ci: track the hub's main instead of the @v1 tag - #21
Merged
Conversation
`v1` is a moving tag, not a rotting pin — but it is a second pointer pushed by hand, and it is one commit behind `main` today, which is the failure it invites: a fix merges and reaches nobody. The hub's own `uses:` examples already say `@main`, and all 166 references to the lab-sotashimozono hub are `@main`; this brings both owners onto one convention. Blast radius is exactly #20 (julia-ci.yml, +84/-0, purely additive, gated on shards>1 and push:main) — nothing else differs between v1 and main. Files: labeler.yml, tagbot.yml Refs lab-sotashimozono/.github#15
`token: ${{ secrets.BOT_PAT }}` with no `BOT_PAT` under `workflow_call.secrets`. An undeclared secret
in a called workflow is EMPTY unless the caller passes `secrets: inherit`; the documented caller does,
which is both why it worked and why nothing reported it.
actionlint has been failing on this repo's `push:main` runs since at least 2026-07-26 for exactly this
— and every pull request passed, because reviewdog reports only inside the diff, so a defect in a file
the PR does not touch cannot surface. Touching tagbot.yml here is what brought it into scope.
Same class of defect as lab-sotashimozono/.github#17 (documentation.yml's undeclared DOCUMENTER_KEY),
found the same way: comparing what a reusable references against what it declares.
Refs lab-sotashimozono/.github#15
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Every caller of the QAtlasHub hub references
@v1.v1is a moving tag currently atv1.0.3,so it is not a rotting pin — but it is a second pointer that has to be pushed by hand, and it is one
release behind
mainright now, which is exactly the failure it invites: a fix merges and reachesnobody.
Measured, 2026-07-30
241 reusable references across the three owners:
lab-sotashimozono@mainQAtlasHub@v1QAtlasHub@mainThose 10 are not callers — they are the
uses:examples inside the reusables' own header comments,which already say
@main. So the hub documents@mainand every real caller uses@v1. The labhub has no
@v1references at all; this brings the two owners onto one convention.Blast radius, exactly
QAtlasHub/.githubmainisv1plus one commit:Purely additive — the
julia-citimings producer, gated onshards > 1andpush:main, with therecording job
continue-on-error. Nothing else differs, so this switch delivers exactly that andnothing more.
It also makes that fix testable:
record-timingsonly runs onpush:main, so no PR anywhere couldexercise it (#19). A
shards: 4caller on@mainis the first real check.Why
@mainrather than movingv1Moving the tag on every merge makes
v1an alias formainwith extra machinery and one more thingto forget.
mainis gated by the hub's ownactionlint, and the standing rule is to track latestrather than pin. After this the tag is unreferenced and can be retired separately.
Refs lab-sotashimozono/.github#15