extended MinProblemSolver::Function to 1) report the space dimensionality, 2) compute gradient if needed

refactored downhill simplex implementation a bit; hopefully, fixed the bug with random failures in the tests

Fixed getContinuousSize() to handle huge matrices properly. This should solve http://code.opencv.org/issues/3232

fixed bug with simplex initialization in Downhill solver (thanks to Toby Rule for the patch): http://code.opencv.org/issues/3990.
re-enabled Downhill solver test, let's see how it goes

fixed http://code.opencv.org/issues/3935
possibly fixed http://code.opencv.org/issues/3594

fixed invalid output of cv::dft when using DFT_ROWS + DFT_COMPLEX_OUTPUT (http://code.opencv.org/issues/3428)

A difference of two pointers was casted to unsigned which can lead to
overflow on 64-bit systems.

2. Algorithm::load/save added (moved from StatModel)
3. copyrights updated; added copyright/licensing info for ffmpeg
4. some warnings from Xcode 6.x are fixed

Moved out of TLS to properly collect data from different threads;
Concurrent access guards were added to collection functions;

... before any exceptions occur.

(cherry picked from commit 08da247a)

(cherry picked from commit 16ce114e)

explained reason why IPP is disabled on 32-bit Linux. disable IPP's minMaxLoc_32f on 32-bit OSes only

2) disable IPP on 32-bit Linux when OpenCV is built as shared libs. Otherwise we get linker errors
3) disable IPP's minMaxIdx 32-bit floating-point flavor in a hope that it fixes some test failures