Commit Graph

11635 Commits

Author SHA1 Message Date
Roman Donchenko
90f8906e1e Merge pull request #1690 from SpecLad:cmake-2.8.12 2013-10-25 19:24:58 +04:00
Roman Donchenko
c016c43d25 Fixed Android SDK build - again. 2013-10-25 19:12:32 +04:00
Andrey Pavlenko
2e5db6e56c fixing typo 2013-10-25 18:53:24 +04:00
Andrey Pavlenko
2b6fca68bf fixing typo 2013-10-25 18:00:46 +04:00
Roman Donchenko
55ede74022 Merge pull request #1680 from alalek:cmake_fix_android_builds 2013-10-25 17:32:34 +04:00
Ilya Lavrenov
e1596d69cc fixed extrapolation in ocl::adaptiveBilateralFilter 2013-10-25 17:29:21 +04:00
Alexander Alekhin
7a8f9fe746 cmake: fix android installation for different NDK ABIs 2013-10-25 17:17:00 +04:00
Ilya Lavrenov
b33a62beb0 fixed separable filter extrapolation 2013-10-25 17:08:44 +04:00
Vadim Pisarevsky
485d36d3c0 changed InputArray's enumeration constant for UMat, since it may potentially conflict with existing OpenCL module. refactored Kernel's setArg API (now UMat is passed to a kernel as a structure, as Alexander A suggested). removed Kernel's cleanup callback from the external API; now each kernel keeps track of used matrices and they are dereferenced after it's complete. 2013-10-25 16:46:03 +04:00
Roman Donchenko
36b5180aa1 Another FFmpeg fix from master.
From commit dd74a851, to be exact. Now cap_ffmpeg.cpp should actually
build if HAVE_FFMPEG is true.

Also modified some gpu sources in a similar manner.
2013-10-25 16:08:14 +04:00
peng xiao
68a8a11161 Rewrite distanceToCenters.
It supports NORM_L1 distance types now and can
use user provided indices.
Also fixed a bug of kmeans where distance pointers should be float instead
 of double.

NORM_L2 changed to NORM_L2SQR, Accuracy and Perf tests are added

added ROI support in accuracy test of distanceToCenters
2013-10-25 15:02:00 +04:00
Roman Donchenko
1188894133 Fixed indentation in CV_FFmpegWriteBigVideoTest::run. 2013-10-25 14:37:54 +04:00
Roman Donchenko
2ca49eef37 Added stupid hacks to make the video tests pass with FFmpeg 2.0.2.
Need to go back at some point and fix this for real.

(cherry picked from commit dfe07df87b)
2013-10-25 14:34:02 +04:00
Roman Donchenko
5a5d569d72 Changed MPEG-2 resolution in the FFmpeg test.
Newer FFmpeg prohibits 4096x4096 MPEG-2, presumably because it violates
the standard.

http://git.videolan.org/gitweb.cgi/ffmpeg.git/?p=ffmpeg.git;a=commit;h=7fb87bc5f24b1be13269109506c05e4c54695b5e

(cherry picked from commit 424a7b0ab0)

Conflicts:
	modules/highgui/test/test_ffmpeg.cpp
2013-10-25 14:31:27 +04:00
Roman Donchenko
260c2e0b08 Decoupled test_ffmpeg.cpp from ffmpeg_codecs.hpp.
(cherry picked from commit 3f3ae33327)
2013-10-25 14:29:37 +04:00
Roman Donchenko
d3af86c0ea Added the missing setting of HAVE_FFMPEG to true on Windows.
While the FFmpeg video IO backend gets compiled even without it,
the tests (for both FFmpeg and video IO) don't. This should fix
that.
2013-10-25 14:22:07 +04:00
Roman Donchenko
76b904b022 Replaced our usage of LINK_PRIVATE with that of LINK_INTERFACE_LIBRARIES.
The reasons for that are twofold:

1) LINK_PRIVATE is only available since CMake 2.8.7.
2) The way it was used generated a warning because of CMake policy CMP0023:

   http://www.cmake.org/cmake/help/v2.8.12/cmake.html#policy:CMP0023

Using LINK_INTERFACE_LIBRARIES actually causes another warning - this time
because of CMake policy CMP0022:

   http://www.cmake.org/cmake/help/v2.8.12/cmake.html#policy:CMP0022

I set the policy to OLD, because NEW means subtle changes when compiling
with CMake 2.8.12, and I don't want to research that this close to release.
:-)

