Sync upstream main into personal - #58
Merged
Merged
Conversation
…pingdotgg#9001) Co-authored-by: Guillermo Casanova <guillermo.casanova.b@gmail.com>
Co-authored-by: maria-rcks <254055478+maria-rcks@users.noreply.github.com>
Co-authored-by: t3-code[bot] <269035359+t3-code[bot]@users.noreply.github.com> Co-authored-by: maria <254055478+maria-rcks@users.noreply.github.com>
Co-authored-by: maria-rcks <254055478+maria-rcks@users.noreply.github.com>
Six repeat contributors remained unvouched. Add their approved accounts to the trust list: - `github:ipanasenko` - `github:nateEc` - `github:maxwellyoung` - `github:tsouth89` - `github:0x4bs3nt` - `github:Adamulek123` This change does not modify runtime or workflow logic. Made by GPT-5.6 Sol in the Codex harness.
…g#9129) Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
Co-authored-by: maria-rcks <254055478+maria-rcks@users.noreply.github.com>
Co-authored-by: Rodrigo Brechard <rodrigo@clubtidy.fr> Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
…ruma WASM engine (pingdotgg#8360) Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
…gdotgg#9152) Takes over pingdotgg#6608 by @x1xhlol. Server: thread.message-sent updates the projection row directly and folds latestUserMessageAt in as a running maximum instead of running refreshThreadShellSummary. Client: thread.activity-appended takes an O(1) append path for in-order, unseen activities on reducer-produced arrays, with a WeakMap id index, and falls back to filter and sort for out-of-order, re-delivered, context-window, and snapshot-loaded cases. Reducer benchmark: 10,000 in-order appends went from 1,180 ms to 70 ms. Change authored by Claude Fable 5.1 running in Claude Code, based on work by @x1xhlol. Co-authored-by: Lucas Valbuena <lucknitelol@proton.me>
WebSocket reconnect catch-up only checked the number of missing events, then loaded and decoded the whole replay range before sending it. A handful of events with large tool payloads could allocate gigabytes and OOM the server. Before replaying, run a SQL preflight that counts the events in the range and sums their serialized payload bytes with octet_length(). If the range is over the existing row limit or an 8 MiB byte budget, fall back to the fresh snapshot path. Applies to both thread and shell subscriptions. Authored by @rcawston. Review, octet_length fix, and merge by Claude Fable 5.1 in Claude Code on behalf of @t3dotgg.
…aming (pingdotgg#9145) Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
Co-authored-by: Theo Browne <me@t3.gg> Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
…nvironments (pingdotgg#9147) Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
Include SQLite conditions and schema issue tags without copying query data. Continue @Sy-D's [pingdotgg#4837](pingdotgg#4837). Add the missing Bun error codes and test the real SQL client. Created with GPT-6 Astra (preview) in Codex. Co-authored-by: Sy-D <8460326+Sy-D@users.noreply.github.com> Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
Co-authored-by: maria-rcks <254055478+maria-rcks@users.noreply.github.com>
…ictation (pingdotgg#9548) Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
…ingdotgg#9547) Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
…g#9534) Follow-up to pingdotgg#9507, carrying over the reset-credit redemption from pingdotgg#9421. Codex grants a reset credit when it has rate-limited an account unfairly (`"Thanks for using Codex! You've been granted one free rate limit reset."`). Redeeming one clears the current 5h/weekly windows. The Limits tab now shows how many are banked and when the next expires, with a confirmed **Use a reset credit** action. ## How it works - `ServerProviderUsageLimits.resetCredits` carries the count and soonest expiry; the Codex probe reads it from the same `account/rateLimits/read` it already makes. - `ProviderInstance.consumeResetCredit` is a new optional hook — account-level, so it sits beside `refreshModels` rather than on the thread-routed adapter. The Codex driver implements it over a short-lived app-server (via `withCodexAppServerClient`, factored out of the status and skills probes which duplicated the setup), then re-probes. - Single-flight per instance with one idempotency key kept until Codex reports an outcome, so a retry after a timeout does not open a second attempt. - New `provider.consumeResetCredit` RPC under the operate scope; the outcome (`reset` / `nothingToReset` / `noCredit` / `alreadyRedeemed`) is shown inline. Only Codex reports credits today. A provider without the hook gets a clear "does not bank reset credits" error; one without credits shows nothing. ## Screenshots The local Codex row with one banked credit (the same account via the CLIProxyAPI hub above it shows no credit, as expected — the hub does not relay them):  Close-up of the row:  Clicking it opens the confirmation; nothing is sent until **Use credit**:  ## Verification - Mapper tests for the credit summary; provider, contract, and Usage page suites pass; typecheck clean. - Verified against a real Codex Pro account holding one credit: summary and expiry render, the confirm dialog opens. **Not redeemed** — that would spend the credit. Written by Claude Fable 5 via Claude Code; design and single-flight approach from @StiensWout's pingdotgg#9421. <!