Repository navigation
fix: backport 0.12.1 batch 58 (25 wide-bar fixes) - #32855
Conversation
) (cherry picked from commit 0006912)
…30492) With Slack tokens set but no KIROCREW_OWNER_ID, or a workspace that fails enterprise validation, init_socket_mode disables Slack and only logs the reason. The connect error stays empty, so System > Services shows a bare Not connected. Record the reason as the connect error so the status surfaces show it. (cherry picked from commit ee76465) Original author: Rohan Khanderia (#30492).
) The spawn note and the Subagent Model/Effort help say a dedicated subagent runtime costs about 0.55 GB, the figure measured in #23510. The Session Sharing help, the spawn tool descriptions, the llm-council skill and the subagent spec still said ~400MB, so one runtime showed two figures. They now all use the measured one. Fixes #30362 (cherry picked from commit f2ac71e) Original author: Zezhen Xu (#30581).
Restoring a full snapshot of a home over 4 GiB without
--max-archive-gib is refused with a message saying such an archive is
one "which no memory bundle produces". A whole-home bundle reaches the
same bound, so the wording reads wrong for the most common case.
Both refusals in _refuse_oversized_archive now name the restore bound
instead: the entry-count one matches the writer-side message ("which
`kirocrew restore` refuses"), and the byte one says the archive is over
the bound restore holds it to, then keeps the --max-archive-gib hint.
Fixes #30398
## Pattern harvest
Not generalizable: a one-off wording slip in two refusal strings.
(cherry picked from commit d34fe7e)
Original author: Matthew Briones (#30647).
…30953) The collapsed rail drew 20px icons in 32px rows, so the icons sat close together and the narrow rail read crowded. Collapsed rows now use the expanded rows' vertical padding, so both are 36px tall. Expanded rows are unchanged. Closes #30932 Co-authored-by: CrysisDeu <crysisdeu@users.noreply.github.com> (cherry picked from commit 722b07e) Original author: Zezhen Xu (#30953).
Each Issue Radar left-rail accordion header opens its section and shows a rotating chevron, but the header button never exposed whether its section was open. Add aria-expanded from the section expanded prop and pin it in the left rail test. (cherry picked from commit a3a0917) Original author: Rohan Khanderia (#31187).
…#31367) The sync shortcut button in MD Notebook settings toggles key recording on and off, but its fixed aria-label hides the Recording text and it carries no pressed state, so a screen reader never hears that recording started. Add aria-pressed from the recording state and pin both states. (cherry picked from commit d8d4b05) Original author: Rohan Khanderia (#31367).
…#31214) The single-app featured card in Discover printed the category id, which is the English CATEGORY_ORDER value, so every language saw Research & Writing or Developer Tools in English. The Discover list row already renders the catalog label through categoryLabel; the featured row now does the same. (cherry picked from commit 8095bc3) Original author: Rohan Khanderia (#31214).
Folder rows in the skill directory browser show and hide their children, but the row button never exposed whether the folder was open. Add aria-expanded from the open state and pin both states in the existing expand test. (cherry picked from commit 92e76f7) Original author: Rohan Khanderia (#31486).
The two buttons that toggle a Knowledge folder source's inline scan panel -- the chevron button and the "N files failed" link -- flip a chevron and swap their label but never set aria-expanded, so a screen reader cannot tell the control is a disclosure or whether its panel is open. Set aria-expanded on both, reflecting the open state, matching the sibling folder-picker button and the SourceGroup folder header. No behaviour or input handling changes. Fixes #31865 (cherry picked from commit 73889c1) Original author: Rohan Khanderia (#31870).
Issue Radar filter rows (Open, Merged, Draft, Assigned to me and the rest) show they are on only by colour, so assistive tech hears a plain button either way. Add aria-pressed from the row active prop and pin the on and off states in the PR filters test. (cherry picked from commit 2217bd2) Original author: Rohan Khanderia (#31683).
…31871) The Design Tweak History toggle and the preview Dimensions dropdown open and close a region, but neither button said whether it was open. Their labels stay the same in both states, so a screen reader user could not tell. Both buttons now carry aria-expanded with their open state. (cherry picked from commit a47c2d4) Original author: Rohan Khanderia (#31871).
Issue Radar label pills filter the list on and off, but a selected pill shows it only by its fill colour and bold text, so assistive tech hears the same button either way. Add aria-pressed from the selected prop and pin both states in the label palette test. (cherry picked from commit 5cae4a9) Original author: Rohan Khanderia (#31947).
…32054) The Issue Radar rail Dashboards rows mark the dashboard on screen only by colour, bold text and an accent icon, so assistive tech cannot tell which one is open. Set aria-current="page" on that row and pin it, and its absence while another view is on screen, in the rail entry tests. (cherry picked from commit f6cfe99) Original author: Rohan Khanderia (#32054).
…2045) The Memory graph tab's group filter is a row of toggle chips -- an "All" button plus one per group. The active chip is shown only through accent styling, with no aria-pressed, so a screen reader announces each as a plain button and cannot convey which filter is applied. Set aria-pressed on every chip, reflecting the active filter ("All" pressed when no group filter is set). No behaviour or layout change. Fixes #32023 (cherry picked from commit 3e3ae0a) Original author: Rohan Khanderia (#32045).
