fix(webview): invalidate live sibling view state on reset and settings import (vps2 F4) - #1562
Conversation
|
Warning Review limit reachedNext included review available in 10 minutes. View limit detailsLimit details: You’ve used all 4 included reviews currently available. You've used all free OSS reviews for now. Wait for the free limit to reset to keep reviewing this public repository. Review configuration: ⚙️ Run configurationConfiguration used: Path: .coderabbit.yaml Review profile: ASSERTIVE Plan: Team Run ID: 📒 Files selected for processing (30)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
Review statusThanks for contributing. This comment tracks the review sequence and the next action. Current step: Required CI passed. Waiting for automated review of the latest commit. If automated review does not start, a maintainer must restart it. Review-state labels are managed by this workflow; do not edit them manually. |
80c147f to
da97f98
Compare
Codecov Report❌ Patch coverage is 📢 Thoughts on this report? Let us know! |
916900f to
d56b60b
Compare
d56b60b to
c2822dc
Compare
…en view-identity tests Track the in-flight tab panel creation with a module-level promise so concurrent openClineInNewTab calls reuse one panel and provider (adds a Promise.all regression test). ClineProvider.spec sets the private view via the public resolveWebviewView() instead of a ts-ignore assignment. registerCommands.spec types evictCurrentTask/refreshWorkspace on the fixture and drops the as any attachment. eslint-suppressions: prune the registerCommands.spec.ts entry (two as any suppressions removed).
…n the concurrency assertion
…-bar posts - openClineInNewTab: extract the unserialized creation body into createTabPanelUnlocked and guard the in-flight slot clear so a settled creation cannot clobber a replacement already stored in the slot. - onDidDispose: clear the tracked tab ref only when the disposing panel is still the tracked one, so a late disposal of a replaced panel cannot clobber the replacement's ref. - MDM lookup failure: log the fallback to the output channel instead of swallowing it silently. - Route the six title-bar button handlers through a shared postActions helper that posts each action in order and logs failures with the handler-specific prefix. - package.json: add the four InTab commands to the command palette, scoped to the active tab panel. - Tests: handler-level regression for openInNewTab + popoutButtonClicked started before the first creation resolves; fresh-creation test for a settled in-flight promise; stale-panel disposal regression; retained panel assertion for disposed tab instances; rightmost-editor column placement assertion; MDM fallback output assertion; %s placeholders for primitive it.each titles. - Stryker directives for the two equivalent setPanel type-literal mutants (setPanel branches only on type === sidebar).
b9e8fb7 to
05f264b
Compare
Replace the weak toBeDefined() assertion in the dispose spec with an identity check against the panel returned during creation, per the CodeRabbit actionable comment on this PR (review run 7c4cfeb3-6dd9-4615- 9a58-70cfc705eca2). The tracked tab is now pinned with toBe(panel) before the dispose assertions, so a wrong or duplicated tracked panel fails the suite instead of passing a defined-only check. Upstream: Zoo-Code-Org#1528 (vps2 F0)
05f264b to
eac3873
Compare
Retain the tracked tab panel in the InTab handler cases and assert that getInstanceForView was called with that exact panel, per the CodeRabbit actionable comment on this PR (review run 4afe1273-8739-4235-90d3-311db5f6ccb9, inline comment 3952466254 on the tabHandlerCases spec). A handler resolving any other view now fails instead of passing on the stubbed provider result alone; the same identity pin is applied to plusButtonClickedInTab. Upstream: Zoo-Code-Org#1528 (vps2 F0)
…States Each ClineProvider instance now owns a unique viewId (renderContext plus a monotonic counter) and registers a stable viewStateId for durable persistence. - Per-view state buffer (viewLocalState) holds mode / currentApiConfigName / apiConfiguration overrides in memory; saveViewState persists the non-secret subset durably under the active view id, rekeyed to the stable id on registration. - viewStates is stored as a map pruned to the newest 50 entries; writes go through a serialized queue so concurrent provider instances merge without lost updates. - setViewStateId sanitizes ids and rejects "__proto__" so a per-view entry can never be keyed through the Object.prototype setter. - postMessageToWebview no longer awaits the webview ack: a remounted or disposed page never acknowledges, and awaiting would wedge task-critical callers. - History restore falls back to the default mode view-locally instead of writing the shared global mode. - GlobalState gains the "viewStates" key and GLOBAL_STATE_KEYS tracks it. Adds F1a coverage in ClineProvider.spec.ts (viewId uniqueness, saveViewState persistence semantics, loadViewState fallback and failure, pruning, the __proto__ guard) and adapts the two history-restore tests in ClineProvider.sticky-mode.spec.ts to the view-local restore. getState() merging of hydrated per-view values and the remaining view-state suites land in the follow-up (F1b).
