ci: record why ecosystem-test's dependencies stay unpinned - #94
Merged
Conversation
Resolves #90. The issue framed this as a tradeoff -- early warning of upstream drift versus a PR queue an upstream release can block -- and said the deciding fact was whether ecosystem-test is a required check. It is not, and neither half of the objection survives checking: * the `protect-main` ruleset carries only `deletion` and `non_fast_forward`. There are NO required status checks, so a red job cannot block a merge. * nothing depends on this job. wheels.yml's release chain gates on `validate-wheels`, in a different workflow, so a failure here cannot stop a release either. So the dilemma rested on a premise this repo does not have. A failure is loud (a red check, a failed run) and cheap (blocks nothing), which is what an early-warning lane should be. Floating is already the right answer and is already in place; what was missing was the reasoning. No behaviour change -- a comment, so the next dependency audit does not "fix" this by pinning it. Records the one condition that would flip the tradeoff: adding required status checks to the ruleset. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
|
Warning Review limit reached
On-demand reviews are free for the next 23 days. After that, they cost $0.25 per reviewed file. Or wait 24 minutes for your next included review. View limit detailsLimit details: You’ve used the included review currently available. Your 73 included PR review attempts over the past 7 days set your current allowance at 1 review per hour. Review configuration: ⚙️ Run configurationConfiguration used: defaults Review profile: CHILL Plan: Pro Run ID: 📒 Files selected for processing (1)
Comment |
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.
Resolves #90.
#90 framed this as a genuine tradeoff — early warning of ecosystem drift, versus a PR queue that an upstream release can block — and named the deciding fact: is
ecosystem-testa required check?It is not, and checking it dissolved the dilemma rather than settling it.
No required status checks at all. And nothing depends on the job either —
wheels.yml's release chain gates onvalidate-wheels, in a different workflow, so a redecosystem-testcannot stop a release any more than it can stop a merge.So the objection that motivated pinning — "a torch release turns every PR red and blocks the queue" — is not a thing that can happen in this repo. A failure here is loud (a red check on the PR, a failed run) and cheap (blocks nothing), which is exactly what an early-warning lane should be.
The resolution is that floating was already right
That makes this the rare audit finding where the status quo wins. Pinning would trade the job's entire value — telling us when a torch or jax release breaks interop — for protection against a failure mode this configuration cannot produce.
What was actually missing was the reasoning. The unpinned installs sat next to six dependencies pinned earlier today for good reasons, with nothing to distinguish deliberate from overlooked. The next person auditing would reasonably have "fixed" them.
So this adds the comment, covering three things:
No behaviour change. Lands as
ci:, not inpatch_tags, so it cuts no release.🤖 Generated with Claude Code