Keep established reviews alive during remote outages - #141
Conversation
Codex Review SummaryThis comment shows the latest Codex review activity on this pull request.
ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings. |
![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: a3797bd61c
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".
|
@codex review |
|
Codex Review: Didn't find any major issues. Keep it up! Reviewed commit: ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
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". |
1 participant
A connection failure during remote refresh currently stops the watcher, local server, and active agent even after a Review is usable. Keep the last valid Snapshot available, report that it has not refreshed, and retry on the existing remote refresh cadence. Recovery resumes normal updates without regenerating notes for an unchanged target.
The change classifies connection failures from Git fetch, remote default-branch lookup, and GitHub PR metadata lookup. HTTP 500, 502, 503, and 504 responses are also retried. Startup failures and permanent target errors still exit nonzero.
Regression coverage uses local remotes and fake providers to exercise repeated outages, unchanged and changed targets after recovery, HTTP availability, completed notes, in-flight agent completion, and shutdown. The recovery tests failed before the fixes, including Git and GitHub CLI HTTP 500 responses.
Validation:
pnpm run lint,pnpm test, andpnpm exec fallow audit --changed-since origin/main.Fixes #140