feat(app-shell): the environment admin's storage-capacity banner, from the served verdict (objectui#10439) - #10910
Merged
objectstack-fleet[bot] merged 4 commits intoSep 28, 2026
Conversation
…er from the served verdict (objectui#10439) The tenant runtime's GET /api/v1/usage/storage already serves the storage verdict its upload / bulk-import guardrail refuses with (warn at 80% and up, blocked at 100%), plus usedMb / limitMb. Nothing rendered it. - useStorageUsageReading reads the flat storage half, contract-first, and classifyStorageUsage resolves it from warn / blocked alone; the two numbers never decide the banner. - StorageUsageBanner renders "X MB of Y MB used" on warn, and the "storage is full: uploads and imports are paused" banner with the upgrade entry on blocked. Mounted in ConsoleShell beside ReadRateBanner, behind the same useWorkspaceAdminStatus gate. - readStorageUsage shares one in-flight request per URL, so the storage banner and the read-rate report cost one request between them. - console.storageUsage.* in all ten locale packs. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_015AUunPkX7UTkCH9e7AdZo1
…geset (objectui#10439) Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_015AUunPkX7UTkCH9e7AdZo1
Contributor
|
changeset-claim-re-read
|
github-actions
Bot
added
documentation
Contributor
✅ Console Performance Budget
The eager closure is every chunk the entry reaches through static imports — what the browser fetches and parses before the app renders. The entry chunk on its own is a small fraction of it. 📦 Bundle Size Report
Size Limits
|
This was referenced Sep 28, 2026
Contributor
Author
Contract reviewServed-tier: Inputs read: card objectui#10439 (body and all 4 comments: 5824958765, 5824965929, claim 5864158793, os-dev-report 5864901900); PR #10910 (body, the 21-file list, the net diff against ① Derived judgmentsAccept-set — what the client admits from
Classifier Rendering against the adjudication the card quotes (at or over 80%: 「已用 X / Y」; at or over 100%: 「已满，上传与导入已暂停」 plus 升级入口):
Public-surface changes in the diff:
Shipped prose, sentence by sentence:
② Semver level
③ Boundary flags
Implemented-by: VERDICT: PASS Generated by Claude Code |
github-merge-queue
Bot
removed this pull request from the merge queue due to failed status checks
Sep 28, 2026
objectstack-fleet
Bot
deleted the
claude/issue-10439-storage-usage-banner
branch
September 28, 2026 09:02
akarma-synetal
pushed a commit
to akarma-synetal/objectui
that referenced
this pull request
Sep 28, 2026
…ts name 23 objectui issues that answer 404, and re-qualify 19 bare objectstack numbers (objectui#10803, batch 6) (objectstack-ai#10914) Part of objectstack-ai#10803 Clause-②: no Dispatched implementation of the `domain:ui` seat objectstack-ai#1 claim (comment `5864334310`) on objectui#10803, batch 6, session `https://claude.ai/code/session_01DuWo5bdP9SdVebamn99GGk`. Citations only: no sentence's claim moves, and every edited pending changeset's frontmatter is byte-identical. The one runtime string that moves is `InputSchema.wrapperClass`'s zod `.describe()` text in `@object-ui/types`, which loses its dead pointer and nothing else (amendment `5860244997`, Q1 = A; `patch` changeset). No test pins any changed text: the literal-anchor sweep below finds no specific anchor, so no test file is edited. This batch carries the release note `5863776648`'s three lists: - the last **23 family numbers**; - the **18 bare objectstack numbers above 10900** from PR objectui#10892's C0 census, plus **#13086**, a 404 the C0 instrument cannot see because it shares a token with #13337 (**Premise**); - the live-but-wrong bare `objectstack-ai#3391` at the two `rowCrudAffordances.ts` sites beside `objectstack#3720`. ## Why `Part of`, not a closing line The order says `Part of`. With this PR the family list (all 115 numbers of amendment `5860244997` Q2) and the C0 above-10900 list both read 0. Whether the card now closes, or carries the dead `objectstack#N` class measured in **Acceptance notes** 1, is the seat's call. ## Premise, re-measured on `origin/main` `9f0c84a44` (the branch point) - **The 23 family numbers.** REST `GET /repos/objectstack-ai/objectui/issues/N`: 23 of 23 answer 404, and a second read of each answers 404 again. `GET .../pulls/N` answers 404 for all 23. Lit controls: objectui#10533 and objectui#7714 answer 200. - **No new family member.** The distinct `objectui#N` citations in the two in-scope classes at the branch point, less those at batch 1's head `6c3ad7c80`, are 24 numbers. Each was read once: 24 of 24 answer 200. So the family list is still batch 5's 23. - **#14026 was a sister-repo card written as objectui's.** objectui#14026 answers 404 as an issue and as a pull. objectstack#14026 answers 301 to objectui#10102, the card it was transferred to. That card's measurement-round claim (branch `claude/issue-14026-import-mapping-selector-probe`) is the session of `ecf14190e`, the commit that added `14026-list-import-mappings-pin.md`, and six sibling in-scope lines already write `objectstack#14026` for the same card. - **C0, the bare-number census** (the instrument of PRs objectui#10875 and objectstack-ai#10892, at the branch point): ``` git grep -hoP '(?:^|(?<=[^\w#&/.\-]))#\d+(?![0-9A-Za-z_])' 9f0c84a -- '.changeset/*.md' 'packages/*/src/**' ':!**/__tests__/**' ':!**/*.test.*' ':!**/*.spec.*' ':!**/test/**' ':!