Repository navigation
fix(dashboard): clearer search status and limit copy for capped tree - #20495
Conversation
d1b7a07 to
a47c5c2
Compare
First Principles Review (Opus 5.5) — ✅ PASSPremise-level review of First-Principles-Verdict: PASS Nothing to check. What this change shipsInventory (4 items) — 4 justifiedIntent: on a project too big to list in full, the folder panel's search status and limit messages should say plainly what is happening. This is a FIX (closes #16743).
[FIRST-PRINCIPLES-REVIEWED] 1fbe749 Description read[DESCRIPTION-READ] 928e777880ee656c2037b2aca3fc2924a584c0132ce625bf19c7c11262c0c624 That sha256 names the PR description this verdict read. Recompute it with |
Design Review (Opus 5.5) — ✅ PASSDesign-level review of Design-Verdict: PASS Check that the "Searching…" row stays hidden on a non-truncated tree while a wider page loads, since [DESIGN-REVIEWED] 1fbe749 Description read[DESCRIPTION-READ] 182436b7db34fe234dbb6862f0e5983ee9eb3c4b5200c56658738a5bcd05ee5a 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 |
UX Review (Opus 5.5) — 🟡 CONCERNSUX-level review of UX-Verdict: CONCERNS The rewritten search line still fails its job: the blind reader had "no idea what it really means" and read it as contradicting the other limit messages. Watch
Suggestions
[UX-REVIEWED] 1fbe749 Description read[DESCRIPTION-READ] edebbf35fea075d7f6cfbe5519994ead11aa75a43de84b87037b958406b9b37e That sha256 names everything this verdict read: the PR description, plus the 7 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: |
a47c5c2 to
7c9b815
Compare
|
|
|
|
|
|
|
|
self-added: no
|
|
|
7c9b815 to
4cb68ec
Compare
|
|
|
|
|
self-added: no
|
|
self-added: no
|
|
self-added: no
|
4cb68ec to
f2209d5
Compare
|
|
|
|
|
f2209d5 to
1fbe749
Compare
|
|
|
|
|
self-added: no
|
|
self-added: no
|
|
|
self-added: no
|
|
|
|
KiroCrewCaptain
left a comment
There was a problem hiding this comment.
Approved by a maintainer who read the review lane notes and accepted them as follow-up work.
Problem / Motivation
Goal: On a project too large to list in full, the folder panel's search status and limit messages say plainly what is happening.
The capped project tree (#16029) left four rough edges, found by the UX review on that PR:
Why it matters
People with big repos hit these surfaces every time they open the Files tab. Copy that contradicts itself or hides that a search is still running makes them think files are missing.
Not a goal
What changed (motivation → approach → change)
website/src/pages/chat/FolderPanel.tsx: the "Searching…" status now sits below the listed matches and also shows while the first server answer is still out, even if the tree's own matches are already listed. It also shows when a refined query still shows the previous answer as placeholder rows (react-queryisPlaceholderData), but not while a wider page loads. It goes away once this query is answered.website/src/pierre/PierreWorkspaceTreeImpl.tsx: the empty capped tree gets its own string (components.workspaceTree.root_truncated): "Workspace too large to list: limit of 10,000 items reached"docs/system-specs/modules/learn-cron-dashboard.md: the Project tree spec quotes the new row and empty-state copy and the new search line.website/src/i18n/locales/(en.manual, zh-CN, hi, es, fr, bn, pt, ru, de, ja, ko, it, en-XA): all 12 catalogs plus the pseudolocale carry the new and changed strings. Two capture scripts (website/scripts/capture-path-chips.mjs,website/scripts/capture-files-tree-empty-folder.mjs) follow the new copy, and the truncated-tree scene now waits for the "Searching…" row.Backwards compatibility
Compatible: copy changes plus one new catalog key; no API, setting or stored data changes.
Removes nothing:
pages.chat.activityViewer.workspace_truncatedis still used by the top-of-tree badge; only the empty state stops reusing it.Tests
FolderPanel.tree.test.tsxnew case "announces the search still running while the truncated tree already lists matches": status row present while the search is pending with tree matches shown, gone after it answers, and back for a refined query whose rows are the previous placeholder. Each half red on its parent condition, green on the fix. 3 shuffled runs (seeds 1-3): 29/29. Red on the parent (status row missing), green on the fix. 5 runs of the file with--sequence.shuffle(seeds 1-5): 29/29 each time.PierreWorkspaceTreeImpl.test.tsxandFolderPanel.tree.test.tsxpins updated to the new copy.root_truncatedstates the fact rather than telling the user to search, because the Files rail shows the same empty state above a local-only "Filter files…" box. Style guides (src/i18n/style/*Style.test.tschanged-values checks, run withI18N_BASE_REF=origin/main: 14 files, 207 tests green).npm run i18n:check(also withI18N_BASE_REF=origin/main),tsc -p tsconfig.app.json, eslint on the changed files, and vitest over the two test files plussrc/i18n(50 files, 1094 tests): all green.Manual verification
Captured with the repo harnesses (
scripts/capture-path-chips.mjsscenefolder-truncated, andscripts/capture-files-tree-empty-folder.mjs --expect-state-rowsagainst a dist built from this branch). Both exited 0 with their assertions passing.Screenshots / video
Search in flight on a capped tree: new line above the results, and "Searching…" shown while the tree's own match is already listed.
Search answered: status gone, server match merged in.
Row under a capped folder (
big) now reads "Some items not shown: limit of 10,000 items reached".Empty capped tree: its own message.
Related Issues
Closes #16743
Pattern harvest
Not generalizable: a one-off set of copy and status polish on surfaces #16029 added; the reused-key case is already covered by the "one key, one meaning" rule in
website/docs/i18n-catalog.md.Checklist