{"side": "before", "runtimeSourceSHA": "53bcb9af7a37dd0b4df11a7d6fa0106e9e5b7356", "testSourceSHA": "17b58caf24f000852f1da7893ae0a21b64f828a4", "cwd": "/home/ubuntu/wt/pr-33624-before/apps/server", "command": "TMPDIR=/tmp/lead-33624-r1/evidence-capture/tmp vp test run src/speech/native.test.ts --reporter=verbose", "nativeSourceSHA256": "7c92949426e3f4ff4d54a850851d151b3233bed8b63b5c2a0cad335b656c0897", "testSourceSHA256": "b0f03762bcb7c10dacead6084922a8a8f8e573625c1e31a8158301a1168cbd08", "isolation": "owned before checkout only; exact source snapshots swapped and restored; shared after checkout untouched"}
note: You are running `vp test` as a Vite+ built-in command. If you meant to run the test npm script, use `vpr test` instead.

 RUN  v5.0.5 /home/ubuntu/wt/pr-33624-before/apps/server

 ✓ src/speech/native.test.ts > returns transcription from an isolated process and disposes it 85ms
 ✓ src/speech/native.test.ts > probes generic vocabulary support and delivers ordered terms to offline inference 91ms
 ✓ src/speech/native.test.ts > ignores IPC messages that do not belong to the speech protocol 43ms
 ✓ src/speech/native.test.ts > terminates hung native inference when the service shuts down 77ms
 ✓ src/speech/native.test.ts > returns streaming previews and final text from an isolated process 45ms
 ✓ src/speech/native.test.ts > kills a hung streaming feed without waiting for native cleanup 48ms
 × src/speech/native.test.ts > skips process exit teardown when the backend dies during native inference 63ms
   → orphaned worker ran exit teardown while native inference was in flight: expected true to be false // Object.is equality
 ✓ src/speech/native.test.ts > filters silence and preserves partial speech through the production worker and resets between recordings 133ms
Silero silence filtering unavailable; preserving raw recording and using bounded ASR chunks.
 ✓ src/speech/native.test.ts > decodes complete stable chunks before Stop and returns the ordered final tail 51ms
Silero silence filtering unavailable; preserving raw recording and using bounded ASR chunks.
Silero silence filtering unavailable; preserving raw recording and using bounded ASR chunks.
 ✓ src/speech/native.test.ts > preserves withheld audio on VAD failure while keeping ASR failures distinct 227ms

⎯⎯⎯⎯⎯⎯⎯ Failed Tests 1 ⎯⎯⎯⎯⎯⎯⎯

 FAIL  src/speech/native.test.ts > skips process exit teardown when the backend dies during native inference
AssertionError: orphaned worker ran exit teardown while native inference was in flight: expected true to be false // Object.is equality

- Expected
+ Received

- false
+ true

 ❯ src/speech/native.test.ts:186:7
    184|       NodeFS.existsSync(marker),
    185|       "orphaned worker ran exit teardown while native inference was in…
    186|     ).toBe(false);
       |       ^
    187|   } finally {
    188|     NodeFS.rmSync(directory, { recursive: true, force: true });

⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯[1/1]⎯


 Test Files  1 failed (1)
      Tests  1 failed | 9 passed (10)
   Start at  01:29:44
   Duration  1.50s (tests 68%, setup 27%, transform 3%, import 1%)


exit_status=1
