Repository navigation
gallery: incoming-message-disclosure plays open the "Received N messages" row first - #32820
Conversation
…lines #32720 gathers delivered messages into one collapsed "Received N messages" row, so the fanout-infra line the fold test waits for is not drawn until that row opens. The test and the gallery entry's plays now open it first. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Agent-Session: https://cmux.com/deeplink/agent?id=aid-2f629b93d5a20c58e6f8
|
cmux status (one comment, edited in place; each line is updated by the job that owns it) Review App of this head (app.zip, kept 1 day)
Gallery before/after for
States
Pick other states with a Passes: CI passes on CI passes on Written by web and size checks: green Shipped in:
|
GitHub Actions
Inactive
|
Review (subagent opus-5.5 medium): SHIP at 18b7c90 Visual verdict: unchanged by this diff. It changes no UI code, only the gallery note and play. Checked: the selector reaches the MessageGroupHead button of the single delivered inbox batch, which has Follow-ups: other gallery plays and tests that wait for delivered |
…se-red-fold-keeps-place # Conflicts: # webviews/test/agent-pane-fold-keeps-place.test.ts
GitHub Actions
Inactive
teamleaderleo
changed the title
|
Review (subagent opus-5.5 medium): SHIP at 6bf3254 Visual verdict: unchanged by this diff (gallery play and note only). The play opens the single delivered "Received 3 messages" batch (idempotent on aria-expanded) before the waiting group and the >=5 Follow-ups: none. |
…waiting group at the end With the "Received 3 messages" row opened first, the plays render again (they timed out on base). Opening the waiting group at the transcript's end then pushes the rows above up 66px (the pinned end, Leo 2026-11-02), measured 0.0777 in gallery health run 41563279796, the same kind of shift a-shipped-open already carries. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Agent-Session: https://cmux.com/deeplink/agent?id=aid-2f629b93d5a20c58e6f8
GitHub Actions
Inactive
|
Review (subagent opus-5.5 medium): SHIP at b5f1fcf Visual verdict: unchanged by this diff (gallery play, note and shift baseline only). Delta since the SHIP at 6bf3254: B/C/D shift baselines go from 0 to 0.0777. On base those plays timed out before the waiting-group step, so the 0 rows never measured it. The measured shift (gallery health run 41563279796) is the deliberate open at the transcript end pushing the rows above up 66px, the same kind of shift that a-shipped-open already carries at 0.16476. D runs the identical play. Follow-ups: lower D's row with |
Gallery health run 41564747382 measured a-shipped-open at 0.2829: the waiting group opening at the end (0.0777, as in B/C/D) plus the report opening (0.2051). On base the play times out, so its row of 0.16476 predates #32720. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Agent-Session: https://cmux.com/deeplink/agent?id=aid-2f629b93d5a20c58e6f8
GitHub Actions
Inactive
|
Review (subagent opus-5.5 medium): SHIP at a1dc1a2 Visual verdict: unchanged by this diff (gallery play, note and shift baseline only). Delta since the SHIP at b5f1fcf: the a-shipped-open row goes from 0.16476 to 0.2829. The old row came from #30256, before #32720. The new total is the 0.0777 push-up from opening the waiting group at the end plus the report opening (0.2051), where only user-6 below the report moves down and the opened line stays put. That is the intended "panel grows under it" behaviour, not a hidden regression. Follow-ups: a per-step baseline for the report-open wait would catch drift without the push-up folded in. If opening at the end is later made to hold still, lower all four rows together. |
|
Found 1 test failure on Blacksmith runners: Failure
|
![Fix with [code]smith](https://pr-comments-assets.blacksmith.sh/codesmith/fix-with-codesmith-light-3x.png)
Bugbot needs on-demand usage enabledBugbot uses usage-based billing for this team and requires on-demand usage to be enabled. A team admin can enable on-demand usage in the Cursor dashboard. |
|
Review (next-review patch-id; carried): SHIP at 603f906 The independent SHIP at 595e24a carries to this head: the PR's own change ( |
|
Review (next-review patch-id; carried): SHIP at 4ebe630 The independent SHIP at 595e24a carries to this head: the PR's own change ( |
…Received N messages" row The plays open #32720's gathered row before they open its lines, so the rows under it move once more; rows set to this PR's measured worst (run 41631275437), as #32820 does for the disclosure entry. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Agent-Session: https://cmux.com/deeplink/agent?id=aid-3bab30503f0b9a7059e5
…er's mark centres on its line (#33752) * gallery: received-row plays open "Received N messages"; a sender's mark centres on its line #32720 gathers received lines next to each other into one closed "Received N messages" row, so plays that waited for .cv-received lines timed out (incoming-message files, files-elsewhere, minimal-open, minimal-fade-count, minimal-inline-name; agent-message-row folded-run-opened). They now open that row first, as #32820 does for the disclosure entry. The row's head draws the one sender's mark (SenderMark) in the tool icon slot; its glyph sat in nested inline spans on a text baseline, 2px above the line's centre (center-line, row received:user-N in message-preview and folded-run). The wrappers now centre the glyph, in the head and in a received line's mark. agent-message-row#both-directions-opened's shift row follows the opened lines' current geometry (0.02068 -> 0.0265). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Agent-Session: https://cmux.com/deeplink/agent?id=aid-3bab30503f0b9a7059e5 * gallery health: shift rows for the received plays that now open the "Received N messages" row The plays open #32720's gathered row before they open its lines, so the rows under it move once more; rows set to this PR's measured worst (run 41631275437), as #32820 does for the disclosure entry. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Agent-Session: https://cmux.com/deeplink/agent?id=aid-3bab30503f0b9a7059e5 --------- Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
