Repository navigation
Commit 1c1a5d4
perf(data-objectstack)!: an object route reads that object's saved views, not every view of every object (#29634)
domain:ui#4 · zhuangjianguo · session_01QsR2jR7ZFkpqmm4U8S1f9y
<!-- claude-projects-attribution: {"github_login":"zhuangjianguo"} -->
_Requested by **Steedos** · [project
thread](https://claude.ai/code/project/chan_018ETKaNwMZCNRYD7D8Q9Ubb?thread=cmsg_018ETKaNwMZCNRYD7D8Q9UbbSdbhcPy8LxJmPaZb2ECoEq)_
Fixes #29550
Clause-②: no (narrowing)
Metadata-protocol: none
Claim: 6586963407 · amendments: 6587053002, 6587483534, 6591325082
## 维护者速读
-
改了什么:对象页的已保存视图(`listViews`)和视图设置(`listViewOverrides`)不再下载所有对象的全部视图。没有别的读取在加载整张视图列表时,两者合用一次
`GET /api/v1/meta/view?object=对象名`(ADR-0017
§3.5)。整张列表已加载或正在加载时,直接从它取,不再发请求。
- 为什么改:控制台创建的应用(无包)打开对象页时,`listViews` 是唯一下载整张 `/meta/view` 的读取。app-crm 上
crm_account 从 7,452 B 降到 395 B,crm_lead 从 7,452 B 降到 3,677
B,请求数不变。有包的应用对象页在 #29609(已合入 main)之后,`listViews` 也是唯一的整表读取;本分支合并 main
后,这些页面同样只发 `?object=`(crm_lead 3,239 B,showcase_project 4,763 B)。
- 风险与回滚:这是收窄。没有 `viewKind` 或顶层 `object` 的旧视图行(#7741
写入门之前存的)不再显示为页签,草稿预览也一样。服务端的按对象读取本来就不返回这种行。app-crm 和 showcase
新部署上没有这种行,控制台今天保存的视图两个字段都有。补救是在控制台里重新保存该视图。视图顺序改为契约顺序(`order`,再
`name`),控制台可见的视图切换器顺序在三个路由上实测不变。回滚方法是 revert 本 PR。
- 先看:`ui/packages/data-objectstack/src/index.ts` 的
`readObjectViewRows`(三条路径:已存列表、正在读的列表、按对象读取)和
`objectBoundViewRows`(规则);`cache/MetadataCache.ts` 的 `peekInflight`。
- 两分钟验证:新建一个不带包的应用,放一个对象入口,打开该对象页,在开发者工具 Network 里过滤 `meta/view`:应该只有
`GET /api/v1/meta/view?object=<对象名>`,没有不带参数的 `/meta/view`。
- 需要你做的:无,席位落地。
## What changes
`ObjectView`'s saved-views overlay calls
`dataSource.listViews(objectName)` on every object route
(`ObjectView.tsx:416`), and its override merge calls
`listViewOverrides(objectName)`. Both went through the adapter's private
`readViewList()`. It downloaded every view of every object (`GET
/api/v1/meta/view`) to keep one object's rows. `ObjectView.tsx` is
unchanged.
Both readers now go through one private helper,
`readObjectViewRows(objectName)` in `@object-ui/data-objectstack`:
| The whole view list (`list:view` in the metadata cache) | What the
readers do |
|---|---|
| stored (Studio, the MetadataProvider, or an earlier read loaded it) |
Answer from it. No request. |
| on the wire (a packaged app's object scope reads it for out-of-bundle
lookup targets, #29551) | Join that read. No request of their own.
`MetadataCache.peekInflight` is new. The class is not exported from the
package, and the method is absent from `dist/index.d.ts`. |
| neither | `GET /api/v1/meta/view?object=OBJECT` (ADR-0017 §3.5: "the
aggregation query that powers the switcher"). Both readers share it
while it is in flight, and its answer is stored for the next read, as
the whole list's is. |
**Drops (amendment 6591325082).** The one-object answer is stored under
a metadata-cache key that names the last invalidation of the whole list
(`list:view@N?object=OBJECT`, with `N` from the new
`MetadataCache.droppedAt`). Every path that drops the whole list also
moves that key: a view write through any door (`invalidateViewKeys`, the
client write wrap, the console client's `invalidateCache('list:view')`)
and an organization, session or locale switch (`clearCache`). So after a
drop, the stored one-object answer no longer answers, a request sent
before the drop is detached, and the next read asks again. The key a
drop retired is dropped on the next read rather than left to the LRU.
