Skip to content

Commit 82e993b

Browse files
committed
Merge remote-tracking branch 'origin/main' into claude/project-thread-ds43h1
2 parents 8269772 + ff88105 commit 82e993b

87 files changed

Lines changed: 3313 additions & 865 deletions

File tree

Some content is hidden

Large Commits have some content hidden by default. Use the searchbox below for content that may be hidden.
Lines changed: 18 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,18 @@
1+
---
2+
'@object-ui/plugin-list': patch
3+
'@object-ui/permissions': patch
4+
---
5+
6+
fix(plugin-list): the list toolbar offers the "In a queue" toggle only when the organization has at least one queue
7+
8+
Clause-②: yes (widening)
9+
Metadata-protocol: none
10+
11+
The "In a queue" toggle showed on every list of an owned object, because the platform adds the `queue` column to every such object. In an organization with no queue it could only ever return an empty list. It now appears when both of these hold:
12+
13+
- **The object carries the platform's `queue` column**, as before.
14+
- **The session says the organization has a queue.** `GET /auth/me/permissions` answers `hasAnyQueue` (true when the organization has a declared queue or a `sys_queue` row). The console already reads that answer once per session, so opening a list sends no new request and reads no queue list.
15+
16+
When the answer is unknown (no permission provider is mounted, the role-based `PermissionProvider` is used, or the session is anonymous), the toggle stays hidden. Turning the toggle on filters exactly as before, and the server's queue behaviour (claim, release, the read rule) does not change.
17+
18+
`@object-ui/permissions`: `PermissionContextValue` gains one optional member, `hasAnyQueue?: boolean`. `MePermissionsProvider` sets it from the authenticated answer. It is `undefined` with no answer.
Lines changed: 11 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,11 @@
1+
---
2+
'@objectstack/skills': patch
3+
---
4+
5+
docs(skills): objectstack-platform says a locale added in the `.custom` pair is served on the base's address, and a locale a dependent adds under the base's name stays under its own package
6+
7+
Clause-②: no
8+
Metadata-protocol: none
9+
10+
- **Before.** The ADR-0146 customization paragraph said a locale the base does not ship is added in the pair without `overrides`, and stopped there: nothing said where that added locale is served, and nothing said what a dependent package's added locale under the base's name reaches. The reads of a base's address now decide an email template member by member, so an author could expect the pair's added locale to stay under `hotcrm.custom` only, or a dependent's to answer on the base's address.
11+
- **Now.** One sentence beside that rule: the base's address serves the pair's added locale too (both lists, the execution list, a send in that locale), and a locale a dependent adds under the base's name is served only under its own `?package=`, never by a base-named or package-less read. The file stays under its token ceiling: the sentence is paid for by tightening four restating clauses elsewhere in the file (one intro clause, one explanatory aside, one restated driver rule, one filler template bullet). No rule is removed, no frontmatter, schema or behaviour changes.
Lines changed: 24 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,24 @@
1+
---
2+
'@objectstack/service-analytics': patch
3+
---
4+
5+
fix(service-analytics)!: the `/analytics/sql` dry run refuses a selection asking one column twice, as `/analytics/query` does, instead of rendering a statement the query refuses
6+
7+
Clause-②: no (narrowing) — `/analytics/sql` (`AnalyticsService.generateSql`) now refuses `INVALID_QUERY` / 400 a selection that names one column twice in its groups (two granularities of one time dimension, the column raw beside its own bucket, or two spellings of it bucketed twice), with the refusal `/analytics/query` gives the same selection, field for field; it rendered a statement for each. Nothing is widened: no export, public member, answer member, error code or status is added.
8+
Metadata-protocol: none
9+
10+
<!-- adr-0087: not-required (no-migration-prescription) No key of any schema is added, removed, renamed or re-shaped, and no export or type of the package root changes: a dry run that rendered a statement the query refuses is refused with the query's refusal, already on the wire. There is nothing for a migration to rewrite and no tombstone. -->
11+
12+
**BREAKING** accept-set narrowing on the `/analytics/sql` dry run (`@objectstack/service-analytics`). `/analytics/query` and `queryDataset` are unchanged.
13+
14+
- **Before.** Measured on real SQLite on `main` `7e60dc23` (native declines every bucketed selection, so the ObjectQL strategy's echo answers). `/analytics/query` refused each selection below, and the dry run did this instead:
15+
- `timeDimensions` `closed_at` by month and by year: rendered one `year` bucket, the month entry dropped.
16+
- `closed_at` raw beside `CUBE.closed_at` by month, or a second dimension over `closed_at` beside its bucket: rendered `GROUP BY closed_at, strftime('%Y-%m', …)`.
17+
- `closed_at` by month beside `CUBE.closed_at` by year: rendered both buckets.
18+
- Beside a cross-object `account.name`: rendered the join and one bucket.
19+
