[FFmpeg-user] 0.10.3 to 0.11 performance regression?

[Carsten Mattner]

On Sun, Jun 3, 2012 at 2:17 PM, Carl Eugen Hoyos <cehoyos at ag.or.at> wrote:
> Carsten Mattner <carstenmattner <at> googlemail.com> writes:
>
>> Another question: when bisecting I have constantly hit git
>> revisions where the build results were avconv other
>> executabled instead of ff* executables. Is that due to
>> temporary merge commits to/from libav.git?
>
> That's because of the merges, yes. (What is temporary about them?)
> I always start with merge commits to find regressions, that is
> easier imo.

Temporary because (not having inspected the merge commits)
I think after the merge is done the existing build infrastructure
for outputting ff* executables will be put in place and therefore
end what I called "temporary". I'm sorry if that was confusing.

> Command line and complete, uncut console output missing.

That was the full command line while replacing filenames
and bitrate with placeholders.

I've extracted the important seeming parts of console output
so that the mail doesn't get even more verbose.

The only other interesting thing may be that ffmpeg is configured
with --enable-libmp3lame. otherwise it's just a random linux
system.

I can mail the full log to anyone interested but had assumed
(maybe falsely) that the behavior difference would be easy to
spot. Sorry.