Repository navigation
fix(ci): pin the Zephyr fork by revision, not by branch name (+ a gate that the pin stays reachable) - #405
Merged
Conversation
… recorded All 15 `west init` sites cloned pulseengine/zephyr by BRANCH name (`--mr gale/sem-replacement`), so a run's Zephyr was whatever that branch held at the time and no gale commit recorded it. gale#401's 38 -> 0 kernel delta spanned exactly such a change (pulseengine/zephyr#1 merged mid-issue): the runs are honest, but not reproducible from gale's history alone. The revision now lives in .github/zephyr-fork-pin, read by every site, with a refusal if the file is empty. Taking a fork change becomes a reviewable bump, like any other pin. Pinned at 9550d256 (current branch head). New kill-criteria gate `zephyr-fork-pin`: the pin must be 40-hex AND still reachable from gale/sem-replacement (compare status identical|ahead), so a rebase or force-push on the fork fails here rather than in every Zephyr job at once. It carries its own negative control — an unreachable revision must be rejected by the same query. Both arms verified against the live fork: identical for the pin, ahead for its parent, 404 for a fabricated sha. The control branches on gh's EXIT STATUS: gh prints the 404 body to stdout, so a string compare would have failed the gate for the wrong reason. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_011QG86sovTbfnPNY9SfhSmo
Codecov Report✅ All modified and coverable lines are covered by tests. 📢 Thoughts on this report? Let us know! |
…a SHA The first version passed the pinned revision to west init. west runs 'git clone --branch <rev>', which takes a branch or tag, so every Zephyr job failed at once: 'fatal: Remote branch 9550d256... not found in upstream origin' (56 checks). Each of the 15 sites now inits on gale/sem-replacement, then fetches and checks out the pinned revision in the manifest repo (<workspace>/zephyr, per the fork's west.yml 'self: path: zephyr') BEFORE west update reads it, and asserts HEAD equals the pin — an unpinned build would otherwise look identical in the log. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_011QG86sovTbfnPNY9SfhSmo
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.
gale CI cloned
pulseengine/zephyrby branch name at all 15west initsites, so each run built against whatevergale/sem-replacementheld at that moment, and no gale commit recorded which. #401's 38 → 0 kernel delta spanned exactly such a change (pulseengine/zephyr#1, merged mid-issue). Those results are honest but not reproducible from gale's history..github/zephyr-fork-pinholds the revision; every site reads it and refuses if it is empty. Pinned at9550d256— today's branch head, so this changes no build.kill-criteriagatezephyr-fork-pin: the pin must be 40-hex and still reachable fromgale/sem-replacement(comparestatusidentical|ahead). A rebase or force-push on the fork now fails here, once, instead of failingwest initin every Zephyr job at the same time.identicalfor the pin,aheadfor its parent, 404 for the fabricated sha. The control branches on gh's exit status — gh prints the 404 body to stdout, so a string compare would have failed the gate for the wrong reason.Taking a fork change is now a reviewable pin bump, like varve's.
🤖 Generated with Claude Code
https://claude.ai/code/session_011QG86sovTbfnPNY9SfhSmo