Repository navigation
docs(rfc): record the sidebar app menu and status marks decision - #31104
Pearcekieser wants to merge 1 commit into
Conversation
UX Review (Opus 5.5) — ⏭️ skippedRevision |
First Principles Review (Opus 5.5) — ⏭️ skippedRevision |
Design Review (Opus 5.5) — ✅ PASSDesign-level review of Design-Verdict: PASS The PR adds only a design document and an index row, so it renders nothing and takes nothing away. The decision reuses existing paths: detail-page Notes
[DESIGN-REVIEWED] 66f3709 Description read[DESCRIPTION-READ] d9d5277461f3eabbfb210f75e22bb7f1a2ddc6bf0cc15c9076076f88d53f8e41 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 |
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: |
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: |
|
Hold at abde113: this needs a human product decision. The document is accurate and well scoped, and every check is green. Approving it is the decision, though. The RFC says a maintainer accepts it by approving this PR (L17), and on #7883 the rest of the menu was described as waiting on a product decision. The menu set is a product choice: it puts Disable and Uninstall on the sidebar row, leaves out Enable and Settings, and adds two status marks. An automated review should not be the acceptance. To clear this, a maintainer who owns the apps and sidebar surface approves the PR, which records acceptance. Until then, the front matter could read Non-blocking note: the Design lane's touch-gesture note is already covered by exit criterion 5. — crew-pr-reviewer (r147) |
|
note: Agreed. A maintainer's approval of this PR is the acceptance, and no automated lane stands in for it. The status line in the document says the same thing. I am keeping @bolichen97, you picked up #7883 on 11-05. Can you review this RFC and approve it if the menu items and the row marks look right to you? The implementation PR comes next and is linked to #7883. |
bolichen97
left a comment
There was a problem hiding this comment.
Design brief for abde113 (RFC, docs only).
What it does: records the decision on #7883 for sidebar app rows: a full actions menu (on right-click, Shift+F10 or the context-menu key, and touch) in place of today's Details and Unpin, plus status marks on the row. It builds on #26362 and verifies today's behaviour at 4f7b1ee782 (AppRailMenu, appNavTarget, the detail-page update flow, the Library's disable and uninstall, the unread count and backend_status.restart_exhausted). It adds a README index row.
Pros: the actions a user already has in the Library become reachable from the row, through flows that exist today (the detail page's consented update, the Library's uninstall preview), so no new action path is designed here. Keyboard and touch entry are stated, and the touch hold is reconciled with the existing drag sensor.
Cons: it is a decision record only; nothing is built yet, so the menu's real behaviour still needs its own implementation PR and review.
Long-term tech debt: no.
Recommendation: sign. As an RFC, the sign-off belongs to the repository owner, so this is a comment rather than an approval. All 96 checks pass.
|
This changes behaviour users already rely on, so it needs a design sign-off from a maintainer before it can merge. A maintainer will post the decision here. To propose a design, follow the request-for-change process in CONTRIBUTING.md (docs/request-for-change/). Review result on abde113: no blocking findings. The diff adds the RFC and one index row. No blocking repository rule applies to either path. The review checked each Current behavior claim against the named files at 4f7b1ee, and they hold. Merging this PR records the decision as accepted, so the approval has to come from a maintainer. Done well: every Current behavior claim names the file and symbol it depends on. The design also keeps two existing safeguards: the new Library requests are accepted from router state only, and Uninstall keeps its confirmation dialog. Optional:
|
abde113 to
580bf91
Compare
|
Thanks. Two of the optional points are in 580bf91.
I left the decider's name and the date for the maintainer who approves, so the record names the person who made the decision. |
bolichen97
left a comment
There was a problem hiding this comment.
Design brief for 580bf91. All 97 checks pass. I compared it with abde113, which my earlier brief covered.
What changed in the RFC text:
- A row's unread count now has two named sources. An app that pushes its own count through
useNavBadge()shows that count; otherwise the count comes from the app's unacknowledged bell notes (appNotificationBadges). - "Mark notifications as read" now shows when the app has unacknowledged bell notes, not whenever a count is shown. A count the app pushes itself is the app's to clear, and the status-mark table says the same.
doc-pris now31104, and the README row moves with main.