**/tests/**' | sort -u | wc -l ``` - 984 distinct at `9f0c84a44`, against 998 at batch 5's `b2683a2c0`. The 14 that batch 5 re-pointed are gone, and none is new. - Above 10900 there are 23: the same 5 CSS colours and the same 18 numbers. Each of the 18 answers 404 as an objectui issue and as an objectui pull on two reads. - In objectstack, 16 of the 18 answer 200 with the sentence's own subject (the **C0 mapping** below). #13033 and #13413 answer 404 there too, as issues and as pulls. Their squash commits exist in objectstack's history (a full, not shallow, clone): `c459da6bc` "narrow the per-option `default` key out of the form-view options vocabulary … (#13033)" and `89448a52b` "remove the inert AUTH_SSO_PROVIDER_SCHEMA export (#13413)". Both are ancestors of objectstack `main` (`git merge-base --is-ancestor`, exit 0). - **The C0 instrument's blind spot.** Its lookbehind refuses a `#` preceded by `/`, `-`, `.` or `&`. That form (`#A/#B`, `-#N`) carries 109 distinct numbers at the branch point. The 17 that no earlier batch read were each read once. 16 answer 200. **#13086 answers 404**, as an issue and as a pull, in objectui and in objectstack. It sits in `6985-wizard-card-r-alignment.md` as "the #13337/#13086 fence", in the same token as #13337, so it rides here. objectstack's `check-yaml-examples.ts` header, at its landing `2ebfe7e9f` (PR objectstack#13267, which answers 200), reads "Check YAML Examples (anti-drift for the AUTHORING format, #13086)" and "Ruled 2026-08-29 (#13086)". - **objectstack-ai#3391.** objectui#3391 answers 200 with an unrelated subject (a record-header api action placeholder). objectstack#3391 answers 200 as the apiMethods whitelist contract card ("跟踪:UI 操作按钮与 apiMethods 白名单一致性契约落地"), and objectstack#3720's own title calls itself "objectstack-ai#3391 遗漏的第四个面". - Every edited changeset is pending: it is present in `.changeset/` on `main`. The checkout is not shallow. ## Census (the enumeration pin for this batch) The instrument PR objectui#10854 printed and PRs objectui#10869 / objectstack-ai#10875 / objectstack-ai#10892 reused, with this batch's 23 numbers substituted (REF = a commit or tree): ``` git grep -nE '(objectui#|#|issues/)(8072|8127|8137|8204|8229|8248|8307|8408|9231|9241|9244|9365|9373|9375|9542|9553|9585|10117|10119|10120|10129|10132|14026)([^0-9]|$)' REF -- '.changeset/*.md' 'packages/*/src/**' ':!**/__tests__/**' ':!**/*.test.*' ':!**/*.spec.*' ':!**/test/**' ':!**/tests/**' | grep -v 'objectstack#' | wc -l ``` - REF = `9f0c84a44` (branch point): **104** lines. - One more line is hidden by the `grep -v 'objectstack#'` filter: `9943-viewtype-totals-page-row-retired.md` names objectui#8127 beside an `objectstack#` citation. - So the true population is **105** lines: 28 changeset lines in 26 files, and 77 src lines in 42 files. - Unfiltered, REF reads 111. The other 6 lines are live `objectstack#14026` lines. - REF = `2b3d2061a` (this head): **0**. Unfiltered it reads 7, all live: the six `objectstack#14026` lines and the re-qualified one. - This head merged with a fresh `main` (`8522396c0`, `git merge-tree --write-tree`, clean, tree `f19d17884`): **0**. REF = `8522396c0` alone: 104. - **C0**, the 19 numbers, bare and not `objectstack#`-qualified: 28 lines in 23 files at the branch point; 0 at this head. At this head C0 reads 966 distinct numbers, and above 10900 only the five CSS colours are left. - Lit control, the printed instrument over live objectui#7714 at this head: 17 lines. Hex-colour false positives (a number followed by a hex letter) at the branch point: 0. - **Out of scope, as it stands** (the 23 numbers, filtered, whole tree at this head): - 94 test lines in 36 files; - 5 `.github` lines in 1 file (`ci.yml`, objectui#9241); - 2 `apps/console` lines in 1 file (`FormPage.tsx`, objectui#8408); - 1 line of the root `vitest.config.mts`; - 0 scripts, 0 governed, 0 `CHANGELOG.md`. - For the 19 C0 numbers: 13 test lines in 10 files, and 1 line of `packages/plugin-tree/README.md`. ## Citation form - **Landing shas.** The 9-character backticked sha of the commit on `main` that landed the change the sentence rests on, as in the earlier batches. All 15 distinct objectui shas below are ancestors of `main`: `git merge-base --is-ancestor`, exit 0 each. Control legs in the same checkout: the head of PR objectui#10902 (`496c63c06`) answers exit 1, and the known ancestor `5f789538d` answers exit 0. `git rev-parse --short=9` returns the same 9 characters for each. - **Own-card pointers** in a changeset are dropped, not replaced. The configured changelog generator (`@changesets/cli/changelog`, per `.changeset/config.json`) prefixes each released entry with the hash of the commit that added the file, which is that landing. - **The claim lived only on the card.** Where the sentence rests on something that lived only on the dead card, the sha only locates it: "the card behind SHA". - **Nothing answers.** The pointer is dropped and the sentence names the card by role (objectstack-ai#9553 only). - **Sister-repo numbers.