| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
| |
This reverts commit 51fbd3cf7db14399bc1c69c4a5fd63f157fd209d.
The long log was truncated due to a formatting error.
Signed-off-by: Randy MacLeod <Randy.MacLeod@windriver.com>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
With opencv-3.2, the configuration detects if the target supports
half-precision floating-point format. This fails to compile for some
Intel targets such as skylake with an error such as:
error: '_mm_cvtph_ps' was not declared in this scope
The configuration worked in opencv-3.1 so revert two commits to
drop the FP16 detection even though it may make opencv slower.
The only change in the configure log is:
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
|
|
|
|
|
|
|
|
|
| |
-Only enable protobuf as dependency if dnn module is enabled
-Download manually xfeatures from 3rd party repo
-Remove uneeded fixgcc60patch
Signed-off-by: Ricardo Ribalda Delgado <ricardo.ribalda@gmail.com>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
|
|
|
|
|
|
|
|
| |
It seems like gcc 6.0 does not have the same behaviour as previous 5.x
with regards isystem flag.
Signed-off-by: Ricardo Ribalda Delgado <ricardo.ribalda@gmail.com>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
|
|
|
|
|
|
|
| |
We should not rely on cmake downloading ipp at configure time.
Signed-off-by: Ricardo Ribalda Delgado <ricardo.ribalda@gmail.com>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
|
|
|
|
|
|
|
|
|
| |
Libdir was pointing to the wrong location
also fix for QA Issue: opencv.pc failed sanity test (tmpdir) in path
Signed-off-by: Ricardo Ribalda Delgado <ricardo.ribalda@gmail.com>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
During the testing of the meta-ros layer, I discovered that opencv
does not cross-compile for the ppc architecture. The error is
reported in the meta-ros issue tracker #355 [1]. To address this
error, this commit applies a patch that is already provided upstream
in the opencv version 3.0.
[1] https://github.com/bmwcarit/meta-ros/issues/355
Signed-off-by: Lukas Bulwahn <lukas.bulwahn@oss.bmw-carit.de>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
|
| |
Upgrade OpenCV to the 2.4.11 release.
Remove the opencv-fix-pkgconfig-generation patch which has been integrated upstream,
be it in modified form.
Disable 1394 support by default to get a deterministic build.
Fix "jasper" dependency, the BUILD_JASPER parameter served only to build an internal
library, while WITH_JASPER actually controls whether jpeg2000 support was desired.
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
|
|
|
|
|
|
|
|
| |
Replace absolute library path with library name spec and library search
path option in pkg-config file.
Signed-off-by: Stefan Herbrechtsmeier <stefan@herbrechtsmeier.net>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
|
|
|
|
| |
Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
|
|
OpenCV <= 2.2 won't compile on oe-core due to the lack of v4l1 headers, and therefore wasn't included in this patch.
Also, the support for newer cv2-style python bindings is a bit hacked up at the moment due to the way OpenCV detects the Python version and Numpy headers (see 0001-Fix-CMakeLists.txt-numpy-detection.patch). In particular, it requires the native & target python to be the same version and have similar site-packages directories, which is true at least for now in OE.
Signed-off-by: Connor Abbott <cwabbott0@gmail.com>
Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
|