diff --git a/config/reliability-gates.jsonc b/config/reliability-gates.jsonc index e288dab4b67..414c41338ce 100644 --- a/config/reliability-gates.jsonc +++ b/config/reliability-gates.jsonc @@ -4118,7 +4118,7 @@ "providers": ["remote-runtime"], "coveredPlatforms": ["macos"], "coveredProviders": [], - "coverageNotes": "Audit6018 removes the whole HtmlDocPreview.toolbar.test.tsx suite. Its last current case exercises real DOM/registry hide, explicit close, recreation and unmount lifetime against supplied native guest boundaries; those alarms are intentionally lost without equivalent replacement. Remove its six compound-selector tokens, testFiles entry and obsolete focus assertionRef (the latter assertions were already absent from the current single case). No complete active dated receipt selects this source; all remain unchanged and do not validate narrowed commands. No fresh runtime/native/import-closure claim. Complete prior account preserved verbatim: Audit5916 removes the entire local-worktree-creation-part-02.spec import from the retained runtime driver, intentionally losing the mocked exact-name create refusal when a selected PR matches a remote branch. The sole zero-auth terminal-handle registrar for another deleted suite is also retired, with no authored credit. Driver selectors retain only their narrowed import scope. Unchanged dated independent receipts are not equivalent replacement or fresh whole-gate validation. Complete prior account retained verbatim: Audit5910 retires the complete use-doc-preview-guest-tools suite: token-scoped pending-card position after inner scroll, load reset and detached guest listener cleanup alarms are intentionally lost without equivalent replacement. Narrow three compound selectors and remove its testFiles entry. All remaining assertions/selectors and complete dated evidenceRuns retain only their original scope; no fresh validation or replacement claim. Complete prior account retained verbatim: Audit5901 removes only the mobile-creation-and-orchestration-part-03.spec import from the retained runtime driver. The fragment's missing-handle wait abort and waiter retirement on reload-exit alarms are intentionally lost. This is a conservative import-closure notice: named command filters may exclude those cases. Driver selectors remain; unchanged historical receipts retain their original scope and establish no fresh whole-gate pass or equivalent replacement. Complete prior account retained verbatim: Audit5896 removes only the paired-settings.spec and dispatch-mailbox-delivery.spec import lines from their retained runtime/mailbox drivers. Those imported suites and alarms are intentionally retired, without equivalent replacement. Driver selectors remain; unchanged dated receipts retain original scope and do not establish a new whole-gate import-closure pass. Complete prior account retained verbatim: Audit5583 removes the entire src/renderer/src/lib/browser/documents/doc-preview-grants.test.ts suite and its active selectors. Intentionally lose stale preview grant rejection and newer-capability preservation alarms without equivalent replacement or fresh gate evidence. Historical prior account follows: Audit4362 intentionally deletes the imported stale-waiter retry and manually assembled OSC9999 mobile-republish cases, plus the mocked missing-worktree cleanup suite (4 authored / 5 finite cases in three suites). Lost alarms include repeated unread worker_done pointers, working-to-waiting mobile status publication, missing-worktree filtering, and scoped daemon shutdown with unrelated duplicate identities. No replacement coverage or new platform proof is claimed. Audit3994 retires two authored callbacks imported by the shared construction test driver: split OSC 9999 carry and rendererless status ownership, plus OpenCode permission text taking priority in status and tui-idle refusal. Remaining imports and selectors keep their scope. No equivalent replacement or fresh gate execution is claimed. Audit3460 final-v2 composes accepted AUTHOR3454/PEER3455 on actual post263 main: intentionally removes99 authored/106 potential expanded cases across16 selected drivers, including7 whole suites, retaining24 authored/26 potential cases. No equivalent replacement or fresh runtime is claimed. The complete mocked session-close routing suite is removed, including the explicit cross-worktree cleanup alarm. Remove its active file, command token and assertion reference; historical selectors and receipts retain their original scope. Audit3403 composes root-adopted3393/3394,3400/3401 and3398/3399: intentionally retires167 authored registrations/196 potential expanded cases across93 whole suites without equivalent replacement or fresh execution. Retained65 registrations/79 potential cases include one conditional missing-input failure sentinel; potential counts are not executions. Historical notes and complete dated receipts retain their exact original scopes. The retained runtime registration driver loses three complete imported fragments: changed-working-label title facts and suppressed-frame tracked snapshots; refusing SSH GitLab MR-base lookup before local Git/glab calls; and missing folder-workspace path creation/update refusal before spawn/store mutation. These alarms are intentionally lost. Other imported fragments and separately rooted dated receipts are unchanged; they are not equivalent replacement coverage. Audit3342 retires the complete mobile editor/diff tab-publication projection suite: editor/diff serialization, workspace-document privacy withholding, pending/focus and capability/options projections are intentionally lost without equivalent replacement. The surviving focused selectors are unexecuted in this source composition. Six complete compound macOS receipts whose command fields select the changed source are retired verbatim as history; their other selected tests receive no promoted evidence. Other dated receipts keep their original scope; no fresh runtime/UI/mobile/mixed-version or cross-platform validation is claimed. Audit3335 retires the complete three-case browser-tabs snapshot assembly suite: canonical snapshot replay/placement/focus, local host-page adoption and unstamped host echo leaving a local page untouched alarms are intentionally lost without equivalent replacement. The surviving focused selector is unexecuted in this source composition. Unrelated dated receipts keep their original scope; no fresh runtime/network, rendered UI, mobile, mixed-version or cross-platform validation is claimed. Audit 3301 removes the caller-chat-target precedence over host selection example from web-session-tabs-sync-focus-intent.test.ts. The snapshot-before-RPC replay and split-sibling-first exact-leaf focus cases remain unchanged and match the existing assertion reference. Preserve current selectors, assertion references, unrelated active receipts and complete archived history. These mocked state-composition survivors provide no equivalent replacement, rendered, native, network or fresh execution proof. Audit3108: doc-preview-grants.test.ts now retains only the stale grant-rejection race: a late rejected grant request must not evict a newer grant registry entry, witnessed by later revocation of grant-2. Retire stable minted-location and outside-workspace POSIX/Windows request-projection alarms without replacement. The direct nonempty grant selector remains unchanged; no real grant-server/file authorization, SSH/HTML delivery or native-focus result is claimed. Audit2133 current-parent preparation: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2095 current-parent preparation2126: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2090 (sources2088/2089): Current reduced suites intentionally retire the alarms listed in knownGaps; prior dated execution and replacement claims do not validate changed source bytes or selectors. Historical prior account (preserved verbatim): Audit2086 (sources2082/2083): Current reduced suites intentionally retire the alarms listed in knownGaps; prior dated execution and replacement claims do not validate changed source bytes or selectors. Historical prior account (preserved verbatim): Audit2076 (sources2073/2074): Intentionally retire these low-value alarms without replacement: src/main/runtime/orca-runtime-tests/terminal-handles-part-02.spec.ts: recovers a moved SSH pane through the tab it now sits in, not the one its lease froze; reports a plain expired orphan; does not report a pane whose only recent lease was superseded; does not report a pane whose only recent lease had its relay id recycled; picks the eligible orphan over a superseded predecessor that matches first; collects the same tabs the per-tab reader reports, in one sweep of the leases; reports no candidate when no lease names any of the worktree tabs. src/main/runtime/orca-runtime-tests/runtime-availability.spec.ts: starts unavailable with no authoritative window; reports the configured Windows terminal shell on status; reports floating workspace availability from settings on status; polls floating tabs with targeted PTY liveness and no repo/provider inventory; hydrates persisted tabs when the store cannot report repos; advertises browser screencast only when a renderer window is available; routes mobile Codex reset consumption through the account mutation coordinator; maps a definite pre-provider rejection into an authoritative current snapshot. Existing unrelated