[codex] Gate Android Talk capture starts in background by NianJiuZst · Pull Request #98055 · openclaw/openclaw

NianJiuZst · 2026-06-30T06:25:00Z

What Problem This Solves

Android Talk push-to-talk capture can start microphone capture and move NodeForegroundService into the microphone foreground-service type. On modern Android, starting a microphone foreground service from the background is restricted, so remote talk.ptt.start / talk.ptt.once invocations should not start a new capture while the app is backgrounded.

Why This Change Was Made

This gates new Talk capture starts at the Android runtime boundary. talk.ptt.once remains a foreground-only dispatcher command, while talk.ptt.start is handled in NodeRuntime so idempotent retries for an already-active capture can still return the current capture id. TalkModeManager.beginPushToTalk(allowNewCapture = false) owns the final new-capture check to avoid a race between an external active-capture precheck and the actual begin operation.

User Impact

Background remote commands no longer try to create a new microphone capture path that Android can reject at foreground-service startup time. Existing active PTT sessions can still be stopped/cancelled, and ptt.start retries for an active session still reconcile to the same capture id.

Evidence

cd apps/android && ./gradlew :app:testThirdPartyDebugUnitTest --tests ai.openclaw.app.node.InvokeCommandRegistryTest --tests ai.openclaw.app.node.InvokeDispatcherTest --tests ai.openclaw.app.GatewayBootstrapAuthTest --tests ai.openclaw.app.voice.TalkModeManagerTest
git diff --check
.agents/skills/autoreview/scripts/autoreview --mode branch --base origin/main

Known unrelated local check gap:

cd apps/android && ./gradlew :app:ktlintCheck currently fails on untouched upstream files: apps/android/app/src/main/java/ai/openclaw/app/GatewayExecApprovals.kt and apps/android/app/src/main/java/ai/openclaw/app/ui/design/ClawSurfaces.kt.

clawsweeper · 2026-06-30T06:58:16Z

Codex review: needs maintainer review before merge. Reviewed July 2, 2026, 9:16 AM ET / 13:16 UTC.

Summary
The PR changes Android Talk push-to-talk handling so background starts cannot create a new microphone capture while preserving active-session retry, stop, and cancel behavior.

PR surface: Other +88. Total +88 across 9 files.

Reproducibility: yes. Current main routes background talk.ptt.start through prepareTalkCapture without a foreground gate before microphone foreground-service promotion, and the PR discussion includes before/after Android 16 emulator proof of that path.

Review metrics: 1 noteworthy metric.

Foreground command policy changed: 1 command metadata change, 1 runtime capture gate. This changes remote node.invoke behavior during upgrade, so maintainers should review the foreground-only policy intentionally.

Merge readiness
Overall: 🐚 platinum hermit
Proof: 🐚 platinum hermit ✨ media proof bonus
Patch quality: 🐚 platinum hermit
Result: ready for maintainer review.

Overall follows the weaker of proof and patch quality, so missing proof can cap an otherwise strong patch.

Rank-up moves:

none.

Risk before merge

[P1] The PR intentionally changes upgrade behavior for any workflow that relied on starting a new Android Talk PTT microphone capture while the app is backgrounded; green CI cannot decide whether maintainers accept that compatibility break.

Maintainer options:

Accept foreground-only Talk starts (recommended)
Merge once maintainers accept that new background Android PTT starts now fail fast instead of attempting microphone capture.
Pause for a compatibility mode
If background PTT starts are considered supported on any target Android path, pause and replace this with an explicit compatibility or API-level policy design.

Next step before merge

No automated repair is needed; the remaining action is maintainer acceptance and merge handling for the intentional compatibility change.

Security
Cleared: The diff changes Android runtime microphone gating, tests, and i18n line metadata only; it does not add dependencies, workflows, permissions, secret handling, or supply-chain execution paths.

Review details

Best possible solution:

Merge the exact-head foreground gate if maintainers accept foreground-only new Android Talk capture starts, keeping the active retry and stop/cancel paths routable as the patch does.

