[rescue] rescue/2026-08-09-review-599 — preserve the PR #599 backfill review snapshot - #691
[rescue] rescue/2026-08-09-review-599 — preserve the PR #599 backfill review snapshot#691EtanHey wants to merge 12 commits into
Conversation
…o feat/indexing-window-backfill
…o feat/indexing-window-backfill
…o feat/indexing-window-backfill
…o feat/indexing-window-backfill
…o feat/indexing-window-backfill # Conflicts: # tests/test_ingest_denylist.py
|
Important Review skippedDraft detected. Please check the settings in the CodeRabbit UI or the ⚙️ Run configurationConfiguration used: Organization UI Review profile: ASSERTIVE Plan: Pro Plus Run ID: You can disable this status message by setting the Use the checkbox below for a quick retry:
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
![macroscopeapp[bot]](https://avatars.githubusercontent.com/in/900172?s=60&v=4){kind=small}
| for test_path in "$TEST_ROOT"/test_cli*.py "$TEST_ROOT"/test_watch_backfill_cli.py; do | ||
| if [ -f "$test_path" ] && ! is_real_db_test_file "$test_path"; then | ||
| append_unique "$test_path" | ||
| mapped=1 | ||
| fi | ||
| done | ||
| ;; |
There was a problem hiding this comment.
🟡 Medium scripts/run_tests.sh:146
A change to src/brainlayer/cli/__init__.py marks the file as fully mapped after selecting only test_cli*.py and test_watch_backfill_cli.py, so a changed-only pre-push run no longer falls back to the full suite. Other tests that import brainlayer.cli — e.g. tests/test_agent_profiles.py, tests/test_doctor.py, tests/test_runtime_store.py, and tests/test_status_truthfulness.py — are skipped on such a change, so edits to their command implementations can pass changed-only checks without running the relevant tests. Either map every test that exercises this module or set changed_source_unmapped=1 (instead of mapped=1) so the full-suite fallback still applies.
src/brainlayer/cli/__init__.py)
- for test_path in "$TEST_ROOT"/test_cli*.py "$TEST_ROOT"/test_watch_backfill_cli.py; do
- if [ -f "$test_path" ] && ! is_real_db_test_file "$test_path"; then
- append_unique "$test_path"
- mapped=1
- fi
- done
+ for test_path in "$TEST_ROOT"/test_cli*.py "$TEST_ROOT"/test_watch_backfill_cli.py; do
+ if [ -f "$test_path" ] && ! is_real_db_test_file "$test_path"; then
+ append_unique "$test_path"
+ fi
+ done
+ changed_source_unmapped=1
;;🚀 Reply "fix it for me" or copy this AI Prompt for your agent:
In file @scripts/run_tests.sh around lines 146-152:
A change to `src/brainlayer/cli/__init__.py` marks the file as fully mapped after selecting only `test_cli*.py` and `test_watch_backfill_cli.py`, so a changed-only pre-push run no longer falls back to the full suite. Other tests that import `brainlayer.cli` — e.g. `tests/test_agent_profiles.py`, `tests/test_doctor.py`, `tests/test_runtime_store.py`, and `tests/test_status_truthfulness.py` — are skipped on such a change, so edits to their command implementations can pass changed-only checks without running the relevant tests. Either map every test that exercises this module or set `changed_source_unmapped=1` (instead of `mapped=1`) so the full-suite fallback still applies.
| def is_legacy_excluded_path(path: str | Path) -> bool: | ||
| """Identify roots blocked by the blanket denylist retired in July 2026.""" | ||
| parts = Path(path).expanduser().parts | ||
| return any( | ||
| _contains_ordered(parts, expected) | ||
| for expected in ( | ||
| (".claude", "projects", "subagents"), | ||
| (".codex", "sessions"), | ||
| (".cursor", "agent-transcripts"), | ||
| (".gemini", "sessions"), | ||
| ) | ||
| ) |
There was a problem hiding this comment.
