On Wed, Apr 5, 2017 at 5:40 PM, wm4 <nfx...@googlemail.com> wrote: > On Wed, 5 Apr 2017 17:22:53 +0200 > Diego Biurrun <di...@biurrun.de> wrote: > >> On Wed, Apr 05, 2017 at 05:16:24PM +0200, wm4 wrote: >> > On Wed, 5 Apr 2017 16:27:48 +0200 >> > Vittorio Giovara <vittorio.giov...@gmail.com> wrote: >> > > On Wed, Apr 5, 2017 at 4:11 PM, wm4 <nfx...@googlemail.com> wrote: >> > > > On Wed, 5 Apr 2017 15:35:26 +0200 >> > > > Vittorio Giovara <vittorio.giov...@gmail.com> wrote: >> > > >> --- a/libavfilter/avfiltergraph.h >> > > >> +++ /dev/null >> > > >> @@ -1,29 +0,0 @@ >> > > >> -#ifndef AVFILTER_AVFILTERGRAPH_H >> > > >> -#define AVFILTER_AVFILTERGRAPH_H >> > > >> - >> > > >> -#include "avfilter.h" >> > > >> -#include "libavutil/log.h" >> > > >> - >> > > >> - >> > > >> -#endif /* AVFILTER_AVFILTERGRAPH_H */ >> > > > >> > > > I guess I should remove it from my code? >> > > >> > > I suppose so, it was well advertised in the APIchange document and >> > > this is breaking season :) >> > >> > How is anyone supposed to notice/remember that? Maybe put a "#warning" >> > into the header when it's deprecated? (Not sure how standard that is, >> > but it could be put under gcc/clang guards.) >> >> It's not standard. That's why we removed all #warning directives from the >> codebase a long time ago. > > You could put in a declaration that uses a deprecated symbol, and thus > causes a deprecation warning. No idea whether that would work either.
I'd be ok with that, if you think delaying removal of an empty header is important for downstream, do you think you could send a patch that implements what you propose? -- Vittorio _______________________________________________ libav-devel mailing list libav-devel@libav.org https://lists.libav.org/mailman/listinfo/libav-devel
