Repository navigation
chore(sync): rebase latest upstream and preserve fork integrations - #62
Merged
Merged
Conversation
…ngdotgg#29922) Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
…29909) Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
…arent's turn (pingdotgg#29752) Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
…and longer fences (pingdotgg#29755) Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
…#26495) Co-authored-by: maria-rcks <maria@kuuro.net>
…ts or subagents (pingdotgg#29917) Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
… without raw provider text (pingdotgg#29877) Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
…ingdotgg#29905) Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
…ect under diff.relative (pingdotgg#29879) Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
…in a subagent or side chat (pingdotgg#29891) Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
…ngdotgg#29912) Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
…ingdotgg#28684) Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com> Co-authored-by: Julius Marminge <julius0216@outlook.com>
…squash-merged (pingdotgg#29919) Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
…ointed parent run (pingdotgg#25662) Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
…ngdotgg#29721) Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
pingdotgg#29580) Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com> Co-authored-by: maria-rcks <maria@kuuro.net>
… the project folder (pingdotgg#29908) Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com> Co-authored-by: Maria <maria-rcks@users.noreply.github.com>
…ngdotgg#29962) Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
…dotgg#29911) Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
…oading it (pingdotgg#29893) Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
…ingdotgg#29902) Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
…intSigner (pingdotgg#30529) Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
…bile (pingdotgg#30540) Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
…ngdotgg#31071) Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
…he git-read timeout (pingdotgg#30554) Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
…pingdotgg#31104) Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Merge pingdotgg/t3code main through 61c3794, retaining fork provider pools, project intelligence, PEA/pstack, and client integrations. Resolve Effect service and MCP registration changes, preserve existing migrations, and reconcile provider refresh, orchestration, authentication, and client behavior.
Merge main's hosted-runner validation and private Linux CI routing while preserving upstream macOS release preparation.
…1-02 chore: merge upstream main and resolve fork integration conflicts
The Windows job gave up when `Build Linux <arch>` was missing from the run for 5 minutes, treating that as a renamed job. Since the validate job was added, the build is unlisted until validate finishes, so a busy hosted queue failed a healthy release (run 41752989226). Treat absence as still waiting, bounded by the step timeout. Raise that to 180 minutes, since that run needed about 50 and exceeded the old 40. [skip ci] so this push does not cancel the in-progress release re-run. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Wait for LegendList to finish its hidden initial layout before recording the heading position. Keep real timers running through the click and expansion so the virtualizer can complete its layout work. Co-authored-by: GPT-6 Astra <noreply@openai.com>
…es (#60) * fix(connection): reconnect paired devices over authenticated LAN routes * fix(connection): adapt LAN credentials to current auth and discovery APIs
* fix(opencode): point agents at T3's browser instead of OpenCode's OpenCode denies its own browser tool until its desktop app attaches, but T3's full-access `*` allow re-enabled it. Agents reached for a browser that can never connect under T3 and never found `preview_*`. Re-deny the browser tool and skill in full-access sessions, and give OpenCode the T3 browser instructions Codex and ACP already get when browser access is enabled. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix(opencode): preserve external browsers and v1 tool guidance Scope native browser denies to T3-managed OpenCode 2 servers and retain external desktop browser access. Add the existing T3 browser instructions to OpenCode 1 turns only when its MCP connection and browser capability are available. Cover managed/external resume and disconnected/disabled browser states. Validated with 170 OpenCode tests and direct provider/server typechecks. Reviewed and repaired with GPT-6 Astra Pro through ChatGPT coding tools. * fix(opencode): retain MCP ownership and native browser permissions * fix(opencode): deny browsers children inherit through full access A child that inherits only a `*` allow kept OpenCode's browser, because stripping explicit browser allows does nothing to that rule. A desktop permission event during session.update could also leave the stale allow list on the server while the local cache moved on. Co-Authored-By: Grok 4.7 <noreply@x.ai> Co-authored-by: Cursor <cursoragent@cursor.com> --------- Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com> Co-authored-by: Grok 4.7 <noreply@x.ai> Co-authored-by: Cursor <cursoragent@cursor.com>
Angle-bracketed link destinations ([label](<t3-thread://...>)) were not parsed as labeled links: the inner <...> matched as an autolink, the label was discarded, and an unresolved title substituted "Thread no longer available" as the destination text. Thread titles were also resolved only from the shell snapshot, which omits idle subagent children and all but the 25 newest settled threads, so live threads cited from chat read as deleted. Mobile and web now mount the cited thread's family read (the on-demand rows search and the Settled shelf already use) until its title lands. Generated with [Devin](https://devin.ai) Co-authored-by: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* refactor(server): resolve the mDNS host name in one place RemoteOpenTargets derived <first-label>.local from os.hostname(), which is wrong on a Mac whose hostname came from DHCP: Bonjour answers to LocalHostName instead. Move the decision into a shared helper that prefers scutil on macOS and falls back to the hostname. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix(server): advertise the mDNS hostname as a LAN route Every direct route was a raw IPv4 address, so a router or Wi-Fi change left clients reconnecting forever. Report <name>.local after the LAN addresses. Clients already save it as a learned LAN route, and the name survives a subnet change. Machines with no LAN address do not report it. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * docs: explain the learned local network name route Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix(server): retry a failed Bonjour name lookup Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix(connection): keep host names off LAN credential routes Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
* feat(evals): evaluate independent thread suggestions * feat(chat): suggest fresh threads for independent drafts * fix(server): thread split suggestions ignore empty message context Most real user messages carry a context object with zero records, and the suggestion service treated any context object as a reason to abstain, so most threads could never receive a suggestion. Abstain only when a prior message actually references context records. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> * fix(chat): dismissed thread split suggestion stays hidden while the draft is edited Dismissal was keyed to the exact draft text, so continuing to type the same draft requested inference again and brought the suggestion back. Dismissal now holds until the draft is sent or cleared, on web and mobile. Mobile also skips requests when the environment has no decision endpoint, matching web. The user guide names the setting as the UI labels it. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> * fix(chat): dismissed thread split suggestion stays hidden across remounts The composer remounts when switching threads or restoring a persisted draft, which reset dismissal to false and brought the suggestion back for a draft the user already declined. Hold dismissal in a per-thread set so it survives remounts until the draft is sent or cleared, on web and mobile. Co-Authored-By: SWE-2 High <noreply@cognition.ai> * chore(evals): ignore python bytecode cache Running the thread-split unittest leaves evals/thread-split/__pycache__ untracked; ignore it alongside eval run output. Co-Authored-By: SWE-2 High <noreply@cognition.ai> * test(evals): measure thread split detection on held-out human conversations --------- Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com> Co-authored-by: SWE-2 High <noreply@cognition.ai>
|
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Syncs the fork with pingdotgg/t3code upstream main at 7f6f47b (v0.1.7). The merge-preserving rebase retains project memory, summaries, Loom/code intelligence, pstack, Traex, provider pools, linked environments and device discovery. A final reconciliation merge includes fork main so this PR can merge normally without rewriting main.
Resolves auth/crypto and client integration conflicts, retains fork migration numbering, repairs the Loom JSON index for legacy-row recovery, regenerates dependencies, and aligns Codex replay fixtures without relaxing protocol assertions or timeouts. Changed checks omit read-only vendored reference paths from formatter/test inputs to avoid macOS E2BIG errors.
Verification:
vp run check:changed e6c4180b7a --concurrency 1: all steps passed, including 9 affected typechecks and 207 tests.Model/harness: GPT-6.1 Sol via T3 Code/Codex.