Selecting a skill in a pending question answer can leave the picker stuck and trigger Maximum call stack size exceeded. The answer callback reads the old editor snapshot and resets the caret before the new answer has rendered, re-entering Lexical with stale token offsets.
Let the editor retain its selection when reporting an answer change. Picker replacements already schedule their own caret placement. This removes the extra snapshot read and focus reset.
Refs #7805. This fixes the reproduced recursion and caret corruption; it preserves the existing skill picker. The broader request to make question answers plain text remains open. Rebased onto main.
Native Electron verification with two synthetic pending questions: the base leaves $bro unconverted, keeps 12 picker options open, and logs a stack overflow. The fix creates the Bro chip, closes the picker, and reports no errors. Typing/backspace works; Next starts empty, Previous preserves the first answer, and removing the request restores the original draft. No final answer was submitted to a provider. Web typecheck and 36 existing editor/answer tests pass; scoped lint exactly matches 17 base warnings.
Before/after images and video are captured. GitHub asset upload remains pending while its background browser connection is unavailable.
Model: GPT-6. Harness: Codex.
Rebased on main 20363c32c9bfdbf49c2716ef11d1f18483fcc01b. Scoped validation rerun after rebasing; existing native and human review limits remain.
Tested in the web app against a dev server on this branch and on main. Claude asked a real question with AskUserQuestion, and I answered it by typing $bro and picking the skill from the picker.
Main: no chip, and typing x lands mid-word ($bxro) | This PR: the skill becomes a chip and typing continues after it |
|---|---|
| Main: clicking the Bro option also leaves plain text | This PR: clicking the Bro option creates the chip |
|---|---|
Regular composer on this PR still works with a skill chip and an @-mention together:
Not reproduced: the "Maximum call stack size exceeded" error on main. Not tested: Next/Previous between several questions (each test question was single). A React flushSync warning shows when picking a skill, on main and on this PR, in both the answer box and the regular composer, so it isn't from this change.
Bugbot uses usage-based billing for this team and has hit its on-demand spend limit.
A team admin can raise the spend limit in the Cursor dashboard, or wait for the next billing cycle to continue.
Verdict: Approved at 1728ebb
Macroscope's review found this PR approvable — This is a focused, single-file fix that removes stale caret restoration during pending-question answer updates while preserving existing answer state handling and explicit replacement behavior. The runtime impact is narrow and readily accounted for, with no schema, deployment, security, billing, or configuration changes.
Notes:
No code changes detected at 88b2083. Prior analysis still applies.
You can add or adjust custom eligibility rules. Learn more.
✅ Thread transfer remains within every enforced ceiling.
| Provider | Metric | Main baseline | This PR | Impact | PR ceiling | |
|---|---|---|---|---|---|---|
| Codex | Total thread wire | 5.0 KiB | 4.9 KiB | −23 B (−0.5%) | 6.8 KiB | ✅ |
| Codex | Thread snapshot wire | 3.8 KiB | 3.8 KiB | 0 B (0.0%) | 4.9 KiB | ✅ |
| Codex | Live turn WebSocket wire | 1.2 KiB | 1.2 KiB | −23 B (−1.9%) | 2.0 KiB | ✅ |
| Codex | Live turn WebSocket decoded | 20.9 KiB | 20.8 KiB | −41 B (−0.2%) | 29.3 KiB | ✅ |
| Codex | Live turn messages | 2 | 1 | −1 (−50.0%) | 8 | ✅ |
| Claude | Total thread wire | 5.0 KiB | 5.0 KiB | 0 B (0.0%) | 6.8 KiB | ✅ |
| Claude | Thread snapshot wire | 3.8 KiB | 3.8 KiB | 0 B (0.0%) | 4.9 KiB | ✅ |
| Claude | Live turn WebSocket wire | 1.2 KiB | 1.2 KiB | 0 B (0.0%) | 2.0 KiB | ✅ |
| Claude | Live turn WebSocket decoded | 21.2 KiB | 21.2 KiB | 0 B (0.0%) | 29.3 KiB | ✅ |
| Claude | Live turn messages | 2 | 2 | 0 (0.0%) | 8 | ✅ |
Baseline: 85272ee · PR result: 88b2083 · Source CI: success
10 historical turns, 5 command tools per turn, 878.9 KiB retained MCP result per historical turn, and a 1.05 MiB retained result in the measured turn.
Updated in place by a trusted workflow. PR artifacts are strictly validated and never executed.
Bugbot uses usage-based billing for this team and has hit its on-demand spend limit.
A team admin can raise the spend limit in the Cursor dashboard, or wait for the next billing cycle to continue.
Note
It looks like this branch is under active development. To avoid overwhelming you with review comments due to an influx of new commits, CodeRabbit has automatically paused this review. You can configure this behavior by changing the reviews.auto_review.auto_pause_after_reviewed_commits setting.
Use the following commands to manage reviews:
![cursor[bot]](https://avatars.githubusercontent.com/u/206951365?s=80)
@coderabbitai resume to resume automatic reviews.@coderabbitai review to trigger a single review.Use the checkboxes below for quick actions:
No actionable comments were generated in the recent review. 🎉
c7173bc3-17dd-4c18-a538-7b9cc53bbf8cReviewing files that changed from the base of the PR and between c5f06b451fe3836594375fc059560cf84cf200e2 and cac90ee06c587c438aad8b70e6edb866c9a8c5a1.
apps/web/src/components/ChatView.tsxIncluded review availability: This review used your included allowance. Your plan provides up to 10 included reviews per hour; 7 remain after this review.
The pending custom-answer callback now accepts the question ID and answer value. It no longer compares composer snapshots or restores focus at cursor offsets.
Changes: Composer Selection Ownership
| Layer / File(s) | Summary |
|---|---|
Simplify custom-answer callback apps/web/src/components/ChatView.tsx | The callback no longer receives cursor metadata or restores composer focus. Selection management is left to the editor. |
Priority: ➖ Normal
Estimated code review effort: 1 (Trivial) | ~5 minutes
Change: Bug fix
Suggested reviewers: matthewferoz
cac90Pending-answer edits retain selection through the editor’s cursor state, and no concrete user-facing regression is established. The change appears ready for normal checks.
| Check name | Status | Explanation |
|---|---|---|
| Title check | Passed | The title clearly and concisely describes the primary change: preventing question answers from resetting the editor caret. |
| Description check | Passed | The description covers the problem, change, scope reference, verification results, limitations, and UI evidence. It reports the relevant tests, real-app checks, known warning, and untested multi-quest… |
| Linked Issues check | Passed | Check skipped because no linked issues were found for this pull request. |
| Out of Scope Changes check | Passed | Check skipped because no linked issues were found for this pull request. |
Comment @coderabbitai help to get the list of available commands.
Added 1 commit