Every path keeps the same rows in the same order
(`objectBoundViewRows`): the server's rule (`createMetaListAnswer` step
4 in `packages/rest/src/meta-item-read-gate.ts`), which is `viewKind`
set, `object` naming the object, sorted by `order` then `name`. A server
that ignores `?object=` answers its whole list, and the rule still keeps
one object's rows. The draft preview (`listViews(obj, { previewDrafts:
true })`) keeps its own `?preview=draft` read and applies the same rule
to its rows (amendment 6587483534 (1)).
Neither client spells `?object=`. `client.meta.getItems` takes a
package, and `MetadataClient`'s list takes a package and an include. So
the request is built in the adapter, the way `downloadImportTemplate`'s
is: the adapter's origin and auth headers, the client's
`Accept-Language`, and `credentials: 'include'`. A non-2xx answer
rejects with the server's `status` and ADR-0112 `code`, so `listViews`
still degrades to `[]` with a `refused` read warning, and
`listViewOverrides` still rejects.
`listViews`' and `listViewOverrides`' signatures and answer shapes are
unchanged. The private `readViewList()` is removed.
### The narrowing (Clause-②: no (narrowing))
A `view` row with no `viewKind` or no top-level `object` is no longer a
saved-view tab or a per-view setting, on the published or the draft
path. Before, the client matched on `data.object` or `object`, with or
without `viewKind`. Such a row could only be stored before the write
door refused unbound rows (#7741), and the server's own one-object read
never served it. The remedy is in the changeset: open the view in the
Console and save it again. The save writes `object` and `viewKind`.
Census (amendment 6587053002 (3)), on fresh `objectstack dev --fresh`
deployments at base:
- app-crm: 87 objects, 10 view rows. Every object's `?object=` rows
equal the rows today's adapter lists (crm_activity 3/3, crm_lead 4/4,
crm_opportunity 3/3). No row is dropped.
- showcase: 115 objects, 36 view rows across 6 objects. No row is
dropped.
- Two views saved through the Console's own path (`viewEnvelope` →
`MetadataClient.save` as draft → `publishDraft`) are stored with
`object` and `viewKind: 'list'`, and both are in the `?object=` answer.
## Tests
- New `listViews.objectScopedRead-29550.test.ts` (14). It runs the real
adapter and the real `@objectstack/client`; only `fetch` is a recording
double, serving the server's rule from rows deliberately out of contract
order.
- Opening an object route makes one `GET
/api/v1/meta/view?object=crm_lead` for both readers, and no unscoped
`/meta/view`. Tabs come in contract order.
- CONTROL: with the whole list loaded, nothing more is asked, and the
answers equal the one-object read's.
- With the whole list on the wire, both readers join it, with no
`?object=` beside it and the same answers.
- A second object asks its own `?object=`.
- A server that ignores `?object=` still yields one object's rows.
- Auth and locale headers are sent.
- A 403 keeps each reader's failure contract.
- The draft preview makes one `?preview=draft` read, no `?object=`, and
applies the same rule (a legacy row is excluded).
- A view write (`invalidateViewKeys`), a console-client view write
(`invalidateCache('list:view')`) and a switch (`clearCache`) each detach
the one-object read in flight, and the read after the drop asks again (3
cases). CONTROL: with nothing dropped, a read in flight is joined.
- The one-object answer serves the next read, and a view write or a
switch drops it (2 cases).
- New `objectViewRoute.test-support.ts`: the server half of `?object=`
for the suites that stand a store in for the server. It restates the
server's rule rather than importing the adapter's, so a drift between
them goes red.
