Repository navigation
fix(test): update slot-reader enumeration for _switch_target_busy centralization - #12715
NicholasRBowers wants to merge 1 commit into
Conversation
GPT 5.6 Review — ✅ no blocking findingsGPT 5.6 completed its review of This comment is updated in place on each push. Review detailsNo findings. False positive or not applicable? A repository writer can comment: |
GPT 5.6 Review —
|
1 similar comment
GPT 5.6 Review —
|
Design Review (Fable 5) — ✅ PASSDesign-level review of Design-Verdict: PASS Confirm Backend Tests shard 4 goes green on this PR's CI run — I verified the new pin against HEAD by AST reading, not execution. [DESIGN-REVIEWED] 300bd3b |
Opus 4.8 Review — ✅ no blocking findingsReviewed Verdict parsed from the review's SHA-scoped output markers for commit False positive or not applicable? A repository writer can comment: |
… centralization The switch-handler refactor moved the slot.running pre-commit refusal out of the four switch handlers (api_chat_slot_agent/model/reasoning_effort/ workspace) into the shared _switch_target_busy helper, but the enumeration pin in test_running_and_turn_running_slot_readers_are_enumerated was not updated, so Backend Tests shard 4 fails on both platforms for every PR in the repo. Remove the four migrated handler entries and add _switch_target_busy as a reservation-semantics reader -- its docstring pins exactly that reading (slot.running is set at dispatch, before provider.start() registers a session), so the helper belongs in the same class its callers vacated. Fixes #12713 Co-authored-by: Kiro Crew <noreply@kirodotdev.github.io>
328f5ec to
300bd3b
Compare
|
Thanks for the quick fix. #12713 is now closed by #12718 (merged as 194f6d5), which keeps the four switch handlers enumerated by crediting callers of |
Pull request was closed
Summary
The switch-handler refactor centralized the
slot.runningpre-commit refusal from four switch handlers (api_chat_slot_agent,api_chat_slot_model,api_chat_slot_reasoning_effort,api_chat_slot_workspace) into the shared_switch_target_busyhelper, buttest_running_and_turn_running_slot_readers_are_enumeratedstill pins the old reader set — Backend Tests shard 4 fails on both platforms for every PR in the repo since the merge landed on main.This updates the pin to match: the four migrated handler entries are removed and
_switch_target_busyis added as a reservation-semantics reader — its own docstring pins exactly that reading (slot.runningis set at dispatch, beforeprovider.start()registers a session), so the helper belongs in the same class its callers vacated.Delta verified by re-running the test's own AST scan: 4 sites out, 1 site in, no semantics changes anywhere else.
Testing
test/test_pending_boundary_consumer_semantics.py— 9 passed on main + this fix (the enumeration test fails on pristine main tip0a8005093)Fixes #12713
Pattern harvest
Rule candidate: review-prompt
Pattern: A refactor that moves a pinned read (an enumeration test asserts every reader site) must update the pin in the same commit; centralizing four handler reads into one helper without touching the pin breaks every PR's CI, not just the refactor's.