Do we have a high-confidence way to reproduce the issue?

Yes. Current main routes background talk.ptt.start through prepareTalkCapture without a foreground gate before microphone foreground-service promotion, and the PR discussion includes before/after Android 16 emulator proof of that path.

Is this the best way to solve the issue?

Yes. Gating new capture at NodeRuntime plus a final TalkModeManager allowNewCapture check is the narrow owner-boundary fix; marking only talk.ptt.once foreground-only preserves active-session retry and stop/cancel control.

AGENTS.md: found and applied where relevant.

Codex review notes: model internal, reasoning high; reviewed against b60e8c48382d.

Label changes

Label changes:

add proof: sufficient: Contributor real behavior proof is sufficient. A maintainer comment provides exact-head Android 16 emulator before/after screenshots and runtime observations for blocked new background capture, active retry, and stop release.
add proof: 📸 screenshot: Contributor real behavior proof includes screenshot evidence. A maintainer comment provides exact-head Android 16 emulator before/after screenshots and runtime observations for blocked new background capture, active retry, and stop release.
add rating: 🐚 platinum hermit: Overall readiness is 🐚 platinum hermit; proof is 🐚 platinum hermit and patch quality is 🐚 platinum hermit.
add status: 👀 ready for maintainer look: ClawSweeper has no concrete contributor-facing blocker left for this PR. Sufficient (screenshot): A maintainer comment provides exact-head Android 16 emulator before/after screenshots and runtime observations for blocked new background capture, active retry, and stop release.
remove rating: 🧂 unranked krab: Current PR rating is rating: 🐚 platinum hermit, so this older rating label is no longer current.
remove status: 📣 needs proof: Current PR status label is status: 👀 ready for maintainer look.

Label justifications:

P2: This is a focused Android runtime bug fix with limited blast radius and no crash-loop, data-loss, or cross-channel emergency signal.
merge-risk: 🚨 compatibility: The PR deliberately makes new background Android PTT capture starts fail fast, which can change existing remote-control workflows after upgrade.
rating: 🐚 platinum hermit: Overall readiness is 🐚 platinum hermit; proof is 🐚 platinum hermit and patch quality is 🐚 platinum hermit.
status: 👀 ready for maintainer look: ClawSweeper has no concrete contributor-facing blocker left for this PR. Sufficient (screenshot): A maintainer comment provides exact-head Android 16 emulator before/after screenshots and runtime observations for blocked new background capture, active retry, and stop release.
proof: sufficient: Contributor real behavior proof is sufficient. A maintainer comment provides exact-head Android 16 emulator before/after screenshots and runtime observations for blocked new background capture, active retry, and stop release.
proof: 📸 screenshot: Contributor real behavior proof includes screenshot evidence. A maintainer comment provides exact-head Android 16 emulator before/after screenshots and runtime observations for blocked new background capture, active retry, and stop release.

Evidence reviewed

PR surface:

Other +88. Total +88 across 9 files.

View PR surface stats

Area	Files	Added	Removed	Net
Source	0	0	0	0
Tests	0	0	0	0
Docs	0	0	0	0
Config	0	0	0	0
Generated	0	0	0	0
Other	9	103	15	+88
Total	9	103	15	+88

What I checked:

