[FFmpeg-devel] [PATCH] seek print NOPTS

Reimar Döffinger Reimar.Doeffinger
Tue Oct 20 21:06:24 CEST 2009


On Tue, Oct 20, 2009 at 04:34:32PM -0200, Ramiro Polla wrote:
> Calm down, this is not worth flaming for...

No, but I'd still like to make some comments, in the hope
to get some frustrated Windows developers to be a bit less
frustrated and alone :-)

> But it's true around here that mentioning "Windows" anywhere blinds
> everyone into thinking patches are worthless, even if the issue was
> just discovered on Windows and is also reproducible on Linux, and even
> if the patch has more uses than just fixing an issue on Windows.

First, I am not convinces that this is necessarily just "Windows", I think
"BSD" or "glibc" can reap fairly similar responses. If you want proof,
I think bktr.c still does not compile on some BSDs because the hacks just
got too much (well, or maybe it is just my fault for speaking up against those
patches), and that's a BSD-only file...
But, as much as most of us probably dislike marketing, I strongly recommend
to emphasize first any other issues your patch improves (possibly searching
for those or doing other things that make it fit more overall) and only "in passing"
mention it fixes OS-specific issues.
In part because we all are humans, in part to avoid flaming, but in particular
for Windows-specific patches because it helps reviewing _a lot_: Chances are,
whoever reviews it doesn't have/use Windows, so can't test that.
If there are enough other reasons to do it like that, that is a non-issue,
if not there almost certainly will be a lot more "clueless discussion" (if
I maintain it that usually means it will end up ignored until I get to personally
test it in Windows and confirm the issue, its description and that it is not
yet another of those cases where the "proper way" is just particularly well
hidden - which I consider quite common for Windows in general and on Linux
at least for audio output stuff).

> Anyways the patch is in already, there's no reason to continue this thread...

Sorry for continuing the thread anyway, I just wanted to show that there
are more than a few possible viewpoints on this (and these kind of things
in general).



More information about the ffmpeg-devel mailing list