)"#,
+        r#"(?i)(^|[^A-Za-z0-9_-])url\(\s*(?:"((?:[^"\\\n\r]|\\.)*)"|'((?:[^'\\\n\r]|\\.)*)'|([^)"'(\s{};]*))\s*\)"#,
     )
     .expect("CSS_URL_RE")
 });
@@ -177,6 +179,12 @@ pub fn rewrite_sheet_urls(css: &str, sheet_dir: &str, page_dir: &str) -> String
 /// `url()` is content, not a comment opener, and an unclosed comment runs to
 /// the end of the sheet. Every span boundary sits on an ASCII byte, so the
 /// spans are valid `str` indices.
+/// A byte that can continue a CSS identifier, so `myurl(` is a custom
+/// function and not the `url(` token.
+fn is_ident_byte(b: u8) -> bool {
+    b.is_ascii_alphanumeric() || b == b'-' || b == b'_' || b >= 0x80
+}
+
 fn comment_spans(css: &str) -> Vec<(usize, usize)> {
     #[derive(Clone, Copy, PartialEq)]
     enum State {
@@ -207,6 +215,7 @@ fn comment_spans(css: &str) -> Vec<(usize, usize)> {
                     && css[i..]
                         .get(..4)
                         .is_some_and(|s| s.eq_ignore_ascii_case("url("))
+                    && !(i > 0 && is_ident_byte(bytes[i - 1]))
                 {
                     i += 4;
                     while i < bytes.len() && bytes[i].is_ascii_whitespace() {
@@ -249,7 +258,8 @@ fn rewrite_code_urls(css: &str, sheet_dir: &str, page_dir: &str) -> String {
     CSS_URL_RE
         .replace_all(css, |caps: &regex::Captures| {
             let whole = caps.get(0).map(|m| m.as_str()).unwrap_or("");
-            let (target, quote) = match (caps.get(1), caps.get(2), caps.get(3)) {
+            let before = caps.get(1).map(|m| m.as_str()).unwrap_or("");
+            let (target, quote) = match (caps.get(2), caps.get(3), caps.get(4)) {
                 (Some(m), _, _) => (m.as_str(), "\""),
                 (_, Some(m), _) => (m.as_str(), "'"),
                 (_, _, Some(m)) => (js::trim(m.as_str()), ""),
@@ -282,7 +292,7 @@ fn rewrite_code_urls(css: &str, sheet_dir: &str, page_dir: &str) -> String {
                     .chars()
                     .any(|c| c.is_whitespace() || matches!(c, '(' | ')' | '"' | '\''));
             let quote = if needs_quotes { "\"" } else { quote };
-            format!("url({quote}{relative}{suffix}{quote})")
+            format!("{before}url({quote}{relative}{suffix}{quote})")
         })
         .into_owned()
 }
diff --git a/crates/html/tests/cascade_units.rs b/crates/html/tests/cascade_units.rs
index 513d40e96..3131834ec 100644
--- a/crates/html/tests/cascade_units.rs
+++ b/crates/html/tests/cascade_units.rs
@@ -406,4 +406,12 @@ fn linked_sheet_urls_are_rewritten_page_relative() {
         out.contains("/* real url( */ .n { background: url(css/n.png) }"),
         "{out}"
     );
+    // `myurl(` is a custom function, not the url token: its comment stays a
+    // comment and its argument is not rewritten.
+    let custom = ".q { mask: myurl(a /* url(x.png) */); background: url(q.png) }";
+    let out = rewrite_sheet_urls(custom, "/site/css", "/site");
+    assert_eq!(
+        out,
+        ".q { mask: myurl(a /* url(x.png) */); background: url(css/q.png) }"
+    );
 }