[FFmpeg-cvslog] bink: no need to increase width twice

[Reimar Döffinger]

On Wed, Mar 21, 2012 at 09:23:41PM +0100, Michael Niedermayer wrote:
> On Wed, Mar 21, 2012 at 08:20:45PM +0100, Reimar Döffinger wrote:
> > On Wed, Mar 21, 2012 at 01:11:00PM +0100, Michael Niedermayer wrote:
> > > On Wed, Mar 21, 2012 at 07:58:25AM +0100, Reimar Döffinger wrote:
> > > > 
> > > > 
> > > > On 21 Mar 2012, at 04:33, git at videolan.org (Paul B Mahol) wrote:
> > > > 
> > > > > ffmpeg | branch: master | Paul B Mahol <onemda at gmail.com> | Wed Mar 21 01:41:50 2012 +0000| [594a3d631583a2f086fac25db6fb3542fd0fb187] | committer: Michael Niedermayer
> > > > > 
> > > > > bink: no need to increase width twice
> > > > > 
> > > > > Signed-off-by: Paul B Mahol <onemda at gmail.com>
> > > > > Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
> > > > > 
> > > > >> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=594a3d631583a2f086fac25db6fb3542fd0fb187
> > > > > ---
> > > > > 
> > > > > libavcodec/bink.c |    2 +-
> > > > > 1 files changed, 1 insertions(+), 1 deletions(-)
> > > > > 
> > > > > diff --git a/libavcodec/bink.c b/libavcodec/bink.c
> > > > > index c8daab3..4ca96e5 100644
> > > > > --- a/libavcodec/bink.c
> > > > > +++ b/libavcodec/bink.c
> > > > > @@ -150,7 +150,7 @@ static void init_lengths(BinkContext *c, int width, int bw)
> > > > > 
> > > > >     c->bundle[BINK_SRC_BLOCK_TYPES].len = av_log2((width >> 3) + 511) + 1;
> > > > > 
> > > > > -    c->bundle[BINK_SRC_SUB_BLOCK_TYPES].len = av_log2(((width + 7) >> 4) + 511) + 1;
> > > > > +    c->bundle[BINK_SRC_SUB_BLOCK_TYPES].len = av_log2((width >> 4) + 511) + 1;
> > > > > 
> > > > 
> > > > Huh? Those don't give the same result when width is e.g. 9.
> > > > At least the commit message is really unclear.
> > > 
> > > it was a bugfix by someone, kostya later fixed it differently by
> > > rounding width up before this code so this become unneeded
> > 
> > Ok, but I'll appreciate it if you can remember to put that kind of
> > thing in the commit message.
> 
> I would have, had i realized the issue at the time of applying the
> patch ...
> 
> Maybe you could help here and keep an eye out for patches with too terse
> commit messages and reply to them if you spot one ?

I'm afraid I'm currently mostly down to reading only -cvslog properly
due to lack of time. But I'll try to at least write something when
I happen so see one before.
And there's still hope that one day they'll fix git to provide an
acceptable solution to this issue (and no, git notes counts as crap,
not as a solution).