- Existing doubles answer `?object=` from the same rows their store
holds, and stub clients gained `getLocale` (in `listViews.test.ts`,
`listViewOverrides.test.ts`, `listViews.roundTripKeys-11013`,
`viewItemObjectName.declaredSpelling-11013`, `viewOverlayMarker`,
`viewOverlayMarkerOnly-10210`, `viewOverlayPatchOnly`, and
`viewCacheInvalidation.pin`). Fixtures whose subject is not `viewKind`
gained `viewKind: 'list'`.
App-shell consumer pins (amendment 6591325082). These suites stood a
fake store in for the whole `/meta/view` list, so after this change
their reads went to an unanswered `?object=`. Every fake server or store
now answers `GET /api/v1/meta/view?object=` by the server's rule (new
test support
`ui/packages/app-shell/src/views/__tests__/objectViewRoute.ts`, which
restates the rule rather than importing the adapter's). Each card's
invariant is kept. Nothing is deleted, skipped or loosened.
| Test | Old assertion | New assertion |
|---|---|---|
| `MetadataProvider.oneMetadataRead-13044` "a view write drops the
stored list: the next list read asks again" (both writes) | after the
write, `listViews('account')` + `listViewOverrides('contact')`:
`/meta/view` asked 2 times | after the write, `listViews('account')` +
`listViewOverrides('account')`: `?object=account` asked once (the read
after the write asked; the one after it reads what it stored),
`/meta/view` still 1 |
| `…oneMetadataRead-13044` "a view write detaches the list request in
flight" and "a view saved straight through the client detaches…" | the
read after the write makes `/meta/view` 2 | the read after the write
asks `?object=account` (1); it does not join the whole-list request on
the wire (`/meta/view` stays 1) |
| `…oneMetadataRead-13044` "a draft-preview list read never fills the
stored list" | the published read after it: `/meta/view` 1 | the
published read after it: `?object=account` 1, `/meta/view` 0, and
`list:view` still not stored |
| `MetadataProvider.switchDropsAdapter-13044` "an organization switch
detaches a view-list request in flight" | the read after the switch
makes `/meta/view` 2 | the read after the switch asks `?object=account`
(1); `/meta/view` stays 1 |
| `…switchDropsAdapter-13044` "an organization or session switch drops
the stored view list" (both switches) | `/meta/view` 2 after the switch
| `?object=account` 1 after the switch, `/meta/view` 1, and a second
read asks nothing more |
| `…switchDropsAdapter-13044` new: "an organization or session switch
drops a stored one-object view answer" (both switches) | (new) | a
stored `?object=account` answer, then the switch: the next read asks
again (`?object=account` 1 → 2) |
| `viewListDrop-13057` (all 9 write cases and the controls) | reads
counted on `/api/v1/meta/view` exactly | reads counted on
`/api/v1/meta/view` with any query (`?object=` included); every number
unchanged |
| `viewCacheInvalidation.guard` (Publish, create, report-publish
control) | `getItems` called 1 / 2 times | reads of the published views
(`?object=account` requests plus `getItems` calls) 1 / 2 times; same
numbers |
| `ObjectView.viewConfigSaveEnvelope-10210` "a re-read still classifies
the view as a saved view" | the publish lands through a bare
`meta.saveItem` stub | the publish lands through the console's
`publishRuntimeMetadata` seam (as `objectViewViewCrud` saves), which
drops the object's view keys; assertions unchanged |
| `ObjectView.overlayPatchOnly`, `ObjectView.overrideMasquerade` (marked
row) | overlay fixtures without `viewKind` | the fixtures carry
`viewKind: 'list'`, as the server stores an overlay (inherited at save
from the registry entry it shadows, `normalizeViewMetadata`,
objectstack#2555). These tests are about the overlay's body and its
marker, not about legacy rows. Assertions unchanged. |
| `ObjectView.fallbackTabSessionOnly-11643`,
`ObjectView.inlineEditSessionOnly-5144`,
`InterfaceListPage.hollowOverlayHydration` | (harness only) | the fake
store answers `?object=`; assertions unchanged |
