Fix app crash when ffmpeg subprocess exits unexpectedly.
FfmpegJob's stderr / stdout reader threads were spun up without a try/catch around the blocking read() loop. When the ffmpeg subprocess exited unexpectedly — cancel() closing the stream from another thread, or Android killing libffmpeg.so as a PhantomProcessRecord during a long transcode — read() threw InterruptedIOException and that propagated as an uncaught exception on a daemon thread, which on Android is a FATAL EXCEPTION that takes down the whole process.
Observed while HLS-streaming a 4K HEVC Main 10 source (1 GB, 204 s). ffmpeg ran for about a minute, Android's phantom-process policy reaped libffmpeg.so, two FATAL EXCEPTION traces hit logcat, and Haven crashed mid-stream.
Both read loops now catch IOException quietly. Job status is still reflected through process.exitValue() and await() picks it up normally — the only visible change is that an unexpected subprocess exit ends the job cleanly instead of killing the app.
Still pending
The underlying phantom-process kill itself isn't addressed by this patch. Android 13+ reaps long-running subprocesses that aren't tied to a foreground service. For large-media HLS streams and transcodes, a FOREGROUND_SERVICE_MEDIA_PROCESSING service is the proper fix — queued as a follow-up.
Full Changelog: v5.24.15...v5.24.16