evidenceRuns retain their original scope; unchanged dated receipt objects do not validate changed source bytes or reduced selectors. No fresh gate execution or historical import-closure equivalence is claimed. Frozen current literal-import graph reaches these changed test bytes; unresolved local and computed edges remain debt, and current reachability is not historical execution or equivalent replacement evidence. Historical prior account (preserved verbatim): Audit2043: Remove the registration-wait, webview-security, preview-authorization, guest-lifecycle and grant-IPC unit suites and the sibling-directory assertion reference. Runtime and browser capability consumers also lose removed imported examples. Remaining document-policy, file-reader and grant-registry assertions do not replace those alarms. Preserve the 27 independent dated receipts at their original unchanged selectors; reduced prospective unit selectors have no fresh result or equivalent replacement claim. Historical prior account (preserved verbatim): Audit1973: Delete the browser RPC forwarding/schema/scalar suite and its prospective selector. Preserve all27 unrelated dated receipts; no actual browser operation alarm is replaced. Whole original changed dated receipts are retired, never narrowed or transferred. This metadata review claims no fresh execution, equivalent replacement, provider/PTY, live remote/mobile, rendered, cross-platform, soak or measured-budget result. Historical prior account (preserved verbatim): Audit1925 (sources1922/1923): The complete terminal-file-link-actions suite is removed. Lose supplied local/remote menu-row inventories, browser label/order/download/reveal callback forwarding, platform-copy and source-group pinning examples. No OS reveal/download, real remote preview or browser navigation was exercised; the other real document-preview sources are not equivalent replacements. Preserve meaningful remaining selectors/assertions and unchanged dated receipts within their original scope. Reduced selectors have no fresh gate execution evidence in this metadata packet. No equivalent replacement, physical watcher/PTY/provider, live SSH/remote/mobile, rendered, cross-platform, soak or measured-budget result is claimed. Historical prior account (preserved verbatim): Audit1899 replays sealed Audit1895: workspace-doc-history.test.ts is entirely removed from its current command and testFiles. Supplied recency/dedupe/cap/title/identity examples and the10000-row getter-read budget alarm are intentionally lost. No current assertionRef or active receipt selects that file. Unrelated active receipts and all prior gap literals remain verbatim; no replacement history/index-work coverage or new browser/paired-preview gate execution is claimed. Historical coverage account before Audit1899: Audit1853 reduces LinkActionPopover.test.tsx and archives its complete 2026-09-06 mixed-selector receipt. The lost copy/open-state and Settings callback-order/omission alarms are accepted without equivalent replacement coverage. No fresh execution of this gate is credited. Audit1709 removes low-value test definitions and retires affected assertion or execution-receipt claims without equivalent replacement coverage. Per-source reviews bound retained behavior and injected/native/remote limits. Unchanged assertions and receipts keep their original scope; gates lacking current active evidence are downgraded. Historical scope before this audit: Audit1687 removes low-value test projections, mock wiring, private-state assertions, inventories and duplicate examples without equivalent replacement coverage. Per-source reviews record retained public behavior and refusal boundaries with injected/native/remote limits. Affected historical receipts are retired; gates without active current evidence are downgraded. Historical scope before this audit: Audit1630 removes low-value Git, GitHub/GitLab and IPC projections, argument inventories and duplicate examples without equivalent replacement coverage. Per-source reviews record retained cancellation, ownership, authorization, filesystem and SQLite boundaries with injected/native/remote limits. Affected historical receipts are retired; gates without active current evidence are downgraded. Historical scope before this audit: Audit1571 removes mock forwarding, capability/schema inventories, private predicates, supplied error/result replays, and filesystem-mocked persistence checks without equivalent replacement claims. Retained changed tests also exercise real local permission-file reread, registry proxy rollback/credential retirement, site display recovery, and pending sender/credential/document admission with injected Electron/UI producers. Retained network tests exercise owned stream credit/cleanup, generation and retirement fences, real loopback proxy/SOCKS bytes and file-backed binding state. SSH/WSL/paired runtime, native sessions, authority producers and storage clears remain injected; simulated relaunch uses filesystem mocks. The unchanged profile-retirement suite was statically reviewed and not rerun. No actual SSH/WSL runtime, native Chromium jar deletion, durable process restart, remote DNS or external site proof. Receipts selecting changed tests are retired and unaffected receipts remain unchanged. Historical scope before this audit: Audit1568 removes API inventories, supplied mock-value/forwarding assertions, implementation-shape tables and negative/setup cases that do not reach their intended boundary, without equivalent replacement claims. Retained changed definitions exercise owned manager lifetime/queue/refcount/clamp and real synthetic process/files/staging boundaries. Seventeen unchanged native Electron definitions across this combined audit are statically reviewed and not rerun. Changed native suites exercise local popup wire/cookie/session and clean/native process-identity behavior; known popup-admission flaws remain unfixed. Unit native/event/session/registry inputs remain injected. No actual OS focus, registry/host execution, CRX verification or durable restart proof. Receipts selecting changed tests are retired and unaffected receipts remain unchanged. Historical scope before this audit: Audit1564 retires copied cookie key recipes, mock bridge/preview forwarding and flag tables, source-spelling/script inventories and setup-only alarms without equivalent replacement coverage. Retained changed cases exercise actual scoped SQLite merge/rollback/schema refusal, owned cancellation and grant authority/canonical containment, ranged response consistency, replacement guest ownership and executed guest-script promise/DOM cleanup behavior. Session, host response, geometry/observer and transport producers remain injected where documented. One unchanged native extension definition is not rerun locally; no actual remote filesystem, native layout/focus or authentication claim. Receipts selecting modified tests are retired; unaffected exact-selector receipts remain unchanged. Historical scope before this audit: Audit1412 changes referenced sources. Retained unchanged-source receipts apply only to their unchanged commands; the reduced selectors have no new execution result. Historical coverage account (not a current claim for changed sources): Audit1386 changes referenced sources; no new execution result validates reduced selectors. Retained unchanged receipts apply only to unchanged commands. Historical coverage: Audit1308 changes referenced sources. Retained unchanged-source receipts apply only to their unchanged commands; the reduced selectors have no new execution result. Historical coverage account (not a current claim for changed sources): A macOS desktop host and a separate paired Electron client exercise the Explorer preview action end to end: the document renders inside the client-local orca-preview guest, the client gains exactly one browser workspace and it is the document one, the host's own page registry and tab snapshot gain nothing, the split does not take focus from the source editor, an explicit click activates the preview, and closing it with its X leaves the editor and terminal intact. Both host oracles are read in a run where an ordinary URL browser tab from the same client does arrive in them, because a host that cannot project browser pages at all would otherwise answer the absence question the same way. A second journey quits that client with a document tab open and relaunches it on the same profile, which is the restore the user's original report turned on: the same workspace row comes back blank and named by the document, renders again, and does so over a grant URL that differs from the one that was quit. The same journey pins the guest URL to the orca-preview scheme, so a regression back to file:// or to a host browser page is visible. The rewritten journey caught a live gap the old one could not see: since previews became browser pages they had no chrome to hand focus to their guest, so main refused every reported link click as coming from an unfocused guest and the one route out of a preview was dead. Main's own log answered it — registered true, grant live, focused false — after the embedder reported the anchor really taking six real presses. The pane now hands focus over when it is the reader's surface, and the journey routes. The preview's own components moved out of the editor folder into browser-pane/workspace-doc/ with this change, since a preview is a browser