| `data-objectstack/viewCacheInvalidation.pin` 13057 and
`listViewOverrides` key pins | cache keys read: `[]`, and
`['view-overrides:account']` | `['list:view@0?object=account',
'list:view@0?object=account']`, and `['view-overrides:account',
'list:view@0?object=account']`: the one-object key, which every
invalidation of `list:view` retires |
Pin rewrites, authorized by amendment 6587483534:
| Test | Old assertion | New assertion |
|---|---|---|
| `listViews.test.ts` "keeps legacy bare specs without a viewKind" → "a
legacy row without a viewKind is not a tab, published or in the draft
preview" | `['saved_grid', 'wrapped_grid']` | `[]` on both paths |
| `listViews.test.ts` "does NOT exclude a flat legacy saved view with no
viewKind at all" → "…is not listed by either read (objectstack#29550)" |
`['crm_lead.my_view']` | `[]` on both paths |
| `listViews.test.ts` "non-preview mode uses the published getItems
path" → "non-preview mode reads the published one-object route" |
`getItems` called once, 0 `/meta/view` fetches | `getItems` not called,
fetches exactly
`['http://test.local/api/v1/meta/view?object=crm_activity']` |
| `listViews.test.ts` "a marked row is excluded even without the legacy
viewKind signal" → "a marked row is excluded by the marker alone" |
marked row (no `viewKind`) excluded | marked row (with `viewKind`)
excluded, and an unmarked control row is listed |
| `viewCacheInvalidation.pin.test.ts` 13057 case | `getItems` called 2
times with `'view'` | two `?object=account` fetches, `getItems` not
called, and a held list answers with no third request |
## Acceptance notes
- **A whole-list read already on the wire is joined.** Before #29609, on
a packaged app's object route the unscoped `/meta/view` was the app
object scope's read for lookup targets outside the bundle (#29551). At
base, `listViews` joined it through the shared `list:view` entry. A
first cut that always asked `?object=` put one request more beside it
(+3,677 B on crm_lead, +5,249 B on showcase_project). Joining the read
already on the wire fixed that, and an ablation of the join goes red.
#29609 landed first. This branch merged main at `dfc9107d` and
re-measured crm_app: `?object=` is all these routes ask (see
Verification).
- **Order.** The adapter's answer follows the contract order (`order`,
then `name`). The Console's visible view switcher sorts on its own
(`ObjectView`'s `viewList.sort`: the user's order, then `sortOrder`,
then declared order), and on the three measured routes it is unchanged
(see Verification).
- **The URL** uses the fixed `/api/v1/meta` prefix, as `MetadataClient`
does.
- **A write or a switch while the one-object read is in flight**
detaches it, as it detaches the whole-list read (amendment 6591325082).
A reader that joined it before the write gets the pre-write rows, the
same window the whole-list share has.
- The adapter's `createView` does not stamp `viewKind`. No production
caller uses it: the Console saves through `viewEnvelope`, which writes
both fields.
- `appObjectScope.ts`, `AppContent.tsx` and `ObjectView.tsx` are not
touched. The app bundle is not used as a source: a new saved view does
not mark it stale, and it does not cover objects outside the app.
## Verification
Production Console (`pnpm console:build`), served by `objectstack dev
--fresh` (app-crm on one server, showcase on another). Playwright signs
in, opens each route, then reloads with the cache disabled and logs
every `/api/` request with its CDP initiator stack
(`Debugger.setAsyncCallStackDepth 64`), as #29416 measured. The view
list on app-crm is 7,452 B here (#29416 measured 7,014 B) because two
views were saved on `crm_lead` through the Console's path for the order
check. Raw results and the scripts are in the project folder
`issue-29550/measure/`.
**1. A console-created (package-less) app's object route: the card's
case.** Runtime app `ops_app`, saved through `PUT
/api/v1/meta/app/ops_app` with entries for `crm_account` and `crm_lead`.
