PreToolUseHookTest#test_edit_judge_event_denies fails intermittently #531

New issue

Open

opened 2026-09-04 15:19:05 +00:00 by jared · 1 comment

jared commented

2026-09-04 15:19:05 +00:00

Owner

VERDICT: CREATE

Context

This ticket comes from work on the os-sdlc architecture refactor (docs/plans/os-sdlc-architecture-refactor-plan.md).

Observed

OsSdlc::PreToolUseHookTest#test_edit_judge_event_denies failed intermittently on 2026-09-04. It failed twice.
The first failure happened on the first suite run after the phase 1 merge, commit 3b7a0ef.
The second failure happened during phase 2 Verdict work, in a git worktree.
A clean rerun of ruby tests/all.rb was green both times. No hook files were touched between the failing run and the passing rerun.

ruby tests/all.rb

Reproduce

The failure is intermittent. It does not reproduce on demand.

ruby tests/all.rb

Expected

def test_edit_judge_event_denies
  # this test must pass on every run, not only on rerun
end

Origin

Trigger: ruby tests/all.rb, twice on 2026-09-04
Improvised this session: reran the suite; treated the rerun as ground truth
Chain: intermittent test failure ← suspected environmental or timing dependency in the pre_tool_use hook test (possibly the rubocop server or a subprocess) ← UNKNOWN, not yet diagnosed
Root candidate: this ticket
Where: plugins/os-sdlc/tests/hooks/pre_tool_use_hook_test.rb, plugins/os-sdlc/hooks/pre_tool_use.rb
Session: 78279828-6e55-428d-a1a8-350cda9e6226
Transcript: /home/jared/.claude/projects/-home-jared-dev-cc-os/78279828-6e55-428d-a1a8-350cda9e6226.jsonl

Co-Authored-By: Claude Fable 5.1 noreply@anthropic.com
Claude-Session: https://claude.ai/code/session_014DmLVHCHtPPZcF9gb5erDD

VERDICT: CREATE ## Context This ticket comes from work on the os-sdlc architecture refactor (docs/plans/os-sdlc-architecture-refactor-plan.md). ## Observed `OsSdlc::PreToolUseHookTest#test_edit_judge_event_denies` failed intermittently on 2026-09-04. It failed twice. The first failure happened on the first suite run after the phase 1 merge, commit 3b7a0ef. The second failure happened during phase 2 Verdict work, in a git worktree. A clean rerun of `ruby tests/all.rb` was green both times. No hook files were touched between the failing run and the passing rerun. ``` ruby tests/all.rb ``` ## Reproduce The failure is intermittent. It does not reproduce on demand. ``` ruby tests/all.rb ``` ## Expected ```ruby def test_edit_judge_event_denies # this test must pass on every run, not only on rerun end ``` ## Origin - Trigger: `ruby tests/all.rb`, twice on 2026-09-04 - Improvised this session: reran the suite; treated the rerun as ground truth - Chain: intermittent test failure ← suspected environmental or timing dependency in the pre_tool_use hook test (possibly the rubocop server or a subprocess) ← UNKNOWN, not yet diagnosed - Root candidate: this ticket - Where: plugins/os-sdlc/tests/hooks/pre_tool_use_hook_test.rb, plugins/os-sdlc/hooks/pre_tool_use.rb - Session: 78279828-6e55-428d-a1a8-350cda9e6226 - Transcript: /home/jared/.claude/projects/-home-jared-dev-cc-os/78279828-6e55-428d-a1a8-350cda9e6226.jsonl Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_014DmLVHCHtPPZcF9gb5erDD

jared added the

plugin/os-sdlc

filed-by/agent

labels

2026-09-04 15:19:05 +00:00

jared commented

2026-09-06 16:40:00 +00:00

Author

Owner

Seen again 2026-09-06 during the phase 3 entity refactor (branch arch-refactor-phase-3-entities), on suite runs of ~1067 tests with parallelize_me!: PreToolUseHookTest#test_edit_judge_event_denies failed once in one of three runs and passed on isolated rerun; a sibling test in the same file flaked once the same way. Separately, one run failed with Expected /[ \t]+$/ to not match "class Offenses " (trailing-whitespace assertion on generated output), not reproduced. Both look like parallel-thread races on shared temp paths, not logic. No fix attempted.

Seen again 2026-09-06 during the phase 3 entity refactor (branch arch-refactor-phase-3-entities), on suite runs of ~1067 tests with parallelize_me!: PreToolUseHookTest#test_edit_judge_event_denies failed once in one of three runs and passed on isolated rerun; a sibling test in the same file flaked once the same way. Separately, one run failed with `Expected /[ \t]+$/ to not match "class Offenses "` (trailing-whitespace assertion on generated output), not reproduced. Both look like parallel-thread races on shared temp paths, not logic. No fix attempted.

No milestone

No project

No assignees

1 participant

Notifications

Due date

The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set

Reference

jared/cc-os#531

No description provided.

Rows
Columns