pane and no longer an editor one; their i18n keys keep the historical auto.components.editor.HtmlDocPreview prefix rather than dropping thirty translated strings across four locales for a tidier key path. Seven gate commands named those files by their old paths and were re-issued at the new ones, and the nine dated runs behind them were retired rather than left naming files that are no longer there, on the same principle this gate already applied twice: vitest treats a missing named file as a filter and exits 0, so such a command reports green while silently running only the half that still exists. Each retired record's finding is carried into the replacement run recorded below it, including the red half of the entry/subresource pair, which was re-measured by re-applying its mutant at the new path rather than being dropped. The fence over the page-management channels was decided channel by channel before it was built, and that decision record is: of the 23 channels that can name a page, 6 are allowed for a workspace-document page — the tool channels the reader drives, which reach it through the one authority that reads both halves of the registry — 9 refuse (download accept and cancel, devtools, certificate proceed, client page metadata publish, viewport override, WebAuthn account response, registration probe, registration repair), 3 cannot name a page at all because they take a profile or a worktree instead, and the rest are the registration and activation doors, refused on membership. What shipped is not the per-channel guard that classification implies: eleven call sites read the browsing map directly, so a guard per channel is exactly the caller-flag seam that has to be remembered at every future call site. The registry was split in two instead, which makes the classification structural — a document guest is absent from the browsing map, so every refusing channel refuses by absence rather than by a guard, and a channel added later inherits the refusal without being classified at all. Unit contracts cover grant scoping and path containment, per-owner read routing, refusal of a truncated or over-cap read, the old-server binary degradation, scheme admission on will-attach-webview, the guest navigation policy, the plan layer's local/SSH/paired/out-of-worktree routing, and the distinct sentence the reader sees for each failure reason. The paired browser-creation surfaces this gate still owns — capability rejection cleanup, canonical page identity, reconciliation rollback, close convergence — keep their existing headed and headless journeys. Linux and Windows live journeys remain gaps. A refused subresource is a notice on a rendered document rather than a failure panel, since only the entry document's failure leaves the reader with nothing; the entry document is served as text by every owner and cannot be refused for its format at all. Grants die with the renderer document that minted them, at window creation and on every reload, because the renderer holds the only record of which preview owns which grant. The mint and revoke channels answer only the trusted renderer, and a preview tab is never republished to mobile clients, which have no grant to read it with. External URLs leave the preview only via a trusted user click on an anchor while the preview is focused; script-driven navigation and popups reach nothing. The guest's navigation sinks are deny-only and route no URL at all, and the guest carries a preview-only preload, pinned by will-attach-webview and denied to every other attach path, whose capture-phase click handler reports an anchor's URL to main; main answers only a report from a preview guest with a live bound grant that is the focused contents, and only for http(s). Two earlier reader-facing records from 2026-08-26, which measured a shape where a subresource failure was filtered out and the panel still carried an unsupported-asset sentence, were retired with that shape rather than left contradicting the 10/10 runs of the command they share. Two more were retired on the same principle when the external-link design changed: a record showing that removing a recent-input gate let the unattended document open two browser tabs, and a record showing that dropping the webview's allowpopups attribute left a target=_blank press with nothing to route. Both measured machinery that no longer exists — the input gate could not tell a script's navigation from the click it followed, which is the case now recorded below, and the preload intercepts the press before Chromium considers a popup at all, so allowpopups was removed. WebRTC is fenced by the guest's IP-handling policy alone: this Chromium answers a webrtc content-security-policy directive as unrecognized and gathers candidates anyway, so the directive is deliberately absent from the served policy and a unit assertion keeps it out. A standalone Electron probe on the same version gathered 7 candidates including a STUN reflexive one without the policy and 0 with it. Two command paths this gate used to carry no longer exist, and their structured evidence was retired with them rather than left naming files that are no longer there: vitest treats a missing named file as a filter and exits 0, so such a command reports green while silently running only the half that still exists. The first is src/renderer/src/lib/file-preview-capability-cleanup.test.ts, deleted in STA-5557 when HTML previews stopped creating paired browser tabs: on 2026-08-12 its five-assertion STA-4063 oracle ran red 5/5 on scan e4c278eb52 and on main bc2e30000b, partially green 2/5 on pre-fix PR #14100, and green 5/5 on candidate 14b5877a75; the surviving half of that oracle is web-runtime-browser-capability-cleanup.test.ts, still commanded above. The second is tests/e2e/paired-remote-html-browser-focus.spec.ts, renamed in STA-5557 to paired-remote-html-preview-local-render.spec.ts because the journey no longer proves anything about browser focus: its 2026-08-12 runs established the owning-runtime routing and no-focus-steal behavior of the then-current browser-tab preview (red on main 346e59c879 and a90a18d43d, green on candidates d05217c40d and b7de9c0961, red again with only the side-preview focus or routing seam disabled), and its 2026-08-26 runs are recorded below under the current path. An adversarial pass after the journey landed found four coverage gaps rather than shipping defects, and all four are pinned here. The page url had a third door: the navigation commit wrote whatever url was reported, unfenced, and while nothing reaches it for a document page today it sat one report away from giving a document page a URL that persistence, the publish boundary and the address bar all read. The group projection was the load-bearing half of the mobile publish boundary that no test had: removing it alone left the whole suite green, because no test in it had tab groups, while the phone payload gained a group whose order and active tab named a document tab it was never sent. The focus handoff was covered only in the narrowing direction, so widening it — dropping the page-activity check, or either half of the reader-surface check — survived both suites. And the handoff itself was offered once and never again, so another app taking the front left the preview's one route out shut until the pane remounted, which is a desktop gap rather than a test-only one. The standalone capability-cleanup and browser-placement ledger suites were removed, including their mixed-version rejection, cleanup transfer, capacity and partition alarms. web-runtime-session.test.ts no longer supplies browser creation, rollback or recovery assertions. The grouped workspace-document exclusion remains; the separate no-group publication example was removed. Existing live/browser-store evidence does not establish equivalent replacement alarms. Earlier descriptions of the removed fixtures and their run results are historical. The client-owned-placement and layout-groups suites are removed. The broad mirror fixture no longer supplies preview placement/reservation/move assertions. Browser rows retain pre-response replay and real local-workspace reuse; focus intent retains response ordering and exact-leaf sibling ordering. Supplied browser-layout and placement/focus projection alarms are lost; retained evidence is narrower and not equivalent. The test-value audit retires renderer grant-release mock ledgers, direct creation/title/session-field copies, notification/onboarding calls and several placement/per-page projections. Canonical-ID rejection, document reopen/restore and preview URL fences remain; they do not prove close-to-registry revocation or replace every lost alarm. Historical execution evidence is unchanged. The test-value audit removed the direct legacy editor-preview hydration example from tabs-hydration.test.ts. Dropping legacy preview chrome while preserving the same-document ordinary editor no longer has that alarm. The surviving corrupt-id deduplication case does not cover that migration, and retained preview journeys are not equivalent replacement coverage. Historical run counts and migration mutation findings describe the earlier suite. Audit1308: src/renderer/src/components/editor/EditorPanelShell.header.test.tsx removed. Header visibility and mocked editor-surface inventory alarms disappear. No equivalent replacement coverage is claimed. Audit1312 removes low-value imported specs from aggregate