-- CURSOR_SUMMARY --> --- > [!NOTE] > **Medium Risk** > Redemption spends real account credits over a new RPC; correctness depends on per-account locking and idempotency, though disabled instances and non-Codex providers are rejected explicitly. > > **Overview** > Adds **end-to-end redemption of banked Codex rate-limit reset credits** from the Limits UI on web and mobile, backed by a new operate-scoped `provider.consumeResetCredit` RPC. > > **Contracts and server:** `ServerProviderUsageLimits` can include `resetCredits` (count + next expiry). Codex probes attach that from `account/rateLimits/read`. Optional `ProviderInstance.consumeResetCredit` is implemented for Codex via a scoped app-server call to `account/rateLimitResetCredit/consume`, then a limits refresh. `CodexResetCreditCoordinator` serializes redemptions per Codex account directory, reuses one idempotency key until Codex returns an outcome, and times out hung requests. `withCodexAppServerClient` is extracted so status, skills, and redemption share the same short-lived app-server setup. > > **Clients:** Limits rows show banked credits and a confirmed **Use a reset credit** action that calls `serverEnvironment.consumeResetCredit` and surfaces outcomes (`reset`, `nothingToReset`, etc.) or errors. > > **Web usage sources (same PR):** Adding/removing CLIProxyAPI hubs and the add dialog target a **selected connected environment** (with picker when several are connected), gated by operate access—not only the primary environment. > > <sup>Reviewed by [Cursor Bugbot](https://cursor.com/bugbot) for commit 98f32e6. Bugbot is set up for automated code reviews on this repo. Configure [here](https://www.cursor.com/dashboard/bugbot).</sup> <!-- /CURSOR_SUMMARY --> Co-authored-by: Wout Stiens <71498452+StiensWout@users.noreply.github.com> Co-authored-by: Claude Code <noreply@anthropic.com>
…otgg#9125) Refresh missing PR associations after agent turns on the thread's current branch. Preserve background policy, known PR caches, and failed-lookup backoff. Serialize status loads and refreshes to prevent stale responses from hiding a PR. Find branches pushed under their own name while still tracking the default branch. Original work by Theo Browne with Claude Fable 5.1 in Claude Code. Takeover fixes created with GPT-6 Astra (preview) in Codex. Co-authored-by: Theo Browne <me@t3.gg> Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
Co-authored-by: Claude Code <noreply@anthropic.com>
Co-authored-by: maria-rcks <254055478+maria-rcks@users.noreply.github.com>
…ands (pingdotgg#9553) Scrolling a long thread to the end with the composer at rest landed flush against the short composer. The expansion that followed then covered the last rows, because the timeline reserves only the live overlay height and does not move for footer growth. The timeline now keeps the expanded composer's height clear while the composer rests, so expanding it again changes nothing above the composer. The composer reports its resting flag from a layout effect and publishes a fresh overlay height whenever that flag changes, so the reservation is always computed from a height that belongs to the same layout. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Co-authored-by: maria-rcks <254055478+maria-rcks@users.noreply.github.com>
…otgg#9562) Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
Co-authored-by: maria-rcks <254055478+maria-rcks@users.noreply.github.com>
Co-authored-by: maria-rcks <254055478+maria-rcks@users.noreply.github.com>
Co-authored-by: maria-rcks <254055478+maria-rcks@users.noreply.github.com>
Co-authored-by: maria-rcks <254055478+maria-rcks@users.noreply.github.com>
Co-authored-by: maria-rcks <254055478+maria-rcks@users.noreply.github.com>
…-sync-20260901 # Conflicts: # apps/mobile/src/lib/threadActivity.ts # apps/server/src/orchestration/Layers/ProviderCommandReactor.ts # apps/server/src/provider/Drivers/CodexDriver.ts # apps/server/src/provider/Layers/CodexProvider.ts # apps/server/src/provider/Layers/CodexSessionRuntime.ts # apps/server/src/provider/ProviderDriver.ts # apps/web/src/components/ChatView.tsx # apps/web/src/components/Sidebar.tsx # apps/web/src/components/chat/ChatComposer.tsx # apps/web/src/components/chat/MessagesTimeline.tsx # apps/web/src/components/preview/PreviewAutomationHosts.tsx # apps/web/src/components/preview/previewAutomationOpenReadiness.test.ts # packages/shared/package.json
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.
Upstream
07891e9569c88457516b44c08c820471762969e8-- updated 2026-09-04.New things to try
Improvements
Bug fixes
Fork compatibility
providerSkillsRPC still drives exact$skilllookup, while upstream workspace snapshots drive slash menus. Structured invocation, unknown-skill errors, nativeturn/steer, and authoritative start-response turn IDs remain intact alongside upstream context compaction and Codex reset-credit support.Validation
/home/openclaw/.local/share/t3code-sync-vite-plus/node_modules/.bin/vp install --frozen-lockfile./home/openclaw/.local/share/t3code-sync-vite-plus/node_modules/.bin/vp check(0 errors; 15 warnings)./home/openclaw/.local/share/t3code-sync-vite-plus/node_modules/.bin/vp run typecheck./home/openclaw/.local/share/t3code-sync-vite-plus/node_modules/.bin/vp run lint:mobile; 11 Swift and 12 Kotlin source files checked, with SwiftLint, ktlint, and detekt unavailable.