** - A bare sister-repo number becomes `objectstack#N` after reading it there. - "objectstack PR #N" becomes "PR objectstack#N", since the qualifier now carries the repository. "framework #N", "upstream #N" and "spec #N" keep their word, as batch 4 kept "framework PR objectstack#6942". - Where the objectstack pull request itself answers 404, the sentence cites its commit in that repository as objectstack `SHA`, the spelling the tree already uses for an objectstack commit ("Measured on objectstack `9bd4344e4`"). - **Runtime text** carries no sha: see **Special cases** 6. ## Mapping, the 23 family numbers Lines / files are the branch-point census for that number. "Method" is how the landing was found; for every source site, `git log -S 'objectui#N' -- FILE` names the commit that wrote the citation, and it is the landing below unless the row says otherwise. | dead number | resolution | method | lines / files | why the commit carries it | |:--|:--|:--|:--|:--| | objectstack-ai#8072 | `c974edf14`; own-card pointer dropped; runtime pointer dropped | changeset's adding commit, subject "(objectui#8072)" | 5 / 3 | it mirrors `wrapperClass` on `InputSchema`, which is what both comments and the `7722` changeset say happened | | objectstack-ai#8127 | `ca3942729`; "the card behind `ca3942729`" twice (**Special cases** 2) | `git log -S` names `ca3942729` for the first citation in all 10 source files; `05a49f2ee` (objectui#9880) and `8a7e09fa1` (the `9943` changeset's landing) wrote three later sentences that cite the same fix | 17 / 11 | it derives `ViewType` from `@objectstack/spec` instead of re-declaring it, and its message records the `page`-degrades-like-a-typo measurement the sentences cite | | objectstack-ai#8137 | own-card pointer dropped (landing `0fa7a9c83`) | changeset's adding commit | 1 / 1 | the only site is its own changeset's pointer | | objectstack-ai#8204 | own-card pointer dropped (landing `580b0fdf4`) | changeset's adding commit | 1 / 1 | the pointer opened a paragraph as its own sentence | | objectstack-ai#8229 | "a separate finding" (changeset); "the finding `8b7ea3945` reconciled" (`flex.tsx`) | `8b7ea3945`'s message: "Reconciles the third face objectui#8229 found" | 2 / 2 | the finding lived on the dead card; `8b7ea3945` is the commit that reconciled it, and it is the `7735` changeset's own landing | | objectstack-ai#8248 | own-card label dropped from the second list item (landing `d02942b0e`) | changeset's adding commit; its message covers objectui#8458 and objectui#8248 | 1 / 1 | the changeset lists the two cards its one landing closed | | objectstack-ai#8307 | `5591f03bd`; own-card pointer dropped | changeset's adding commit | 11 / 4 | it makes a lane header over a windowed fetch say `77+`, and it wrote every comment that cites the card | | objectstack-ai#8408 | own-card pointer dropped; "seam 2 of the card behind `8241a4400`" | changeset's adding commit; `8241a4400`'s message names no seams | 2 / 2 | the seam list lived on the dead card | | objectstack-ai#9231 | `383502b23`; own-card pointer dropped | changeset's adding commit | 3 / 2 | it gives the create dialog's confirm control its own accessible name, the rule both `i18n.ts` comments state | | objectstack-ai#9241 | own-card pointer dropped (landing `250429c8f`) | changeset's adding commit | 1 / 1 | the only site is its own changeset's heading | | objectstack-ai#9244 | `bd0995738` | `git log --grep`: its message names the card; it wrote all three comments | 3 / 2 | it emits one col-span class per breakpoint tier, not one for the widest | | objectstack-ai#9365 | own-card pointer dropped (landing `0970a0e00`) | changeset's adding commit | 1 / 1 | the only site is its own changeset's pointer | | objectstack-ai#9373 | `c1006ed8e`; own-card pointer dropped | changeset's adding commit | 2 / 2 | it resolves the inline locale map before the interpolation options, the `ListView` comment's subject | | objectstack-ai#9375 | own-card pointer dropped (landing `e427e9c00`) | changeset's adding commit | 1 / 1 | the only site is its own changeset's pointer | | objectstack-ai#9542 | `43c0d1710`; own-card pointer dropped; "filed as a separate card" (**Special cases** 3) | changeset's adding commit; `git log -S` names it for the `action-button` / `action-icon` comments too | 9 / 6 | it accepts and resolves an inline `I18nLabel` on `resultDialog` and makes `ResultDialogSpec` derive its label members | | objectstack-ai#9553 | nothing answers: "a separate card carried that census" | `git log --grep` finds only `40f34b4ba`, which points at the card; no commit lands it | 1 / 1 | see **Special cases** 4 | | objectstack-ai#9585 | "`ee70287e4`'s pin measures it NOT GATED"; in its own changeset, "a pin in this change measures it" | no commit names it; `ee70287e4` adds the test "NOT GATED: the renderer paints the very node the mirror refuses, without parsing it" | 2 / 2 | see **Special cases** 5 | | objectstack-ai#10117 | `4c6f549ef`; own-card pointer dropped | changeset's adding commit | 3 / 2 | it resolves a lookup title candidate in `page:header` and adds the record-key safety net both comments describe | | objectstack-ai#10119 | `73a3c89af`; own-card pointer dropped | changeset's adding commit | 4 / 2 | it makes a nav ancestor's gates reach its subtree and stops a group outliving its children | | objectstack-ai#10120 | `80c54122e`; own-card pointer dropped | changeset's adding commit; for each file `git log -S` names it, or a later commit that cites its gate: `e0f820246` (objectui#10563) or `b809375ac` (objectui#10163) | 16 / 13 | it makes a form neither submit nor offer a field the caller may read but not edit, the FLS half every site names | | objectstack-ai#10129 | `6cc910b6d`; own-card pointer dropped | changeset's adding commit | 10 / 5 | it routes a field-backed action param to its record picker and refuses an unreadable one | | objectstack-ai#10132 | `061f5e829`; own-card pointer dropped | changeset's adding commit | 8 / 4 | it