test drivers; old whole-driver execution receipts are archived. Retained registrations are unchanged. No equivalent replacement coverage is claimed. Audit1383: src/main/runtime/orca-runtime-tests/worktree-native-chat-setup.spec.ts reduced. Lose command-list/receipt/default/background activation matrices whose provider spawn returns canned IDs. Retained test proves actual tab allocation separation, not terminal startup. No equivalent replacement coverage is claimed. Existing exact unaffected receipts and their platform/provider/budget claims remain scoped to unchanged selectors. Changed optional selectors have no new execution result. Audit1386: src/renderer/src/lib/doc-preview-grants.test.ts reduced. Repeated-mount mint count, ordinary release/remint and failed-mint cache retry standalone alarms removed; no actual grant server/access test. No equivalent replacement coverage is claimed. Audit1401: src/main/runtime/orca-runtime-tests/mobile-session-tabs-duplicate-pty-repair.spec.ts removed. Loses listing-on-already-repaired-data alarm; it never established runtime repair invocation. src/main/runtime/orca-runtime-tests/mobile-session-tabs-split-agent-tab.spec.ts removed. Loses actual mobile split sibling shell versus recorded-agent resume command alarms; no equivalent coverage asserted. src/main/runtime/orca-runtime-tests/worktree-native-chat-setup.spec.ts removed. Loses separate setup/startup tab result in that manually assembled path. src/main/runtime/orca-runtime.test.ts reduced. Driver registration/import of a whole-deleted selected spec is removed; directly authored cases remain unchanged. No extra authored removal credit. No equivalent replacement coverage is claimed. The existing primary-selector receipt and its claims remain scoped to their unchanged sources; the broad optional command has no new execution result. Audit1412: src/main/browser/guest-lifecycle/manager/browser-manager-guest-policy-profile.test.ts removed. Loses workspace-document in-grant navigation/popup refusal, cleanup/re-attach, cross-registry guest authorization and browsing-map separation alarms. No equivalent browser policy coverage is claimed. No equivalent replacement coverage is claimed. Audit1421 removes browser-address-bar-suggestions.test.ts. Document suggestion title/path matching, recency merging and path-prefix ranking alarms are lost. No equivalent replacement coverage is claimed. Historical receipts remain scoped to unchanged selectors; the reduced command has no new execution result. Audit2452 removes the whole fake-effect browser-close routing suite. Active selectors and assertion references no longer select it. Its requested-worktree default close, sibling-page close and explicit missing-page typed reply alarms are intentionally retired without equivalent replacement. Prior dated receipts remain unchanged and scoped to their original sources; no fresh execution of the reduced gate, rendered browser, live remote, mobile or cross-platform validation is claimed.", + "coverageNotes": "Audit6088 removes the whole LinkActionPopover.test.tsx suite. Benign duplicate-copy suppression, toast feedback and retry-admission alarms are intentionally lost without equivalent replacement. Remove its standalone command and testFiles entry; all dated receipts and historical records retain their original scope. No native clipboard, fresh runtime or whole-gate validation is claimed. Complete prior account preserved verbatim: Audit6018 removes the whole HtmlDocPreview.toolbar.test.tsx suite. Its last current case exercises real DOM/registry hide, explicit close, recreation and unmount lifetime against supplied native guest boundaries; those alarms are intentionally lost without equivalent replacement. Remove its six compound-selector tokens, testFiles entry and obsolete focus assertionRef (the latter assertions were already absent from the current single case). No complete active dated receipt selects this source; all remain unchanged and do not validate narrowed commands. No fresh runtime/native/import-closure claim. Complete prior account preserved verbatim: Audit5916 removes the entire local-worktree-creation-part-02.spec import from the retained runtime driver, intentionally losing the mocked exact-name create refusal when a selected PR matches a remote branch. The sole zero-auth terminal-handle registrar for another deleted suite is also retired, with no authored credit. Driver selectors retain only their narrowed import scope. Unchanged dated independent receipts are not equivalent replacement or fresh whole-gate validation. Complete prior account retained verbatim: Audit5910 retires the complete use-doc-preview-guest-tools suite: token-scoped pending-card position after inner scroll, load reset and detached guest listener cleanup alarms are intentionally lost without equivalent replacement. Narrow three compound selectors and remove its testFiles entry. All remaining assertions/selectors and complete dated evidenceRuns retain only their original scope; no fresh validation or replacement claim. Complete prior account retained verbatim: Audit5901 removes only the mobile-creation-and-orchestration-part-03.spec import from the retained runtime driver. The fragment's missing-handle wait abort and waiter retirement on reload-exit alarms are intentionally lost. This is a conservative import-closure notice: named command filters may exclude those cases. Driver selectors remain; unchanged historical receipts retain their original scope and establish no fresh whole-gate pass or equivalent replacement. Complete prior account retained verbatim: Audit5896 removes only the paired-settings.spec and dispatch-mailbox-delivery.spec import lines from their retained runtime/mailbox drivers. Those imported suites and alarms are intentionally retired, without equivalent replacement. Driver selectors remain; unchanged dated receipts retain original scope and do not establish a new whole-gate import-closure pass. Complete prior account retained verbatim: Audit5583 removes the entire src/renderer/src/lib/browser/documents/doc-preview-grants.test.ts suite and its active selectors. Intentionally lose stale preview grant rejection and newer-capability preservation alarms without equivalent replacement or fresh gate evidence. Historical prior account follows: Audit4362 intentionally deletes the imported stale-waiter retry and manually assembled OSC9999 mobile-republish cases, plus the mocked missing-worktree cleanup suite (4 authored / 5 finite cases in three suites). Lost alarms include repeated unread worker_done pointers, working-to-waiting mobile status publication, missing-worktree filtering, and scoped daemon shutdown with unrelated duplicate identities. No replacement coverage or new platform proof is claimed. Audit3994 retires two authored callbacks imported by the shared construction test driver: split OSC 9999 carry and rendererless status ownership, plus OpenCode permission text taking priority in status and tui-idle refusal. Remaining imports and selectors keep their scope. No equivalent replacement or fresh gate execution is claimed. Audit3460 final-v2 composes accepted AUTHOR3454/PEER3455 on actual post263 main: intentionally removes99 authored/106 potential expanded cases across16 selected drivers, including7 whole suites, retaining24 authored/26 potential cases. No equivalent replacement or fresh runtime is claimed. The complete mocked session-close routing suite is removed, including the explicit cross-worktree cleanup alarm. Remove its active file, command token and assertion reference; historical selectors and receipts retain their original scope. Audit3403 composes root-adopted3393/3394,3400/3401 and3398/3399: intentionally retires167 authored registrations/196 potential expanded cases across93 whole suites without equivalent replacement or fresh execution. Retained65 registrations/79 potential cases include one conditional missing-input failure sentinel; potential counts are not executions. Historical notes and complete dated receipts retain their exact original scopes. The retained runtime registration driver loses three complete imported fragments: changed-working-label title facts and suppressed-frame tracked snapshots; refusing SSH GitLab MR-base lookup before local Git/glab calls; and missing folder-workspace path creation/update refusal before spawn/store mutation. These alarms are intentionally lost. Other imported fragments and separately rooted dated receipts are unchanged; they are not equivalent replacement coverage. Audit3342 retires the complete mobile editor/diff tab-publication projection suite: editor/diff serialization, workspace-document privacy withholding, pending/focus and capability/options projections are intentionally lost without equivalent replacement. The surviving focused selectors are unexecuted in this source composition. Six complete compound macOS receipts whose command fields select the changed source are retired verbatim as history; their other selected tests receive no promoted evidence. Other dated receipts keep their original scope; no fresh runtime/UI/mobile/mixed-version or cross-platform validation is claimed. Audit3335 retires the complete three-case browser-tabs snapshot assembly suite: canonical snapshot replay/placement/focus, local host-page adoption and unstamped host echo leaving a local page untouched alarms are intentionally lost without equivalent replacement. The