fixed a few compile problems in SparseMat code; fixed iteration over an empty sparse matrix (ticket #1587)

Input/Output arrays are updated to support fixed-type and fixed-size semantic for all compatible types

minor build system changes (now cuda code in opencv_core is compiled using CUDA_ARCH* cmake variables)

e.g. "n += abs(sign(value))"  or  "n += sign(cv::norm(value) > eps_64"

Fixed some bugs with checking for CV_64FC1 type.
Added eigen tests for 1*1 source matrix. 

fixed bag in the checking size of matrix and vector for eigen test; all eigen test execution was checked: it works correctly; added first version of homography test for C++ interface

Optimized version of 32F Norm function demands bigger MaxErr value, because it calculates norm in floats instead of double.

added COVAR_ and SORT_ enums to core.hpp; fixed many, many VS2005, VS2010 and MinGW (GCC 4.5.2) warnings