fix(benchmark): survive release version bumps without orphaning evidence - #498
Merged
Conversation
The dependency-lock digest covers package.json, so the 0.138.0 version bump invalidated the published CodeTraceBench evidence pin and blocked the release. Split the pin: the current constant tracks the live lockfiles, the evidence keeps its creation-time digest and package version, and a new test recomputes the lock digest with the version stamped back to prove the two differ by the version field alone. A real dependency change still fails that proof and forces a new benchmark run or explicit retirement.
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
The v0.138.0 publish failed in verify: the dependency-lock digest covers
package.json, so the release version bump changed the computed digest (1e03f2…→d06e3e…) and the benchmark pin check refused. The design had never survived a release — 0.138.0 is the first one since the benchmark evidence landed.Fix: split the pin.
ANALYST_BENCHMARK_DEPENDENCY_LOCK_SHA256now tracks the current lockfiles (checked bycheck-analyst-benchmark-implementation.mjs, unchanged).ANALYST_BENCHMARK_EVIDENCE_DEPENDENCY_LOCK_SHA256+ANALYST_BENCHMARK_EVIDENCE_PACKAGE_VERSIONfreeze what the published receipt was created against; the receipt binding tests compare against these, so the sealed evidence is never edited.package.jsonversion substituted back to the evidence version and requires it to equal the evidence digest — proving the only drift is the version stamp. Any real dependency change fails this proof and forces a new benchmark run or explicit retirement of the evidence.Proof: digest check valid, full suite 4,606 passed / 0 failed (exit 0), typecheck + lint green. Tag
v0.138.0moves to the merge commit after this lands (the original tag never published — verify failed before both publish jobs).