surviving focused selector is unexecuted in this source composition. Unrelated dated receipts keep their original scope; no fresh runtime/network, rendered UI, mobile, mixed-version or cross-platform validation is claimed. Audit 3301 removes the caller-chat-target precedence over host selection example from web-session-tabs-sync-focus-intent.test.ts. The snapshot-before-RPC replay and split-sibling-first exact-leaf focus cases remain unchanged and match the existing assertion reference. Preserve current selectors, assertion references, unrelated active receipts and complete archived history. These mocked state-composition survivors provide no equivalent replacement, rendered, native, network or fresh execution proof. Audit3108: doc-preview-grants.test.ts now retains only the stale grant-rejection race: a late rejected grant request must not evict a newer grant registry entry, witnessed by later revocation of grant-2. Retire stable minted-location and outside-workspace POSIX/Windows request-projection alarms without replacement. The direct nonempty grant selector remains unchanged; no real grant-server/file authorization, SSH/HTML delivery or native-focus result is claimed. Audit2133 current-parent preparation: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2095 current-parent preparation2126: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2090 (sources2088/2089): Current reduced suites intentionally retire the alarms listed in knownGaps; prior dated execution and replacement claims do not validate changed source bytes or selectors. Historical prior account (preserved verbatim): Audit2086 (sources2082/2083): Current reduced suites intentionally retire the alarms listed in knownGaps; prior dated execution and replacement claims do not validate changed source bytes or selectors. Historical prior account (preserved verbatim): Audit2076 (sources2073/2074): Intentionally retire these low-value alarms without replacement: src/main/runtime/orca-runtime-tests/terminal-handles-part-02.spec.ts: recovers a moved SSH pane through the tab it now sits in, not the one its lease froze; reports a plain expired orphan; does not report a pane whose only recent lease was superseded; does not report a pane whose only recent lease had its relay id recycled; picks the eligible orphan over a superseded predecessor that matches first; collects the same tabs the per-tab reader reports, in one sweep of the leases; reports no candidate when no lease names any of the worktree tabs. src/main/runtime/orca-runtime-tests/runtime-availability.spec.ts: starts unavailable with no authoritative window; reports the configured Windows terminal shell on status; reports floating workspace availability from settings on status; polls floating tabs with targeted PTY liveness and no repo/provider inventory; hydrates persisted tabs when the store cannot report repos; advertises browser screencast only when a renderer window is available; routes mobile Codex reset consumption through the account mutation coordinator; maps a definite pre-provider rejection into an authoritative current snapshot. Existing unrelated evidenceRuns retain their original scope; unchanged dated receipt objects do not validate changed source bytes or reduced selectors. No fresh gate execution or historical import-closure equivalence is claimed. Frozen current literal-import graph reaches these changed test bytes; unresolved local and computed edges remain debt, and current reachability is not historical execution or equivalent replacement evidence. Historical prior account (preserved verbatim): Audit2043: Remove the registration-wait, webview-security, preview-authorization, guest-lifecycle and grant-IPC unit suites and the sibling-directory assertion reference. Runtime and browser capability consumers also lose removed imported examples. Remaining document-policy, file-reader and grant-registry assertions do not replace those alarms. Preserve the 27 independent dated receipts at their original unchanged selectors; reduced prospective unit selectors have no fresh result or equivalent replacement claim. Historical prior account (preserved verbatim): Audit1973: Delete the browser RPC forwarding/schema/scalar suite and its prospective selector. Preserve all27 unrelated dated receipts; no actual browser operation alarm is replaced. Whole original changed dated receipts are retired, never narrowed or transferred. This metadata review claims no fresh execution, equivalent replacement, provider/PTY, live remote/mobile, rendered, cross-platform, soak or measured-budget result. Historical prior account (preserved verbatim): Audit1925 (sources1922/1923): The complete terminal-file-link-actions suite is removed. Lose supplied local/remote menu-row inventories, browser label/order/download/reveal callback forwarding, platform-copy and source-group pinning examples. No OS reveal/download, real remote preview or browser navigation was exercised; the other real document-preview sources are not equivalent replacements. Preserve meaningful remaining selectors/assertions and unchanged dated receipts within their original scope. Reduced selectors have no fresh gate execution evidence in this metadata packet. No equivalent replacement, physical watcher/PTY/provider, live SSH/remote/mobile, rendered, cross-platform, soak or measured-budget result is claimed. Historical prior account (preserved verbatim): Audit1899 replays sealed Audit1895: workspace-doc-history.test.ts is entirely removed from its current command and testFiles. Supplied recency/dedupe/cap/title/identity examples and the10000-row getter-read budget alarm are intentionally lost. No current assertionRef or active receipt selects that file. Unrelated active receipts and all prior gap literals remain verbatim; no replacement history/index-work coverage or new browser/paired-preview gate execution is claimed. Historical coverage account before Audit1899: Audit1853 reduces LinkActionPopover.test.tsx and archives its complete 2026-09-06 mixed-selector receipt. The lost copy/open-state and Settings callback-order/omission alarms are accepted without equivalent replacement coverage. No fresh execution of this gate is credited. Audit1709 removes low-value test definitions and retires affected assertion or execution-receipt claims without equivalent replacement coverage. Per-source reviews bound retained behavior and injected/native/remote limits. Unchanged assertions and receipts keep their original scope; gates lacking current active evidence are downgraded. Historical scope before this audit: Audit1687 removes low-value test projections, mock wiring, private-state assertions, inventories and duplicate examples without equivalent replacement coverage. Per-source reviews record retained public behavior and refusal boundaries with injected/native/remote limits. Affected historical receipts are retired; gates without active current evidence are downgraded. Historical scope before this audit: Audit1630 removes low-value Git, GitHub/GitLab and IPC projections, argument inventories and duplicate examples without equivalent replacement coverage. Per-source reviews record retained cancellation, ownership, authorization, filesystem and SQLite boundaries with injected/native/remote limits. Affected historical receipts are retired; gates without active current evidence are downgraded. Historical scope before this audit: Audit1571 removes mock forwarding, capability/schema inventories, private predicates, supplied error/result replays, and filesystem-mocked persistence checks without equivalent replacement claims. Retained changed tests also exercise real local permission-file reread, registry proxy rollback/credential retirement, site display recovery, and pending sender/credential/document admission with injected Electron/UI producers. Retained network tests exercise owned stream credit/cleanup, generation and retirement fences, real loopback proxy/SOCKS bytes and file-backed binding state. SSH/WSL/paired runtime, native sessions, authority producers and storage clears remain injected; simulated relaunch uses filesystem mocks. The unchanged profile-retirement suite was statically reviewed and not rerun. No actual SSH/WSL runtime, native Chromium jar deletion, durable process restart, remote DNS or external site proof. Receipts selecting changed tests are retired and unaffected receipts remain unchanged. Historical scope before this audit: Audit1568 removes API inventories, supplied mock-value/forwarding assertions, implementation-shape tables and negative/setup cases that do not reach their intended boundary, without equivalent replacement claims. Retained changed definitions exercise owned manager lifetime/queue/refcount/clamp and real synthetic process/files/staging boundaries. Seventeen unchanged native Electron definitions across this combined audit are statically reviewed and not rerun. Changed native suites exercise local popup wire/cookie/session and clean/native process-identity behavior; known popup-admission flaws remain unfixed. Unit native/event/session/registry inputs remain injected. No actual OS focus, registry/host execution, CRX verification or durable restart proof. Receipts selecting changed tests are retired and unaffected receipts remain unchanged. Historical scope before this audit: Audit1564 retires copied cookie key recipes, mock bridge/preview forwarding and flag tables, source-spelling/script