Merged
Conversation
…e files_embedded - Replace hardcoded timeout=600 with FLOW_CODEX_TIMEOUT env var (default 600s) - Print WARNING to stderr when codex resume falls back to new session - Force files_embedded=True in plan-review when sandbox prevents disk reads Task: fn-6-v2-reliability-and-performance-hardening.5
- compat.py: use msvcrt.locking() on Windows instead of no-op flock,
with warning fallback for unknown platforms
- adversarial.py: warn on unconsumed {{...}} placeholders after
template substitution
- phases.md: replace raw `codex exec` with `flowctl codex exec` to
match guard allowlist
Write runtime state before spec/receipt in cmd_done so a mid-write crash still marks the task as done (runtime is authoritative). Cache get_repo_root() per cwd to avoid repeated subprocess calls, following the same pattern as get_state_dir(). Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Status is now managed exclusively by runtime state files. Definition JSON files no longer have status written to them in cmd_task_reset, cmd_restart, cmd_skip, and cmd_split. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- cmd_start: 11 tests (claim, deps, blocked, force, resume, invalid ID) - cmd_done: 10 tests (evidence, spec update, duration, cross-actor, force) - cmd_ready: 7 tests (unblocked, dep done, skipped-as-done, in_progress, blocked) - cmd_next: 6 tests (ready task, resume, all-done, completion/plan review gates) - cmd_restart: 6 tests (cascade, dry-run, force, skip-todo, invalid/missing) - cmd_block: 3 tests (status, done-fails, spec update) Task: fn-6-v2-reliability-and-performance-hardening.2
- Thread use_json parameter through all rp.py helpers (require_rp_cli, run_rp_cli, parse_windows, parse_builder_tab) and cmd_* functions - All ~15 error_exit calls now use args.json instead of hardcoded False - Add 1MB log rotation in ralph-guard _debug_log_path() (single-file .1) - Bump RALPH_GUARD_VERSION to 0.15.0 Task: fn-6-v2-reliability-and-performance-hardening.7
- gather_context_hints: collect all symbols, batch into ≤5 git grep calls instead of N per-symbol subprocess calls - find_dependents: load all task files once upfront, then BFS over in-memory dict instead of re-globbing per BFS level - cmd_next: use load_all_tasks_with_state(epic_id) single-scan batch load instead of per-file load_task_with_state loop Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
z23cc
added a commit
that referenced
this pull request
Apr 9, 2026
P0 fixes (state loss — root cause of 5 issues): - get_flow_dir() now walks up directory tree (FLOW_STATE_DIR env → walk-up → CWD) Fixes: #1 state loss, #3 state not persistent, #5 worker parallel fail, #9 .flow symlink issues. Same pattern as git finding .git. - flowctl recover --epic <id> [--dry-run]: rebuilds task completion status from git log. Fixes #11 no recovery after state loss. P1 fixes (guard + review): - Guard graceful fallback: missing tools → "skipped" (not "failed"). Only actual failures block pipeline. Fixes #8. - Review-backend availability check: if rp-cli/codex not in PATH, auto-fallback to "none" with warning. Fixes #7. P2 fixes (UX): - Slug max length 40→20 chars. "Django+React platform with account management" → "fn-3-django-react-plat" not 40-char monster. Fixes #2 #12. - Brainstorm auto-skip: trivial tasks (≤10 words, contains "fix"/"typo"/etc) skip brainstorm entirely. Fixes #6. - --interactive flag: pause at key decisions. Fixes #10. 370 tests pass. Zero new dependencies. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Second optimization round based on deep audit of v0.1.18. Focuses on business logic reliability, performance hot paths, and review system robustness.
P0 — Critical reliability
P1 — High priority fixes
lock-check(read-only) replaced withlock --task(acquire-or-fail) — eliminates check-then-act raceFLOW_CODEX_TIMEOUT, resume fallback logs warning,files_embeddedrespects sandbox modeP2 — Performance + DX
gather_context_hintsuses alternation pattern (500 subprocess → ~10)load_all_tasks_with_state()instead of per-file readsargs.jsonP3 — Cross-platform + prompts
msvcrt.locking()instead of no-op{{placeholders}}codex exec→flowctl codex exec(matches guard allowlist)Test plan
python3 -m pytest scripts/flowctl/tests/ -v— 255 passed (43 new)bash scripts/smoke_test.sh— 90/90 passed🤖 Generated with Claude Code