fix(voice): 30 s and Manual outlast silent Web Speech runs

_onWebSpeechEnd counted every run that ended without a result toward the
two-retry cap, however long the run lasted. Under 30 s or Manual, each
no-speech window of the engine is one silent run, so the third ended the
recording at about 18 s of silence, before the deadline, and Manual
stopped on its own. Both recognition.start() sites now stamp the run start;
an empty run counts toward the cap only when it ended within a second of
starting (a recognizer that cannot listen), and any other run resets it.
The 250 ms backoff is unchanged.

The vm harness in test/voice-silence-timeout.test.ts now passes the test's
Date into the context, whose own Date ignores fake timers.

Also from review: the VoiceInput stubs of the four settings-ui.js tests
carry SILENCE_TIMEOUT_SECONDS; Settings-Reference lists Stop after silence;
the touchscreen Linux fallback leaves out ChromeOS and says a touch Linux
laptop still matches; the dead _webSpeechStream teardown is gone; the two
voice tests use vitest expect instead of node:assert.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

.changeset/quiet-voice-pauses.md

@@ -6,4 +6,4 @@ Add a configurable voice silence timeout and manual-stop dictation so thinking p
 
 Provider-default Deepgram and Web Speech final events now wait for the 3-second silence deadline instead of stopping immediately. Claude keeps recording across utterance endpoints and preserves final words after Stop, with a bounded wait on failed connections.
 
-Fix repeated words and lost final transcripts on Chrome for Android, animate browser speech activity without a competing microphone capture, and preserve partial text when reconnecting. Bound repeated empty browser recognition restarts.
+Fix repeated words and lost final transcripts on Chrome for Android, animate browser speech activity without a competing microphone capture, and preserve partial text when reconnecting. Bound restarts of a browser recognizer that keeps ending instantly without a result, while silent runs of normal length keep 30-second and Manual dictation going.