8000 [pull] main from can1357:main by pull[bot] · Pull Request #11 · ericrihm/oh-my-pi · GitHub
[go: up one dir, main page]

Skip to content

[pull] main from can1357:main - #11

Merged
pull[bot] merged 238 commits into
ericrihm:mainfrom
can1357:main
Aug 16, 2026
Merged

pull[bot] merged 238 commits into
ericrihm:mainfrom
can1357:main

Conversation

@pull
@pull pull Bot commented Aug 16, 2026 •
Copy link
Copy Markdown

See Commits and Changes for more details.


Created by pull[bot] (v2.0.0-alpha.4)

Can you help keep this open source service alive? 💖 Please sponsor : )

roboomp and others added 30 commits August 11, 2026 11:17
Applied OpenCode Go DeepSeek tool-choice compat to both OpenAI APIs, so the Responses route drops named selectors while preserving tools.

Added generated-policy and request-payload regression coverage.

Fixes #8243
…ding

Enter in a multi-select ask question toggled the focused option exactly
like Space, with the only submit path hidden behind the Submit tab that
appears only for multi/multi-question dialogs. Enter now submits the
current selection (matching single-select), Space still toggles, and the
footer hint reflects it.

Fixes #8252
Abort an in-flight model turn before unregistering Vibe tools so it cannot emit stale worker calls after cancellation.

Add a regression test covering the live streaming transition.

Fixes #8326
abort()'s stranded-queue drain could restart a queued user turn on the still-live Vibe tools before teardown removed them. Run the exit teardown with the queued-message drain suppressed, then re-arm it so the queued turn resumes as a clean non-Vibe turn.

Fixes #8326
Replace the resettable queued-message block with a nested mode-exit suppression depth that steer and follow-up queueing cannot clear.

Cover a steer arriving after abort settles but before Vibe teardown releases.

Fixes #8326
Clear the scheduled marker when the queued-message continuation claims its run so later idle steers and follow-ups can schedule another drain.

Add coverage for two successive successful idle drains.

Fixes #8326
- Multi-question dialogs now advance on Enter in multi-select mode
  instead of submitting every question at once (roboomp blocking)
- Ask tool accepts an empty multi-select submission as a "select none"
  answer instead of aborting as a cancellation (codex P2)
- Footer hint reflects advance vs submit for multi-question dialogs
- Move the #8252 changelog entry to Unreleased (codex P2)

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Cursor's backend can pin a per-conversation rejection (bare
resource_exhausted, zero tokens) to one wire conversationId forever,
making the session unusable until /fork mints a new id. On the first
such failure the id is rotated once and the cached state migrates, so
the caller's retry loop starts a fresh conversation — the same recovery
/fork performs. Repeated failures keep the rotated id so a genuine
account-level exhaustion is not hidden (#8345).

The regex intentionally drops the global flag: test() would otherwise
share lastIndex across calls and silently miss every second message.

Covered by an end-to-end HTTP/2 fixture: a real local server rejects the
first conversation and accepts the next, asserting the wire
conversationId differs; a second fixture fails every request and asserts
the rotation happens exactly once.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
completeSimple reports a transient provider failure (Anthropic
overloaded_error / rate_limit_error, HTTP 429/500/502/503/529) by
RESOLVING with stopReason "error" rather than throwing, so callers that
wrapped it in a try/catch-based retry never actually retried.

The helper handles both shapes, classifies with the existing AIError
predicates so the retryable set stays defined in one place, honors
retry-after / x-ratelimit-reset* from response headers or error text, and
returns/rethrows the failure unchanged once attempts are exhausted so
existing caller fallbacks keep working.

Aborts stay aborts: a cancel observed at an attempt boundary surfaces
that attempt's own result, while a cancel during backoff rejects with the
abort reason instead of being relabelled as the provider failure.
instrumentedCompleteSimple is the single funnel for every oneshot LLM
call in the agent, so retry belongs here rather than in a try/catch above
each caller - the failure arrives as a resolved AssistantMessage.

