Repository navigation
Conversation
|
@codex review Please review the batching boundary (individual choices versus confirmed manuscript delivery), interrupted decision recovery, and the embedded toolbar's loading/error behavior. The changes are based on current main and include real-browser/LaTeX regression coverage. |
Codex Review Summary: ✅ CompletedThis 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: e06592320a
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 Both findings were fixed in 5d3ee17 with regressions verified red before the fix and green afterward. Please review the updated head, especially legacy batch completion and reconciliation. The real browser/LaTeX test passes with zero builds during choices and one after application; the batching and toolbar regressions now also run in CI. |
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 5d3ee1794c
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 The deferred commit finding is fixed in 3b7e013. Both save endpoints now use the same confirmed-delivery check, with the commit-hash fallback limited to legacy responses. Added coverage through the real commitTask adapter; all batching route regressions and typecheck pass. Please review the latest head. |
|
Codex Review: Didn't find any major issues. Can't wait for the next one! 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". |
|
@codex review Please review the latest head f0c85de. The previous review completed without findings on 3b7e013. The only additional change makes the toolbar browser regression portable across Linux/macOS fonts: wait for the settled non-building toolbar and assert compilation preserves its measured height instead of assuming 44px. The full local Chromium review-flow check passes. CI has been restarted for this exact head. |
|
Codex Review: Didn't find any major issues. Nice work! 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". |
|
@codex review The user found two additional review-flow defects. Please review latest head 65732f2, focusing on note advancement after both synchronous decisions and asynchronous alternative saves, selection/preview identity during failed or racing requests, and the bounded read-only reconnect. Both original UI defects were reproduced by browser regressions on f0c85de. The fixes pass at 1440px and 390px; the existing note-results browser suite and 24 related unit tests pass. CI now includes both note-results and the new failure/race/advancement browser script. No user desktop app was operated. |
cu1ch3n
changed the title
|
Codex Review: Didn't find any major issues. Can't wait for the next one! 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". |
Review the existing generic request-failure translations in their added note-preview context. Preserve their text and generation provenance while refreshing context fingerprints.
|
@codex review Please review final head 16c13a5. Production code is unchanged from 65732f2, which completed review without findings. This follow-up only refreshes nine translation-context fingerprints after confirming the existing generic request-failure translations still apply to note previews, and adds the drawer screenshot. The full i18n checks now pass; CI is rerunning. |
|
Codex Review: Didn't find any major issues. Another round soon, please! 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". |
|
@codex review Please review head 041b05d. Updating the static demo assertions exposed another reproducible issue: the embedded renderer overwrites the host's selected-note classes when it repaints after a decision. ReviewNoteConnections now observes class changes only on linked change elements and restores those markers without re-scrolling or re-focusing; comparisons before toggling avoid observer feedback loops. Please check observer lifetime and focus behavior. The full static playground browser test now passes, including the real renderer's pending-note marker, drawer retention, explicit outside-click closure and wraparound to an earlier pending note. Its 14 unit tests pass. The ordinary-editor note-results, note-options-flow (1440px/390px) and review-flow browser suites also pass. |
|
Codex Review: Didn't find any major issues. Keep them coming! 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". |
|
Independent web-only audit completed on
One existing behavior was recorded outside this change: Escape on a drawer tab is consumed by the review tool shortcut handler, so it does not close the drawer. The relevant handler and dismissal code are unchanged from the base; outside-click dismissal passes. No unrelated shortcut change was added. Current-head CI remains successful and Codex review is completed with no findings; all three older threads remain resolved. No user desktop app or real writing project was operated. No merge, release, or deployment. |
|
@codex review Please perform a fresh review of the full PR at head 041b05d. Focus on confirmed decision batching across all write routes, option/baseline request races and retry identity, drawer advancement during asynchronous saves, and the lifetime of renderer highlight observers. The user requested another review/fix loop. The prior exact-head CI and audit are green; report actionable regressions, including gaps not covered by the existing browser tests. |
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 041b05d3d2
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 Please review head f19fcb1. Both findings from the latest review are fixed, with failing-before/passing-after regressions. The legacy batch fix also handles an accepted alternative finishing in a resumed build, using one workspace listener and the same zero-pending legacy predicate; deferred tasks must still compile only on application. Confirmed saves now retire their request identity so a retained final drawer can decide again after undo, while failed retries preserve their identity. Local validation: 202 app/workspace/note tests, typecheck/build, 1440px/390px drawer flow and real-LaTeX browser batching/failure recovery pass. Please check the full PR and these completion/identity boundaries. |
|
Codex Review: Didn't find any major issues. 🚀 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". |
|
@codex review Please perform another fresh full-PR review at head f19fcb1, as requested by the user. All prior findings are fixed and exact-head CI is green. Look for remaining concrete regressions: asynchronous/recovered option completion and legacy manuscript flushing, listener lifetime across multiple server instances, response-loss/retry/undo identities in the retained drawer, and navigation while saves or preview reads are in flight. Please validate findings against actual reachable paths rather than repeating resolved issues. |
|
Codex Review: Didn't find any major issues. What shall we delve into next? 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". |
What and why
Review choices could trigger manuscript compilation before application, and a duplicate build badge could wrap below the toolbar. Accepting a note also closed its drawer; a failed option switch could leave the previous option’s PDF under the newly selected tab.
How it was checked
f0c85de.f19fcb1; fresh Codex review completed with no new findings after another full-PR review. All five review threads are resolved, including the two findings fixed inf19fcb1.f0c85depassed CI and review. Its full local UI run exercised all 30 scripts: 28 passed; Git timeout assertions atcheck-git-workspace.mjs:153andcheck-git-panel.mjs:249failed identically on untouched base9dcf5ca. These baseline failures are unrelated to this PR.