Repository policy read: Root AGENTS.md and scoped apps/android/AGENTS.md were read fully; root review guidance affected the proof, compatibility-risk, and owner-history checks, while the scoped Android release policy was not directly relevant to this code review. (AGENTS.md:1, b60e8c48382d)
Current main still has the bug path: On current main, talk.ptt.start and talk.ptt.once both route through runPreparedTalkPttCommand, and prepareTalkCapture promotes NodeForegroundService to TalkMode without first checking foreground state. (apps/android/app/src/main/java/ai/openclaw/app/NodeRuntime.kt:1244, b60e8c48382d)
Foreground service microphone path confirmed: NodeForegroundService.setVoiceCaptureMode starts the foreground service for TalkMode, and foregroundServiceTypesForVoiceMode adds FOREGROUND_SERVICE_TYPE_MICROPHONE for TalkMode. (apps/android/app/src/main/java/ai/openclaw/app/NodeForegroundService.kt:206, b60e8c48382d)
PR gates new background PTT starts at runtime: The PR makes background talk.ptt.start call beginPushToTalk(allowNewCapture = false), adds a foreground check before capture preparation, and keeps foreground starts on the prepared capture path. (apps/android/app/src/main/java/ai/openclaw/app/NodeRuntime.kt:1244, ae446c11c8e5)
Race-sensitive active retry handled in TalkModeManager: The PR makes TalkModeManager return an existing active capture when new capture is disallowed, while rejecting a background call that would create a new capture. (apps/android/app/src/main/java/ai/openclaw/app/voice/TalkModeManager.kt:246, ae446c11c8e5)
Dispatcher metadata keeps the intended split: The PR marks talk.ptt.once foreground-only while leaving talk.ptt.start, stop, and cancel routable to the runtime/manager state gate. (apps/android/app/src/main/java/ai/openclaw/app/node/InvokeCommandRegistry.kt:156, ae446c11c8e5)

Likely related people:

steipete: Peter Steinberger introduced Android Talk mode and the native node Talk handoff in the central runtime/dispatcher/manager paths, and also provided exact-head runtime proof on this PR. (role: feature introducer and recent PR verifier; confidence: high; commits: 86d897cfaa6a, 466f7183207d; files: apps/android/app/src/main/java/ai/openclaw/app/NodeRuntime.kt, apps/android/app/src/main/java/ai/openclaw/app/node/InvokeDispatcher.kt, apps/android/app/src/main/java/ai/openclaw/app/voice/TalkModeManager.kt)
joshavant: Recent work on Android microphone foreground-service handling touched the service/runtime boundary that this PR hardens. (role: recent foreground-service contributor; confidence: high; commits: 8ab36e4308df; files: apps/android/app/src/main/java/ai/openclaw/app/NodeForegroundService.kt, apps/android/app/src/main/java/ai/openclaw/app/NodeRuntime.kt)
Ayaan Zaidi: Recent Android Talk/realtime and foreground-service work touched TalkModeManager, NodeForegroundService, and adjacent voice tests used to validate this behavior. (role: recent Android Talk area contributor; confidence: medium; commits: 70614f88cc05, e52a3b31e47d, 6d7eb9bb8483; files: apps/android/app/src/main/java/ai/openclaw/app/voice/TalkModeManager.kt, apps/android/app/src/test/java/ai/openclaw/app/voice/TalkModeManagerTest.kt, apps/android/app/src/main/java/ai/openclaw/app/NodeForegroundService.kt)

What the crustacean ranks mean

🦀 challenger crab: rare, exceptional readiness with strong proof, clean implementation, and convincing validation.
🦞 diamond lobster: very strong readiness with only minor maintainer review expected.
🐚 platinum hermit: good normal PR, likely mergeable with ordinary maintainer review.
🦐 gold shrimp: useful signal, but proof or patch confidence is still limited.
🦪 silver shellfish: thin signal; proof, validation, or implementation needs work.
🧂 unranked krab: not merge-ready because proof is missing/unusable or there are serious correctness or safety concerns.
🌊 off-meta tidepool: rating does not apply to this item.

Shiny media proof means a screenshot, video, or linked artifact directly shows the changed behavior. Runtime, network, CSP, and security claims still need visible diagnostics.

How this review workflow works

ClawSweeper keeps one durable marker-backed review comment per issue or PR.
Re-runs edit this comment so the latest verdict, findings, and automation markers stay together instead of adding duplicate bot comments.
A fresh review can be triggered by eligible @clawsweeper re-review comments, exact-item GitHub events, scheduled/background review runs, or manual workflow dispatch.
PR/issue authors and users with repository write access can comment @clawsweeper re-review or @clawsweeper re-run on an open PR or issue to request a fresh review only.
Maintainers can also comment @clawsweeper review to request a fresh review only.
Fresh-review commands do not start repair, autofix, rebase, CI repair, or automerge.
Maintainer-only repair and merge flows require explicit commands such as @clawsweeper autofix, @clawsweeper automerge, @clawsweeper fix ci, or @clawsweeper address review.
Maintainers can comment @clawsweeper explain to ask for more context, or @clawsweeper stop to stop active automation.

