GPU support on Mac OS X 10.7.3 - Cuda 4.1.28 (Bug #1877)

We don't have such problems. We tested it with CUDA4.2 MacOS 10.6.7.

Anyway I added namespaces in trunk and 2.4 branch.

Is it still reproduced?

I'm experiencing this issue as well:

OSX: 10.7.3
CUDA Toolkit: 4.2

/Users/bertini/Documents/workspace/libs_c/OpenCV-2.4.0/modules/gpu/src/imgproc.cpp:373:5: error: no matching function for call to 'buildWarpPlaneMaps'
buildWarpPlaneMaps(dst_roi.tl().x, dst_roi.tl().y, map_x, map_y, K_Rinv.ptr<float>(), R_Kinv.ptr<float>(),
^_{~~~~~~~~~~~~~~}~
/Users/bertini/Documents/workspace/libs_c/OpenCV-2.4.0/modules/gpu/src/imgproc.cpp:357:15: note: candidate function not viable: no known conversion from 'int' to 'Size'
(aka 'Size_<int>') for 1st argument;
void cv::gpu::buildWarpPlaneMaps(Size src_size, Rect dst_roi, const Mat &K, const Mat& R, const Mat &T,
^
/Users/bertini/Documents/workspace/libs_c/OpenCV-2.4.0/modules/gpu/src/imgproc.cpp:405:5: error: no matching function for call to 'buildWarpCylindricalMaps'
buildWarpCylindricalMaps(dst_roi.tl().x, dst_roi.tl().y, map_x, map_y, K_Rinv.ptr<float>(), R_Kinv.ptr<float>(), scale, StreamAccessor::getStream(stream));
^_{~~~~~~~~~~~~~~~~~~~~}~
/Users/bertini/Documents/workspace/libs_c/OpenCV-2.4.0/modules/gpu/src/imgproc.cpp:390:15: note: candidate function not viable: no known conversion from 'int' to 'Size'
(aka 'Size_<int>') for 1st argument;
void cv::gpu::buildWarpCylindricalMaps(Size src_size, Rect dst_roi, const Mat &K, const Mat& R, float scale,
^
/Users/bertini/Documents/workspace/libs_c/OpenCV-2.4.0/modules/gpu/src/imgproc.cpp:437:5: error: no matching function for call to 'buildWarpSphericalMaps'
buildWarpSphericalMaps(dst_roi.tl().x, dst_roi.tl().y, map_x, map_y, K_Rinv.ptr<float>(), R_Kinv.ptr<float>(), scale, StreamAccessor::getStream(stream));
^_{~~~~~~~~~~~~~~~~~~}~
/Users/bertini/Documents/workspace/libs_c/OpenCV-2.4.0/modules/gpu/src/imgproc.cpp:422:15: note: candidate function not viable: no known conversion from 'int' to 'Size'
(aka 'Size_<int>') for 1st argument;
void cv::gpu::buildWarpSphericalMaps(Size src_size, Rect dst_roi, const Mat &K, const Mat& R, float scale,
^
/Users/bertini/Documents/workspace/libs_c/OpenCV-2.4.0/modules/gpu/src/imgproc.cpp:1146:33: error: cannot initialize an array element of type 'Caller' (aka 'void ()(const
PtrStep<cufftComplex>, const PtrStep<cufftComplex>, DevMem2D_<cufftComplex>, cudaStream_t)') with an lvalue of type 'void (const cv::gpu::GpuMat &, const
cv::gpu::GpuMat &, cv::gpu::GpuMat &, int, bool, cv::gpu::Stream &)'
static Caller callers[] = { mulSpectrums, mulSpectrums_CONJ };
^_~~~~~~~~~
/Users/bertini/Documents/workspace/libs_c/OpenCV-2.4.0/modules/gpu/src/imgproc.cpp:1175:33: error: cannot initialize an array element of type 'Caller' (aka 'void ()(const
PtrStep<cufftComplex>, const PtrStep<cufftComplex>, float, DevMem2D_<cufftComplex>, cudaStream_t)') with an lvalue of type 'void (const cv::gpu::GpuMat &, const
cv::gpu::GpuMat &, cv::gpu::GpuMat &, int, float, bool, cv::gpu::Stream &)'
static Caller callers[] = { mulAndScaleSpectrums, mulAndScaleSpectrums_CONJ };
^_{~~~~~~~~~~~~~~~~}~
[ 69%] Built target opencv_legacy
5 errors generated.
make²: * [modules/gpu/CMakeFiles/opencv_gpu.dir/src/imgproc.cpp.o] Error 1
make¹: * [modules/gpu/CMakeFiles/opencv_gpu.dir/all] Error 2

Compiling with:

C++ Compiler: /usr/bin/c++
nirvana:build bertini$ /usr/bin/c++ --version
Apple clang version 3.1 (tags/Apple/clang-318.0.58) (based on LLVM 3.1svn)
Target: x86_64-apple-darwin11.3.0
Thread model: posix

I confirm that to eliminate the first three errors (those about no matching functions) it's enough to fully qualify the names of the functions using cv::gpu::device::imgproc

All the errors are eliminated by adding the namespaces.