The Mutex manager registered with ffmpeg must reset the mutex to NULL after destruction, otherwise ffmpeg will give the invalid mutex to the next mutex manager when it asks it to CREATE a new mutex.
See ffmpeg code: http://git.videolan.org/?p=ffmpeg.git;a=blob;f=libavcodec/utils.c;h=28c5785398fcf11a3d3c70a8cd09e9df798e2734;hb=HEAD#l3423

Cherry picked from head (file has moved but issue is the same).
Conflicts:
	modules/videoio/src/cap_ffmpeg_impl.hpp

Currently, FFMPEG source files are included in highgui project file regardless of CMake WITH_FFMPEG option.
After applying this PR FFMPEG files are included only if WITH_FFMPEG option is enabled.

Taken from 3f076552

In C99 'inline' is not a hint to the compiler to inline the function,
it is an attribute that affects the linkage of the function.  'inline'
functions are required to have a definition in a different compiliation
unit, so compilers are free to delete 'inline' functions if they want to.

This issue can be seen in Clang when compiling at -O0.  Clang
will sometimes delete 'inline' functions which creates an invalid
program.

Issue 3746: http://code.opencv.org/issues/3746

Merge branch 'bugfix_brute_force_match_cl' of https://github.com/WilhelmHannemann/opencv into bugfix_brute_force_match_cl

Bugfix brute_force_match.cl (see http://code.opencv.org/issues/2837): wrong results for non-float descriptors in OpenCL BruteForceMatcher

Bugfix brute_force_match.cl (Bug #2837): wrong results for non-float descriptors in OpenCL BruteForceMatcher

use different fix for the problem, embedded right into the BFMatcher.

This reverts commit ed2cdb71.

Check sure that we're not already below required leaf false alarm rate before continuing to get negative samples.