chore(sync): bring in upstream 52a769d on top of the 42da166 integration

[Hylouis233]

Continues the upstream sync started by #16, which landed 42da166 as 5ba8c88.

What this brings in

Six upstream commits between 42da166 and 52a769d:

• 52a769d4 fix(web): text selection keeps going under the turn minimap (fix(web): text selection keeps going under the turn minimap pingdotgg/t3code#16748)
• 2d943723 fix(server): Cursor threads in Auto use Cursor's Auto-review (fix(server): Cursor threads in Auto use Cursor's Auto-review pingdotgg/t3code#19712)
• d2f2b7cd fix(mobile): pull request lists fill the available scroll viewport (fix(mobile): pull request lists fill the available scroll viewport pingdotgg/t3code#18767)
• 38d9760c fix(mobile): agent-spawned threads can be mentioned from the @ picker (fix(mobile): agent-spawned threads can be mentioned from the @ picker pingdotgg/t3code#15841)
• d09a9ead fix(mobile): new threads no longer get stuck on Starting (fix(mobile): new threads no longer get stuck on Starting pingdotgg/t3code#15870)
• 98bc998e fix(prs): bound refreshes by each tracker read (fix(prs): bound refreshes by each tracker read pingdotgg/t3code#19403)

25 files, 244 insertions, 110 deletions — all under apps/web, apps/mobile, packages/shared and packages/provider-cursor.

Why it merged cleanly

I checked these 25 paths against this fork's customized paths before merging: zero overlap. None of them touch apps/server/src/orchestration-v2/ or the provider/ACP layer, so the fork's sibling-buffer ordering, ExternalAcpAdapterV2 / ExternalAcpDriver / ExternalAcpPolicy, and ProviderRuntimeLifetime are untouched. No conflict resolution was required.

Fork state verified intact

Measured against the new upstream tip, this fork now retains 38 customized paths — the 37 that survived #16 (scripts/opent3code-*.cjs, opent3code-release.yml, opent3code-sync.yml, ExternalAcp*, ProviderRuntimeLifetime, EventSink.runtime-request-races.test.ts, integrations/dsh-opent3code/, the opent3code docs and assets) plus .coderabbit.yaml, which #16's merge had dropped and which this branch restores.

Release isolation is unchanged: mobile-android-apk.yml still gates on if: github.repository == 'pingdotgg/t3code', so fork CI cannot publish.

Not in this PR

Deliberately left alone: haisi platform pingdotgg#90/pingdotgg#86/pingdotgg#89, miniprogram #11, Step-Code PR #5 / issue pingdotgg#219, and the 35-branch Step-Code fork.

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has reached their quota limit.

[github-actions]

Thread transfer impact

✅ Thread transfer remains within every enforced ceiling.

ℹ️ The exact PR base did not have a successful artifact. Baseline uses the latest successful main measurement shown below.

Provider	Metric	Main baseline	This PR	Impact	PR ceiling
Codex	Total thread wire	4.9 KiB	5.0 KiB	+82 B (+1.6%)	6.8 KiB	✅
Codex	Thread snapshot wire	3.7 KiB	3.8 KiB	+81 B (+2.1%)	4.9 KiB	✅
Codex	Live turn WebSocket wire	1.2 KiB	1.2 KiB	+1 B (+0.1%)	2.0 KiB	✅
Codex	Live turn WebSocket decoded	20.4 KiB	20.9 KiB	+441 B (+2.1%)	29.3 KiB	✅
Codex	Live turn messages	2	2	0 (0.0%)	8	✅
Claude	Total thread wire	4.9 KiB	5.0 KiB	+100 B (+2.0%)	6.8 KiB	✅
Claude	Thread snapshot wire	3.7 KiB	3.8 KiB	+82 B (+2.2%)	4.9 KiB	✅
Claude	Live turn WebSocket wire	1.2 KiB	1.2 KiB	+18 B (+1.5%)	2.0 KiB	✅
Claude	Live turn WebSocket decoded	20.7 KiB	21.2 KiB	+441 B (+2.1%)	29.3 KiB	✅
Claude	Live turn messages	1	1	0 (0.0%)	8	✅

Baseline: 5cd52d7 · PR result: 935cc84 · Source CI: success

Scenario and decoded snapshot size

10 historical turns, 5 command tools per turn, 878.9 KiB retained MCP result per historical turn, and a 1.05 MiB retained result in the measured turn.

• Codex decoded thread snapshot: 108.5 KiB
• Claude decoded thread snapshot: 108.8 KiB

Updated in place by a trusted workflow. PR artifacts are strictly validated and never executed.

[Hylouis233]

Reviewed the complete 26-file diff at 935cc843254c583dde7f03f8a8936e4e5c5448c7: 25 upstream application/shared/provider paths plus the restored .coderabbit.yaml. The mobile readers consistently defer server detail subscriptions until the selected detail ref exists; context-ID normalization and its regressions move to the shared module; PR refresh sizing is scoped to the actual read unit; Cursor Auto keeps auto-review enabled without changing explicit approval overrides. No fork ACP/DSH implementation or publishing workflow is changed in this diff.

Verified the actual jobs and executed steps in CI run 38781027687: server test shards, general tests, lint/types, desktop build/preload, Rust, release smoke, OpenT3Code Maintenance, transfer artifacts and PR Gate succeeded. Mobile fingerprint run 38781027600 also succeeded. No unresolved inline review threads were returned. Copilot's quota-exhausted response is not a completed review. Transfer measurements meet ceilings, but use an older successful main baseline rather than the exact PR base; no exact-base performance comparison or native/mobile/live-provider acceptance is claimed.

No blocking issue found in this integration diff. This is an AI-assisted review on the owner's behalf, not independent maintainer approval. Merge remains subject to repository protections and the pinned head.

[Hylouis233]

Merged after pinned-head verification

Merged into main as be4dab86c815e26de111102425c17441fa542aa4 using a history-preserving merge. The reviewed head was 935cc843254c583dde7f03f8a8936e4e5c5448c7.

The complete PR is 26 files, 253 additions and 110 deletions: the 25 upstream paths described above plus the restored .coderabbit.yaml. Actual executed steps in CI 38781027687 passed server shards, general tests, lint/types, desktop build/preload, Rust, release smoke, OpenT3Code Maintenance and PR Gate. Mobile fingerprint also passed. The complete diff was reviewed and there were no unresolved inline threads immediately before merging. Copilot's exhausted quota was not counted as a review.

This lands the six-commit batch through upstream 52a769d4; #16 already landed the preceding 42da166 integration and #11 was separately closed. It does not claim to include later upstream commits. Transfer measurements passed their ceilings but used an older successful baseline, not the exact PR base. Integrated browser/mobile/native-device acceptance and live provider verification are not established by these CI results. No deployment or release was requested by this merge action.