fix: preserve STT EOS timestamp anchor

[nightcityblade]

Summary

• keep the previous STT end-time anchor when an END_OF_SPEECH event has no timestamp
• add a regression test for no-VAD STT turn detection preserving the final transcript timestamp

Fixes #5876

Testing

• uv run pytest tests/test_agent_session.py -k "stt_eos_without_timestamp_preserves_previous_stt_anchor or stt_eos"
• uv run ruff check livekit-agents/livekit/agents/voice/audio_recognition.py tests/test_agent_session.py
• uv run ruff format --check livekit-agents/livekit/agents/voice/audio_recognition.py tests/test_agent_session.py

[CLAassistant]

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.

---

nightcityblade seems not to be a GitHub user. You need a GitHub account to be able to sign the CLA. If you have already a GitHub account, please add the email address used for this commit to your account.
_{You have signed the CLA already but the status is still pending? Let us recheck it.}

[devin-ai-integration]

✅ Devin Review: No Issues Found

Devin Review analyzed this PR and found no potential bugs to report.

View in Devin Review to see 2 additional findings.