On Wed, Jun 07, 2017 at 01:52:58PM +0200, Hendrik Leppkes wrote:
> On Wed, Jun 7, 2017 at 1:28 PM, Diego Biurrun <di...@biurrun.de> wrote:
> > On Wed, Jun 07, 2017 at 01:18:19PM +0200, wm4 wrote:
> >> On Wed, 7 Jun 2017 13:09:15 +0200 Diego Biurrun <di...@biurrun.de> wrote:
> >> > On Tue, Jun 06, 2017 at 06:51:07PM +0200, wm4 wrote:
> >> > > --- a/configure
> >> > > +++ b/configure
> >> > > @@ -2166,7 +2166,7 @@ zmbv_encoder_deps="zlib"
> >> > >
> >> > > # hardware accelerators
> >> > > d3d11va_deps="d3d11_h dxva_h ID3D11VideoDecoder"
> >> > > -dxva2_deps="dxva2api_h DXVA2_ConfigPictureDecode"
> >> > > +dxva2_deps="dxva2api_h DXVA2_ConfigPictureDecode ole32"
> >> >
> >> > This looks wrong. Why does dxva2 suddenly depend on ole32?
> >>
> >> Because it suddenly does. You can read the C code.
> >
> > Try removing it. "ole32" is not a thing in configure. Its presence in
> >
> > avconv_suggest="dxva2_lib ole32 psapi shell32"
> >
> > is therefore a bug. Introduced by yours truly it seems...
>
> Before the recent rework the dxva2_lib thing caused a linkage against
> ole32, for it not breaking - most windows toolchains have a list of
> default libraries they link all binaries against, which includes ole32
> - however, that list is not guaranteed for every toolchain, ie. if you
> use clang-c2 that list is empty, so being explicit in the required
> libraries would be rather nice.
dxva2_lib still links against ole32, around line 4890:
enabled dxva2 &&
check_lib dxva2_lib windows.h CoTaskMemFree -lole32
Patch 6/6 in this series removes that check so there is no longer linkage
against ole32 after that. I gather from what you said that this is not
the intended end result. So .. what is the intended end result?
Diego
_______________________________________________
libav-devel mailing list
libav-devel@libav.org
https://lists.libav.org/mailman/listinfo/libav-devel