Opt-in rather than default-on: oneshotKind is free-form and callers may
pass arbitrary ctx.tools, so the funnel cannot itself prove a request is
replay-safe.

Response headers are captured per attempt and cleared between attempts,
so a stale retry-after can never be reused for a later failure.
Each is a single, side-effect-free completion whose result is parsed after
it resolves, so one transient provider failure previously aborted the
whole operation - for /compact that left the user's context full.

Adds SummaryOptions.oneshotRetry, because both compaction paths call the
same generateSummary and the policy therefore cannot be a constant inside
it. Manual /compact has no outer loop and gets retry by default;
auto-compaction passes false because session-maintenance already retries
the whole attempt, and nesting would multiply the budget (10 outer x 3
inner) while stacking each outer wait on an inner backoff.
Session title generation, TTS speech enhancement, commit-message
generation, the auto-thinking and unexpected-stop classifiers, memory
extraction/consolidation, the commit analysis/summary/changelog/map/reduce
passes and the mnemopi LLM callback each aborted or degraded on the first
transient blip. Several returned null, which made a provider overload
indistinguishable from a legitimate empty result.

The commit analysis, summary, changelog and reduce passes also fed a
provider error message straight into their response parsers: they never
checked stopReason, so a failed request produced garbage instead of
surfacing the error. The map phase's own retry helper only caught thrown
errors, so a resolved stopReason "error" bypassed it entirely.

Sites deliberately left unwrapped: the Anthropic web_search provider
(server-side execution, billed per search - a re-request duplicates a real
side effect), the legacy streamSimple shim (already-emitted events would
duplicate), the auth-gateway health probe (a probe must report the state
it observed), and the script paths whose own retry logic tracks per-attempt
usage or re-requests only the failed subset.
userMessageText and customMessageText were empty and inherited the
terminal default over light paper card backgrounds, so chat bubbles
could wash out on dark terminals. Pin those tokens to forestFloor.

Leave selectedBg and toolTitle adaptive: a dark selectedBg regresses
light terminals when text is terminal-default, and toolTitle is also
used on unpainted inline tool rows.
The welcome banner (WelcomeComponent) captures the session's active
model name at init time but never receives updates when the model
changes. Its setModel() method already existed but was never called
after construction.

This wires the model_changed agent session event to a new
#updateWelcomeModel() method in InteractiveMode, mirroring the
existing #updateWelcomeLspServers() pattern. The banner now reflects
the live model after:
- delayed config/modelRoles load (the startup race where an
  alphabetically-first provider is picked before config resolves)
- explicit /model switches
- retry-fallback model swaps
Address review feedback: init-time model switches (#reconcileModeFromSession,
#enterPlanMode for plan.defaultOnStartup) fire model_changed before the
subscription exists, leaving the banner stale for those cases too.

Adds a catch-up #updateWelcomeModel() call right after the subscription is
installed, mirroring how the sessionName listener is registered before
those same init steps (line 1094-1095 comment).

Also adds a regression test for the catch-up path and comments tying the
component test to the wiring gap it covers.
setModel() invalidates the render cache but doesn't schedule a paint.
The initial forced startup paint (line 1104) may have already rendered
the old model, so the catch-up #updateWelcomeModel() must explicitly
request a render — matching #updateWelcomeLspServers().
The watchdog discarded every overshoot longer than `sleepMs` (60s) as
system sleep, on the rationale that "the process could not have run JS
during the missed interval". That is true for a suspended process and
false for a CPU-bound wedge, which is exactly the case where JS ran the
whole interval at 100% CPU.

Duration cannot separate the two: both produce an arbitrarily large gap.
Keying the suppression on magnitude therefore dropped the most severe
stalls and kept only the mild ones, so the probe went quiet precisely
when it had something to report. Issue #5372 shows an 82,391ms block that
16.4.8 logged and current builds do not.

Classify by mechanism instead. `cpuNow()` reads process CPU time, and an
over-`sleepMs` gap is suppressed only when the process also burned less
than half of it on CPU. A resumed process shows a gap it did not spend
CPU on; a wedged one shows a gap it spent almost entirely on CPU. The
observed CPU time is included in the log line so a reader can tell which
kind of stall they are looking at.

