Skip to content

Commit 314f3ed

Browse files
authored
Merge pull request #1999 from lidge-jun/codex/cursor-call-release-note-2
docs(devlog): record the real workflow state at the gated SHA
2 parents aad8e26 + 737ae8c commit 314f3ed

1 file changed

Lines changed: 40 additions & 6 deletions

File tree

‎devlog/_plan/260818_cursor_call_integration/060_release_readiness.md‎

Lines changed: 40 additions & 6 deletions
Original file line numberDiff line numberDiff line change
@@ -69,6 +69,32 @@ work would be wrong.
6969
npm { preview: '2.23.0-preview.20260816', latest: '2.24.2' }
7070
releases v2.24.2 (latest), v2.24.1, v2.24.0
7171

72+
`dev` advanced to `aad8e26014dab0921c768a639f9d324af6f1fa27` when this note itself
73+
merged. That commit adds this file and nothing else
74+
(`git diff --name-only 4f72d6f755 aad8e2601` → one path), so the gate evidence above
75+
still describes the code on `dev`.
76+
77+
## Workflow state at the gated SHA — measured, not assumed
78+
79+
The objective asked whether the two workflows `scripts/release.ts` waits for exist
80+
for this SHA. They do not, and the reasons differ:
81+
82+
| Workflow | State at `4f72d6f755` | Why |
83+
|----------|----------------------|-----|
84+
| Cross-platform CI | `in_progress` (run `32104616258`, started 05:53:03Z) | started automatically on the merge; still running when this note was written |
85+
| Service lifecycle | **never ran** | correct: it triggers only on `src/service.ts`, `src/cli.ts`, `src/cli/index.ts`, `src/lib/bun-runtime.ts`, `package.json`, `bun.lock`, or its own workflow file (`.github/workflows/service-lifecycle.yml:6-16`). This campaign touched none of them — verified with `git diff --name-only` across the whole merge range |
86+
87+
**This is not a blocker, and the distinction matters.** `release.ts:397-401` waits
88+
for Service lifecycle at the RELEASE commit, and the release commit always bumps
89+
`package.json` — which is a trigger path. So the workflow will fire when the release
90+
is cut. Its absence here means the campaign did not touch service-lifecycle surface,
91+
not that a gate was skipped.
92+
93+
What this does mean: **no cross-platform evidence exists for this code yet.** The CI
94+
run is unfinished and nothing has waited on it. Windows and macOS remain unverified
95+
for this diff, and the Linux-only argument above is the whole of the platform
96+
evidence.
97+
7298
## Recommendation: 2.25.0
7399

74100
A minor rather than a patch, because the externally observable behaviour of a failed
@@ -93,11 +119,20 @@ The maintainer decides. This note recommends.
93119

94120
Preconditions before running either:
95121

96-
1. Cross-platform CI green at `4f72d6f755` (currently in progress).
97-
2. Service lifecycle green at the same SHA, since `package.json` is a trigger path
98-
for it and the release gate requires it.
99-
3. A version decision. The promotion itself does not bump; `scripts/release.ts` owns
100-
that and is the only sanctioned publish path — never a direct `npm publish`.
122+
1. **Cross-platform CI green.** Run `32104616258` was still `in_progress` at the
123+
time of writing; check it before promoting:
124+
125+
gh run view 32104616258 --json status,conclusion
126+
127+
This is the only outstanding technical precondition, and it is the one the
128+
owner's CI waiver covered for the merge but does NOT cover for a release —
129+
`release.ts` blocks on it regardless of what a human waived earlier.
130+
2. **Service lifecycle** needs nothing here: it did not run at this SHA because the
131+
campaign touched none of its trigger paths, and it will fire on the release
132+
commit's `package.json` bump. Do not wait for it on `dev`.
133+
3. **A version decision** (see the recommendation above). The promotion itself does
134+
not bump; `scripts/release.ts` owns that and is the only sanctioned publish path
135+
— never a direct `npm publish`.
101136

102137
## Open follow-ups a reader would otherwise assume were fixed
103138

@@ -129,4 +164,3 @@ sweep with explicit notes about what the merge does and does not cover — #1992
129164
injected-policy prose is untouched, #1938's integer-for-string coercion is a separate
130165
path, and #1527's rate-limit asymmetry is unaddressed while its silent-collapse half
131166
is fixed. Each says so and invites a reopen.
132-

0 commit comments

Comments
 (0)