…lude viewStates from settings transfer
6027b27 to
97b8877
Compare
…ions through the view-local buffer
…overrides Fold ClineProvider viewLocalState on top of ContextProxy values in getState() (mode, apiConfiguration, and all per-view fields) so each webview reports its own selections while falling back to shared global state for everything else. Ports the getState-merging and local-state-isolation spec coverage from the superseded vps2 source. Also pins the full default surface of the merged read path, including the apiConfiguration provider fill-in when provider settings sanitize the raw value away (mutation-diff gate).
…tore deleteProviderProfile only rewrote the UI-facing listApiConfigMeta and currentApiConfigName in ContextProxy, leaving the profile's settings in the ProviderSettingsManager store (context.secrets). Per-mode mappings (modeApiConfigs) that still pointed at the deleted profile re-activated its stale settings on the next handleModeSwitch, clobbering the active configuration: the subtask child profile's gpt-4.1-mini leaked into ask-mode tasks, breaking downstream e2e suites (60s timeouts on search_files no-match and terminal reuse after zero-chunk shell race). Purge the profile from the store on delete so dangling mode mappings can no longer resolve it: listConfig().find(id) fails and handleModeSwitch continues with the current configuration. The F3 mode/profile isolation commit further up the chain introduces the same purge plus per-view pin handling. Regression test: sticky-profile spec "deleteProviderProfile removes the stored profile so a dangling mode mapping can no longer re-activate it".
… state WMH webviewDidLaunch persists the webview view-state id via provider.setViewStateId and re-pins the view-local currentApiConfigName through provider.saveViewState when the view-local profile is missing but the shared global selection is still valid. updateSettings is routed through provider.setValue so view-local buffer and pin sync stay consistent with the other mutation paths. The webview VSCodeAPIWrapper gains a stable getViewStateId persisted via setState (with an in-memory fallback) and the launch effect posts the id with the webviewDidLaunch message.
…ate ids getViewStateId now trims and rewrites unsafe characters before reuse, mirroring ClineProvider.setViewStateId, and rejects whitespace-only and __proto__ values by generating a fresh id. Regression coverage: normalized reuse, whitespace-only, and __proto__.
…d dispose retention
Port vps2 F3 (mode/profile writes) from upstream 978, hunk-by-hunk against the F1c..CS residual: - ClineProvider: add repointPersistedViewStates() to re-point currentApiConfigName across per-view entries when a profile is renamed or replaced, and prune orphaned entries; validate handleModeSwitch slugs against the custom-modes manager and no-op with a log on unknown modes; drop the as-any cast in delegateParentAndOpenChild. - Task: route mode switches through provider.handleModeSwitch(task) and keep the submitted message on failure instead of setMode(). - SwitchModeTool and specs: durable per-view mode writes. - webviewMessageHandler: no change vs base - the kimi-code OAuth hunk in the residual is CS-only divergence (not-ported register item 1), not part of Zoo-Code-Org#978/Zoo-Code-Org#979. - webviewMessageHandler.spec: ported only Zoo-Code-Org#979's 4 mock fields + defaultModeSlug import; the stack-side legacy-repair test, Key-aware getValue mock and em-dash comment are retained (register item 4). Zoo-Code-Org#979's mode-routing WMH.spec describe ("routes mode messages through handleModeSwitch instead of writing ContextProxy directly") exists in neither CS nor the stack and is ported by no unit (open question, logged). - Tests: H3/H4 durable handleModeSwitch writes in ClineProvider.spec.ts; profile-mutation, profile-activation and handleModeSwitch-integration describes (incl. A4 non-focused-target regression and new mutation-killing tests) in ClineProvider.parallelMode.spec.ts; sticky-mode and webviewMessageHandler spec updates; retain the setViewStateId __proto__ guard + spec test - shipped F1a hardening; the residual's guard removal is lineage divergence, not F3 content. - eslint-suppressions.json: no-explicit-any counts decrease for core/webview/ClineProvider.ts (12 -> 11) and core/webview/__tests__/ClineProvider.sticky-mode.spec.ts (36 -> 33). Upstream: Zoo-Code-Org#978 (vps2 F3) - issue Zoo-Code-Org#978; content ported hunk-by-hunk from the F1c..CS residual, cross-checked against upstream PR