makes the two declared-translatable dashboard surfaces resolve, including the axis `title` forward | | #14026 | re-qualified `objectstack#14026` | see **Premise** | 1 / 1 | it is the card the pin's measurement answered | ## Mapping, the 19 bare sister-repo numbers | number | resolution | what objectstack says | lines / files | |:--|:--|:--|:--| | objectstack-ai#11289 | `objectstack#11289` ("upstream") | "record:details sections cannot survive an empty record: `hideEmpty` / `collapsible` / `showBorder` are honoured by the renderer but undeclared" | 2 / 2 | | #11662 | PR `objectstack#11662` | "feat(spec): declare hideEmpty / collapsible / showBorder on record:details sections" | 1 / 1 | | #12616 | PR `objectstack#12616` | "feat(spec): declare record:details section headerColor as a closed six-token enum" | 1 / 1 | | #12718 | PR `objectstack#12718` | "feat(spec): retire preview mode — the RuntimeMode 'preview' value and the whole PreviewModeConfig block" | 1 / 1 | | #13033 | objectstack `c459da6bc` | the pull request answers 404; its squash commit narrows the per-option `default` key out of the form-view options vocabulary, the ruling `form-spec.ts` says was executed upstream | 1 / 1 | | #13337 | `objectstack#13337` | "docs(objectui): layout-dsl teaches only shapes the live schemas accept" | 1 / 1 | | #13086 | `objectstack#13267` | the card answers 404; PR objectstack#13267 landed its ruled YAML-examples gate (see **Premise**) | the same line | | #13413 | objectstack `89448a52b` | the pull request answers 404; its squash commit removes an inert schema export and leaves a note recording the absence as a choice, the shape the `base.zod.ts` note cites as precedent | 1 / 1 | | #13632 | `objectstack#13632` | "[spec] `FieldSchema` accepts a `lookup`/`master_detail` with no `reference` target …", the card 17.3.0's refinement answered | 3 / 3 | | #13733 | PR `objectstack#13733` | "feat(spec): wizard view v1 — declaration-and-refusal tightening of FormViewSchema type:'wizard' (Card S)" | 1 / 1 | | #13855 | `objectstack#13855` | the field-grouping decision card whose option B is a section `group` reference, landed by `39404f3d9` (#13897) "a layout section can reference a declared field group" | 1 / 1 | | #13906 | `objectstack#13906` ("framework") | "two more `computeExecCtx` seams read "failed" and "not wired" as one value …" | 6 / 3 | | #14274 | PR `objectstack#14274` | "fix(sdui-parser): refuse an authored `type` attribute on the html tier …" | 1 / 1 | | #14945 | `objectstack#14945` | "A flow cannot REFUSE with per-record text …", honoured by `cca699149` "the flow `end` node honours `outcome: 'refused'`" | 1 / 1 | | #15469 | `objectstack#15469` ("spec") | "`GanttConfigSchema` is `strictObject(...).passthrough()` …", landed as `9c270bba0` "close the gantt/tree config .passthrough() windows … (#15469)" | 2 / 2 | | #15948 | `objectstack#15948` | "fix(plugin-auth)!: session payload `positions[]` is the security axis, not the better-auth role scalar" | 1 / 1 | | #17493 | `objectstack#17493` | "three residues of #17322's node-door refusal …", landed as `2c1011b01` "refuse a blank string in a flow node's predicate slot" | 1 / 1 | | #20051 | `objectstack#20051` | "judge a flattened view overlay's top-level `options.KIND` …", whose door half is `6a4aec71d` | 1 / 1 | | #20160 | PR `objectstack#20160` | "fix(spec): a joined report refuses a top-level dataset / rows / columns / values, pointing each onto blocks[]" | 2 / 2 | The live-but-wrong number: "(`/me/permissions` `apiOperations`, objectstack-ai#3391)" becomes "… `apiOperations`, objectstack#3391)" at the two `rowCrudAffordances.ts` sites. ## Special cases (the judgement calls) 1. **#14026 is re-qualified, not dropped as an own-card pointer.** The card behind `14026-list-import-mappings-pin.md` was never an objectui card: it was objectstack#14026, now objectui#10102. Dropping the pointer would lose which hypothesis the pin refuted. Re-qualifying it matches the six sibling lines. 2. **objectstack-ai#8127, the card by role twice.** - `CreateViewDialog.tsx`: "one of the sites the card records as "drifted, …"" becomes "one of the sites the card behind `ca3942729` records as …". That record lived on the card. - `plugin-list` `ViewSwitcher.tsx`: "objectui#8127 was filed against the two in `@object-ui/types`, and the maps below were described there as total …" becomes "the card behind `ca3942729` was filed against …". - `normalize-list-view.ts`: "The bug objectui#8127 records" becomes "The bug `ca3942729` records". That commit's message carries the measurement: `page` resolved "to exactly what it resolved a typo to". 3. **objectstack-ai#9542 in `8648-ui-action-four-undeclared-keys.md`.** - The superseded paragraph keeps its pre-landing voice: "so it is filed as a separate card and ⛔ not guessed at here". - The superseding paragraph reads "`43c0d1710` landed the derivation". - "(Noted here by the objectui#9542 seat, …)" becomes "(Noted here by the seat that landed `43c0d1710`, …)". 4. **objectstack-ai#9553, nothing answers.** `base.zod.ts` said "No count of authored `events` keys is stated here — objectui#9553 carries that census." No commit landed that card. `40f34b4ba` only points at it, and the `AGENTS.md` fix `7550728a6` names a different card. So the sentence now reads "— a separate card carried that census". The census is not restated, per AGENTS.md objectstack-ai#9. 5. **objectstack-ai#9585, the measurement's instrument.