On a harness without mid-turn steer, a block in a turn that already answered is explained afterwards by the refusal continuation. Its finished-task instruction asked for one short line noting the block, so that line became the user's last message and buried the summary; collapse-all mode showed only the block. The finished task now opens with a one-line recap of the outcome, named as the one restatement the "do not repeat" rule allows, then one short line noting the block. Closes #21962 (cherry picked from commit f980022)
The Workflows library shows the saved workflows as a list of buttons, and marks the open one with accent styling alone -- no aria-pressed. A screen reader announces each as a plain button and cannot convey which workflow is open in the editor, including the row auto-selected on load. Set aria-pressed on each list button, reflecting the selected definition. No behaviour or layout change. Fixes #32161 (cherry picked from commit a1d4896) Original author: Rohan Khanderia (#32166).
First Principles Review (Opus 5.5) — ✅ PASSPremise-level review of First-Principles-Verdict: PASS Nothing blocks. Every item is a cherry-pick of a fix already merged on What this change shipsInventory (10 items) — 10 justifiedIntent: a FIX. It brings 25 small fixes from
[FIRST-PRINCIPLES-REVIEWED] db12756 Description read[DESCRIPTION-READ] 6c7e0af7ee96ba77789cd3276d4116038229db80b45cef3277068de6d5faa1dd That sha256 names the PR description this verdict read. Recompute it with |
UX Review (Opus 5.5) — ✅ PASS (a BLOCK with no Trigger and no Outcome does not block)UX-level review of UX-Verdict: PASS cannot evaluate: missing screenshots of the visible frontend picks. This PR attaches and commits no image, so the blind read was not performed. Evidence gaps
[UX-REVIEWED] db12756 Description read[DESCRIPTION-READ] f4367737c4a972c9fafc47e0a7c7d386883d7e2d414ab480c08fd6d28ccc47ca That sha256 names everything this verdict read: the PR description, plus the 3 evidence file(s) the workflow collected from it and handed this review. It covers that evidence because the media strip replaces every attachment URL with the same placeholder, so an edit swapping one attachment for another would leave the description's captured text identical while what this review judged changed. Recompute it with |
Design Review (Opus 5.5) — ✅ PASSDesign-level review of Design-Verdict: PASS Nothing blocks: these are clean cherry-picks of fixes already merged on Notes
[DESIGN-REVIEWED] db12756 Description read[DESCRIPTION-READ] bc7148aa4a68a415902d53f2650e6d0508f73f14b2562285a67620a861943968 That sha256 names everything this verdict read: the PR description, plus the 1 evidence file(s) the workflow collected from it and handed this review. It covers that evidence because the media strip replaces every attachment URL with the same placeholder, so an edit swapping one attachment for another would leave the description's captured text identical while what this review judged changed. Recompute it with |
Opus 5.5 Review — ✅ no blocking findingsReviewed Verdict parsed from the review's SHA-scoped output markers for commit False positive or not applicable? A repository writer can comment: |
GPT 6.1 Review — ✅ no blocking findingsGPT 6.1 completed its review of This comment is updated in place on each push. Review detailsNo findings. False positive or not applicable? A repository writer can comment: |
Problem / Motivation
Goal: Bring 25 small fixes that merged on
mainafterv0.12.0ontorelease/0.12.0, one commit each, in a PR of their own.Why it matters
Each row is a user-visible fix from the 0.12.1 wide-bar list (Ray, 11-07). Every pick applied cleanly onto release.
Not a goal
main; nothing is changed only here.CHANGELOG.mdchange.What changed (motivation → approach → change)
Each main squash commit was cherry-picked with
-xontorelease/0.12.0@aa7dfb4aae, in merge order onmain, one commit each. Every pick applied cleanly, with no prerequisite and no hand-resolved hunk; every commit has added and removed lines identical to its main commit.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 commits carried a non-public author email or an internal co-author trailer, which the push guard refuses, so they are authored with the public identity and name their original author in the message; internal trailer lines are removed. The other 3 keep their original author.
builder-security checked: each change is the one already merged on
main; no bind, listener, new route, auth mechanism or storage grant is added.Backwards compatibility
Compatible: each change is the one already merged on
main, and its compatibility note applies unchanged.Removes nothing: no setting, key or caller is removed.
Tests
Proof: not provable -- other: a backport of main commits whose proofs ran on their own PRs
release/*, so the direct test modules were run here: every touched test file plus every test file that imports a touched module, 55 files,-n 4 --dist loadgroup. 2684 passed, 0 failed.tsc -p tsconfig.app.json --noEmitpasses.vitest runon every touched test file and every touched component name: 62 files, 686 passed, 3 failed. The 3 are insrc/pages/SchedulePage.secrets.test.tsxand fail the same way on a cleanrelease/0.12.0@aa7dfb4aae.Manual verification
N/A - each fix was verified on its own main PR. The 0.12.1 RC soak is the end-to-end check.
Screenshots / video
Why no screenshot: each frontend pick carries the same rendered change as its main PR, whose screenshots show it.
Related Issues
Backport to the 0.12 line for 0.12.1 of #30394, #30492, #30523, #30581, #30647, #30577, #31021, #30953, #31187, #31010, #31367, #31499, #31228, #31214, #31486, #31870, #31683, #31871, #31947, #31927, #32054, #32045, #31776, #32276, #32166.
Sibling sweep
Siblings: none -- each source PR carried its own sibling sweep; this PR only carries those fixes.
Pattern harvest
Not generalizable: this PR only carries fixes already on
main. The source PRs have their own harvest.Checklist
feat|fix|docs|style|refactor|perf|test|chore|ci|build|revert: ...)Contribution License Agreement
Contributions are accepted under the project's Apache-2.0 license; see the Licensing section of CONTRIBUTING.md.