Re: [AD] 5.0.4

[Peter Wang <novalazy@xxxxxxxxxx>]

On 2011-08-09, John-Kim Murphy <leftium@xxxxxxxxxx> wrote:
> On Tue, Aug 9, 2011 at 9:21 AM, Peter Wang <novalazy@xxxxxxxxxx> wrote:
> 
> > >    - Current ogg/vorbis/vorbisfile MSVC sln files produce libraries
> > prefixed
> > >    with "lib" (i.e. libogg.lib + libogg.dll)
> > >    - I tried modifying the CMake files to account for the actual library
> > >    name, but it didn't help.
> > >       - CMake (v2.8.5) is trying to be smart and taking off that "lib"
> > >       prefix before searching for libraries?
> > >       - Generates/builds, but acodec addon is disabled
> >
> > Does it work if you edit the find_library lines in
> > cmake/FindVorbis.cmake?
> >
> 
> Those are precisely the lines I tried modifying. I tried adding "libogg
> libogg_static", moving them in front of the original arguments, and even
> removing the original "ogg ogg_static" arguments. (All three sets of
> find_library calls, of course)

It worked for me.  I have added additional names to SVN.

> > > Finally, I tweaked the CMake files so the d3dx9 libraries are
> > auto-detected
> > > and only searched for once.
> >
> > Is this much of a problem?
> >
> 
> I actually falsely assumed I was getting video card acceleration because
> d3d9 was detected even though d3dx9 was not detected (Perhaps I was getting
> some, but not all possible acceleration?) Ignoring that reason, it's just a
> nice convenience, especially when you are re-running CMake as often I was.
> Also, my tweak prefers the latest version of d3d9 instead of the older one
> that comes by default on Windows 7/Visual Studio 2010. Finally, while
> debugging the CMakefile.txt I was confused why the d3dx9 library was
> detected twice, in two different files/variables.
> 
> I've been swamped with work, so I haven't been able to confirm on the 5.0
> branch. I attached a patch with my changes. It also includes some fixes that
> allows the skater demo to compile/work on Windows 7 + MSVS 2010. I think
> some of the CMake if conditions can be collapsed even further, and the
> vorbis m.lib part  has to be made conditional based on MSVC/GCC, but it's a
> good start.

Thanks, I incorporated most of your changes.

Peter