** The card measured the render path NOT GATED, and no commit names it. But the commit that wrote both citing sentences, `ee70287e4`, adds a pin that re-measures exactly that: "NOT GATED: the renderer paints the very node the mirror refuses, without parsing it". - `disclosure.ts` now reads "(`ee70287e4`'s pin measures it NOT GATED)". - Its own changeset, `8236-collapsible-open-intercept-retire.md`, whose landing is `ee70287e4`, reads "(a pin in this change measures it NOT GATED)". 6. **The runtime string.** Only the listed text moves. | file | member | before | after | |:--|:--|:--|:--| | `types/src/zod/form.zod.ts` | `InputSchema.wrapperClass` `.describe()` | "Classes on the wrapper div around the input and its label (objectui#8072)" | "Classes on the wrapper div around the input and its label" | No test, doc or changeset quotes it: a whole-tree search for the old string returns only the source line. 7. **objectstack-ai#8229 in `flex.tsx`.** "(objectui#8229, folded into objectui#7735's ruling)" becomes "(the finding `8b7ea3945` reconciled, folded into objectui#7735's ruling)". objectui#7735 answers 200 and stays. 8. **#13413's sentence.** "Precedent of the same shape: objectstack#12009 / PR #13413." becomes "… objectstack#12009 / objectstack `89448a52b`." objectstack#12009 answers 404 too, but it is a qualified sister-repo number outside this batch's lists, so it is left and listed (**Acceptance notes** 1). 9. **#20160 across a line break.** In `10746-joined-report-clears-binding.md`, "(objectstack PR" ends one line and "#20160: …" opens the next. They become "(PR" and "objectstack#20160: …". ## The literal-anchor sweep (both test-pin classes, ruling `5861900779`) - **Instrument.** Every string, template, numeric and regex literal in all 4068 tracked test and script files (119016 distinct literals), parsed with TypeScript. - **Candidate filter.** A literal is a candidate only if it matches the diff's removed lines with two lines of context, raw or comment-flattened. That leaves 2839. - **Test applied to each candidate.** Does its occurrence count DROP between `9f0c84a44` and `2b3d2061a` in any of the 92 files this PR changes, in raw text or in a comment-flattened form? 229 distinct literals drop. - **Every one is a generic token.** Digits and short numbers, punctuation, single letters, `objectui#`, `objectui#1`, and whole-tree scanners such as `/#\d+/` and `/objectui#\d+/`. The three `objectui#\d+` matchers (`ActionRunner.disabledGate`, `registry-inputs-spec-parity`, `catalog-gallery-render`) assert over their own test data and read no changed file. - **No specific anchor.** None is a batch number, a changed phrase, or the changed `.describe()` string. So no test pins changed runtime text (class one), and no source-reading test pins a changed comment or docblock citation (class two). - Test titles and comments that name these numbers are out of the card's classes and stay (**Census**, out of scope). ## Held **By the serial rule: nothing.** Re-mapped before the push, on 9 open PRs: objectui#10910, objectstack-ai#10908, objectstack-ai#10907, objectstack-ai#10906, objectstack-ai#10901, objectstack-ai#10891, objectstack-ai#10777, objectstack-ai#10278 and the release PR objectstack-ai#5400 (objectstack-ai#10902 and objectstack-ai#10904 had merged since the claim). Two of them share files with this PR. Their hunks were read against their merge-bases; each file is identical at its merge-base and at `9f0c84a44`, so the lines map directly. - **objectui#10907, `types/src/zod/form.zod.ts`.** It inserts a refusal constant before `InputSchema`, two tombstone members further down `InputSchema`, and a comment in the `InputShorthandSchema` arm. This PR's three edited lines (the `wrapperClass` comment and description, and the arm's "shrank that row" comment) sit outside every one of its hunks and their 3-line context. - **objectui#10906, `metadata-admin/i18n.ts` and `previews/ViewPreview.tsx`.** Its `i18n.ts` insertions are far from the two `createDraft` comments, and its `ViewPreview.tsx` hunks are far from the `options` fold docblock. Trial merges with this head (`git merge-tree --write-tree`): - clean for objectstack-ai#10910, objectstack-ai#10908, objectstack-ai#10907, objectstack-ai#10906, objectstack-ai#10901, objectstack-ai#10891 and objectstack-ai#10777; - objectstack-ai#10278 (`eab4c8e52`) conflicts in `ObjectGrid.tsx`, `plugin-grid/README.md` and `content/docs/plugins/plugin-grid.mdx`, and conflicts identically against `9f0c84a44` alone; - objectstack-ai#5400 (Version Packages) regenerates and is not a hold. objectui#10891 shares no file with this PR. ## Changesets - `.changeset/10803-dead-citation-sweep-sixth-batch.md`, EMPTY frontmatter. It covers the comment-only edits in 15 released packages; no published behaviour changes through them. It points at the second file for `@object-ui/types`' runtime text. - `.changeset/10803-sixth-batch-runtime-strings.md`, `'@object-ui/types': patch`: the `wrapperClass` description loses its pointer. No key, path, issue code, accept set, refusal or severity moves. ## Proof of prose-only (C4), against `9f0c84a44`, on this head `2b3d2061a` - **Source.** Each of the 53 touched `.ts` / `.tsx` files was parsed at `9f0c84a44` and at this head with TypeScript 6.0.3's `createSourceFile`, and re-printed by `createPrinter({ removeComments: true })`. - 52 of 53 prints are identical. - `form.zod.ts` is equal once the one listed substitution (**Special cases** 6) is applied to the base print, matched once. - 0 parse diagnostics. - Lit controls on the same instrument: editing a string literal moves the print; re-spacing a comment does not. - **Changesets.