- At a non-UTC `timezone`: answered `NOT_IMPLEMENTED` / 501, saying the query is served.
20+
- **Now.** The echo judges the keys `execute()` hands the engine with the same dataset door (`dataset-group-column-door.ts`, #28798) before it renders a bucket, so it answers the query's `INVALID_QUERY` / 400 with the same message, `object`, `param`, `path` and `fields`. On a selection that routes to the cross-object path, those are that path's keys. The echo's one-granularity-per-dimension map is gone: it reads the granularities the query's key builders read.
21+
- **Unchanged.** Every echo the query serves renders byte for byte (pinned): one bucket of a column, buckets of two columns, one granularity asked twice, raw columns, a cross-object dimension. The 501 for a bucket the engine groups in memory is unchanged when the selection names no column twice. No new refusal is added; the query's own now reaches the dry run.
22+
- **Who sees it.** A caller of `/analytics/sql` sending one of those selections now gets the query's refusal, instead of SQL for a query that does not run.
23+
24+
**The remedy.** Ask each column once: drop the raw dimension or the bucket's granularity, or ask one granularity. The refusal names the members.
Lines changed: 14 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,14 @@
1+
---
2+
'@objectstack/objectql': patch
3+
---
4+
5+
fix(objectql): an object extension whose `extend` folds to a legacy `ns__name` key is filed under that key, and never opens a second, owner-less object entry (#28858)
6+
7+
Clause-②: no — no accept set moves and nothing is exported; an extension that already registered now lands on the object it names. Measured reach: no package or example in this repository or in hotcrm spells an `extend` that folds.
8+
Metadata-protocol: none
9+
10+
`getObject` reads a short name as a legacy `<ns>__<name>` key when that is the only object it can mean (an object a package still ships as `crm__case`, which `ObjectSchema` accepts). The registry used to file an extension under the name it is written with. So a manifest `objectExtensions` entry or a stored `object_extension` row with `extend: 'case'` opened a second entry, `case`, with no owner, beside `crm__case`. Its fields reached no read of the object, and nothing checked the entry.
11+
12+
- **Now.** Every entry that registers an extension (the manifest loop and the stored row's hydration, both through `registerObject`) files it under the key `getObject` reads its `extend` as. The fields land on `crm__case`, and no `case` entry opens. The spelling is named once per spelling and package at `warn`, with the key to write.
13+
- **Never a boot refusal.** The manifest loop and the hydration run while packages are still registering. The `object_extension` item door keeps refusing the folded spelling at save (#28834).
14+
- **Unchanged.** An `extend` spelled with the registered key, or naming an object registered under that exact name, registers as before, with no line. `computeFQN` and `parseFQN` are untouched. A legacy owner that registers after the extension is not yet there to fold to, so that extension is filed as written, as before.
Lines changed: 14 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,14 @@
1+
---
2+
'@object-ui/plugin-detail': patch
3+
---
4+
5+
fix(plugin-detail): on a desktop the record page's enlarged related tab draws the child object's list grid, as the phone already does
6+
7+
Clause-②: no
8+
Metadata-protocol: none
9+
10+
The enlarged related tab (the record page's active tab) drew a generic `data-table` on a desktop: rows about 65 px tall where the child object's list page draws 41 px, a name in plain text that did not open the record, and the count twice (on the 全部 chip and on the line under the toolbar). It now renders `object-grid` over the rows the related list already read, the branch the phone has taken since the phone card rows landed, so the tab is the child object's list face: the list page's row height and density, the name as a link that opens the record, and the grid's row actions.
11+
12+
The related list stays the source of truth. Its window, page and sort reach the grid as host props (the ones the list view hands the same grid), so paging and sort are still the server's `$top` / `$skip` / `$orderby`, and its own pager stays the one pager. The filter box, the 全部 / 我的 / 逾期 chips, 新建 and 导出 are unchanged. The host's row actions (`list_item`, and `record_related` inside a parent record) ride to the grid's row menu and still dispatch through the host against the child object and its row, which refreshes the list. While the 全部 chip is on and nothing narrows the rows, the line under the toolbar no longer repeats the chip's count.
13+
14+
One behaviour changes: while the filter box has text, the grid offers no header sort on a relational column (its stored value is an id), as on every list page; the data-table sorted that column by its label in that state. Compact panels in the right column and lists that are not enlarged are unchanged.
Lines changed: 18 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,18 @@
1+
---
2+
'@object-ui/plugin-detail': patch
3+
'@object-ui/app-shell': patch
4+
'@objectstack/spec': patch
5+
---
6+
7+
fix(plugin-detail): the record page's highlight strip is the object's declared `highlightFields` only; the path's stage and the owner each show once, under one word for the owner (#28868)
8+
9+
