[FFmpeg-devel] [PATCH 2/8] ffprobe/avtextformat: Rename and move writer functions and options

[Soft Works]

> -----Original Message-----
> From: ffmpeg-devel <ffmpeg-devel-bounces at ffmpeg.org> On Behalf Of Soft
> Works
> Sent: Freitag, 28. Februar 2025 02:31
> To: FFmpeg development discussions and patches <ffmpeg-devel at ffmpeg.org>
> Subject: Re: [FFmpeg-devel] [PATCH 2/8] ffprobe/avtextformat: Rename and
> move writer functions and options
> 
> 
> 
> > -----Original Message-----
> > From: ffmpeg-devel <ffmpeg-devel-bounces at ffmpeg.org> On Behalf Of
> > Michael Niedermayer
> > Sent: Freitag, 28. Februar 2025 02:23
> > To: FFmpeg development discussions and patches <ffmpeg-
> devel at ffmpeg.org>
> > Subject: Re: [FFmpeg-devel] [PATCH 2/8] ffprobe/avtextformat: Rename
> and
> > move writer functions and options
> >
> > On Thu, Feb 27, 2025 at 02:01:34PM +0000, softworkz wrote:
> > > From: softworkz <softworkz at hotmail.com>
> > >
> > > Signed-off-by: softworkz <softworkz at hotmail.com>
> > > ---
> > >  fftools/ffprobe.c                   | 910 ++++++-------------------
> --
> > -
> > >  libavutil/Makefile                  |   1 +
> > >  libavutil/avtextformat.h            |  50 ++
> > >  libavutil/textformat/avtextformat.c | 620 +++++++++++++++++++
> > >  4 files changed, 848 insertions(+), 733 deletions(-)
> > >  create mode 100644 libavutil/textformat/avtextformat.c
> >
> > this and patch 1 crash
> >
> > ==1574970== Invalid read of size 1
> > ==1574970==    at 0x483EF46: strlen (in /usr/lib/x86_64-linux-
> > gnu/valgrind/vgpreload_memcheck-amd64-linux.so)
> > ==1574970==    by 0x7683D14: __vfprintf_internal (vfprintf-
> > internal.c:1688)
> > ==1574970==    by 0x7696F99: __vsnprintf_internal (vsnprintf.c:114)
> > ==1574970==    by 0x1239AA1: av_vbprintf (in ffmpeg/ffprobe_g)
> > ==1574970==    by 0x1256473: av_log_default_callback (in
> > ffmpeg/ffprobe_g)
> > ==1574970==    by 0x1256162: av_log (in ffmpeg/ffprobe_g)
> > ==1574970==    by 0x125CB4B: av_opt_set_defaults2 (in
> ffmpeg/ffprobe_g)
> > ==1574970==    by 0x127611C: avtext_context_open (in ffmpeg/ffprobe_g)
> > ==1574970==    by 0x2F9672: main (in ffmpeg/ffprobe_g)
> > ==1574970==  Address 0xffffffff is not stack'd, malloc'd or (recently)
> > free'd
> > ==1574970==
> > ==1574970==
> > ==1574970== Process terminating with default action of signal 11
> > (SIGSEGV)
> > ==1574970==  Access not within mapped region at address 0xFFFFFFFF
> > ==1574970==    at 0x483EF46: strlen (in /usr/lib/x86_64-linux-
> > gnu/valgrind/vgpreload_memcheck-amd64-linux.so)
> > ==1574970==    by 0x7683D14: __vfprintf_internal (vfprintf-
> > internal.c:1688)
> > ==1574970==    by 0x7696F99: __vsnprintf_internal (vsnprintf.c:114)
> > ==1574970==    by 0x1239AA1: av_vbprintf (in ffmpeg/ffprobe_g)
> > ==1574970==    by 0x1256473: av_log_default_callback (in
> > ffmpeg/ffprobe_g)
> > ==1574970==    by 0x1256162: av_log (in ffmpeg/ffprobe_g)
> > ==1574970==    by 0x125CB4B: av_opt_set_defaults2 (in
> ffmpeg/ffprobe_g)
> > ==1574970==    by 0x127611C: avtext_context_open (in ffmpeg/ffprobe_g)
> > ==1574970==    by 0x2F9672: main (in ffmpeg/ffprobe_g)
> > ==1574970==  If you believe this happened as a result of a stack
> > ==1574970==  overflow in your program's main thread (unlikely but
> > ==1574970==  possible), you can try to increase the size of the
> > ==1574970==  main thread stack using the --main-stacksize= flag.
> > ==1574970==  The main thread stack size used in this run was 8388608.
> >
> > testcase: ./ffprobe_g -v 99 ~/tickets/3679/wgetmjpegdump
> >
> > tell me if this file is not in the ticket and ill send it to you
> 
> Hi Michael,
> 
> only patch 8/8 is expected to work. Since this is a RFC for the moment,
> I spared the time for making sure that every single commit builds.
> (please see cover letter notes in 0/8)
> 
> Once the direction is clear, I'll prepare a patchset with all commits
> working of course.
> 
> Thanks

Did you try 8/8 - does it pass for you? It's fine locally on x64 Linux, on AzDevops Linux x64 and on Patchwork andriy/make_fate_x86.
But what bothers me a bit is that it failed on the Chinese CPU system.

sw