** The frontmatter block of every one of the 37 edited changesets is byte-identical at `9f0c84a44` and this head (37 of 37, by md5). The overwrite gate below agrees. - **Scope of the diff:** 92 files, +175 / −140. That is 37 edited and 2 new changesets, and 53 non-test source files in 15 released packages. No test file. ## Gates, on this head `2b3d2061a` Each line is the gate's own verdict and exit code, captured by redirect-then-`$?`. - `node scripts/check-changeset-presence.mjs`, exit 0: "53 source file(s) of 15 released package(s) changed, and this change declares 2 changeset(s): .changeset/10803-dead-citation-sweep-sixth-batch.md, .changeset/10803-sixth-batch-runtime-strings.md." - `pnpm changeset:check`, exit 0: "All workspace packages are in the changeset fixed group." / "No changeset declares a `major` bump." - `node scripts/check-changeset-overwrite.mjs` (report-only), exit 0: "2 changeset(s) added, 37 modified, 0 deleted". `declared at base` equals `declares now` for 37 of 37. - `pnpm check:changeset-claims` (report-only), exit 0: - "Every one of those 1 address(es) either names the tree it was read from, or points at a line this change does not move"; - "Every package declared across those 31 body(ies) is either not negated …"; - the standing notice "75 pending changeset(s) describe a file this change touches". - Read against the diff: a pending changeset quoting a replaced pointer would itself carry the dead number and sit in the census, which reads 0. No pending changeset quotes the changed description. - `pnpm check:control-bytes`, exit 0: "check-control-bytes: OK (scanned 9178 tracked text file(s); skipped 85 binary)." - `pnpm check:new-line-citations`, exit 0: "VERDICT new-cross-file-line-citations: 0 new citation(s), enforcement report-only -> exit 0". - `pnpm check:pending-changeset-literals`, exit 0: "No test source names a pending changeset." - Also run over the touched comments: - `pnpm check:installed-pin-claims`, exit 0 ("OK"); - `pnpm check:comment-mask-corpus`, exit 0 (1 disagreeing file, within the ceiling objectui#7882 holds open); - `pnpm check:handler-key-reads`, exit 0 ("every judged read is a declared member of it"). - The governed-surface predicate over the 92 paths, exit 0: "NOT GOVERNED — 92 path(s) checked against 5 governed surface(s); none matched." Lit control `AGENTS.md`: exit 3. **Tests**, through the shared verify lock, on `2b3d2061a`. Each is `VERDICT command-exit 0`. - `scripts/__tests__/`, the whole directory, whose whole-tree scanners read the touched files: `Test Files 177 passed | 2 skipped (179)`, `Tests 5322 passed | 2 skipped (5324)`. The two skipped files are the network-escape fixtures that run only as a child. - `packages/types/`, the whole package, whose runtime text moved: `Test Files 264 passed (264)`, `Tests 5854 passed (5854)`. `pnpm --filter @object-ui/types type-check`, exit 0. - **The pins nearest the re-pointed text:** - in `types`: `input-wrapper-class-mirrored-8072`, `zod-mirror-parity`, `wrapper-class-declared-7722`, `zod-mirror-authors-no-defaults-7735`, `collapsible-open-refusal-8236`, `tree-view-config-readers-8253` and `layout-default-jsdoc-7361`; - in `components`: `collapsible-open-intercept-8236`, `registration-defaults-match-renderer-8229` and `action-undeclared-keys-8648`; - in `plugin-view`: `ViewSwitcher` and `ViewSwitcher.viewTypeTotalsBothLegs-9943`; - in `plugin-kanban`: `laneCountHonesty-8307`; - in `plugin-grid`: `rowCrudAffordances` and `rowCrudEffectiveOps`. - Result: `Test Files 15 passed (15)`, `Tests 381 passed (381)`. - No red leg: the sweep found no anchor to move, so there is no pin whose old copy should fail. **Declared narrowing.** NOT MEASURED locally: - the full suites and type-check of the 14 other touched packages, and eslint. - Reason: the comment-stripped syntax tree of 52 of 53 touched source files is identical to `main`, and the 53rd differs only by the listed literal. - CI runs the full farm. ## Acceptance notes 1. **A sister-repo class of the same defect: 30 `objectstack#N` citations answer 404 in objectstack.** - **Measurement.** Every distinct `objectstack#N` in the two in-scope classes at the branch point is 359 numbers. Each was read once: 328 answer 200, 1 answers 301 (objectstack#14026 to objectui#10102), and 30 answer 404, each confirmed by a second read. - **The 30:** objectstack-ai#5970 objectstack-ai#5976 objectstack-ai#6038 objectstack-ai#6124 objectstack-ai#6281 objectstack-ai#6331 objectstack-ai#6450 objectstack-ai#6483 objectstack-ai#6515 objectstack-ai#9933 objectstack-ai#9934 objectstack-ai#10354 objectstack-ai#10485 objectstack-ai#10695 objectstack-ai#11330 objectstack-ai#11507 objectstack-ai#11513 #11658 #11703 #11753 #11846 #12009 #12868 #13117 #13670 #16126 #17147 #17762 #17987 #18012. - **Sites at this head:** 82 lines in 63 files, 26 of them changeset lines. - **Three of them sit in sentences this PR edits, and are left as they are:** - objectstack#11846, "(objectstack#11846, landed as PR objectstack#12718)" in `6748-preview-mode-provenance-ratchet.md`; - objectstack#12009, in `base.zod.ts`'s precedent line; - objectstack#12868, in `form-spec.ts`'s "RULED 2026-08-28 (objectui#6263 / objectstack#12868, …)". - **Why left.