[FFmpeg-devel] [PATCH] Why does this break FATE?

Soft Works softworkz at hotmail.com
Thu Sep 9 20:51:42 EEST 2021



> -----Original Message-----
> From: ffmpeg-devel <ffmpeg-devel-bounces at ffmpeg.org> On Behalf Of
> Andreas Rheinhardt
> Sent: Thursday, 9 September 2021 11:50
> To: ffmpeg-devel at ffmpeg.org
> Subject: Re: [FFmpeg-devel] [PATCH] Why does this break FATE?
> 
> Soft Works:
> > Test seek-lavf-asf failed. Look at tests/data/fate/seek-lavf-
> asf.err for details.
> > make: *** [/build/ffmpeg-git/tests/Makefile:256: fate-seek-lavf-
> asf] Error 139
> >
> > $ cat tests/data/fate/seek-lavf-asf.err
> > /build/ffmpeg-git/tests/fate-run.sh: line 78: 21786 Segmentation
> fault      $target_exec $target_path/"$@"
> >
> >
> > It's the same on both Windows/MSYS2 and Linux. Let's see how
> patchwork results will be...
> >
> >
> > Signed-off-by: softworkz <softworkz at hotmail.com>
> > ---
> >  libavcodec/version.h | 6 +++---
> >  1 file changed, 3 insertions(+), 3 deletions(-)
> >
> > diff --git a/libavcodec/version.h b/libavcodec/version.h
> > index 83db2b242a..d162607f4b 100644
> > --- a/libavcodec/version.h
> > +++ b/libavcodec/version.h
> > @@ -27,9 +27,9 @@
> >
> >  #include "libavutil/version.h"
> >
> > -#define LIBAVCODEC_VERSION_MAJOR  59
> > -#define LIBAVCODEC_VERSION_MINOR   7
> > -#define LIBAVCODEC_VERSION_MICRO 102
> > +#define LIBAVCODEC_VERSION_MAJOR  60
> > +#define LIBAVCODEC_VERSION_MINOR   0
> > +#define LIBAVCODEC_VERSION_MICRO 100
> >
> >  #define LIBAVCODEC_VERSION_INT
> AV_VERSION_INT(LIBAVCODEC_VERSION_MAJOR, \
> >
> LIBAVCODEC_VERSION_MINOR, \
> >
> It segfaults because asf_read_pts() uses an uninitialized stack
> packet
> which av_read_frame() unrefs as soon as FF_API_INIT_PACKET is 0
> (which
> your bump did). (And actually you could have easily found this out
> yourself by just using a debugger.)

I don't have a setup for debugging tests (yet), but I should have 
looked a bit down in version.h
Due to the fact that it was just a single test failing, I hadn't 
expected a cause at a global level like it turned out to be.

Thanks for analysing,
softworkz




More information about the ffmpeg-devel mailing list