inventories and setup-only alarms without equivalent replacement coverage. Retained changed cases exercise actual scoped SQLite merge/rollback/schema refusal, owned cancellation and grant authority/canonical containment, ranged response consistency, replacement guest ownership and executed guest-script promise/DOM cleanup behavior. Session, host response, geometry/observer and transport producers remain injected where documented. One unchanged native extension definition is not rerun locally; no actual remote filesystem, native layout/focus or authentication claim. Receipts selecting modified tests are retired; unaffected exact-selector receipts remain unchanged. Historical scope before this audit: Audit1412 changes referenced sources. Retained unchanged-source receipts apply only to their unchanged commands; the reduced selectors have no new execution result. Historical coverage account (not a current claim for changed sources): Audit1386 changes referenced sources; no new execution result validates reduced selectors. Retained unchanged receipts apply only to unchanged commands. Historical coverage: Audit1308 changes referenced sources. Retained unchanged-source receipts apply only to their unchanged commands; the reduced selectors have no new execution result. Historical coverage account (not a current claim for changed sources): A macOS desktop host and a separate paired Electron client exercise the Explorer preview action end to end: the document renders inside the client-local orca-preview guest, the client gains exactly one browser workspace and it is the document one, the host's own page registry and tab snapshot gain nothing, the split does not take focus from the source editor, an explicit click activates the preview, and closing it with its X leaves the editor and terminal intact. Both host oracles are read in a run where an ordinary URL browser tab from the same client does arrive in them, because a host that cannot project browser pages at all would otherwise answer the absence question the same way. A second journey quits that client with a document tab open and relaunches it on the same profile, which is the restore the user's original report turned on: the same workspace row comes back blank and named by the document, renders again, and does so over a grant URL that differs from the one that was quit. The same journey pins the guest URL to the orca-preview scheme, so a regression back to file:// or to a host browser page is visible. The rewritten journey caught a live gap the old one could not see: since previews became browser pages they had no chrome to hand focus to their guest, so main refused every reported link click as coming from an unfocused guest and the one route out of a preview was dead. Main's own log answered it — registered true, grant live, focused false — after the embedder reported the anchor really taking six real presses. The pane now hands focus over when it is the reader's surface, and the journey routes. The preview's own components moved out of the editor folder into browser-pane/workspace-doc/ with this change, since a preview is a browser pane and no longer an editor one; their i18n keys keep the historical auto.components.editor.HtmlDocPreview prefix rather than dropping thirty translated strings across four locales for a tidier key path. Seven gate commands named those files by their old paths and were re-issued at the new ones, and the nine dated runs behind them were retired rather than left naming files that are no longer there, on the same principle this gate already applied twice: vitest treats a missing named file as a filter and exits 0, so such a command reports green while silently running only the half that still exists. Each retired record's finding is carried into the replacement run recorded below it, including the red half of the entry/subresource pair, which was re-measured by re-applying its mutant at the new path rather than being dropped. The fence over the page-management channels was decided channel by channel before it was built, and that decision record is: of the 23 channels that can name a page, 6 are allowed for a workspace-document page — the tool channels the reader drives, which reach it through the one authority that reads both halves of the registry — 9 refuse (download accept and cancel, devtools, certificate proceed, client page metadata publish, viewport override, WebAuthn account response, registration probe, registration repair), 3 cannot name a page at all because they take a profile or a worktree instead, and the rest are the registration and activation doors, refused on membership. What shipped is not the per-channel guard that classification implies: eleven call sites read the browsing map directly, so a guard per channel is exactly the caller-flag seam that has to be remembered at every future call site. The registry was split in two instead, which makes the classification structural — a document guest is absent from the browsing map, so every refusing channel refuses by absence rather than by a guard, and a channel added later inherits the refusal without being classified at all. Unit contracts cover grant scoping and path containment, per-owner read routing, refusal of a truncated or over-cap read, the old-server binary degradation, scheme admission on will-attach-webview, the guest navigation policy, the plan layer's local/SSH/paired/out-of-worktree routing, and the distinct sentence the reader sees for each failure reason. The paired browser-creation surfaces this gate still owns — capability rejection cleanup, canonical page identity, reconciliation rollback, close convergence — keep their existing headed and headless journeys. Linux and Windows live journeys remain gaps. A refused subresource is a notice on a rendered document rather than a failure panel, since only the entry document's failure leaves the reader with nothing; the entry document is served as text by every owner and cannot be refused for its format at all. Grants die with the renderer document that minted them, at window creation and on every reload, because the renderer holds the only record of which preview owns which grant. The mint and revoke channels answer only the trusted renderer, and a preview tab is never republished to mobile clients, which have no grant to read it with. External URLs leave the preview only via a trusted user click on an anchor while the preview is focused; script-driven navigation and popups reach nothing. The guest's navigation sinks are deny-only and route no URL at all, and the guest carries a preview-only preload, pinned by will-attach-webview and denied to every other attach path, whose capture-phase click handler reports an anchor's URL to main; main answers only a report from a preview guest with a live bound grant that is the focused contents, and only for http(s). Two earlier reader-facing records from 2026-08-26, which measured a shape where a subresource failure was filtered out and the panel still carried an unsupported-asset sentence, were retired with that shape rather than left contradicting the 10/10 runs of the command they share. Two more were retired on the same principle when the external-link design changed: a record showing that removing a recent-input gate let the unattended document open two browser tabs, and a record showing that dropping the webview's allowpopups attribute left a target=_blank press with nothing to route. Both measured machinery that no longer exists — the input gate could not tell a script's navigation from the click it followed, which is the case now recorded below, and the preload intercepts the press before Chromium considers a popup at all, so allowpopups was removed. WebRTC is fenced by the guest's IP-handling policy alone: this Chromium answers a webrtc content-security-policy directive as unrecognized and gathers candidates anyway, so the directive is deliberately absent from the served policy and a unit assertion keeps it out. A standalone Electron probe on the same version gathered 7 candidates including a STUN reflexive one without the policy and 0 with it. Two command paths this gate used to carry no longer exist, and their structured evidence was retired with them rather than left naming files that are no longer there: vitest treats a missing named file as a filter and exits 0, so such a command reports green while silently running only the half that still exists. The first is src/renderer/src/lib/file-preview-capability-cleanup.test.ts, deleted in STA-5557 when HTML previews stopped creating paired browser tabs: on 2026-08-12 its five-assertion STA-4063 oracle ran red 5/5 on scan e4c278eb52 and on main bc2e30000b, partially green 2/5 on pre-fix PR #14100, and green 5/5 on candidate 14b5877a75; the surviving half of that oracle is web-runtime-browser-capability-cleanup.test.ts, still commanded above. The second is tests/e2e/paired-remote-html-browser-focus.spec.ts, renamed in STA-5557 to paired-remote-html-preview-local-render.spec.ts because the journey no longer proves anything about browser focus: its 2026-08-12 runs established the owning-runtime routing and no-focus-steal behavior of the then-current browser-tab preview (red on main 346e59c879 and a90a18d43d, green on candidates d05217c40d and b7de9c0961, red again with only the side-preview focus or routing seam disabled), and its 2026-08-26 runs are recorded below under the current