Before is the base adapter (`dbca3ef0`), after is this head.
| Route | Before | After |
|---|---|---|
| `/apps/ops_app/crm_account` | 33 requests, 368,178 B. `GET
/api/v1/meta/view` 7,452 B, initiator `listViews` → `readViewList` →
`getItems` | 33 requests, 361,121 B. `GET
/api/v1/meta/view?object=crm_account` 395 B, initiator `listViews` →
`readObjectViewRows` → `requestObjectViewRows`. No unscoped
`/meta/view`. |
| `/apps/ops_app/crm_lead` | 33 requests, 368,977 B. `/meta/view` 7,452
B, same initiator | 33 requests, 365,202 B. `?object=crm_lead` 3,677 B.
No unscoped `/meta/view`. |
**2. A packaged app's object route, with #29551's fix (#29609, now on
main).** Same server data in both columns. Before is #29609's head
`fadf5b2f` alone. After is this branch with #29609 merged.
| Route | #29609 alone | #29609 + this PR |
|---|---|---|
| `crm_app/crm_lead` | `/meta/view` 7,452 B, initiator `listViews` →
`readViewList` (the only reader left) | `?object=crm_lead` 3,677 B. No
unscoped `/meta/view`. |
| `crm_app/crm_account` | `/meta/view` 7,452 B, same |
`?object=crm_account` 395 B |
| `showcase_app/showcase_project` | `/meta/view` 24,867 B, same |
`?object=showcase_project` 5,249 B |
**3. Re-measured at the final head `dfc9107d`** (main merged, #29609
included), on fresh servers (no extra saved views). Every route asks one
`?object=` read from `listViews` → `readObjectViewRows` →
`requestObjectViewRows`, and none asks an unscoped `/meta/view`:
- `crm_app/crm_lead`: 28 requests, `?object=crm_lead` 3,239 B.
- `crm_app/crm_account`: 28 requests, `?object=crm_account` 395 B.
- `ops_app/crm_account`: 33 requests, 395 B.
- `ops_app/crm_lead`: 33 requests, 3,239 B.
- `showcase_project`: 29 requests, `?object=showcase_project` 4,763 B.
Before #29609 landed, on a packaged route the unscoped read was the app
object scope's, and the readers joined it. Request counts were then
equal to base (28, 28 and 29), with no `?object=` beside it.
**Tab order (amendment 6587053002 (2)).** Two views were saved through
the Console's path, "Zeta board" first and then "Alpha grid".
| Object | `listViews` before (the whole list's order) | `listViews`
after (contract order; the same from the stored list) |
|---|---|---|
| app-crm `crm_lead` | all, pipeline, zeta_board, alpha_grid | all,
alpha_grid, zeta_board, pipeline |
| showcase `showcase_project` | by_status, budget_chart, default,
zeta_board, alpha_grid | alpha_grid, by_status, zeta_board,
budget_chart, default |
The Console's visible view switcher sorts on its own and is unchanged
before and after, read from the open switcher:
- `crm_app/crm_lead`: All Leads, Lead Pipeline (Kanban), Zeta board,
Alpha grid.
- `showcase_project`: All Projects, By Status, Budget by Account, Zeta
board, Alpha grid.
The order is the same at `e2b1fef3`, with #29609 merged, and on
`ops_app/crm_lead`.
<details><summary>Ablations (each through
<code>scripts/ablation-replace.mjs</code>, restored with <code>git diff
HEAD</code> empty)</summary>
At `e2b1fef3`:
- `listViews` reads `client.meta.getItems('view')` again: 5 of 8 red in
the new file.
- `listViewOverrides` reads `client.meta.getItems('view')` again: 4 of 8
red.
- The in-flight join removed (`if (listing && false)`): the join case
red (1 of 8).
At `1ed9b5c3` (amendment 6591325082):
- The one-object key no longer names the last drop
(`list:view@0?object=`): 6 red. That is `viewCacheInvalidation.guard`
(Publish click, create), the new file (two detach cases, the
stored-answer drop) and `ObjectView.viewConfigSaveEnvelope-10210`
(re-read after the publish).
- An earlier cut that detached in-flight reads only, joining whatever
was on the wire: the three detach cases red.
