// on 32 bit osx a deadlock seems to occur between the
// initialisation of the native library and AWT. This
// seems to fix it...
I saw the same issue on my 64 bit osx MacBook and noticed that if I load the AWT native library first(e.g. create a window before calling VideoCapture), then it doesn't have the problem.
I'm wondering what was the work around in OpenIMAJ and since I'm still seeing it in osx, was this issue really fixed?(sorry can't tell from svn log).
Thanks,
Anthony
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Looking back through the logs I would guess that the fix was to force the native library to load immediately in the constructor by calling VideoCapture.getVideoDevices(). As time progressed, the constructor was changed to make use of the output of that call to get the default device, rather than just ignoring it.
I'm fairly sure the issue is fixed - certainly the two of us using it on a daily basis on 64-bit OSX haven't seen any problems irrespective of whether a window is created before using VideoCapture. If you're having repeatable problems it would be helpful to know some details about what you're doing and what version of java you're using.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
In VideoCapture, it says
// on 32 bit osx a deadlock seems to occur between the // initialisation of the native library and AWT. This // seems to fix it...I saw the same issue on my 64 bit osx MacBook and noticed that if I load the AWT native library first(e.g. create a window before calling VideoCapture), then it doesn't have the problem.
I'm wondering what was the work around in OpenIMAJ and since I'm still seeing it in osx, was this issue really fixed?(sorry can't tell from svn log).
Thanks,
Anthony
Looking back through the logs I would guess that the fix was to force the native library to load immediately in the constructor by calling VideoCapture.getVideoDevices(). As time progressed, the constructor was changed to make use of the output of that call to get the default device, rather than just ignoring it.
I'm fairly sure the issue is fixed - certainly the two of us using it on a daily basis on 64-bit OSX haven't seen any problems irrespective of whether a window is created before using VideoCapture. If you're having repeatable problems it would be helpful to know some details about what you're doing and what version of java you're using.