perf(normalize): scan verification labels once

CHANGELOG.md

@@ -5,6 +5,8 @@ this file. The format follows Keep a Changelog, and versioned releases follow
 Semantic Versioning where the repository publishes a release.
 
 ## [Unreleased]
+- Review summaries now report test and documentation coverage separately and
+  consistently. No action is required.
 - Route Strix cross-provider fallbacks to explicit direct-OpenAI models
   (`openai-direct/...`) through the OpenAI inference endpoint instead of
   inheriting a provider-specific primary base: the workflow now provisions

scripts/ci/opencode_review_normalize_output.py

@@ -133,9 +133,14 @@
     "security/privacy:",
 )
 
-APPROVAL_VERIFICATION_PATTERNS = {
-    label: re.compile(re.escape(label)) for label in APPROVAL_VERIFICATION_LABELS
-}
+# Match the longest label first so a future label cannot shadow a more specific
+# label that starts at the same position (for example, docstring coverage:).
+ANY_LABEL_PATTERN = re.compile(
+    "|".join(
+        re.escape(label)
+        for label in sorted(APPROVAL_VERIFICATION_LABELS, key=len, reverse=True)
+    )
+)
 
 SOURCE_LIKE_CHANGED_FILE_EXTENSIONS = frozenset(
     {
@@ -959,32 +964,22 @@ def mentions_verification_posture(reason: str, summary: str) -> bool:
 def label_section(text: str, label: str) -> str:
     """Return text after a verification label until the next known label."""
 
-    def label_starts(candidate: str) -> list[int]:
-        """Return exact verification-label starts without suffix collisions."""
-        starts = []
-        pattern = APPROVAL_VERIFICATION_PATTERNS.get(candidate)
-        if pattern is None:
-            pattern = re.compile(re.escape(candidate))
-        for match in pattern.finditer(text):
-            index = match.start()
-            if (
-                candidate == "coverage:"
-                and text[max(0, index - 10) : index] == "docstring "
-            ):
-                continue
-            starts.append(index)
-        return starts
-
-    starts = label_starts(label)
+    actual_matches = [
+        (match.start(), match.group(0)) for match in ANY_LABEL_PATTERN.finditer(text)
+    ]
+    if label in APPROVAL_VERIFICATION_LABELS:
+        starts = [index for index, candidate in actual_matches if candidate == label]
+    else:
+        # Preserve repository-specific labels without rescanning catalogue labels.
+        starts = [match.start() for match in re.finditer(re.escape(label), text)]
     if not starts:
         return ""
+
     start = starts[-1] + len(label)
     next_starts = [
-        candidate_start
-        for candidate in APPROVAL_VERIFICATION_LABELS
-        if candidate != label
-        for candidate_start in label_starts(candidate)
-        if candidate_start >= start
+        index
+        for index, candidate in actual_matches
+        if candidate != label and index >= start
     ]
     end = min(next_starts) if next_starts else len(text)
     return text[start:end]

tests/test_opencode_review_normalize_output.py

@@ -1386,6 +1386,7 @@ def test_label_and_full_coverage_detection(tmp_path, monkeypatch):
     assert not norm.mentions_full_coverage("", no_source_summary)
     assert norm.contradicts_changed_file_kinds("", no_source_summary)
 
+
     changed_files = tmp_path / "opencode-changed-files.txt"
     changed_files.write_text("README.md\n", encoding="utf-8")
     monkeypatch.setenv("OPENCODE_CHANGED_FILES_FILE", str(changed_files))
@@ -1434,6 +1435,26 @@ def test_label_and_full_coverage_detection(tmp_path, monkeypatch):
     )
 
 
+def test_label_section_scans_labels_once_and_prefers_longest_label():
+    """Parse adjacent labels without rescanning the input for every label."""
+    text = "docstring coverage: 100% coverage: 98% performance: measured"
+
+    assert norm.label_section(text, "docstring coverage:") == " 100% "
+    assert norm.label_section(text, "coverage:") == " 98% "
+    assert norm.label_section(text, "performance:") == " measured"
+    alternatives = norm.ANY_LABEL_PATTERN.pattern.split("|")
+    assert alternatives.index(re.escape("docstring coverage:")) < alternatives.index(
+        re.escape("coverage:")
+    )
+
+
+def test_label_section_preserves_custom_label_lookup():
+    """Repository-specific labels retain the helper's backwards-compatible path."""
+    text = "custom evidence: measured coverage: 98%"
+
+    assert norm.label_section(text, "custom evidence:") == " measured "
+
+
 def test_check_structural_approval_rejects_invalid_or_unsafe_approvals(
     tmp_path, monkeypatch
 ):