steipete · 2026-07-02T13:03:33Z

Maintainer verification on exact head ae446c11c8e565d13188bd9770e8c8e2f1bd16a8:

Made Talk capture policy explicit: talk.ptt.once is foreground-only; talk.ptt.start may reconcile an already-active capture in background but cannot create a new one; stop/cancel remain routable.
Passed 76 focused registry/dispatcher/runtime/manager tests, Android ktlint, native i18n inventory, debug APK assembly, git diff --check, and fresh exact-tree autoreview (no actionable findings).
Exact-head hosted CI is green: 63 passed, 12 skipped.

Live Android 16 emulator proof:

Before (current main): background talk.ptt.start succeeded, promoted the existing service to microphone type, started recognition, and left RECORD_AUDIO running.
After: the same new background start returned NODE_BACKGROUND_UNAVAILABLE; no recognizer start, microphone service transition, running AppOps record, or privacy chip.
Active retry: a foreground-started capture remained addressable from background. Retry returned the existing capture without a second recognizer start, and background stop released the microphone.

This matches Android's current while-in-use foreground-service restriction while preserving lifecycle control for an already-running capture.

Before: background start records	After: new background start blocked	After: active capture retry still works

steipete · 2026-07-02T13:22:30Z

Merged via squash.

Prepared head SHA: ae446c11c8e565d13188bd9770e8c8e2f1bd16a8
Landed commit: e3f46d0926da2192a8d238781aa7f70e74c499e7

* fix: gate Android talk capture starts * refactor(android): make Talk capture policy explicit --------- Co-authored-by: NianJiuZst <180004567+users.noreply.github.com> Co-authored-by: Peter Steinberger <[email protected]>

openclaw-barnacle Bot added app: android App: android size: S labels Jun 30, 2026

NianJiuZst marked this pull request as ready for review June 30, 2026 06:54

steipete self-assigned this Jul 2, 2026

NianJiuZst and others added 2 commits July 2, 2026 14:01

fix: gate Android talk capture starts

3b7947a

refactor(android): make Talk capture policy explicit

ae446c1

steipete force-pushed the codex/android-talk-foreground-gate branch from 29ce660 to ae446c1 Compare July 2, 2026 13:02

steipete merged commit e3f46d0 into openclaw:main Jul 2, 2026
105 of 112 checks passed

github-actions Bot mentioned this pull request Jul 2, 2026

📡 Upstream Digest — 2026-07-02 13:41 UTC curtismercier/openclaw-mods#1189

Open

clawsweeper Bot mentioned this pull request Jul 2, 2026

Security: harden gateway timeouts and auth store sealing #39059

Closed

18 tasks

xialonglee mentioned this pull request Jul 4, 2026

fix(android): stop talk mode audio capture on background #99840

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Uh oh!

[codex] Gate Android Talk capture starts in background#98055

[codex] Gate Android Talk capture starts in background#98055
steipete merged 2 commits into
openclaw:mainfrom
NianJiuZst:codex/android-talk-foreground-gate

NianJiuZst commented Jun 30, 2026

Uh oh!

clawsweeper Bot commented Jun 30, 2026 •

edited

Loading

Uh oh!

steipete commented Jul 2, 2026 •

edited

Loading

Uh oh!

Uh oh!

steipete commented Jul 2, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Uh oh!

Uh oh!

Conversation

NianJiuZst commented Jun 30, 2026

What Problem This Solves

Why This Change Was Made

User Impact

Evidence

Uh oh!

clawsweeper Bot commented Jun 30, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

steipete commented Jul 2, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

steipete commented Jul 2, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

clawsweeper Bot commented Jun 30, 2026 •

edited

Loading

steipete commented Jul 2, 2026 •

edited

Loading