The direction is the same as before: one row menu with Update, Details, Open in Library, Mark notifications as read, Disable, Unpin and Uninstall, plus the update arrow and attention triangle. The edit is a correctness fix, since a menu item that cannot clear an app-pushed count is no longer offered for it.
- Pros: the menu no longer offers an action that would leave the badge in place.
- Cons: with an app-pushed count and no bell notes, the person has no menu way to clear the badge. That is by design, since the app owns that count.
- Tech debt: none added. It records an existing split, and no code changes.
The sign-off stays with the owner.
580bf91 to
d46a0d9
Compare
4e8d29b to
74710ab
Compare
|
Thanks. Push 74710ab records that trade-off as accepted. On the phone chat page a failing backend shows only once the nav drawer is open, because a triangle on a row with no menu would leave the failure in a tooltip that touch cannot open. The sentence sits under Status marks, right after the one about the phone chat rail. |
74710ab to
6cd6d54
Compare
Code change approver recommendation: A person should decideWhy: the decider estimates a 7% chance that this change introduces a defect, within the 12% limit for approving on its own, but it is held at tier 2: design-doc-gate: D. Docs-only RFC recording the sidebar app menu items and row status marks, plus its index row. I checked the Current behavior claims against head: AppRailMenu (Details/Unpin, registry ErrorNotice, trigger disabled during drag), appNavTarget excluding disabled/unavailable apps, LaunchpadTile's Disable/Uninstall (non-locked)/Sync rules, LibraryPage's uninstall preview dialog, TOUCH_HOLD_ACTIVATION 250ms/5px, restart_exhausted shown only on Dev Fleet, and AppDetailPage's router-state-only autoAction. All of them are accurate. The design reuses existing paths, and its new trust surface (Library accepts Disable/Uninstall from router state only) follows the established autoAction rule, which cross-site links cannot reach. Uninstall keeps its confirmation. Landing at status Tier: 2, held for a person: design-doc-gate: D. An automated recommendation, not a review: a maintainer still decides. It approves on its own only when the estimated chance of a defect is at most 12% and the change has at most 400 changed lines. Decider v3c (217fe607) · commit |
6cd6d54 to
620c1b2
Compare
620c1b2 to
66f3709
Compare
Problem / Motivation
Goal: Record which actions the sidebar app menu carries and which status marks a sidebar app row shows, so the implementation has an accepted design to follow.
#7883 asks for app management from the sidebar. #26362 added a right-click menu with Details and Unpin, and the rest of the menu waited on a product decision. Pearce (Pearcekieser) asked for the full menu and for marks such as "update available" on the sidebar row.
Why it matters
Without an accepted RFC on main, the implementation PR changes what a sidebar row does for every user with no recorded decision, and the First Principles lane blocks it. Today a user must go through Apps and Library to update, disable or uninstall an app, and the sidebar does not show that an update is waiting.
Not a goal
What changed (motivation → approach → change)
This PR adds
docs/request-for-change/rfc-sidebar-app-menu.mdwith statusacceptedand adds its row to the RFC index. The document records the menu items in order and when each shows: Update to vX, Details, Open in Library, Mark notifications as read, Disable, Unpin and Uninstall…. It also records the two new row marks and their labels, the "N more" toggle mark, keyboard access, and how a touch hold opens the menu. Each item reuses an existing path. Update goes through the detail page and its trust consent. Disable and Uninstall run on the Library page, which already shows their results and errors, and Uninstall keeps its confirmation dialog. Disable and Unpin each carry a one-line hint that says what they do. Open, Enable and Settings are left out, and so is the Library tile's Sync, and the document says why. When the registry list fails to load, the menu keeps today's failure notice and agent hand-off as its last group. A backend that keeps failing to start gets the same pair, and the phone chat rail, which has no menu, leaves that mark out.Backwards compatibility
Compatible: the PR adds a document and an index row and changes no code.
Removes nothing: no code, file or key is removed.
Tests
Proof: not provable -- docs or config only
Manual verification
N/A -- a document only. I verified the "Current behavior" claims against
4f7b1ee782by reading the files that the document names.Related Issues
Part of #7883
Checklist
feat|fix|docs|style|refactor|perf|test|chore|ci|build|revert: ...)