only use HLS for live tv - otherwise use progressive mkv

[mueslimak3r]

Changes

• revert to using progressive mkv for everything but live tv which will still use HLS
• added a condition to getBufferedPosition() that excludes live-tv. The logic for getting the current position when playing live tv is kinda broken, especially if the player is paused and then resumed.

Issues

• using mkv again should resolve 0.13beta1 track transcoding bug #1410 though I'm not sure about live-tv contexts

• after seeking using HLS, the video would be slightly early but the timestamp wouldn't reflect it.
  This caused:

  • subtitles to be out of sync
  • the last few seconds of a video looping

  According to ffmpeg docs, the flag -noaccurate_seek, which jellyfin uses, supplies a stream starting at the key-frame prior to the requested timestamp. This is incompatible with exoplayer's frame-accurate seeking.
  The ffmpeg docs (https://trac.ffmpeg.org/wiki/Seeking):

  -ss is now also "frame-accurate" even when used as an input option. Previous behavior (seeking only to the nearest
  preceding keyframe, even if not precisely accurate) can be restored with the -noaccurate_seek option

  To resolve this and use HLS in contexts that allow seeking, the version of exoplayer we use will need to support SeekParameters for HLS. As far as I can tell, the flag we'd use is SeekParameters.PREVIOUS_SYNC
  The relevant issue is here:
  Support key-frame accurate seeking in HLS google/ExoPlayer#2882
  And here is the merged PR that should allow this once the next exoplayer release comes out:
  Implements SeekParameters.*_SYNC variants for HLS google/ExoPlayer#9536

Notes

• using HLS has made the player far more stable. If there's a way to synchronize the subtitles despite the player being unaware of the desync then that would be much better than abandoning HLS.

[mueslimak3r]

I'm leaving this as a draft pending this server issue since it seems to only happen with mkv, and while testing this PR I encountered it a bunch:
jellyfin/jellyfin#6957

I've been digging around for a client-side cause of this but haven't found anything