Tooling: admit Node 26 alongside Node 24 - #794
Merged
Merged
Conversation
epeicher
force-pushed
the
tooling/allow-node-26
branch
from
September 11, 2026 08:34
5d4b4f6 to
a2d67cc
Compare
The devEngines pin (>=24 <25, onFail: error) refuses `npm ci` on a Node 26 machine before anything installs, so a reviewer on the current release line cannot run the vitest suite at all. The pin was added to fail fast instead of silently churning the lockfile on a different Node major; measured on Node 26.8.2, the lockfile rewrite is byte-for-byte the same one Node 24.16.0 produces, and build, lint, typecheck and the full vitest suite pass on both. Widen `engines` and `devEngines` to `^24.0.0 || ^26.0.0`: the two current release lines, skipping the end-of-life odd line in between. `onFail: error` stays, so an unvalidated major still fails fast. `.nvmrc` and CI stay on 24. The lockfile regeneration also picks up the package name that was renamed to `openstation` in package.json but never in the lock, which is what made any `npm install` on trunk dirty the lockfile.
epeicher
force-pushed
the
tooling/allow-node-26
branch
from
September 11, 2026 08:40
a2d67cc to
6ed488d
Compare
This was referenced Sep 11, 2026
epeicher
added a commit
that referenced
this pull request
Sep 11, 2026
`PR Preview Publish` resolves the PR from the trusted `workflow_run` payload and requires the PR's head to still be the exact commit the build ran on. When a second push lands while the first build is still running, that commit is superseded before the publish job starts, the lookup matches nothing, and the step calls `core.setFailed`. Nothing is actually broken when that happens: the newer commit's own build/publish pair posts the preview, the job attaches no check run to the PR, and it is not in the trunk ruleset's required checks. But `workflow_run` runs are attributed to the default branch, so each race paints a red run in the trunk Actions tab. It fired twice in two days (#794, and a `codex/agent-async-jobs` push the day before), and it will fire whenever anyone pushes twice in quick succession. Warn and skip instead. The four steps that need a PR number are gated on it, so the job ends green with an annotation saying the head moved on. Publishing the superseded build under the current PR was never the alternative: a reviewer clicking that preview would get older code. The trust model is unchanged. A preview is still only ever published to a PR positively identified from the payload, and the loud failures stay loud: a malformed head SHA or incomplete head data still fails, and an API or permission error throws out of `paginate` rather than returning an empty list.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
What it does
Widens the Node pin in
package.jsonfrom>=24 <25to^24.0.0 || ^26.0.0(bothenginesanddevEngines).Rationale
devEngineswithonFail: errormakes npm refusenpm cion a Node 26 machine before anything installs, so a reviewer on the current release line could not run the vitest suite at all. The pin exists to fail fast instead of silently churning the lockfile on a different Node major. Measured on Node 26.8.2, the lockfile rewrite is byte-for-byte the same one Node 24.16.0 produces, and build, lint, typecheck and the full vitest suite pass on both, so the reason for excluding 26 does not hold.The range admits the two current release lines and skips Node 25, the odd line that is already end-of-life.
onFail: errorstays, so an unvalidated major (27) still fails fast.Implementation
package.json:enginesanddevEngines.runtime.versionbecome^24.0.0 || ^26.0.0..nvmrcand CI stay on 24, so shipped bundles are unchanged.package-lock.json: regenerated. Besides mirroring the newenginesvalue, this picks up the package name that was renamed toopenstationinpackage.jsonbut never in the lock, which is what made anynpm installon trunk dirty the lockfile.Testing instructions
nvm use):npm ci && npm run build && npm run lint && npm run typecheck && npm run test:jsall pass.nvm install 26 && nvm use 26):npm cinow succeeds where it previously failed withEBADDEVENGINES; the same four gates pass (500 test files, 6091 tests).npm install --package-lock-onlyleaves the lockfile clean.