[FFmpeg-devel] Politics

Daniel Cantarín canta at canta.com.ar
Sun Dec 12 17:29:56 EET 2021


 > yesterday, it happened for the 4th and 5th times that another developer
 > called my patchset a “hack”.

Hope it wasn't me. If I did, I'm sorry, didn't wanted to imply bad code
or lack of skills, or anything: I was trying to understand if there's some
way to actually unblock your patchset. And the quicker the way, the
better.


 > subtitle_pts

LOL... that's some elegant answer.

I have my own doubts about subtitle_pts, but that's exactly why I
wouldn't call it "hack": because I don't fully grasp it yet.
I've read some devs arguments (not all) against it, but it also
wasn't clear for me back then why the strong backlash. Some
seemed just conservative and/or idealists, others didn't seem to
value live-streaming as a first class citizen between use cases.
Can't tell if they were right or wrong, or myself at reading them.
But mixed feelings for sure.

Howerver, *if* subtitle_pts is truly the last point against softworkz
patchset, then it's very close to see the light, and that's exciting.

I have a strong bias towards softworkz code, because I value
non-breaking newer use cases over quality of design. This is what
I see:
   - The problems softworkz mentions with live subtitles are VERY real.
   - I can't defend subtitle_pts, at least yet, because I don't get it.
   - If it solves live-streaming subtitles problems, and doesn't break
     anything else, it's fine by me.

But if there's some consensus between established community
figures against that property and that blocks the release, well...
removing the property sounds like the quickest path. Will test it
during week.


Now, the question remains: are we sure subtitle_pts is the last
barrier here?




More information about the ffmpeg-devel mailing list