Clause-②: no
10+
Metadata-protocol: none
11+
12+
**No guessed strip.** An object that declares no `highlightFields` now has no highlight strip, so every field stays in the details body. This holds on the default record page, the phone record header and the record hover card. `deriveHighlightFields` (`@object-ui/plugin-detail`) used to pick up to six fields from a preferred-name list (`owner` / `owner_id` first, then amount, rating, industry, …) and hide them from the body. That guess is deleted. The function keeps its signature and returns the declared list minus the record's title field, at most six, or nothing. A declared strip is unchanged.
13+
14+
**The stage once.** When the page draws `record:path`, its `statusField` joins the body's `hideFields`, so the stage is not drawn again as a body field. The body's hide list is also cut at the six fields the strip shows: a declared entry past the sixth is now drawn by the body instead of by neither.
15+
16+
**The owner once, under one word.** When the strip carries `owner_id`, the meta line under the details leaves the owner out. The platform label of the injected `owner_id` column (`@objectstack/spec`'s system-field fallback) now reads 「负责人」 in zh-CN, the word the ui's `detail.owner` already used. It was 「所有者」. en (`Owner`), ja-JP (`所有者`) and es-ES (`Propietario`) already matched the ui packs. A label an author wrote is still never overridden.
17+
18+
**What an app sees.** An object without declared `highlightFields` loses its strip. Declare the fields the header should show: `highlightFields: ['amount', 'close_date', 'owner_id']`.
Lines changed: 12 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,12 @@
1+
---
2+
'@object-ui/plugin-grid': patch
3+
---
4+
5+
fix(plugin-grid): the list grid's row ⋯ trigger is visible at rest, in a muted tone, instead of appearing only under the pointer
6+
7+
Clause-②: no
8+
Metadata-protocol: none
9+
10+
The Actions column header is always drawn, but each row's ⋯ trigger was hidden by opacity on a fine pointer until the pointer reached the row, so a desktop list showed a headed column of cells that looked empty and read as "no actions available". The trigger is now drawn on every row at rest in the muted foreground tone, on every pointer type alike, and turns full strength on row hover, focus within the row, a selected row, its own keyboard focus, while its menu is open and on a coarse pointer. The menu's items, their gates and their order are unchanged. No key, option or prop is added.
11+
12+
What to do: nothing.
Lines changed: 20 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,20 @@
1+
---
2+
'@objectstack/plugin-email': patch
3+
---
4+
5+
fix(plugin-email): a templated email to a named locale folds it against the template's locales and tries the stack's declared fallback before en-US (#28880)
6+
7+
Clause-②: no
8+
Metadata-protocol: none
9+
Skills-docs: none — no file under `skills/**` teaches the email template locale ladder.
10+
11+
`sendTemplate` and `renderTemplate` used to try a named recipient locale exactly, then `en-US`, then fail with `TEMPLATE_NOT_FOUND`. A recipient whose tag differed from a template row only by region (`en-GB` against an `en` row, `zh-TW` against `zh-CN`) or by case got nothing, and the stack's `i18n.fallbackLocale` was never read, so those deliveries dead-lettered.
12+
13+
With the built-in template loader, a call that names a locale now resolves, in order:
14+
15+
1. that locale against the template's own locales by the platform's BCP-47 rules (`resolveBundleLocale`): exact, case-insensitive, base language (`en-GB` → `en`), then variant expansion (`zh-TW` → `zh-CN`), taking the first match in ascending tag order;
16+
2. the stack's declared fallback (`i18n.fallbackLocale`, else `defaultLocale`, from the i18n service), folded the same way;
17+
3. `en-US`;
18+
4. `TEMPLATE_NOT_FOUND`, still a permanent failure.
19+
20+
An exact match answers as before. A call that names no locale keeps its ladder (`en-US`, then the lowest tag) and does not read the declared fallback. A replacement `TemplateLoader` keeps the exact ladder: the fold is a capability of the built-in loader, and no exported type changes.
Lines changed: 11 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,11 @@
1+
---
2+
'@objectstack/spec': patch
3+
'@objectstack/metadata-protocol': patch
4+
---
5+
6+
fix(spec): the connector read withholds root keys the entry shape does not declare
7+
8+
Clause-②: no
9+
Metadata-protocol: none
10+
11+
Read-side hardening of the built-in `connector` redactor. A stored connector body is served without any top-level key the authored connector entry (`DeclarativeConnectorEntrySchema`) does not declare — including a cleared one (`null`, `''`, `{}`) and one whose own name carries a dot; the declared set is read from that schema's own shape, and the read decorations (`_diagnostics`, `_draft`) are left alone. Under `providerConfig`, an array, or a key whose own name carries a dot, that holds a URL with credential parts is withheld whole rather than served. Each withheld position is named in `redactedKeys`, and the metadata protocol's carry-forward keeps the stored value at rest when the served form is saved — resolving a dotted key by the whole-string fallback when the dotted split resolves nowhere. The runtime's execution read is unchanged, and nothing an author can write changes.

