Add a new variant of detectMultiScale with output arguments 'objects' and 'numDetections'; factor most of detectMultiScale's code into 2 protected methods

Python 3 requires METH_VARARGS | METH_KEYWORDS, not just METH_KEYWORDS.

In these cases, just using print() is enough to ensure py2/3
compatibility.

Added a header with defines to enable the module to compile with either
python 2 or 3 without changes.

Merged pull requests:
	#890 from caorong:patch-1
	#893 from jet47:gpu-arm-fixes
	#933 from pengx17:2.4_macfix_cont
	#935 from pengx17:2.4_filter2d_fix
	#936 from bitwangyaoyao:2.4_perf
	#937 from bitwangyaoyao:2.4_fixPyrLK
	#938 from pengx17:2.4_surf_sample
	#939 from pengx17:2.4_getDevice
	#940 from SpecLad:autolock
	#941 from apavlenko:signed_char
	#946 from bitwangyaoyao:2.4_samples2
	#947 from jet47:fix-gpu-arm-build
	#948 from jet47:cuda-5.5-support
	#952 from SpecLad:jepg
	#953 from jet47:fix-bug-3069
	#955 from SpecLad:symlink
	#957 from pengx17:2.4_fix_corner_detector
	#959 from SpecLad:qt4-build
	#960 from SpecLad:extra-modules

Conflicts:
	modules/core/include/opencv2/core/core.hpp
	modules/gpu/CMakeLists.txt
	modules/gpu/include/opencv2/gpu/device/vec_math.hpp
	modules/gpu/perf/perf_video.cpp
	modules/gpuimgproc/src/cuda/hough.cu
	modules/ocl/include/opencv2/ocl/ocl.hpp
	modules/ocl/src/pyrlk.cpp
	samples/gpu/driver_api_multi.cpp
	samples/gpu/driver_api_stereo_multi.cpp
	samples/ocl/surf_matcher.cpp

'dst' should be treated as an input arg (epilog not needed)

now DeviceInfo provides full information about device (from cudaDeviceProp)

* switch to InputArray/OutputArray
* add Stream support