Skip to content

Commit 04a5702

Browse files
authored
ci(adapter-test): gate adapter-test off pull_request trigger (#1522)
Per @WAWQAQ direction (DM): trim PR-time CI to fast-feedback only. adapter-test (~30-60s) is the next-largest PR wait after e2e-headed (which #1521 just removed). Adapter authors typically run focused tests locally before pushing (`npm run test:adapter`); CI duplication adds queue latency without catching new classes of bugs. PR-time CI surface now: - typecheck / unit (~1 min) - lint gates (typed-error / silent-column-drop) - build × 3 platforms Adapter test guards (still strict): - push to main / dev - nightly cron - workflow_dispatch (manual when an adapter-heavy PR really wants the signal before merge) Same gate as smoke-test (`if: github.event_name == 'push' || schedule || workflow_dispatch`) for consistency.
1 parent fd438c2 commit 04a5702

1 file changed

Lines changed: 5 additions & 1 deletion

File tree

.github/workflows/ci.yml

Lines changed: 5 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -110,8 +110,12 @@ jobs:
110110
- name: Run unit tests under Bun
111111
run: bun vitest run --project unit --reporter=verbose
112112

113-
# Adapter tests are pure unit tests — OS doesn't affect results.
113+
# Adapter tests are pure unit tests — OS doesn't affect results. Gated off
114+
# `pull_request` to keep PR CI under ~2 minutes; adapter authors run focused
115+
# tests locally before pushing, and `push` to main / nightly cron / manual
116+
# dispatch still guard the merged state.
114117
adapter-test:
118+
if: github.event_name == 'push' || github.event_name == 'schedule' || github.event_name == 'workflow_dispatch'
115119
runs-on: ubuntu-latest
116120
needs: build
117121
steps:

0 commit comments

Comments
 (0)