path. An adversarial pass after the journey landed found four coverage gaps rather than shipping defects, and all four are pinned here. The page url had a third door: the navigation commit wrote whatever url was reported, unfenced, and while nothing reaches it for a document page today it sat one report away from giving a document page a URL that persistence, the publish boundary and the address bar all read. The group projection was the load-bearing half of the mobile publish boundary that no test had: removing it alone left the whole suite green, because no test in it had tab groups, while the phone payload gained a group whose order and active tab named a document tab it was never sent. The focus handoff was covered only in the narrowing direction, so widening it — dropping the page-activity check, or either half of the reader-surface check — survived both suites. And the handoff itself was offered once and never again, so another app taking the front left the preview's one route out shut until the pane remounted, which is a desktop gap rather than a test-only one. The standalone capability-cleanup and browser-placement ledger suites were removed, including their mixed-version rejection, cleanup transfer, capacity and partition alarms. web-runtime-session.test.ts no longer supplies browser creation, rollback or recovery assertions. The grouped workspace-document exclusion remains; the separate no-group publication example was removed. Existing live/browser-store evidence does not establish equivalent replacement alarms. Earlier descriptions of the removed fixtures and their run results are historical. The client-owned-placement and layout-groups suites are removed. The broad mirror fixture no longer supplies preview placement/reservation/move assertions. Browser rows retain pre-response replay and real local-workspace reuse; focus intent retains response ordering and exact-leaf sibling ordering. Supplied browser-layout and placement/focus projection alarms are lost; retained evidence is narrower and not equivalent. The test-value audit retires renderer grant-release mock ledgers, direct creation/title/session-field copies, notification/onboarding calls and several placement/per-page projections. Canonical-ID rejection, document reopen/restore and preview URL fences remain; they do not prove close-to-registry revocation or replace every lost alarm. Historical execution evidence is unchanged. The test-value audit removed the direct legacy editor-preview hydration example from tabs-hydration.test.ts. Dropping legacy preview chrome while preserving the same-document ordinary editor no longer has that alarm. The surviving corrupt-id deduplication case does not cover that migration, and retained preview journeys are not equivalent replacement coverage. Historical run counts and migration mutation findings describe the earlier suite. Audit1308: src/renderer/src/components/editor/EditorPanelShell.header.test.tsx removed. Header visibility and mocked editor-surface inventory alarms disappear. No equivalent replacement coverage is claimed. Audit1312 removes low-value imported specs from aggregate test drivers; old whole-driver execution receipts are archived. Retained registrations are unchanged. No equivalent replacement coverage is claimed. Audit1383: src/main/runtime/orca-runtime-tests/worktree-native-chat-setup.spec.ts reduced. Lose command-list/receipt/default/background activation matrices whose provider spawn returns canned IDs. Retained test proves actual tab allocation separation, not terminal startup. No equivalent replacement coverage is claimed. Existing exact unaffected receipts and their platform/provider/budget claims remain scoped to unchanged selectors. Changed optional selectors have no new execution result. Audit1386: src/renderer/src/lib/doc-preview-grants.test.ts reduced. Repeated-mount mint count, ordinary release/remint and failed-mint cache retry standalone alarms removed; no actual grant server/access test. No equivalent replacement coverage is claimed. Audit1401: src/main/runtime/orca-runtime-tests/mobile-session-tabs-duplicate-pty-repair.spec.ts removed. Loses listing-on-already-repaired-data alarm; it never established runtime repair invocation. src/main/runtime/orca-runtime-tests/mobile-session-tabs-split-agent-tab.spec.ts removed. Loses actual mobile split sibling shell versus recorded-agent resume command alarms; no equivalent coverage asserted. src/main/runtime/orca-runtime-tests/worktree-native-chat-setup.spec.ts removed. Loses separate setup/startup tab result in that manually assembled path. src/main/runtime/orca-runtime.test.ts reduced. Driver registration/import of a whole-deleted selected spec is removed; directly authored cases remain unchanged. No extra authored removal credit. No equivalent replacement coverage is claimed. The existing primary-selector receipt and its claims remain scoped to their unchanged sources; the broad optional command has no new execution result. Audit1412: src/main/browser/guest-lifecycle/manager/browser-manager-guest-policy-profile.test.ts removed. Loses workspace-document in-grant navigation/popup refusal, cleanup/re-attach, cross-registry guest authorization and browsing-map separation alarms. No equivalent browser policy coverage is claimed. No equivalent replacement coverage is claimed. Audit1421 removes browser-address-bar-suggestions.test.ts. Document suggestion title/path matching, recency merging and path-prefix ranking alarms are lost. No equivalent replacement coverage is claimed. Historical receipts remain scoped to unchanged selectors; the reduced command has no new execution result. Audit2452 removes the whole fake-effect browser-close routing suite. Active selectors and assertion references no longer select it. Its requested-worktree default close, sibling-page close and explicit missing-page typed reply alarms are intentionally retired without equivalent replacement. Prior dated receipts remain unchanged and scoped to their original sources; no fresh execution of the reduced gate, rendered browser, live remote, mobile or cross-platform validation is claimed.", "motivatingLinks": [ "https://linear.app/stably/issue/STA-5557/remote-html-preview-render-workspace-files-locally-over-the-file", "https://linear.app/stably/issue/STA-3872/unable-to-open-html-files-in-remote-server", @@ -4142,7 +4142,6 @@ "pnpm exec vitest run --config config/vitest.config.ts src/main/browser/document-preview/doc-preview-guest-policy.test.ts", "pnpm exec vitest run --config config/vitest.config.ts src/main/browser/guest-lifecycle/manager/viewport/browser-manager-annotation-bridge.test.ts src/shared/document-preview/doc-preview-scheme.test.ts", "pnpm exec vitest run --config config/vitest.config.ts src/main/browser/guest-lifecycle/manager/viewport/browser-manager-annotation-bridge.test.ts src/main/browser/backends/offscreen-browser-backend-lifecycle.test.ts src/main/browser/document-preview/doc-preview-guest-policy.test.ts src/shared/document-preview/doc-preview-scheme.test.ts", - "pnpm exec vitest run --config config/vitest.config.ts src/renderer/src/components/link-actions/LinkActionPopover.test.tsx", "pnpm exec vitest run --config config/vitest.config.ts src/main/browser/guest-lifecycle/manager/viewport/browser-manager-annotation-bridge.test.ts src/main/browser/backends/offscreen-browser-backend-lifecycle.test.ts src/main/browser/document-preview/doc-preview-guest-policy.test.ts src/shared/document-preview/doc-preview-scheme.test.ts", "pnpm exec vitest run --config config/vitest.config.ts src/renderer/src/store/slices/browser/pages/browser-page-conversion.test.ts src/renderer/src/runtime/sync-runtime-graph/sync-runtime-graph-conversion-publish.test.ts", "SKIP_BUILD=1 ORCA_E2E_WEB_CLIENT=1 ORCA_E2E_FORCE_HEADFUL=1 pnpm exec playwright test tests/e2e/paired-preview-address-bar-convergence.spec.ts --config tests/playwright.config.ts --project electron-headless --workers=1 --reporter=list", @@ -4166,7 +4165,6 @@ "src/main/browser/backends/offscreen-browser-backend-lifecycle.test.ts", "src/shared/document-preview/doc-preview-scheme.test.ts", "src/renderer/src/components/browser-pane/workspace-doc/doc-preview-external-link-confirmation.test.ts", - "src/renderer/src/components/link-actions/LinkActionPopover.test.tsx", "src/renderer/src/store/slices/browser/pages/browser-page-conversion.test.ts", "src/renderer/src/runtime/sync-runtime-graph/sync-runtime-graph-conversion-publish.test.ts", "tests/e2e/paired-remote-html-preview-local-render.spec.ts", diff --git a/config/scripts/workflow/source-routing/pr-feature-area-e2e-source-routes.mjs b/config/scripts/workflow/source-routing/pr-feature-area-e2e-source-routes.mjs index cdd9cf836ab..9de5f89e64f 100644 --- a/config/scripts/workflow/source-routing/pr-feature-area-e2e-source-routes.mjs +++ b/config/scripts/workflow/source-routing/pr-feature-area-e2e-source-routes.mjs @@ -16,11 +16,7 @@ function featureRoute(id, specs, prefixes) { // Feature areas whose specs the nightly suite alone used to select (gap audit, Oct 2026). export const PR_FEATURE_AREA_E2E_SOURCE_ROUTES = [ featureRoute('feature.sparse-checkout', ['sparse-explorer-proof'], [`${COMPONENTS}sparse/`]), - featureRoute( - 'feature.onboarding', - ['onboarding', 'onboarding-language-catalog-retry'], - [`${COMPONENTS}onboarding/`] - ), + featureRoute('feature.onboarding', ['onboarding'], [`${COMPONENTS}onboarding/`]), featureRoute( 