Refs #5372, #7328, #6145
runWorkspaceDiagnostics awaited proc.exited but discarded the exit
status, then treated empty output as success. A checker that never
ran - missing toolchain, crash, OOM kill - writes nothing to either
pipe, so the agent was told "No issues found" when nothing had been
checked at all.

Capture the exit code and, when the checker produced no output *and*
exited non-zero, report the failure instead. A non-zero exit with
output is how tsc/cargo/pyright normally report diagnostics and is
unaffected. This mirrors the exit-status gate that
resolveGoWorkspaceDiagnosticsCommand already applies in this file.

Refs #8386
The conditional fits in the 120-column budget at indentWidth 3, so the
formatter wants it on a single line.
SessionManager.open writes title+session before createAgentSession
claims the id. Agent Hub parked that stub, so the spawn CAS failed
with "already owned by another session generation" and the row
could not be revived (no session_init).
Drives real subprocesses so the 137 under test is a genuine SIGKILL
rather than a hand-written constant: `kill -9 $$` for the OOM-killer
shape, and a watchdog-killed `sleep` for the timeout shape.
can1357 and others added 27 commits August 16, 2026 02:43
… dialogs (@Seljuke)

# Conflicts:
#	packages/coding-agent/test/extensions-runner.test.ts
# Conflicts:
#	packages/coding-agent/src/session/session-loader.ts
#	packages/coding-agent/src/session/session-manager.ts
#	packages/coding-agent/test/session-loader-stream.test.ts
…llback (@roboomp)

# Conflicts:
#	packages/coding-agent/src/session/turn-recovery.ts
# Conflicts:
#	packages/coding-agent/test/interactive-mode-vibe-toggle.test.ts
# Conflicts:
#	packages/coding-agent/src/session/agent-session.ts
#	packages/coding-agent/src/session/session-manager.ts
#	packages/coding-agent/test/agent-session-empty-stop-guard.test.ts
#	packages/coding-agent/test/session-manager-immediate-persist.test.ts
Opened the current Bailian API-key management page from the China interactive login flow and updated its user guidance.

Added regression coverage for the emitted auth URL and instructions.

Fixes #8691
BiomeClient#parseJsonOutput expected a stale --reporter=json shape
(location.path.file, byte-offset span, description), so the shape guard
dropped every diagnostic against Biome 2.x output and lint() returned []
with no warning. Parse the current schema instead: string location.path,
1-indexed location.start/end {line,column}, and message. Drop the now-dead
offsetsToPositions helper and warn when a non-empty diagnostics array has
no recognizable location, so future schema drift surfaces instead of
silently reporting "no lint issues".

Fixes #8694
Reverted the retry-token continuation flow (merge 3462852 and prompt
follow-up 446e745): seen-line rejections resend the full patch again.
…es as template literals

- Replaced `join("\\n") + "\\n"` concatenations with template-literal interpolation in mid-spawn registry fixtures.
- Updated the three Bun.write calls in `persisted-mid-spawn-stub.test.ts` to use the template form.
…argo native mode

- Replaced the `setup` script in `package.json` with a call to `bun scripts/setup.ts`.
- Added `scripts/setup.ts` to run install, native build, coding-agent linking, and `link-omp` steps in order with step-level failure handling.
- Extended `scripts/bazel-natives.ts` with a `--cargo` flag, forbidding incompatible `--source`/bazel arg combinations, and using it to force local host backend selection.
…y snapshots

- Added live tracking and stale status warnings for agent activity snapshots.
- Fixed text wrapping with ANSI escape sequences to defer style open sequences after whitespace.
- Added VirtualRenderScheduler for deterministic virtual-clock rendering tests.
@pull pull Bot locked and limited conversation to collaborators Aug 16, 2026
@pull pull Bot added the ⤵️ pull label Aug 16, 2026
@pull
pull Bot merged commit 37eee71 into ericrihm:main Aug 16, 2026
4 checks passed
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.

Projects

None yet

Development

Successfully merging this pull request may close these issues.

0