🟠 High brainlayer/backfill.py:54
is_legacy_excluded_path matches paths that contain the expected segments as an ordered subsequence anywhere in the path, so it returns True for paths like ~/.codex/other/sessions/file.jsonl or ~/.gemini/other/sessions/file.jsonl where sessions is not directly under .codex/.gemini. The retired denylist required .codex/sessions, .gemini/sessions, and .claude/projects/.../subagents, so --legacy-excluded-only ingests transcripts that were never excluded by the legacy policy. Use consecutive path segment matching (e.g., a sliding window over parts) instead of _contains_ordered.
| def is_legacy_excluded_path(path: str | Path) -> bool: | |
| """Identify roots blocked by the blanket denylist retired in July 2026.""" | |
| parts = Path(path).expanduser().parts | |
| return any( | |
| _contains_ordered(parts, expected) | |
| for expected in ( | |
| (".claude", "projects", "subagents"), | |
| (".codex", "sessions"), | |
| (".cursor", "agent-transcripts"), | |
| (".gemini", "sessions"), | |
| ) | |
| ) | |
| def is_legacy_excluded_path(path: str | Path) -> bool: | |
| """Identify roots blocked by the blanket denylist retired in July 2026.""" | |
| parts = Path(path).expanduser().parts | |
| expected_patterns = ( | |
| (".claude", "projects", "subagents"), | |
| (".codex", "sessions"), | |
| (".cursor", "agent-transcripts"), | |
| (".gemini", "sessions"), | |
| ) | |
| for expected in expected_patterns: | |
| window_size = len(expected) | |
| for i in range(len(parts) - window_size + 1): | |
| if parts[i:i + window_size] == expected: | |
| return True | |
| return False |
🚀 Reply "fix it for me" or copy this AI Prompt for your agent:
In file @src/brainlayer/backfill.py around lines 54-65:
`is_legacy_excluded_path` matches paths that contain the expected segments as an ordered subsequence anywhere in the path, so it returns `True` for paths like `~/.codex/other/sessions/file.jsonl` or `~/.gemini/other/sessions/file.jsonl` where `sessions` is not directly under `.codex`/`.gemini`. The retired denylist required `.codex/sessions`, `.gemini/sessions`, and `.claude/projects/.../subagents`, so `--legacy-excluded-only` ingests transcripts that were never excluded by the legacy policy. Use consecutive path segment matching (e.g., a sliding window over `parts`) instead of `_contains_ordered`.
|
Lead disposition (Etan delegated): holding open as REFERENCE alongside #690 (its review snapshot). Closes together with it when Wave 4 supersedes. — brainlayerClaude lead (Fable 5) 🤖 Generated with Claude Code |
1 participant
Rescue context
This preserves the earlier review-stage snapshot of transcript-window backfill: idempotent replay, watcher/denylist integration, CLI and test-runner mapping, and the first integrity fixes.
The work came from the retained
review-599worktree. Raw ahead and verified-unlanded counts are both 12 againstmain; original PR #599 was CLOSED without merge. This snapshot overlapsbl-599but records the distinct review-stage head that was rescued.It was rescued instead of merged because the original PR closed without landing and Etan's ruling requires each real rescued survivor to remain visible for later disposition.
— maintenanceCodex (worker) · codex/gpt-5.6-sol
Note
Preserve backfill review snapshot with time-windowed, idempotent
watch-backfillreplays--since/--untiloptions to thewatch-backfillCLI command to scope backfills to a half-open UTC window, with per-window registries named viawindow_registry_suffix.--legacy-excluded-onlyto restrict backfills to transcript roots previously blocked by a retired blanket denylist, usingis_legacy_excluded_pathandis_legacy_backfill_denylisted.WindowedFlushin backfill.py as anon_flushadapter that filters entries by time window and source predicate, advances watermarks for excluded/malformed lines, and rejects synthesized timestamps.JSONLWatcherandJSONLTailerin watcher.py to support configurable denylist policies, custom predicates, and watermark-only progress markers for blank or malformed lines.📊 Macroscope summarized 4dfcc6b. 5 files reviewed, 0 issues evaluated, 0 issues filtered, 0 comments posted
🗂️ Filtered Issues
No issues evaluated.