fix(engine): stop recognition before switching engines to prevent segfault (fixes #350) (#355)

When switching from whisper.cpp to VOSK via the settings dialog, the
_auto_apply_settings path called reconfigure() without first stopping
active recognition. This allowed background audio/transcription threads
to continue using the old engine's native C resources (whisper.cpp model)
while they were being freed by Python's GC, causing a segmentation fault.

Changes:
- Add stop_recognition() guard in reconfigure() itself when the engine
  or model changes and recognition is active — this makes the API safe
  regardless of which caller triggers the reconfiguration
- Add defense-in-depth stop_recognition() call in _auto_apply_settings
  to match the existing pattern in _apply_settings_internal

src/vocalinux/speech_recognition/recognition_manager.py

@@ -2073,6 +2073,15 @@ class SpeechRecognitionManager:
 
         if restart_needed:
             logger.info("Engine or model changed, re-initializing...")
+
+            # Stop any active recognition before switching engines.
+            # This is critical to prevent segfaults when the old engine's
+            # native resources (e.g. whisper.cpp C model) are freed while
+            # a background thread is still using them.
+            if self.state != RecognitionState.IDLE:
+                logger.info("Stopping active recognition before engine switch...")
+                self.stop_recognition()
+
             # When reconfiguring from UI, allow downloads
             old_defer = self._defer_download
             self._defer_download = not force_download

src/vocalinux/ui/settings_dialog.py

@@ -1916,6 +1916,13 @@ class SettingsDialog(Gtk.Dialog):
 
             self.config_manager.update_speech_recognition_settings(settings)
             self.config_manager.save_settings()
+
+            # Stop recognition before reconfiguring to avoid segfaults when
+            # switching between engines with native resources (e.g. whisper.cpp)
+            was_running = self.speech_engine.state != RecognitionState.IDLE
+            if was_running:
+                self.speech_engine.stop_recognition()
+
             self.speech_engine.reconfigure(**settings)
             logger.info("Settings auto-applied successfully")
         except Exception as e: