The Linux fallback owner polls its process group every 15 ms after the
direct child exits. Remember the live member found by the last scan and
read only its stat on later polls, so a long-lived background member no
longer costs a full synchronous /proc enumeration per poll.
Fixes#7600
* fix(ui-primitives): keep tall menus inside the mobile viewport
* fix(client): anchor model picker with Floating UI
Use Floating UI for fixed model-menu placement and automatic viewport updates. Preserve top-end alignment and defer automatic focus until placement is ready. Guard asynchronous results across updates and menu teardown.
* fix(client): preserve menu blur under pinned group titles
Keep the current provider title outside the native scroll viewport so
model rows cannot bleed through during fast scrolling. Preserve the
enclosing menu fill and backdrop filter, and update both model pickers,
lifecycle tests and ARIA goldens.
Validation: doc-sync, full build, final type checks, focused lint and
all 12 model-picker browser checks pass. The full unit run has 42,597
passing tests and four failures in unchanged gateway, third-party-
notices and Inspector tests. The Inspector case passes in isolation.
* fix(client): stabilize floating panels during browser zoom
* fix(client): keep portal menus aligned during Safari zoom
* fix(client): commit floating menu positions before paint
* fix(client): unify floating panel positioning
* fix(client): validate grouped menu headings
* fix(client): hide inactive pinned menu headings
* test(client): await floating menu layout across consumers
* test(ui-chat): await folded reply menu positioning
* test(client): align current consumers with asynchronous panel layout
* fix(client): prevent mobile model search focus zoom
* refactor(client): inject mobile device state from layout
* fix(client): keep anchored panel layout synchronous
* fix(client): address panel review coverage and documentation
* fix(client): keep viewport changes focused on iOS zoom
* fix(client): cover anchored panel lifecycle and scrollbar colors
* fix(client): preserve zoomed model menu scrolling and alignment
* fix(client): hide empty model menus and stabilize viewport pan tests
* test(web): use native wheel input for zoomed viewport panning
* fix(client): keep automation pickers scrollable when zoomed
* fix(client): allow horizontal gestures across clock columns
* fix(client): ellipsize overflowing floating-panel labels
* test(web): make clock overflow fixture independent of system fonts
---------
Co-authored-by: Zrs000 <[email protected]>
This pull request lands the tool; switching the shipped defaults is a
separate pull request stacked on it. Every shipped `tool-fs` row states
`applyPatch: false`, the `fs-observation-policy` row stays mounted, the
write, edit, and apply_patch guidance names the policy, and every recorded
expectation outside the apply-patch scenarios returns to the master
baseline. The upgrade guide, the parity exception, and the optional
classification of the policy package move to the stacked change.
The SDK/Desktop parity check requires the `standard` tool sets to match, so
the base row enables apply_patch like the Web standard preset; the SDK PTC
overlay keeps that row while the Desktop PTC preset lists its own, which the
parity exceptions now record. fs-observation-policy, no longer mounted by
default, is classified as an optional product package.
Every headless, SDK, and ACP recording follows: sidecars and recorded headers
gain the apply_patch schema, the fixed-budget compaction scenarios pin
`applyPatch: false` so their recorded thresholds hold (model-switch-notice and
session-resume own their schema sidecars again), the parent presets that
borrow the default pin mirror the base row, and the pwsh and Python PTC pins
carry the schema and guidance.
Keeps the apply_patch toolview registration with master's file-menu inject,
takes master's shortened slot JSDoc, re-adds the seven filesystem type-equiv
manifest entries, regenerates the event matrix and client catalogs, moves the
fs-observation-policy upgrade guide under v0.2.2-alpha.2 with the matching
fs-patch version, and refreshes the two SDK scenarios master added for the
write and edit guidance change.
Evaluation on the shipped Web composition decided the defaults: the
`fs-observation-policy` row in dsh-base ships `disabled: true`, so write,
edit, and apply_patch mutate unconditionally in every profile; the Web
`standard` preset lists apply_patch; and the tool-fs `applyPatchGuard` switch
is gone, apply_patch always consulting the fs/* intent waterfalls like the
other mutating tools.
The write and edit guidance stops citing the policy as the default, and the
apply_patch guidance asks for enough hunk context instead of a prior read.
The two recordings that capture a policy refusal (fs-policy-reject and
apply-patch-rejected) re-enable the row through a per-scenario
runtime.cordis.yml; every other sidecar follows the prompt change, and the
Web standard-preset recordings gain the apply_patch schema. An upgrade guide
records the default change.
* fix(web): prioritize model ids in narrow rows
Keep the model ID in the compact row and expose its display name when expanded. Preserve wide layouts and input focus while resizing.
Fixes#7511
* test(desktop): wait for supervised network recovery
* fix(web): keep folding rows visible through the viewport
* test(web): measure fold movement within the viewport
* test(snapshot): await cancelled job settlement before retry
* fix(web): stabilize model name focus across responsive layouts
* revert: remove unrelated CI fixes from model layout PR
Revert d1c8ef1de1a, 1050ece1980, 6f2f0906341 and 8f7efbca6b0. Preserve the responsive model layout and its review fixes.
* test(web): align DeepSeek Messages model row snapshot
Turtle