‎.claude/skills/pm-dispatch/SKILL.md‎

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -117,7 +117,7 @@ PM 的工作是循环:选卡 → 认领 → 派发 → 收集 → 复核 → 报
117117
| `pm:epic`(父单或 sub-issue) | 已由 epic PM 保留;其它 PM 永不取;⛔ 永不与 `pm:queue` 同挂 |
118118
| `pm:seat` | 座位登记贴:协议载体,不是待分诊的工作 |
119119
| `priority:p0` | 插队:同轮立即派发,越过 `batch`、深度等待(`LOCK_DEPTH_HOLD`)与一切并发上限;无槽即暂停低级在飞 dev 腾位(机制 `references/execution-duties.md`);⛔ 不豁免同文件串行与认领协议 |
120-
| `area:*` | 功能轴:每卡恰一个;分诊首触打,与 `domain:*` 同笔;工具卡打它所护的轴;⛔ 不回填存量。四读者:维护者按轴看板;定级继承该轴清单项;父单分组;候选顺序,即「功能点位次」= 父单在「路上的功能点」上的次序;同轴至多一个在飞,除非文件面不相交 |
120+
| `area:*` | 功能轴:每卡恰一个;分诊首触打,与 `domain:*` 同笔;工具卡打它所护的轴;⛔ 不回填存量。四读者:维护者按轴看板;定级继承该轴清单项;父单分组;候选顺序,即「功能点位次」= 父单在「Features on the road」上的次序;同轴至多一个在飞,除非文件面不相交 |
121121
| open PR 引用该单 | 已实现,复核中 |
122122
| merged PR 带 `Fixes #n` | 完成(GitHub 关单) |
123123

0 commit comments

Comments
 (0)