[FFmpeg-devel] [PATCH 2/6] Implement IOCTL_ERROR macro and use it.
Mon Apr 26 01:20:39 CEST 2010
On date Monday 2010-04-26 00:36:12 +0200, Luca Abeni encoded:
> On Mon, 2010-04-26 at 00:26 +0200, Michael Niedermayer wrote:
> > > > > I am not sure if this kind of changes really improve the readability.
> > > >
> > > > I generalized the previous patch, now we have a log_errno() function
> > > > which is used whenever we need to print the string corresponding to an
> > > > errno.
> > >
> > > If we really want to have a log_errno(), then I think it should not be
> > > private to v4l2.c (I expect that libav* contains some other code that
> > > want to print error messages followed by an errno description).
> > >
> > > I suspect it could go to libavutil/log.c... If Michael approves it, then
> > > I am ok with the av_log() -> log_errno() contained in this patch.
> > you are trying to sidestep saying the patch is uhm by pushing it my
> > direction?
> No, sorry... :)
> I wanted to say that _if_ we want this kind of change, then log_errno()
> should not be a v4l2.c feature. But since I do not maintain log.c, I
> cannot approve putting log_errno() there.
> So, I think there are various questions:
> 1) Is log_errno() useful/wanted? I am undecided about this (other
> people's opinions are welcome)
I believe it is useful, and indeed I'm going to use it in other files
> 2) If the answer to 1) is yes, where should log_errno() go? My opinion
> is "not in v4l2.c". I think log.c is more suitable.
If we don't want to make it public (and there are good reasons for
that), then it may go in some internal header (for example
libavutil/internal.h) and be called ff_log_errno().
> 3) If putting log_errno() in log.c is ok, would this implementation be
> acceptable for the maintainer? (looks ok to me, but I do not maintain
> log.c... This is why I mentioned you).
FFmpeg = Freak and Fancy Minimalistic Puristic Elastic Game
More information about the ffmpeg-devel