Repository navigation
fix(publish): keep per-package and baseline tags in --input - #684
Merged
Merged
Conversation
Contributor
|
Standing release PR: #614 · v0.42.0 queued (4 packages) · open 82d 4h 13m · ✅ ready to merge Release Preview — v0.41.2 → v0.41.3
These changes will be added to the release PR (#614) when merged: Changelogmonorepo 0.41.2 → 0.41.3
After merge — predicted release
Updated automatically by ReleaseKit |
The publish input schema used z.object, which strips undeclared keys, so `releasekit publish --input` dropped updates[].tag (per-package mode) and baselineTags (baseline tag creation), along with every other VersionOutput field the schema didn't list. Switch the schemas to z.looseObject, declare tag and baselineTags, and add a round-trip test over a fully populated VersionOutput so the schema can't fall behind the type again. Closes #675 Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…h, in per-package mode Per-package mode pushed each package's own tag inside the publish loop and skipped the final push, so baseline tags never reached the remote, and with --skip-publish nothing was pushed and no GitHub releases were created. Now that publish --input keeps per-package tags, that path is reachable from the CLI too. Run the final push in per-package mode as well: after every package publishes it adds the remaining tags, and under --skip-publish it pushes everything as batch mode does. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
goosewobbler
force-pushed
the
fix/675-publish-input-passthrough
branch
from
October 22, 2026 22:11
e5bdb65 to
50f83f5
Compare
Owner
Author
|
@greptileai review plz |
The push stage ran `git push --tags`, publishing every local tag. Push each tag the run recorded by explicit ref instead, so an unrelated or conflicting local tag can't be published or fail the push after packages are out. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The round-trip test fails if the schemas go back to stripping fields. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
8 of 20 tasks
This was referenced Oct 22, 2026
Merged
13 of 20 tasks
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
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.
Description
publish --inputvalidated its input withz.objectschemas, which strip undeclared keys.updates[].tagandbaselineTagswere lost, so per-package mode never ran and baseline tags were never created. Any otherVersionOutputfield the schema didn't list was lost too.releaseandstanding-pr publishpass the output in memory and were unaffected.The schemas now use
z.looseObjectand declaretagandbaselineTags. A round-trip test checks that a fully populatedVersionOutputcomes back fromparseInputunchanged.Keeping the tags lets
publish --inputreach per-package mode, which had two gaps: baseline tags were never pushed, and under--skip-publishnothing was pushed and no GitHub releases were created. The final push stage now also runs in per-package mode, after every package has published, so it adds the remaining tags (or pushes everything under--skip-publish). That also fixes baseline tags forreleaseandstanding-pr publishin per-package mode, which had the same gap.The push stage also pushed every local tag (
git push --tags), in batch mode too. It now pushes only the tags this run recorded, each by explicit ref.Related Issue(s)
Closes #675
Type of Change
feat— New featurefix— Bug fixdocs— Documentation changerefactor— Code refactoringtest— Adding or updating testschore— Maintenance, dependencies, toolingScope
@releasekit/version@releasekit/notes@releasekit/publish@releasekit/release@releasekit/config@releasekit/coreChecklist
pnpm lint)pnpm test)pnpm typecheck)Testing Notes
New
input.spec.tstests cover per-package and baseline tags survivingparseInput, rejection of a non-stringtag, and a full round-trip; all three fail on the old schema. The round-trip fixture is typed withRequired<…>, so a newVersionOutputfield is a type error in the editor until the fixture includes it. CI won't catch that, because publish'stypecheckexcludes test files.pipeline.spec.ts: per-package mode runs the final push once, after the last package tag, and not when a package fails; per-package input withskipPublishstill pushes and creates GitHub releases. Both fail with the old guard.parseInput's body is untouched, so #616 should rebase cleanly.🤖 Generated with Claude Code


Confidence Score: 5/5
The PR appears safe to merge; the unrelated-tag push is fixed and no new blocking issues were found.Summary
The PR keeps per-package and baseline tags when reading publish input, then pushes baseline tags after all packages succeed.
Reviews (3) · Last reviewed commit: "fix(publish): push only this release's t..." · Reviewed by Greptile