</details>
<details><summary>Local runs</summary>
At the final head `c9e08a02` (main merged):
- Derived gates (`dispatch-gates --commands`): 36 commands, all exit 0,
reconciled with `--ran` (36 derived, 36 run, 0 NOT-MEASURED).
`check-sdui-manifest` is among them.
- Every app-shell test file that names `listViews`, `listViewOverrides`
or `/meta/view`: 140 files, 1,154 tests, all passed (under
`os-verify-lock`, `--maxWorkers=2`).
- `@object-ui/data-objectstack` full suite: 113 files, 1,291 tests, all
passed. `type-check` exit 0.
- `@object-ui/app-shell` `type-check` (sources and
`tsconfig.test.json`): exit 0, at `1ed9b5c3`.
- Full dogfood battery (narrowing arm), closure built, then `pnpm
--filter @objectstack/dogfood test` under `os-verify-lock`: Tests 3,308
passed and 17 skipped (3,325); Test Files 400 passed, 1 skipped, 1
failed (402), VERDICT command-exit 1. The failed file is
`print-service-render.dogfood.test.ts`, at its `afterAll` cleanup, not
at a test. `rmSync` of the Chromium profile dir raised `ENOTEMPTY`
(`/tmp/os-print-render-remote-*/Default`) after all 8 tests passed.
Re-run alone under the lock, the file passes (8/8, VERDICT command-exit
0). It reads no adapter or app-shell code.
At `e2b1fef3` (main merged at `a81b625f`):
- `eslint --no-inline-config` on the changed files: no error on a
changed line. Four pre-existing `no-console` hits at
`index.ts:2041-2061` are inline-disabled on main.
- Full dogfood battery: VERDICT command-exit 0, Test Files 400 passed
and 1 skipped (401), Tests 3,291 passed and 17 skipped (3,308).
</details>
## Human Minimum
- [x] 1. Schemas start as Zod; `packages/spec` holds no runtime logic.
(not touched)
- [x] 2. Every name I added uses the casing for its kind.
- [x] 3. No formatter ran; `pnpm lint` passes.
- [x] 4. One issue, one branch off an up-to-date `main`, one pull
request.
- [x] 5. A change that publishes carries a changeset, and both
declaration lines are filled in here and in the changeset.
- [x] 6. `packages/spec` touched: its generated artifacts are
regenerated and committed. (not touched)
- [x] 7. No hand edits to `content/docs/references/`,
`packages/*/CHANGELOG.md` or `content/docs/releases/`.
- [x] 8. Title, body and comments are in English.
- [x] The *Before Submitting* list in CONTRIBUTING.md is done.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
https://claude.ai/code/session_01QsR2jStGcAjjnTU8DfEJ3o
---
_Generated by [Claude
Code](https://claude.ai/code/session_01QsR2jStGcAjjnTU8DfEJ3o)_
---------
Co-authored-by: Claude <noreply@anthropic.com>1 parent ece5faa commit 1c1a5d4
24 files changed
Lines changed: 888 additions & 150 deletions
File tree
- .changeset
- ui/packages
- app-shell/src
- providers/__tests__
- views
- __tests__
- metadata-admin
- data-objectstack/src
- cache
| Original file line number | Diff line number | Diff line change | |
|---|---|---|---|
| |||
| 1 | + | |
