I actually did this because it makes it a bit easier to use with a
build from cross-compiled windows FFMpeg/x264 DLLs. When compiling on
linux I just have it compile to a custom prefix directory for the
specific windows arch, then I just copy the arch dir back to my windows
HD when complete. Adding this to the cmake allows me to use the compile
arch directories directly without modification.
On fedora, the freetype cmake module would not find the correct
location. This is because it wasn't using the pkg-cofig include
directory, which is different from the main ft2build.h header directory
(which is /usr/include).
This modification may be a good thing to do on all the cmake packages,
actually. Currently, they all just search for the location of a
specific header, and ignore the pkg-config includes if they find it in
specific hint locations.