fix(executor): extend opencode self-review timeout#2420
fix(executor): extend opencode self-review timeout#2420alekspetrov merged 1 commit intoqf-studio:mainfrom
Conversation
|
Thanks! The diff is much larger than a self-review timeout change should need (+114/−109 in runner.go). Looks like a rebase artifact or formatter pass. Could you rebase off main and resubmit with a minimal diff covering just: (a) backend-aware self-review timeout selection, (b) the new test in runner_test.go? Happy to merge once the runner.go diff is scoped to the timeout logic. |
Self-review uses a hardcoded 2-minute timeout. OpenCode runs are genuinely slower than Claude Code (server-managed session, larger streaming overhead), and the cap cancels review while the backend is still working — surfaces as a false regression. Add backend-aware selfReviewTimeout() helper: 10m for OpenCode, 2m otherwise. Wire it into runSelfReview's context.WithTimeout. Closes qf-studio#2416
alekspetrov
force-pushed
the
upstream-self-review-timeout
branch
from
23f90e0 to
e0a5ab5
Compare
|
Rebased onto current main and force-pushed (was stale by ~v2.99 work). The previous +114/−109 in runner.go was gofmt churn against an older base plus removal of post-branch work. New diff is +49/−2: Will need workflow approval on the Checks tab to run CI on the new push. |
alekspetrov
left a comment
alekspetrov
left a comment
There was a problem hiding this comment.
LGTM. Backend-aware self-review timeout (10m OpenCode / 2m otherwise) with focused tests. Rebased onto current main.
|
Codecov Report✅ All modified and coverable lines are covered by tests. 📢 Thoughts on this report? Let us know! |
3 participants
Summary
Verification
go test ./internal/executor/...Closes #2416