Skip to content

chore: sync upstream through ade473018d - #284

Merged
andrewcai8 merged 70 commits into
mainfrom
sync/upstream-ade473018d
Nov 9, 2026
Merged

andrewcai8 merged 70 commits into
mainfrom
sync/upstream-ade473018d

Conversation

@andrewcai8

Copy link
Copy Markdown
Owner

Why

We want the fork back on upstream's newest commit with green CI, ade473018d, without losing fork behaviour. The fork was 69 commits behind. No new migrations.

Notable upstream changes

Conflict resolutions (9 files)

  • Cursor usage (CursorUsageAccounts.ts, usage.ts, UsageLimits.tsx, docs/user/usage.md): takes upstream's removal of CLI-login and Keychain reads, since we no longer use Cursor. This drops the fork's per-login multi-account reading. The fork's stored history stays, so accounts with no key left still show on Usage. sourcePath stays in the contract.
  • Scheduled tasks (web and mobile editors, scheduledTaskDraft.ts): keeps upstream's thread move and the fork's cloud targets. Mobile hides Destination on a cloud target, and web already forces new_thread there. The mobile cloud check now reads the draft's thread, so unbinding a thread makes cloud targets available again.
  • settings.ts: keeps both sides.
  • agentContextBudget.json: set to measured values. All three are lower than the fork's main.

Fixes the merge needed

  • Deadlock. Upstream's releaseForLinkedEnvironments runs inside open's lock, and the fork's release takes that same lock. Changing Linked environments would have hung the next open of an exclusive runtime. It now passes holdsOpenLock. The existing ProviderSessionManager test timed out before this and passes now.
  • Removed references to the deleted CursorKeychain in a fork script and two fork tests, and an unused mobile import.
  • UsageService.cursorLogins.test.ts is now UsageService.cursorHistory.test.ts. The two CLI-login cases are gone with the feature, and the removed-account history case uses API keys.

Verification

  • vp run typecheck: contracts, provider-cursor, client-runtime, server, web, mobile and desktop report 0 errors and 0 warnings.
  • Full vp test run: contracts, provider-cursor, client-runtime and mobile pass. The server suite passes apart from one case below.
  • vp lint on all 697 files that differ from upstream, vp run knip:check, and vp i --frozen-lockfile pass.
  • vp run check:changed passes apart from the same case. AntigravityInstallation "accepts the agy CLI" fails on this Mac because APFS is case-insensitive. Those files are identical to upstream.
  • An independent verifier on Opus reviewed the resolutions and found no defects from this merge.

Not verified

  • The full web and desktop suites locally. Only scheduledTasksSettings.logic.test.ts ran for web. PR CI covers the rest.
  • A pass in a real client. The scheduled task and Usage screens were checked through code and tests only.

Follow-ups

  • With no key left, stored Cursor history keeps the first cache entry per account instead of the newest. The fork's earlier code did the same. Low priority, since we no longer use Cursor.

Model: Claude Opus 5.5. Harness: Claude Code.

🤖 Generated with Claude Code

shivamhwp and others added 30 commits November 7, 2026 08:39
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: maria-rcks <maria@kuuro.net>
…29355)

Co-authored-by: PR Batch Tester <agent@local.test>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: GPT-6.1 Sol <noreply@openai.com>
Co-authored-by: maria-rcks <maria@kuuro.net>
…gdotgg#28530)

Co-authored-by: PR Batch Tester <agent@local.test>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
…onto an empty ~/.t3 (pingdotgg#23878)

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
pingdotgg#31769)

Co-authored-by: Aditya Vardhan Sharma <adityaradhika13@gmail.com>
Co-authored-by: Lakshmi Tanmay <lakshmi@voltcrash.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
pingdotgg#32035)

Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
…#32325)

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
…2269)

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
…32263)

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
…24202)

Co-authored-by: PR Batch Tester <agent@local.test>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: maria-rcks <maria@kuuro.net>
)

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
…--repository (pingdotgg#29586)

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: GPT 6.1 Sol <codex@openai.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: GPT 6.1 Sol <codex@openai.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
…ingdotgg#32403)

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
… of rejected (pingdotgg#32405)

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
arifali123 and others added 24 commits November 7, 2026 15:53
…#27941)

Co-authored-by: Orin <orin@orin.ws>
Co-authored-by: maria-rcks <maria@kuuro.net>
…s the thread (pingdotgg#28454)

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
)

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
…gg#27541)

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
…etails card (pingdotgg#27582)

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
…in folders (pingdotgg#25883)

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
)

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
…#32266)

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
…g#28255)

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
…tgg#26737)

Co-authored-by: PR Batch Tester <agent@local.test>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: maria-rcks <maria@kuuro.net>
…ges (pingdotgg#32833)

Agents can pass `path` to an absolute .html file instead of retyping the
page as inline `html`. The server reads the file only for full-access
threads, with the existing disk-read guards, and snapshots it at publish.
ACP render recovery matches path calls by path. The mcp-tool-use eval
grades pages passed by path.

Takes over pingdotgg#32817.

Co-authored-by: Stoli Geyer <stoli@tractorbeam.ai>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…otgg#32242)

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: PR Batch Tester <agent@local.test>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
…31379)

Co-authored-by: maria-rcks <maria@kuuro.net>
Co-authored-by: shivam <91240327+shivamhwp@users.noreply.github.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: Bryan Tobing <bryantobing0@gmail.com>
…their own updater (pingdotgg#21524)

Mise-owned providers requested as latest now update with mise upgrade <tool>; pinned tools, version paths and mise x scripts stay manual.

Co-authored-by: Ryan Hughes <ryan@heyoodle.com>
Co-authored-by: Julius Marminge <julius0216@outlook.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@github-actions github-actions Bot added vouch:trusted PR author is trusted by repo permissions or the VOUCHED list. size:XXL 📱 Native Change Changes the native fingerprint; merging blocks production OTAs until a new store build ships. labels Nov 8, 2026
@github-actions

github-actions Bot commented Nov 8, 2026

Copy link
Copy Markdown

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 5.0 KiB 5.0 KiB 0 B (0.0%) 6.8 KiB ✅
Codex Thread snapshot wire 3.8 KiB 3.8 KiB 0 B (0.0%) 4.9 KiB ✅
Codex Live turn WebSocket wire 1.2 KiB 1.2 KiB 0 B (0.0%) 2.0 KiB ✅
Codex Live turn WebSocket decoded 21.5 KiB 21.5 KiB 0 B (0.0%) 24.4 KiB ✅
Codex Live turn messages 1 1 0 (0.0%) 8 ✅
Claude Total thread wire 5.0 KiB 5.0 KiB 0 B (0.0%) 6.8 KiB ✅
Claude Thread snapshot wire 3.8 KiB 3.8 KiB 0 B (0.0%) 4.9 KiB ✅
Claude Live turn WebSocket wire 1.2 KiB 1.2 KiB 0 B (0.0%) 2.0 KiB ✅
Claude Live turn WebSocket decoded 21.9 KiB 21.9 KiB 0 B (0.0%) 24.4 KiB ✅
Claude Live turn messages 1 1 0 (0.0%) 8 ✅

Baseline: 649ab7f · PR result: f71bca3 · 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.

@andrewcai8
andrewcai8 merged commit f0ad252 into main Nov 9, 2026
25 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

📱 Native Change Changes the native fingerprint; merging blocks production OTAs until a new store build ships. size:XXL vouch:trusted PR author is trusted by repo permissions or the VOUCHED list.

Projects

None yet

Development

Successfully merging this pull request may close these issues.