** They were not in this batch's lists, and the class is the seat's to scope. objectstack's own card for dead tracker citations in its tree is objectstack#19123, whose landing `66e266c93` counts #12868 among the dead numbers it measured. Triage item 3 would put a dead number in these classes on this card. The fix shape measured here for such a number is the objectstack commit, as with `c459da6bc`. 2. **The rest of the bare `objectstack-ai#3391` population.** Only the two sites the claim names were re-qualified. 28 more in-scope lines in 10 files write the apiMethods whitelist card as a bare `objectstack-ai#3391`, which resolves to the unrelated objectui#3391: - `ObjectDataPage.tsx`, `ObjectView.tsx` (app-shell); - `managedBy.ts`; - `MePermissionsProvider.tsx`, `PermissionContext.ts`, `PermissionProvider.tsx`; - `fieldWriteGate.ts`; - `ImportWizard.tsx`, `ObjectGrid.tsx`; - `ListView.tsx`. It is live, not a 404, so it is outside the family pin. Two of those lines pair it with a bare 3546 that means objectstack#3546 ("detail/form 面的 edit/delete 按钮接入服务端 effective 操作集"), while objectui#3546 is an unrelated i18n card. Carrier: none. 3. **A stale claim beside a re-pointed sentence.** The same `base.zod.ts` paragraph says "AGENTS.md's abridged protocol sketch shows `events?: Record…` and its action-system commandment authors one". `7550728a6` removed both from `AGENTS.md`, so that sentence is now false. It is not a citation, and this PR does not touch it. Carrier: none. 4. **The C0 blind spot, for any later census.** The bare-number instrument's lookbehind hides a number written after `/`, `-`, `.` or `&` (`#13337/#13086`, `-objectstack-ai#2231`). This batch read the 17 such numbers no batch had read, and #13086 was the only 404. A later census can narrow the lookbehind to refuse only a word character, `#` or `&` before the `#`, and drop URL fragments by hand. 5. **Filenames are not citations.** Pending changeset and test FILENAMES carry several of these numbers. They stay, as in PRs objectui#10707, objectstack-ai#10797, objectstack-ai#10854, objectstack-ai#10869, objectstack-ai#10875 and objectstack-ai#10892. --- _Generated by [Claude Code](https://claude.ai/code/session_01DuWo5bdP9SdVebamn99GGk)_ Co-authored-by: Claude <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Fixes #10439
What this does
The tenant runtime's
GET /api/v1/usage/storagealready serves the storage verdict that its upload and bulk-import guardrail refuses with. Nothing in the console rendered it, so an environment reached 80% with nothing shown and found out it was full when an upload failed. This PR renders the environment admin's banner from that verdict. It follows the product adjudication quoted on the card:blocked: truewarn: trueok,unknown(the endpoint's miss answer),unlimiteduseStorageUsageReading(new, inpackages/app-shell/src/hooks/) reads the flat storage half of the response:state,warn,blocked,usedMbandlimitMb. Parsing is contract-first:statemust be one of the five values the producer declares, andwarnandblockedmust be booleans.usedMbandlimitMbmust be finite, non-negative numbers when present.warnorblockedverdict that arrives without both numbers is off-contract. The producer'sevaluateStorageQuotaraises either flag only after both numbers are known, and the banner copy is made of those two numbers.unavailable. It is never coerced and never rendered with a stand-in number.classifyStorageUsagedecides the banner fromblockedfirst, thenwarn, and from nothing else. It never comparesusedMbwithlimitMb, and never comparesfractionwithwarnFractionor a literal line. NeitherfractionnorwarnFractionis modelled. This is the card's rule: "Read the verdict, do not re-derive the threshold."StorageUsageBanner(new, inpackages/app-shell/src/layout/) is mounted inConsoleShellbesideReadRateBanner, behind the sameuseWorkspaceAdminStatusgate. A non-admin session renders nothing and issues no request.readStorageUsage(new, instorageUsageEndpoint.ts) shares one in-flight request per URL between the two hooks. It forgets the request when it settles, which is the same shape as theINFLIGHTmap inuseApproverDirectory. It keeps no answer, sorefetchstill issues a new request.useReadRateReadingnow reads throughreadStorageUsage.resolveRuntimeApiBasemoved into the shared module and is re-exported from its old home. The three separate "no banner" answers fromclassifyReadRate(unmeasured, unavailable, ok) and all existing read-rate tests are unchanged.console.storageUsage.*keys are added to all ten locale packs. Each pack uses its own unit spelling: MB, Mo (French), МБ (Russian) or ميغابايت (Arabic)..changeset/10439-storage-usage-banner.mdbumps@object-ui/app-shelland@object-ui/i18nasminor.Nothing is added to the package entry (
src/index.ts). The new hook and banner are re-exported only from the internalhooks/andlayout/barrels, the same route the read-rate pair takes. SoClause-②: noholds as the claim stated it.Premises measured, not taken from the card
The cloud side was read at
objectstack-ai/cloudmain96eb092, read-only:Wire keys. The response carries
usedMbandlimitMb. The card body'sstorageUsedMbandstorageLimitMbare the registry-row fields the handler reads, as the claim's correction says. Every miss answers 200 withstate: 'unknown'and both flags false.Nothing read the storage half. On objectui
mainat the fork point,git grep -c -E "usedMb|warnFraction|STORAGE_LIMIT_REACHED|limitMb" -- packages appsfound nothing and exited 1. The control,usage/storage, hit four files.Audience.