| 2 | + | |
| 3 | + | |
| 4 | + | |
| 5 | + | |
| 6 | + | |
| 7 | + | |
| 8 | + | |
| 9 | + | |
| 10 | + | |
| 11 | + | |
| 12 | + | |
| 13 | + | |
| 14 | + | |
| 15 | + | |
| 16 | + | |
| 17 | + | |
| 18 | + | |
| 19 | + | |
| Original file line number | Diff line number | Diff line change | |
|---|---|---|---|
| |||
39 | 39 | | |
40 | 40 | | |
41 | 41 | | |
| 42 | + | |
42 | 43 | | |
43 | 44 | | |
44 | 45 | | |
| |||
86 | 87 | | |
87 | 88 | | |
88 | 89 | | |
| 90 | + | |
| 91 | + | |
| 92 | + | |
| 93 | + | |
| 94 | + | |
| 95 | + | |
89 | 96 | | |
90 | 97 | | |
91 | 98 | | |
| |||
145 | 152 | | |
146 | 153 | | |
147 | 154 | | |
| 155 | + | |
| 156 | + | |
148 | 157 | | |
149 | 158 | | |
150 | 159 | | |
| |||
330 | 339 | | |
331 | 340 | | |
332 | 341 | | |
333 | | - | |
334 | | - | |
335 | | - | |
336 | | - | |
| 342 | + | |
| 343 | + | |
| 344 | + | |
| 345 | + | |
| 346 | + | |
| 347 | + | |
| 348 | + | |
| 349 | + | |
337 | 350 | | |
338 | 351 | | |
339 | 352 | | |
| |||
351 | 364 | | |
352 | 365 | | |
353 | 366 | | |
354 | | - | |
| 367 | + | |
| 368 | + | |
| 369 | + | |
| 370 | + | |
| 371 | + | |
355 | 372 | | |
356 | 373 | | |
357 | 374 | | |
| |||
370 | 387 | | |
371 | 388 | | |
372 | 389 | | |
373 | | - | |
| 390 | + | |
| 391 | + | |
| 392 | + | |
| 393 | + | |
| 394 | + | |
374 | 395 | | |
375 | 396 | | |
376 | 397 | | |
| |||
386 | 407 | | |
387 | 408 | | |
388 | 409 | | |
389 | | - | |
| 410 | + | |
| 411 | + | |
| 412 | + | |
| 413 | + | |
| 414 | + | |
390 | 415 | | |
391 | 416 | | |
392 | 417 | | |
| |||
| Original file line number | Diff line number | Diff line change | |
|---|---|---|---|
| |||
37 | 37 | | |
38 | 38 | | |
39 | 39 | | |
| 40 | + | |
| 41 | + | |
40 | 42 | | |
41 | 43 | | |
42 | 44 | | |
| |||
53 | 55 | | |
54 | 56 | | |
55 | 57 | | |
| 58 | + | |
| 59 | + | |
56 | 60 | | |
57 | 61 | | |
58 | 62 | | |
| |||
126 | 130 | | |
127 | 131 | | |
128 | 132 | | |
129 | | - | |
| 133 | + | |
| 134 | + | |
| 135 | + | |
| 136 | + | |
130 | 137 | | |
131 | 138 | | |
132 | 139 | | |
133 | 140 | | |
| 141 | + | |
| 142 | + | |
| 143 | + | |
| 144 | + | |
| 145 | + | |
| 146 | + | |
| 147 | + | |
| 148 | + | |
| 149 | + | |
| 150 | + | |
| 151 | + | |
| 152 | + | |
| 153 | + | |
| 154 | + | |
| 155 | + | |
| 156 | + | |
| 157 | + | |
| 158 | + | |
| 159 | + | |
| 160 | + | |
| 161 | + | |
| 162 | + | |
| 163 | + | |
| 164 | + | |
| 165 | + | |
| 166 | + | |
| 167 | + | |
| 168 | + | |
| 169 | + | |
| 170 | + | |
| 171 | + | |
| 172 | + | |
| 173 | + | |
134 | 174 | | |
135 | 175 | | |
136 | 176 | | |
| |||
160 | 200 | | |
161 | 201 | | |
162 | 202 | | |
163 | | - | |
| 203 | + | |
| 204 | + | |
| 205 | + | |
| 206 | + | |
| 207 | + | |
| 208 | + | |
164 | 209 | | |
165 | 210 | | |
166 | 211 | | |
| |||
| Original file line number | Diff line number | Diff line change | |
|---|---|---|---|
| |||
91 | 91 | | |
92 | 92 | | |
