Improve v1.0.2 performance and reliability - #10
Conversation
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: b797b65948
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| ); | ||
| let mut scan_overrides = overrides.clone(); | ||
| scan_overrides.codegraph = codegraph.synchronized_count > 0; | ||
| scan_overrides.codegraph = codegraph.changed_count > 0; |
There was a problem hiding this comment.
When csgraph sync is run with an initialized CodeGraph index that is already current, changed_count is 0 even though CodeGraph synchronization/corroboration is enabled by default. Setting scan_overrides.codegraph from only changed_count makes the following native scan skip run_codegraph_corroboration, so the first publish after enabling CodeGraph (or any native-only graph change while the index is current) produces no CodeGraph evidence or affected-test enrichment despite reporting the index as synchronized; avoid using index changes as the sole gate for whether the scan may use CodeGraph.
Useful? React with 👍 / 👎.
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: b797b65948
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| run_codegraph_sync( | ||
| &binary, | ||
| path, | ||
| Duration::from_millis(policy.max_codegraph_sync_wall_time_ms_per_repo), |
There was a problem hiding this comment.
For a stale CodeGraph index where codegraph status itself runs close to maxCodeGraphSyncWallTimeMsPerRepo, this gives the subsequent codegraph sync another full per-repository budget, so a single repository can consume almost 2x the configured limit before the scan begins. The execution policy defines this as the maximum wall time for one repository-local CodeGraph synchronization, so share a single deadline or pass only the remaining time into run_codegraph_sync.
Useful? React with 👍 / 👎.
| /// The compatibility wrapper currently returns `Ok`; the result shape is retained so existing | ||
| /// callers do not require an API migration in the patch release. | ||
| pub fn link_http_boundaries(boundaries: &[HttpBoundary]) -> Result<Vec<Edge>, LinkError> { | ||
| Ok(link_http_boundaries_with_ambiguities(boundaries).edges) |
There was a problem hiding this comment.
For library callers that still use the public link_http_boundaries API, duplicate exact providers now produce Ok with the ambiguous edge silently omitted; there is no degradations field in this return type, so callers can publish or use an incomplete graph while their existing LinkError::AmbiguousProvider handling is bypassed. Keep this wrapper fail-closed and reserve the new _with_ambiguities API for callers that actually consume the ambiguity report.
Useful? React with 👍 / 👎.
Align packaged release and installation guidance with the version being published.
1 participant
No description provided.