test(anchor): harden JS helper extraction

Co-Authored-By: Paperclip <noreply@paperclip.ing>

tests/test_anchor_fallback_ownership.py

@@ -44,12 +44,20 @@ def _run_node_script(script: str) -> str:
     node = shutil.which("node")
     if not node:
         pytest.skip("node executable is required for JavaScript behavior checks")
-    result = subprocess.run(
-        [node, "-e", script],
-        cwd=ROOT,
-        text=True,
-        capture_output=True,
-    )
+    try:
+        result = subprocess.run(
+            [node, "-e", script],
+            cwd=ROOT,
+            text=True,
+            capture_output=True,
+            timeout=10,
+        )
+    except subprocess.TimeoutExpired as exc:
+        pytest.fail(
+            "node behavior check timed out"
+            f"\nstdout:\n{exc.stdout or '<empty>'}"
+            f"\nstderr:\n{exc.stderr or '<empty>'}",
+        )
     if result.returncode:
         pytest.fail(
             "node behavior check failed"
@@ -60,6 +68,54 @@ def _run_node_script(script: str) -> str:
     return result.stdout.strip()
 
 
+def _is_js_identifier_char(char: str) -> bool:
+    return char.isalnum() or char in {"_", "$"}
+
+
+def _previous_significant_js_token(src: str, idx: int) -> str:
+    idx -= 1
+    while idx >= 0 and src[idx].isspace():
+        idx -= 1
+    if idx < 0:
+        return ""
+    if _is_js_identifier_char(src[idx]):
+        end = idx + 1
+        while idx >= 0 and _is_js_identifier_char(src[idx]):
+            idx -= 1
+        return src[idx + 1 : end]
+    return src[idx]
+
+
+def _looks_like_js_regex_literal_start(src: str, idx: int) -> bool:
+    if src[idx] != "/" or src.startswith(("//", "/*"), idx):
+        return False
+    previous = _previous_significant_js_token(src, idx)
+    if previous in {"return", "throw", "case", "delete", "typeof", "void", "yield"}:
+        return True
+    return previous in {"", "(", "[", "{", "=", ":", ",", ";", "!", "?", "&", "|"}
+
+
+def _skip_js_regex_literal(src: str, idx: int) -> int:
+    assert src[idx] == "/", f"expected regex literal at {idx}"
+    idx += 1
+    in_class = False
+    while idx < len(src):
+        if src[idx] == "\\":
+            idx += 2
+            continue
+        if src[idx] == "[":
+            in_class = True
+        elif src[idx] == "]":
+            in_class = False
+        elif src[idx] == "/" and not in_class:
+            idx += 1
+            while idx < len(src) and src[idx].isalpha():
+                idx += 1
+            return idx
+        idx += 1
+    raise AssertionError("JavaScript regex literal did not close")
+
+
 def _skip_js_string_or_comment(src: str, idx: int) -> int:
     if src.startswith("//", idx):
         end = src.find("\n", idx + 2)
@@ -68,6 +124,8 @@ def _skip_js_string_or_comment(src: str, idx: int) -> int:
         end = src.find("*/", idx + 2)
         assert end != -1, "JavaScript block comment did not close"
         return end + 2
+    if _looks_like_js_regex_literal_start(src, idx):
+        return _skip_js_regex_literal(src, idx)
     quote = src[idx]
     if quote == "`":
         return _skip_js_template_literal(src, idx)
@@ -120,10 +178,46 @@ def _matching_delimiter(src: str, open_idx: int, opener: str, closer: str) -> in
     raise AssertionError(f"{opener}{closer} delimiter did not close")
 
 
+def _skip_js_whitespace(src: str, idx: int) -> int:
+    while idx < len(src) and src[idx].isspace():
+        idx += 1
+    return idx
+
+
+def _find_function_declaration(src: str, name: str) -> tuple[int, int]:
+    idx = 0
+    while idx < len(src):
+        next_idx = _skip_js_string_or_comment(src, idx)
+        if next_idx != idx:
+            idx = next_idx
+            continue
+        if not src.startswith("function", idx):
+            idx += 1
+            continue
+        before = src[idx - 1] if idx else ""
+        after_keyword = idx + len("function")
+        after = src[after_keyword] if after_keyword < len(src) else ""
+        if _is_js_identifier_char(before) or _is_js_identifier_char(after):
+            idx += 1
+            continue
+        name_start = _skip_js_whitespace(src, after_keyword)
+        name_end = name_start + len(name)
+        if src[name_start:name_end] != name:
+            idx += 1
+            continue
+        after_name = src[name_end] if name_end < len(src) else ""
+        if _is_js_identifier_char(after_name):
+            idx += 1
+            continue
+        params_open = _skip_js_whitespace(src, name_end)
+        if params_open < len(src) and src[params_open] == "(":
+            return idx, params_open
+        idx += 1
+    raise AssertionError(f"{name} not found")
+
+
 def _function_source(src: str, name: str) -> str:
-    start = src.find(f"function {name}")
-    assert start != -1, f"{name} not found"
-    params_open = src.find("(", start)
+    start, params_open = _find_function_declaration(src, name)
     assert params_open != -1, f"{name} params not found"
     params_close = _matching_delimiter(src, params_open, "(", ")")
     brace = src.find("{", params_close)
@@ -164,6 +258,25 @@ def test_function_extractor_handles_nested_template_literal_interpolation():
     assert "function afterSample" not in body
 
 
+def test_function_extractor_matches_exact_declarations_outside_comments():
+    source = """
+    // function sample(){ return 'comment'; }
+    function samplePrefix(){ return 'prefix'; }
+    function sample(){
+      const hasBrace = /\\{[^}]+\\}/.test(text);
+      return hasBrace ? 'target' : 'fallback';
+    }
+    function afterSample(){ return false; }
+    """
+
+    body = _function_body(source, "sample")
+
+    assert "return hasBrace ? 'target' : 'fallback';" in body
+    assert "return 'comment';" not in body
+    assert "return 'prefix';" not in body
+    assert "function afterSample" not in body
+
+
 def test_transparent_raw_content_helper_is_fallback_only_when_anchor_scene_absent():
     helper = _function_body(_ui_js(), "_transparentStreamOrderedParts")