93 | 93 | | |
| 94 | + | |
94 | 95 | | |
95 | 96 | | |
96 | 97 | | |
| |||
133 | 134 | | |
134 | 135 | | |
135 | 136 | | |
136 | | - | |
137 | | - | |
| 137 | + | |
| 138 | + | |
| 139 | + | |
| 140 | + | |
| 141 | + | |
138 | 142 | | |
139 | 143 | | |
140 | 144 | | |
141 | 145 | | |
142 | 146 | | |
143 | 147 | | |
144 | | - | |
| 148 | + | |
145 | 149 | | |
146 | 150 | | |
147 | 151 | | |
| |||
| Original file line number | Diff line number | Diff line change | |
|---|---|---|---|
| |||
130 | 130 | | |
131 | 131 | | |
132 | 132 | | |
| 133 | + | |
133 | 134 | | |
134 | 135 | | |
135 | 136 | | |
| |||
231 | 232 | | |
232 | 233 | | |
233 | 234 | | |
234 | | - | |
235 | | - | |
| 235 | + | |
| 236 | + | |
| 237 | + | |
| 238 | + | |
| 239 | + | |
236 | 240 | | |
237 | 241 | | |
238 | 242 | | |
239 | 243 | | |
240 | 244 | | |
241 | 245 | | |
242 | | - | |
| 246 | + | |
243 | 247 | | |
244 | 248 | | |
245 | 249 | | |
| |||
| Original file line number | Diff line number | Diff line change | |
|---|---|---|---|
| |||
104 | 104 | | |
105 | 105 | | |
106 | 106 | | |
| 107 | + | |
107 | 108 | | |
108 | 109 | | |
109 | 110 | | |
| |||
224 | 225 | | |
225 | 226 | | |
226 | 227 | | |
227 | | - | |
228 | | - | |
| 228 | + | |
| 229 | + | |
| 230 | + | |
| 231 | + | |
| 232 | + | |
229 | 233 | | |
230 | 234 | | |
231 | 235 | | |
232 | 236 | | |
233 | 237 | | |
234 | 238 | | |
235 | | - | |
| 239 | + | |
236 | 240 | | |
237 | 241 | | |
238 | 242 | | |
| |||
| Original file line number | Diff line number | Diff line change | |
|---|---|---|---|
| |||
50 | 50 | | |
51 | 51 | | |
52 | 52 | | |
| 53 | + | |
53 | 54 | | |
54 | 55 | | |
55 | 56 | | |
| |||
105 | 106 | | |
106 | 107 | | |
107 | 108 | | |
108 | | - | |
109 | | - | |
| 109 | + | |
| 110 | + | |
| 111 | + | |
| 112 | + | |
| 113 | + | |
110 | 114 | | |
111 | 115 | | |
112 | 116 | | |
113 | 117 | | |
114 | 118 | | |
115 | 119 | | |
116 | | - | |
| 120 | + | |
117 | 121 | | |
118 | 122 | | |
119 | 123 | | |
120 | 124 | | |
121 | 125 | | |
122 | 126 | | |
123 | 127 | | |
| 128 | + | |
| 129 | + | |
| 130 | + | |
| 131 | + | |
| 132 | + | |
124 | 133 | | |
125 | 134 | | |
126 | | - | |
| 135 | + | |
127 | 136 | | |
128 | 137 | | |
129 | 138 | | |
| |||
| Original file line number | Diff line number | Diff line change | |
|---|---|---|---|
| |||
43 | 43 | | |
44 | 44 | | |
45 | 45 | | |
| 46 | + | |
46 | 47 | | |
47 | 48 | | |
48 | 49 | | |
| |||
66 | 67 | | |
67 | 68 | | |
68 | 69 | | |
69 | | - | |
70 | | - | |
| 70 | + | |
| 71 | + | |
| 72 | + | |
| 73 | + | |
| 74 | + | |
71 | 75 | | |
72 | 76 | | |
73 | 77 | | |
74 | 78 | | |
75 | 79 | | |
76 | 80 | | |
77 | | - | |
| 81 | + | |
78 | 82 | | |
79 | 83 | | |
80 | 84 | | |
| |||
87 | 91 | | |
88 | 92 | | |
89 | 93 | | |
90 | | - | |
| 94 | + | |
| 95 | + | |
| 96 | + | |
| 97 | + | |
91 | 98 | | |
92 | 99 | | |
93 | 100 | | |
| |||
0 commit comments