'feature.diff-notes', [ diff --git a/src/renderer/src/components/link-actions/LinkActionPopover.test.tsx b/src/renderer/src/components/link-actions/LinkActionPopover.test.tsx deleted file mode 100644 index 5c35e2e7b26..00000000000 --- a/src/renderer/src/components/link-actions/LinkActionPopover.test.tsx +++ /dev/null @@ -1,119 +0,0 @@ -// @vitest-environment happy-dom -import type { ReactNode } from 'react' -import { cleanup, fireEvent, render, screen, waitFor } from '@testing-library/react' -import { afterEach, describe, expect, it, vi } from 'vitest' -import type { LinkActionRequest } from './link-action-request' - -const mocks = vi.hoisted(() => ({ - openSettingsPage: vi.fn(), - openSettingsTarget: vi.fn(), - toastError: vi.fn(), - toastSuccess: vi.fn(), - writeClipboardText: vi.fn() -})) - -vi.mock('sonner', () => ({ - toast: { error: mocks.toastError, success: mocks.toastSuccess } -})) - -vi.mock('@/store', () => ({ - useAppStore: ( - selector: (state: { - openSettingsPage: () => void - openSettingsTarget: (target: unknown) => void - }) => unknown - ) => - selector({ - openSettingsPage: mocks.openSettingsPage, - openSettingsTarget: mocks.openSettingsTarget - }) -})) - -vi.mock('@/components/ui/tooltip', () => ({ - Tooltip: ({ children }: { children: ReactNode }) => children, - TooltipTrigger: ({ children }: { children: ReactNode }) => children, - TooltipContent: ({ children }: { children: ReactNode }) => {children} -})) - -vi.mock('@/components/ui/popover', () => ({ - Popover: ({ - children, - open, - onOpenChange - }: { - children: ReactNode - open: boolean - onOpenChange: (open: boolean) => void - }) => - open ? ( -
-
- ) : null, - PopoverAnchor: () => null, - PopoverContent: ({ children, className }: { children: ReactNode; className?: string }) => ( -
{children}
- ) -})) - -import { LinkActionPopover } from './LinkActionPopover' - -afterEach(() => { - cleanup() - vi.clearAllMocks() - vi.unstubAllGlobals() -}) - -describe('LinkActionPopover', () => { - it('ignores duplicate copy clicks while the clipboard write is in flight', async () => { - vi.stubGlobal('navigator', { userAgent: 'Macintosh' }) - Object.assign(window, { api: { ui: { writeClipboardText: mocks.writeClipboardText } } }) - let resolveWrite: (() => void) | undefined - mocks.writeClipboardText.mockReturnValue( - new Promise((resolve) => { - resolveWrite = resolve - }) - ) - const request: LinkActionRequest = { - anchorX: 100, - anchorY: 200, - destination: 'https://example.com/hidden-destination', - kind: 'url', - primary: { label: 'Open link', run: vi.fn() }, - restoreFocus: vi.fn() - } - - render() - const copyButton = screen.getByRole('button', { name: 'Copy link' }) - fireEvent.click(copyButton) - fireEvent.click(copyButton) - - expect(mocks.writeClipboardText).toHaveBeenCalledOnce() - resolveWrite?.() - await waitFor(() => expect(mocks.toastSuccess).toHaveBeenCalledOnce()) - fireEvent.click(copyButton) - await waitFor(() => expect(mocks.writeClipboardText).toHaveBeenCalledTimes(2)) - }) - - it('shows a failure toast when copying fails', async () => { - vi.stubGlobal('navigator', { userAgent: 'Macintosh' }) - Object.assign(window, { api: { ui: { writeClipboardText: mocks.writeClipboardText } } }) - mocks.writeClipboardText.mockRejectedValue(new Error('denied')) - const request: LinkActionRequest = { - anchorX: 100, - anchorY: 200, - destination: 'https://example.com/hidden-destination', - kind: 'url', - primary: { label: 'Open link', run: vi.fn() }, - restoreFocus: vi.fn() - } - - render() - fireEvent.click(screen.getByRole('button', { name: 'Copy link' })) - - await waitFor(() => expect(mocks.toastError).toHaveBeenCalledWith('Failed to copy link')) - fireEvent.click(screen.getByRole('button', { name: 'Copy link' })) - await waitFor(() => expect(mocks.writeClipboardText).toHaveBeenCalledTimes(2)) - }) -}) diff --git a/tests/e2e/onboarding-language-catalog-retry.spec.ts b/tests/e2e/onboarding-language-catalog-retry.spec.ts deleted file mode 100644 index 988aec70a3f..00000000000 --- a/tests/e2e/onboarding-language-catalog-retry.spec.ts +++ /dev/null @@ -1,122 +0,0 @@ -import { test, expect } from './helpers/orca-app' - -test.use({ dismissOnboarding: false, seedTestRepo: false }) - -test('recovers rejected language catalogs without reloading', async ({ - electronApp, - orcaPage -}, testInfo) => { - const startupLanguage = await orcaPage.evaluate(async () => { - const detection = await window.api.app.getSystemLanguageDetection?.([navigator.language]) - return detection?.locale - }) - const candidates = ( - [ - { locale: 'fr', name: 'Français', title: 'Langue d’affichage' }, - { locale: 'ja', name: '日本語', title: '表示言語' }, - { locale: 'es', name: 'Español', title: 'Idioma de la interfaz' } - ] as const - ).filter((candidate) => candidate.locale !== startupLanguage) - const [first, second] = candidates - if (!first || !second) { - throw new Error('Two unloaded languages are required') - } - await orcaPage.evaluate(async () => { - const store = window.__store - if (!store) { - throw new Error('Renderer store unavailable') - } - const settings = await window.api.settings.get() - if (!settings.telemetry) { - throw new Error('Telemetry settings unavailable') - } - await store.getState().updateSettingsOrThrow({ - uiLanguage: 'en', - telemetry: { ...settings.telemetry, onboardingLanguageCohort: 'control' } - }) - }) - const language = first.locale - const catalogPattern = new RegExp(`/${language}-[^/]+\\.(?:json|js)(?:\\?|$)`) - await electronApp.evaluate(({ session }, selectedLanguage) => { - const pattern = new RegExp(`/${selectedLanguage}-[^/]+\\.(?:json|js)(?:\\?|$)`) - session.defaultSession.webRequest.onBeforeRequest((details, callback) => { - callback({ cancel: pattern.test(details.url) }) - }) - }, language) - const rejected = orcaPage.waitForEvent('requestfailed', { - predicate: (request) => catalogPattern.test(request.url()) - }) - const timeOrigin = await orcaPage.evaluate(() => performance.timeOrigin) - await orcaPage.evaluate(async (selectedLanguage) => { - const store = window.__store - if (!store) { - throw new Error('Renderer store unavailable') - } - await store.getState().updateSettingsOrThrow({ - uiLanguage: selectedLanguage - }) - }, language) - const failedRequest = await rejected - await orcaPage.getByRole('button', { name: /^Language/ }).click() - // Why by title: onboarding is a dialog too; the picker's title follows the active language. - const picker = orcaPage.getByRole('dialog', { - name: new RegExp(`^(Display language|${first.title}|${second.title})$`) - }) - // Why a prefix match: a row's name also carries its muted English gloss and state text. - const row = (name: string) => picker.getByRole('button', { name: new RegExp(`^${name}`) }) - await expect(picker.getByRole('alert')).toContainText(first.name) - await orcaPage.screenshot({ path: testInfo.outputPath('catalog-request-rejected.png') }) - - await electronApp.evaluate(({ session }) => { - session.defaultSession.webRequest.onBeforeRequest(null) - }) - expect( - await orcaPage.evaluate(async (url) => (await fetch(url)).status, failedRequest.url()) - ).toBe(200) - await picker.getByRole('alert').getByRole('button', { name: 'Retry', exact: true }).click() - - await expect(row(first.name)).toHaveAttribute('aria-current', 'true') - await expect(orcaPage.getByRole('dialog', { name: first.title, exact: true })).toBeVisible() - await expect(picker.getByRole('alert')).toHaveCount(0) - await expect - .poll(() => orcaPage.evaluate(() => window.api.settings.get().then((s) => s.uiLanguage))) - .toBe(language) - expect(await orcaPage.evaluate(() => performance.timeOrigin)).toBe(timeOrigin) - await orcaPage.screenshot({ path: testInfo.outputPath('catalog-retry-recovered.png') }) - - const nextLanguage = second.locale - const nextName = second.name - const previousName = first.name - const nextPattern = new RegExp(`/${nextLanguage}-[^/]+\\.(?:json|js)(?:\\?|$)`) - await electronApp.evaluate(({ session }, selectedLanguage) => { - const pattern = new RegExp(`/${selectedLanguage}-[^/]+\\.(?:json|js)(?:\\?|$)`) - session.defaultSession.webRequest.onBeforeRequest((details, callback) => { - callback({ cancel: pattern.test(details.url) }) - }) - }, nextLanguage) - const nextRejected = orcaPage.waitForEvent('requestfailed', { - predicate: (request) => nextPattern.test(request.url()) - }) - await row(nextName).click() - await nextRejected - // Why on the row: a failed switch reports under the language that was chosen. - await expect(row(nextName)).toHaveAttribute('aria-describedby', /.+/) - await expect(picker.getByRole('alert')).toBeVisible() - await expect(row(previousName)).toHaveAttribute('aria-current', 'true') - await expect - .poll(() => orcaPage.evaluate(() => window.api.settings.get().then((s) => s.uiLanguage))) - .toBe(language) - - await electronApp.evaluate(({ session }) => { - session.defaultSession.webRequest.onBeforeRequest(null) - }) - // Why the row: choosing the failed language again is the retry. - await row(nextName).click() - await expect(row(nextName)).toHaveAttribute('aria-current', 'true') - await expect(orcaPage.getByRole('dialog', { name: second.title, exact: true })).toBeVisible() - await expect(picker.getByRole('alert')).toHaveCount(0) - await expect - .poll(() => orcaPage.evaluate(() => window.api.settings.get().then((s) => s.uiLanguage))) - .toBe(nextLanguage) - expect(await orcaPage.evaluate(() => performance.timeOrigin)).toBe(timeOrigin) -})