[vlc-devel] [PATCH 0/3] Fixing input/decoder/sout deadlock with chromecast

Thomas Guillem thomas at gllm.fr
Tue Feb 13 08:50:46 CET 2018


> This has ALL the flags of a patch that must not be merged to a stable branch:
> - It is rushed.

I didn't push it, I proposed the patch, I thought about it a lot and I concluded that my PTS_DELAY hack was really bad. I don't see the rush.

> - It is badly designed.

I'm using the API you created, for me your API and my usage is well desgned.

> - It affects a central piece of code.

Maybe, the patch 1 can have an impact I didn't see. As I say, we can limit it to the sout case. But I don't need the patch 1 anymore (I still believe it's a correct patch)

> - It is invasive.

Yes, the new lock from the i11e_register are invasive.

> - It has no proper tests, and affects a piece of code without proper tests.

Like 80% of code we push.

> - It comes not only after the freeze, but even after the release.

Did you really expect that the VLC 3.0 was a one shot release ?

> Except for one:
> - It does not change the source and binary interfaces.


More information about the vlc-devel mailing list