Route the CS API setConfiguration write through ClineProvider.setValues instead of the raw context proxy so the view-local subset stays in sync with the sidebar view's state. A direct context-proxy write leaves the view-local apiConfiguration stale; the mode-switch / profile-activation paths re-apply the current provider settings to the focused task on every user message, so a task started after an API setConfiguration runs with the stale configuration (wrong model / reasoning flags / endpoint). That is the deterministic 13-test e2e provider-suite failure (72/13) on this head: the provider suites configure the provider via setConfiguration and the mock serves fixtures by the model the request actually carries. Ported from the F5 head (vps2 F5) where the e2e provider suites pass; includes the unit spec asserting the routing. Upstream: Zoo-Code-Org#982 (vps2 F3)
… mocks API.setConfiguration now routes through ClineProvider.setValues (8576b18 "fix(api): route setConfiguration through ClineProvider.setValues"), but the pre-existing spec still shaped the provider mock around the old contextProxy.setValues call. Both tests in this file failed with "this.sidebarProvider.setValues is not a function" on the platform-unit-test runs (F3 Zoo-Code-Org#1560 windows unit, job 101863508169, run 34161293841). Add the provider-level setValues mock to both tests to match the new routing. Upstream: Zoo-Code-Org#1560 (vps2 F3)
The F1a buffer routing and the new delete-time activation branch interacted in two ways: the unrelated-pin branch re-wrote currentApiConfigName through setValue although the view-local buffer already holds the surviving pin (drop the write: it only triggered a viewStates prune pass and could clobber the pin with the shared slot value), and deleting the current profile now activates the fallback, which reads its settings via providerSettingsManager, so the spec test uses a test double like its siblings in the describe.
…s import
- ClineProvider: new broadcastResetToAllInstances() clears each live instance's
view-local cache and issues the single global contextProxy setValue("viewStates", undefined)
write (single write-queue clear; no secrets involved, no prune-cap regression).
- resetState: awaits broadcastResetToAllInstances() before the final postStateToWebview
so parallel tabs do not keep stale durable/in-memory per-view state.
- importExport: ImportWithProviderOptions.provider gains optional
broadcastResetToAllInstances?(); importSettingsWithFeedback calls it in a guarded
try/catch (log-only) after a successful import, so a failing broadcast never fails the import.
- importExport spec: 3 new tests (broadcast called when available / skipped when missing /
import result preserved when broadcast throws, console.warn asserted; the skip test also
asserts the broadcast-failure warn is NOT reached). Provider identifiers use
providerIdentifiers.* per the zoo/no-raw-provider-identifiers rule (lint-required
adaptation from Zoo-Code-Org#981's raw-string casts; no semantic change).
- parallelMode spec: appends the CS source-of-record describes (multi-instance isolation,
_clearViewLocalState) — 5 new tests.
- ClineProvider spec: forward fix of the F3 resetState sentinel (F4's global viewStates
clear removes the key; the F3-era toEqual({}) expectation is replaced by toBeUndefined())
plus a new cross-instance resetState test pinning the multi-instance broadcast path
(sibling view-local cache cleared; sibling and caller each post state exactly once).
- webviewMessageHandler.ts was NOT edited: the importSettings case already passes the full
ClineProvider, which structurally satisfies the extended provider type and reaches the real
broadcast method — Zoo-Code-Org#981's structural wrapper hunk is redundant in this stack.
Upstream: Zoo-Code-Org#980 / PR Zoo-Code-Org#981 (vps2 F4)
97b8877 to
9004f5b
Compare
Draft PR — vps2 unit F4 (cross-instance reset + import invalidation).
Tracking: easonLiangWorldedtech#41 (vps2 series ledger). Upstream issue: #1561 (this series' gap record; the original upstream bug is #980). Port source: upstream PR #981 (fix(webview): invalidate per-view state after reset and import) — closed draft, superseded by the vps2 series; the #41 ledger names #981 as the F4 port source.
Scope
5 files, 342 insertions, 1 deletion (measured vs stack base 8da5c6e):
Budget
Port fidelity (coordinator-verified)
Structural note on the parallelMode spec (coordinator-verified)
The CS parallelMode.spec.ts is 1790 lines: a shared preamble (L1-672), an F1-series test section (L673-1363: viewId uniqueness, local state isolation, saveViewState, stale temporary-id load), the F3 describes (L1364-1673), a blank separator (L1674), the F4 multi-instance describes (L1675-L1789), and the file's final top-level close (L1790). In this series the F1-series section lives in ClineProvider.spec.ts (shipped by F1a/F1b/F1c — the deliberate F1-series describe restructure), and the F2 unit shipped the persisted-pruning and #1065 retention tests inside the parallelMode file (absent from the CS parallelMode file). Both placements are behaviorally covered; the divergence is structural, not a coverage gap.
Series mechanics