Break infinite loop in FFMpeg::read_frame()

[dankamongmen]

Description

Codecs which set the CODEC_CAP_DELAY capability led to infinite
loops during reading (see #2571). Once AVERROR_EOF is returned,
no more packets are going to be received, regardless of internal
buffering. break out of the loop in that case.

I've verified that this resolves the livelock following decoding
of the first packet in the Notcurses AVI test case which
illustrated this problem. I've also verified that said AVI
advertises the CODEC_CAP_DELAY attribute, matching the proposed
fix. In addition, I verified that OIIO continues to decode all
my other test cases without trouble. I hand-inspected the media,
and verified that the frame counts were precisely correct.

It is legal to pass NULL/0 into the decoder regardless of
CODEC_CAP_DELAY, and thus this whole conditional can really
be eliminated (I have no mention of it in my FFmpeg wrappers).
I haven't bothered to do that here, but it's a thought for the
future. Closes #2571.

Tests

I haven't added any, but you're welcome to test with

https://github.com/dankamongmen/notcurses/blob/master/data/samoa.avi

This failed with oiiotool -a --dumpdata before, but does not now. Instead, we get all expected frames, followed by successful termination.

Checklist:

• I have read the contribution guidelines.
• If this is more extensive than a small change to existing code, I
  have previously submitted a Contributor License Agreement
  (individual, and if there is any way my
  employers might think my programming belongs to them, then also
  corporate).
• I have updated the documentation, if applicable.
• I have ensured that the change is tested somewhere in the testsuite
  (adding new test cases if necessary).
• My code follows the prevailing code style of this project.

This transitively resolves dankamongmen/notcurses#547.

[lgritz]

Awesome! Thanks for tracking this down. Definitely something that would have taken me a long, long time with my lack of familiarity with ffmpeg API.

[dankamongmen]

I've got your back, homie =]. Since you're part of my critical infrastructure on at least Fedora, I'm hoping to do a deep dive through src/ffmpeg.imageio sometime this month doing something of an audit. FFmpeg has some definite gotchas. I also hope to improve video decoding performance, which (as I alluded to in an earlier bug) is kinda surprisingly bad, no offense. Using my FFmpeg backend directly, I'm able to decode frames quickly enough that I need carefully wait for frame timings to move on; using OIIO, the videos I'm testing with require ~10x realtime to decode, with sustained 100% CPU burn. I'll get on top of it for ya. =]

[lgritz]

That would be great. It was contributed several years back with the main aim being just being able to pull individual frames out, not with speed. It has received little attention since then -- just the usual touch-ups for new ffmpeg releases, new compiler releases with new warnings, that kind of thing. Definitely needs a good scrubbing.