chore: follow the vibecook-dev transfer and consume Ghosttea from npm - #5
Merged
Merged
Conversation
The repo moved from jamesyong-42/chopsticks to vibecook-dev/chopsticks. Git kept working through GitHub's redirect, so the stale references stayed silent. The load-bearing one is trust-packages.mjs: npm trusted publishing binds to a literal owner/repo on the registry side and does not follow a transfer. Actions now mints OIDC tokens claiming vibecook-dev/chopsticks, which no longer matched any binding, so the next release would have failed at publish with no missing secret to explain it. The nine registry-side bindings have been re-pointed. CHANGELOG links are left alone: they redirect, and release-please generates new ones from GITHUB_REPOSITORY. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
The sibling was transferred to vibecook-dev/ghosttea and renamed every package directory after what it publishes (terminal-* -> ghosttea-*). Both apps still pointed at packages/terminal-*, so all ten references — five file: deps per app plus the link: overrides — resolved to directories that no longer exist and the node_modules symlinks dangled. Pin the five packages to 0.4.0 from the registry instead. The only commits touching the sibling's packages/ since its v0.4.0 tag are that rename and its own transfer chore, so the published tarballs are equivalent to the checkout. minimumReleaseAgeExclude moves to 0.4.0 because those versions are days old and the supply-chain policy would otherwise reject them. build.mjs now resolves bridge-entry through the package exports map. The remaining ghostteaRoot paths are deliberate: the native tabs addon and the cargo-built ghosttead daemon live in the sibling's app rather than its packages, so neither is on npm and both still need the local checkout. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
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.
Two independent fixes, one per commit, both fallout from recent repo transfers.
chore:follow the repository transferThe repo moved
jamesyong-42/chopsticks→vibecook-dev/chopsticks. Git kept working through GitHub's redirect and CI stayed green, so 17 stale references went unnoticed.The load-bearing one was
scripts/trust-packages.mjs. npm trusted publishing binds to a literalowner/repoon the registry side — it doesn't follow a transfer and isn't stored in the repo. Actions now mints OIDC tokens claimingvibecook-dev/chopsticks, which matched no binding, so the next release would have failed at publish with no missing secret to explain it.The nine registry-side bindings have already been re-pointed out of band and verified:
vibecook-dev/chopsticks(all 9)release.ymlcreatePackagejamesyong-42entriesCHANGELOG.mdlinks are deliberately untouched — they redirect, and release-please generates new ones fromGITHUB_REPOSITORY.build(apps):consume Ghosttea from npmThe sibling was transferred to
vibecook-dev/ghostteaand renamed every package directory after what it publishes (terminal-*→ghosttea-*). Both apps still pointed atpackages/terminal-*, so all ten references — fivefile:deps per app plus thelink:overrides — resolved to directories that no longer exist, and thenode_modulessymlinks dangled. The apps were already broken before this PR.Pinned to exact
0.4.0from the registry instead. The only commits touching the sibling'spackages/since itsv0.4.0tag are that rename and its own transfer chore, so the published tarballs are equivalent to the local checkout.minimumReleaseAgeExcludemoves to0.4.0; those versions are days old and the supply-chain policy rejects them otherwisebuild.mjsresolvesbridge-entrythrough the package exports map rather than a source-tree pathbuild:sdkhalf ofprebundleis dropped — the npm tarballs ship prebuiltdist/The remaining
ghostteaRootpaths are deliberate and now carry a comment saying so: the native tabs addon and the cargo-builtghostteaddaemon live in the sibling's app, not its packages, so neither is on npm and both still need the local checkout.GHOSTTEAD_BINoverrides the daemon.Verification
Local, since CI only covers
packages/**:pnpm test— exit 0,fail 0pnpm format:check,pnpm pack:check— both passbridge-entry.js,terminal-render.worker.js, anddist/native/ghosttea_native_tabs.nodeall stagedMerge order
Worth landing before release PR #4. That PR predates this work, so merging it first would publish 0.1.4 with the old
jamesyong-42URLs baked into every package'srepositorymetadata and provenance.🤖 Generated with Claude Code