Message ID | 20161112004458.4156-1-h.leppkes@gmail.com |
---|---|
State | Accepted |
Commit | 0539ff0e1b2f4e388bf4de16d00846ed5cebe811 |
Headers | show |
On Sat, Nov 12, 2016 at 1:44 AM, Hendrik Leppkes <h.leppkes@gmail.com> wrote: > Fixes building on Windows with MSVC using the link libraries distributed with the CUDA SDK. > > check_lib2 is required here because it includes the header to get the full signature of the > function, including the stdcall calling convention and all of its arguments, which enabled > the linker to determine the fully qualified object name and resolve it through the import > library, since the CUDA SDK libraries do not include un-qualified aliases. Something the message didn't say, this only applies to x86 where stdcall is used, x64 uses only one calling convention for everything, and this problem does not occur. - Hendrik
Seems like I never tested on any 32bit platform. lgtm
On Sun, Nov 13, 2016 at 1:31 AM, Timo Rothenpieler <timo@rothenpieler.org> wrote: > Seems like I never tested on any 32bit platform. > Applied, thanks.
diff --git a/configure b/configure index be7e925..c00c591 100755 --- a/configure +++ b/configure @@ -5659,10 +5659,10 @@ fi enabled avfoundation_indev && { check_header_objcc AVFoundation/AVFoundation.h || disable avfoundation_indev; } enabled avfoundation_indev && { check_lib2 CoreGraphics/CoreGraphics.h CGGetActiveDisplayList -framework CoreGraphics || check_lib2 ApplicationServices/ApplicationServices.h CGGetActiveDisplayList -framework ApplicationServices; } -enabled cuda && { check_lib cuda.h cuInit -lcuda || +enabled cuda && { check_lib2 cuda.h cuInit -lcuda || die "ERROR: CUDA not found"; } enabled cuvid && { add_cflags -I$source_path; - check_lib "compat/cuda/cuviddec.h" cuvidCreateDecoder -lnvcuvid || + check_lib2 "compat/cuda/cuviddec.h" cuvidCreateDecoder -lnvcuvid || die "ERROR: CUVID not found"; } && { enabled cuda || die "ERROR: CUVID requires CUDA"; }