|
take it. The gallery-health base-red owner may carry this fix on their own branch: the |
|
Carried on #33987 (your commits kept with authorship). It also fixes the one state still red here, Agent session: agent session |
|
Review (next-review patch-id; carried): SHIP at 01584d1 The independent SHIP at 595e24a carries to this head: the PR's own change ( |
…he plays measure With each opening its own clicked step, gallery health run 41665103842 measured b-received-wording, c-line-count and d-sender-only at 0 (back to base's 0, from #32820's 0.0777) and a-shipped-open at 0.16622/0.16639 (0.1664, from #32820's 0.2829; base's 0.16476 predates #32720's received row, whose opening now lands in the same play). The report line is waited for before it is clicked. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Agent-Session: https://cmux.com/deeplink/agent?id=aid-3bab30503f0b9a7059e5
… clicked step (carries #32820) (#33987) * test(agent pane fold): open the "Received N messages" row before its lines messages" row, so the fanout-infra line the fold test waits for is not drawn until that row opens. The test and the gallery entry's plays now open it first. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Agent-Session: https://cmux.com/deeplink/agent?id=aid-2f629b93d5a20c58e6f8 * gallery: B/C/D disclosure variants keep the push-up from opening the waiting group at the end With the "Received 3 messages" row opened first, the plays render again (they timed out on base). Opening the waiting group at the transcript's end then pushes the rows above up 66px (the pinned end, Leo 2026-11-02), measured 0.0777 in gallery health run 41563279796, the same kind of shift a-shipped-open already carries. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Agent-Session: https://cmux.com/deeplink/agent?id=aid-2f629b93d5a20c58e6f8 * gallery: a-shipped-open carries the waiting group's push-up too Gallery health run 41564747382 measured a-shipped-open at 0.2829: the waiting group opening at the end (0.0777, as in B/C/D) plus the report opening (0.2051). On base the play times out, so its row of 0.16476 predates #32720. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Agent-Session: https://cmux.com/deeplink/agent?id=aid-2f629b93d5a20c58e6f8 * gallery: incoming-message-disclosure plays open each fold as its own clicked step The plays opened the "Received 3 messages" row and the waiting group with raw element clicks in one task, then waited. At the transcript's end each opening pushes the rows above up by design (#32644), and the box-shift gate exempts that only while the step's own target sits in the row marked data-bottom-fold (#33784). A raw click is no step and has no target, so the push-up landed in the following wait, scored as shift (c-line-count 0.1328, user-1 and assistant-2 moving -21.9px). Each opening is now ctx.click, settled before the next, and fanout-infra's report opens the same way. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Agent-Session: https://cmux.com/deeplink/agent?id=aid-3bab30503f0b9a7059e5 * gallery health: incoming-message-disclosure shift rows back to what the plays measure With each opening its own clicked step, gallery health run 41665103842 measured b-received-wording, c-line-count and d-sender-only at 0 (back to base's 0, from #32820's 0.0777) and a-shipped-open at 0.16622/0.16639 (0.1664, from #32820's 0.2829; base's 0.16476 predates #32720's received row, whose opening now lands in the same play). The report line is waited for before it is clicked. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Agent-Session: https://cmux.com/deeplink/agent?id=aid-3bab30503f0b9a7059e5 * gallery health: keep base's other shift rows; only the four disclosure rows change Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Agent-Session: https://cmux.com/deeplink/agent?id=aid-3bab30503f0b9a7059e5 --------- Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com> Co-authored-by: cmux-bot[bot] <348531757+cmux-bot[bot]@users.noreply.github.com>
Base-red-job: web / react-apps-check
The base red
webviews/test/agent-pane-fold-keeps-place.test.ts:80(TimeoutError: waitFor: Timeout 30000ms exceeded) was fixed on base by #32832 while this PR was open. This PR now carries only the matching fix in the gallery entry that the test renders.Cause: #32720 ("a turn's inbox notices and received lines are one 'Received N messages' row") deliberately gathers delivered messages from other sessions into one collapsed "Received N messages" row (Leo, 2026-11-02: "can't we simply just display most of these as recieved x messages and all of them get revealed then we can expand the messages"). In the
agent-pane.incoming-message-disclosuregallery fixture, the three delivered lines (dep-fleet, fanout-infra, nx-sidebar) are now inside that row, so the fanout-infra line's toggle the test waits for is not drawn until the row opens.The behaviour change was intended, so the gallery follows it rather than the product changing:
IncomingMessageDisclosure.gallery.ts:openWaiting(used by thea-shipped-openplay and the B/C/D variants, which otherwise wait for 5.cv-receivedlines that never draw) opens that row too, and thea-shippednote describes the row.Risk: test and gallery only; no shipped code changes.
Agent session: agent session
🤖 Generated with Claude Code