On Wed, 23 Jul 2014 13:20:22 +0200, Diego Biurrun <[email protected]> wrote: > On Wed, Jul 23, 2014 at 12:56:54PM +0200, Anton Khirnov wrote: > > > > On Tue, 22 Jul 2014 15:55:24 -0700, Diego Biurrun <[email protected]> wrote: > > > This allows unconditionally including the public header and cleaning up > > > some duplicated declarations. It also allows cleanly eliminating > > > redefined typedefs for fft-related context structs. > > > --- > > > > > > This is the cleanest solution to the redefined typedefs that I found. > > > All headers work standalone now and some silly duplication is eliminated. > > > > > > It modifies a public header, so some version bump might be necessary, > > > maybe also some prefixing of defines. > > > > > > libavcodec/avfft.h | 12 ++++++++++++ > > > libavcodec/dct.h | 1 + > > > libavcodec/fft.h | 13 +++---------- > > > 3 files changed, 16 insertions(+), 10 deletions(-) > > > > > > diff --git a/libavcodec/avfft.h b/libavcodec/avfft.h > > > index e2e727d..86821e7 100644 > > > --- a/libavcodec/avfft.h > > > +++ b/libavcodec/avfft.h > > > @@ -32,8 +32,20 @@ > > > * @{ > > > */ > > > > > > +#ifndef FFT_FLOAT > > > +#define FFT_FLOAT 1 > > > +#endif > > > + > > > +#if FFT_FLOAT > > > > I don't like this. > > Putting such ugliness into public headers to fix some _internal_ problem is > > EVIL > > After staring at those headers for a considerable amount of time this > is the cleanest I could come up with. It also makes the headers more > straightforward, so I consider this a net gain. >
IMO, it's better to have more ugliness internally, than let it leak to the outside. I have not investigated what you're solving in detail yet, but cannot you achieve what you're trying to do by using an internal wrapper header around avfft.h? -- Anton Khirnov _______________________________________________ libav-devel mailing list [email protected] https://lists.libav.org/mailman/listinfo/libav-devel