I also removed the setting of CMP0003, because it's set by
cmake_minimal_version anyway.
2013-10-25 14:19:02 +04:00
Roman Donchenko
e1c8f5d723 Removed a stray comma in the Android toolchain file.
It was actually acting as an additional argument, breaking the error message.
2013-10-25 13:50:22 +04:00
Ilya Lavrenov
9ea6001d08 enabled ocl::bilateralFilter ROI testing 2013-10-25 13:32:35 +04:00
Roman Donchenko
31f0ab6c80 Merge pull request #1681 from LeszekSwirski:fix-ninjabuild 2013-10-25 13:30:52 +04:00
Andrey Pavlenko
f177e65871 Merge pull request #1683 from alalek:ocl_restore_showdiff 2013-10-25 13:19:02 +04:00
yao
dd0fa63ca8 fix the bug of ocl::bruteForceMatcher 2013-10-25 16:01:41 +08:00
Vadim Pisarevsky
6416c33020 fixed some bugs in OpenCL branch of UMat implementation; maybe now the tests will finally pass?! (please ;( ) 2013-10-25 11:19:40 +04:00
Vadim Pisarevsky
ace0d4bad0 hopefully, the last test failures are fixed now 2013-10-24 23:31:00 +04:00
Vadim Pisarevsky
5920bd4f2e fixed whitespace errors 2013-10-24 23:18:05 +04:00
Vadim Pisarevsky
caa33b2242 added first very simple test for UMat<=>Mat conversion 2013-10-24 22:33:41 +04:00
Roman Donchenko
3b08bf6b62 Merge pull request #1682 from SpecLad:Wmissing-declarations 2013-10-24 22:19:17 +04:00
Leszek Swirski
6e4fb8f275 Replace $(Configuration) with $<CONFIGURATION> 2013-10-24 18:20:16 +01:00
Roman Donchenko
496859a944 Merge pull request #1643 from alalek:cmake_fix_java_link_rules 2013-10-24 19:38:43 +04:00
Roman Donchenko
7804240528 Merge pull request #1679 from ilya-lavrenov:ocl_equalizeHist 2013-10-24 19:12:35 +04:00
Alexander Alekhin
4b17d073c0 cmake: fix linker dependencies for opencv_java
Linker dependencies to all OpenCV modules are invalid.
We should not include other bindings in this list (like "opencv_python").
2013-10-24 19:04:59 +04:00
Roman Donchenko
f51e490bad Fixed a Wmissing-declarations warning when compiling with MinGW. 2013-10-24 18:35:08 +04:00
Vladislav Vinogradov
e5947f581a fix CUDA 5.5 support (npp, arm cross compilation) in CMake scripts:
The patch was submitted to CMake and might be available
in the next CMake release.

But until we have the fix in CMake we should add workaround in our scripts.
2013-10-24 18:17:36 +04:00
Vadim Pisarevsky
b1aa4a4913 yet another attempt to fix test failures 2013-10-24 18:00:15 +04:00
Vadim Pisarevsky
e14171fd33 fixed crash in Python bindings 2013-10-24 17:37:33 +04:00
Roman Donchenko
7e1ded0ebd Merge pull request #1664 from SpecLad:merge-2.4 2013-10-24 17:27:30 +04:00
Vadim Pisarevsky
a9065b7d0c trying to fix failures in tests 2013-10-24 17:23:48 +04:00
Vadim Pisarevsky
303df78cca removed refcount field from Mat; that helped to find and eliminate some memory leaks. perf tests should probably pass now 2013-10-24 16:08:15 +04:00
Roman Donchenko
f0bc253d0e Merge pull request #1667 from SpecLad:detect-openmp 2013-10-24 15:00:27 +04:00
Roman Donchenko
119cdc57e9 Merge pull request #1623 from ImAlsoGreg:2.4-pkgconfig-dash-l 2013-10-24 14:18:12 +04:00
Roman Donchenko
dfe07df87b Added stupid hacks to make the video tests pass with FFmpeg 2.0.2.
Need to go back at some point and fix this for real.
2013-10-24 14:09:08 +04:00
Ilya Lavrenov
dab3000778 fixed bug in ocl::equalizeHist 2013-10-24 13:59:25 +04:00
Roman Donchenko
c7408003d2 Merge pull request #1678 from wejeus:ios-documentationfix-2.4 2013-10-24 13:55:30 +04:00
Roman Donchenko
f5bd8fecc9 Merge pull request #1674 from alalek:cmake_fix 2013-10-24 13:54:31 +04:00
Samuel Wejéus
eddaaa9643 Removed incorrect release of obtained colorspace. This fixes bug #3318 (updated commit for 2.4 branch) 2013-10-24 10:36:01 +02:00
Andrey Pavlenko
2279c209c8 Merge pull request #1676 from ilya-lavrenov:ocl_resize_test 2013-10-24 11:23:55 +04:00
Andrey Pavlenko
931a2ae14f Merge pull request #1673 from ilya-lavrenov:ocl_norm 2013-10-24 11:23:36 +04:00
Andrey Pavlenko
e7f39eb46e Merge pull request #1672 from ilya-lavrenov:ocl_arithm 2013-10-24 11:23:19 +04:00
Andrey Pavlenko
4cf598a862 Merge pull request #1669 from ilya-lavrenov:ocl_copyMakeBorder 2013-10-24 11:23:01 +04:00
Peng Xiao
4b712eeded Update imgproc_canny.cl
Reordering condition checking.
2013-10-24 12:07:54 +08:00