Repository navigation
fix(benchmark): use stored keys and retain replayable fixtures - #719
Merged
zzwong merged 2 commits intoNov 10, 2026
Conversation
zzwong
commented
Nov 10, 2026
This was referenced Nov 10, 2026
zzwong
marked this pull request as ready for review
November 10, 2026 14:21
zzwong
merged commit Nov 10, 2026
b566e94
into
zzwong/shared-identity-experiment-core
12 checks passed
This was referenced Nov 10, 2026
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
Stacked on #718. Resolve identity-provider API keys through explicit credential-store references and fixed provider key names. Offline suites use
credential: {store, name}for direct and saved-candidate selection; setup stays incr set-credential. The unshipped environment-key field and flag are removed together with no fallback.Preflight the largest complete retained fixture before reading a key or sending requests. If serialization would exceed replay’s limit, retain the valid original input and report the size failure with zero provider calls. Compact fixture encoding also avoids unnecessary HTML expansion. Regression tests inspect each actual cohort request and preserve unknown costs for nonzero cache writes.
This addresses the credential, retention and test findings on #714 and #715, and supplies the contracts needed by the dependent review diagnostic.
Validation
make checkpassed: tidy, formatting, lint, the complete test suite and build.