On Wed, Oct 7, 2015 at 6:10 PM, Derek Buitenhuis <derek.buitenh...@gmail.com> wrote: > On 10/7/2015 4:57 PM, wm4 wrote: >> A static variable in a header file? This doesn't look right. You should >> define it in h264.c instead. > > [17:03] <+wm4> Daemon404: what exactly is the point of declaring the > init_once in the .h file? > [17:03] <@Daemon404> i didnt know the right place to put it > [17:03] <+wm4> h264.c > [17:03] <@Daemon404> there's no obviously correct place > [17:04] * Daemon404 will send the libav patchset once rsync actually > finishes and he can run fate > [17:06] <+wm4> there needs to be exactly one instance of this variable, so it > should be in h264.c, unless > h264.c is somehow used for freaky preprocessor template stuff > (which I think it isn't) > [17:07] <@Daemon404> and the function it uses isnt from h264.c, which is why > i had put it in the .h > [17:07] <@Daemon404> i have no strong opinion > > Currently moved to h264.c locally. >
You could put it right above the pthread_once call, inside the function even, as long as there is only one place its ever called. But we should probably come to a more general consensus for functions that may be called in more than one place, where the variable may even need to be referenced in a header somewhere so two files can reference it - or you create a new wrapper that wraps the pthread_once for those cases. - Hendrik _______________________________________________ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
