CMake cannot find 'pthread_getconcurrency’, ‘pthread_setconcurrency’ and ‘pthread_setaffinity_np' functions in 'pthread' library because when linking internal cmake test did’t specifed ‘-l pthread’.
when building for iOS, Xcode allows developers to specify to enable or disable the 'bitcode' compilation option. There's not a preferred way to go and the choice is really up to the developer but considering that:
1. Currently the generated project defaults the option to YES
2. There are almost 90 projects targets that should be modified if one wants to disable the bitcode option (which considerably reduces the footprint of the app)
3. Even though one can select all the 90+ targets and set the option to NO for all of them, the updates could take a few seconds and could be error prone because one could miss to select some targets
I propose to add a CMake setting that is displayed only when building for iOS. By setting this option "before" the project generation would speed up things for developers and would avoid errors at compiling time.
Sorry about this mixup, I was not aware that this particular directory was to be considered a separate project and must not rely on any dependencies from the rest of the OSG project. All OSGNOTIFY messages have been removed and the previous printf statements have been put back.
OpenSceneGraph and the OpenThreads library.
The changes in the file simply remove a few ifndef's that currently
do not allow Linux systems to fully utilize the PThread real-time
scheduling API.
Since Linux now fully supports the PThread scheduling API it would
be beneficial to have it available to use as necessary. I have
been testing this change since OSG release 3.3.7 and have not seen
any ill affects.
The Priority scheduling api is further protected by another ifdef:
#ifdef ALLOW_PRIORITY_SCHEDULING
that only appears to be defined in the pthreads implementation as
well. This would make it unlikely that anyone would be affected
by this unless they are intentionally wanting to run with priority
scheduling. In which case on Linux they would need to make
these same modifications themselves to utilize it to its full extent.
Attached file is for the current trunk as of this date.
"
git-svn-id: http://svn.openscenegraph.org/osg/OpenSceneGraph/trunk@15140 16af8721-9629-0410-8352-f15c8da7e697
My changes:
-------------------
- I changed the cmake files and added a toolchain for building OSG in Android. The toolchain is based on the one used at OpenCV. For building OSG for android you just need to do:
mkdir build_android_static_gles2 && cd build_android_static_gles2
cmake .. -DANDROID_NDK=<path-to-the-android-ndk>
-DCMAKE_TOOLCHAIN_FILE=../PlatformSpecifics/Android/android.toolchain.cmake
-DOPENGL_PROFILE="GLES2"
-DDYNAMIC_OPENTHREADS=OFF
-DDYNAMIC_OPENSCENEGRAPH=OFF
-DANDROID_NATIVE_API_LEVEL=15 # optional
-DANDROID_ABI=armeabim #optional
-DCMAKE_INSTALL_PREFIX=<path-to-the-install-path> #optional
make -j 8
make install
The OPENGL_PROFILE works as expected, changing it to "GLES1" it builds and links OSG using GLES1.
The DYNAMIC_OPENTHREADS/DYNAMIC_OPENSCENEGRAPH parameters also allows to build the dynamic libraries
- I also added some build fixes for android related to the texture formats and added some missing USE_OSG_SERIALIZER_WRAPPER in the osg serializer library to support loading osgb files in static."
git-svn-id: http://svn.openscenegraph.org/osg/OpenSceneGraph/trunk@14514 16af8721-9629-0410-8352-f15c8da7e697
pd->tid.set( (void*)_beginthreadex(NULL,static_cast<unsigned>(pd->stackSize),ThreadPrivateActions::StartThread,static_cast<void *>(this),0,&ID));
the method "pd->tid.set" sets the thread id, however via the startup function "ThreadPrivateActions::StartThread" that thread id is used (see further down the call hierarchy the line "int status = SetThreadPriority( pd->tid.get(), prio);".
Until now I never ran into any problem in debug or release builds, though. It seems that furtunately the tid.set method was executed always before the tid.get method in the startup code. However, this may make trouble in the furture. A simple solution is the following: just replace the line above with following two lines:
pd->tid.set( (void*)_beginthreadex(NULL,static_cast<unsigned>(pd->stackSize),ThreadPrivateActions::StartThread,static_cast<void *>(this),CREATE_SUSPENDED,&ID));
ResumeThread(pd->tid.get());
The trick is just starting the thread in suspended mode so the StartThread function does not get executed and we can safely store the tid by pd->tid.set. Then start the Thread by calling ResumeThread."
This behavior is also described in the pthreads man page (http://man7.org/linux/man-pages/man3/pthread_create.3.html):
>
> Linux-specific details
> The new thread inherits copies of the calling thread's capability
> sets (see capabilities(7)) and CPU affinity mask (see
> sched_setaffinity(2)).
>
To prevent this behaviour I wrote a patch that explicitly sets the affinity mask to all cores of the system, if no specific affinity was defined with PThread::setProcessorAffinity(unsigned int) .
Thank you!
"
