[FFmpeg-devel] [PATCH] define _BSD_SOURCE for bktr.c

[Reimar Döffinger]

On Sat, Dec 13, 2008 at 10:05:25PM +0000, Jacob Meuser wrote:
> On Sat, Dec 13, 2008 at 10:04:32PM +0100, Reimar D?ffinger wrote:
> > On Sat, Dec 13, 2008 at 08:49:29PM +0000, Jacob Meuser wrote:
> > > On Sat, Dec 13, 2008 at 01:07:15PM +0100, Reimar D?ffinger wrote:
> > > > I don't think patch-libavcodec_dsputil_c is in any way acceptable,
> > > > actually it is a reason to reject bug reports from OpenBSD users.
> > > 
> > > without that, there would be no bug reports for ffmpeg, because
> > > ffmpeg would not be usable.
> > 
> > So far it hasn't crashed, and it runs almost all files from the
> > FATE-testsuite just fine.
> 
> ok, then something somewhere got fixed.  thanks for pointing this out.

Well, it might just be good luck on my part, and I am only testing OpenBSD 4.4...
Btw. these are the configure options I used:
And the only remaining issues I am aware if are
1) need to set LD_LIBRARY_PATH if running regression tests with the libs
outside the usual linker paths.
2) quite a few of the regression tests fail with wrong md5 and wrong
filesize (note that I am running under VirtualBox, which might cause some
bugs, when I had AMD-V support disabled I got crashes all the time).
3) endianness detection does not work due to "od" being incompatible
4) as on (IIRC) OSX, cavs DSP functions do not compile with PIC enabled.
5) the cabac asm code will not compile on amd64, because it uses 32 bit
registers in address calculation, which as 2.15 can not handle (no idea
why as versions >= 2.18 can handle it...)

The used configuration was
configure --disable-decoder=cavs --disable-debug --enable-gpl --enable-postproc
--enable-avfilter --enable-avfilter-lavf --enable-shared --enable-swscale
--nm=true

Greetings,
Reimar D?ffinger