1 May
2023
1 May
'23
5:49 a.m.
Patch 4/9 seems underdocumented at least. It's not how wmvcore or quartz operate, and it's not immediately clear just from reading the patch why it's necessary or an improvement.
Patch 5/9 is a bit suspicious. Those calls to wg_parser_stream_get_buffer() will eventually exit anyway, where "eventually" should in practice never be more than 30 ms or so. Interrupting that might seem like it'll help the parser exit faster (although it's not clear just from reading the patch whether that's your goal or not), but in actuality that delay will just get pushed to when the GStreamer element is destroyed.
This is a bit of a large series to review all at once, so I haven't reviewed patches 6-9 yet. I'll try to get to them quickly...