useWorkspaceAdminStatusconfirms admin through any of four sources:owneroradminrole on the active organization's membership (either spelling);user.role;isPlatformAdminrung;org_ownerororg_admininpositions.This is the gate
ReadRateBanneralready uses for the same audience ("the environment's own admin"). Cloud names the audience 「环境管理员」 instorage-quota.tsand in the read-rate changeset, and defines no narrower tenant-console role. So no second gate was invented.The upgrade entry needs no plan field. The guardrail's refusal envelope for the same verdict,
StorageLimitRefusal, declaresdetails.upgradeas the literal typetrue. An unlimited plan (limit 0) never reachesblocked. The link iscloudConsoleUrl(), the control plane's origin. It is left out when that returns an empty string (no upstream cloud), which is the rule from objectui#7253 that the AI usage indicator's call to action also follows. No cloud-side field gap was found.Verification
All runs are at HEAD
efa30fc, after mergingorigin/mainat7ea8118, with heavy runs going throughos-verify-lock.sh.turbo run build --filter="@object-ui/app-shell^..." --concurrency=2pnpm --filter @object-ui/app-shell type-check(tsc --noEmit && tsc -p tsconfig.test.json)--listFilesshows both new test files in the test projectpnpm exec vitest run --reporter=verboseover 28 targets: the 4 touched test files,cloudConsoleUrl-7253, the 22 test files that renderConsoleShell, andpackages/i18n/all-locales-key-parity32/32,entry-locale-shape-74798/8,consoleUserPreferenceBudget3/3pnpm exec eslint --format jsonon the 19 touched.ts/.tsxfilespnpm check:i18n-keys/check:i18n-drift/check:i18n-dead-keysenvalues changed; dead-keys lists nostorageUsagekeypnpm check:vi-mock-specifiers/check:vi-mock-inherit/check:vi-mock-override-shapepnpm check:test-path-roots/check:self-import/check:phantom-deps/check:unreferenced-sourcespnpm check:new-line-citations/check:control-bytesnode scripts/check-changeset-presence.mjs/check-changeset-no-major.mjspnpm check:changeset-claims/check:pending-changeset-literalsconsole.navrows) stays truenode scripts/check-governed-queue-guard.mjs --teston the 21 changed pathsAblation: one-shot, restored, no permanent test file
Each mutation went through
ablation-replace.mjsin WRAP mode. The anchor hit exactly once, the blob changed on disk, and afterwards the blob matched HEAD again withgit diff HEADempty. No build leg was needed: the tests import the source by relative path, sodist/is not on the resolution path.usedMb / limitMb >= 0.8in place of thewarn/blockedcheck).useStorageUsageReading.test.tsxwent red: 3 failed, 15 passed. The failures were the verdict pin, the classifier table, and the unlimited-plan case, because re-deriving divides by a zero limit and would show a banner on an unlimited plan.if (pending) return pending;removed). The same file went red: 1 failed, 17 passed. The failure was the one-request pin (toHaveBeenCalledTimes(1)).Both went red, which is the direction expected before the runs.
Acceptance notes
check:eager-locale-catalogues. It reported PREREQUISITE NOT MET because it reads a built console bundle. This diff only adds rows inside the existing catalogue object literals. It touches no import, nolocales/index.tsand noregistry.ts. The source-shape pinentry-locale-shape-7479passed 8 of 8.check:readme-exports. It exited 1 because 86 self-imports across packages could not be judged: theirdist/index.d.tsis not built. The README diff adds no fenced block and no import binding, so it cannot change this gate's verdict..ts/.tsxfiles instead of the packages'eslint ., because the shared lock was held by another agent's full app-shell vitest run for about 20 minutes. Why this still covers every file the diff can affect:--format jsonoutput, none reported as ignored.eslint.config.jssets noparserOptions.projectorprojectServiceand loads no import-resolving plugin, and no local rule ineslint-rules/reads the disk. So this diff cannot change the verdict for any file it does not touch.react-hooks/set-state-in-effecton the new hook's synchronousloadingset. This is the same pattern, and the same warning, thatuseReadRateReadingalready carries at base.lint.ymlsets no--max-warnings. The warning counts forConsoleShell.tsx(8) anduseReadRateReading.ts(1) are unchanged from base.content/docs/guidepage covers the console's chrome banners: read-rate, impersonation, or now storage. The app-shell README section is the doc surface. No one is picking this up..changeset/9954-read-rate-banner.mdsaysuseReadRateReadingandReadRateBanner"are exported from@object-ui/app-shell". The package entry exports neither; they are re-exported only from the internal barrels. This is reported to the seat and not edited here.origin/mainwas merged twice before opening this PR.Session:
https://claude.ai/code/session_015AUunPkX7UTkCH9e7AdZo1, by anos-devagent dispatched by the domain:ui seat 4 PM.Generated by Claude Code