Hi, I'm working with Jorge trying to get a native osg on Android.
The main problem that we found was the migration of OpenThreads, it's not really complicated but there are some missing api in pthreads (pthread_cancel as mentioned before) and some changes (sleep returns void :O). We made the changes on CMake and OpenThreads to support this changes but when we tried to build against a desktop environment again then things didn't work :(. We wanted to send the modified files once things get working, but maybe is time that other eyes get a look to this. Jorge, could you attach to this thread a bundle with all the changes we made to get things compiling and working on android? I'm attaching a simple (and little ugly) example of a viewer for android, all made in native (except the minimal parts required in Java). Currently we got things building against crystax-ndk-r4. This example "worked" for my Samsung Galaxy S (android 2.1update1). The viewer starts, and shows the nice blue osg screen! :) but I wasn't be able to get draw any geometry (I built osg with gles2 support). To compile it, you need the crystax ndk r4, and osg compiled against this version. Then you need to modify the osgviewerandroid/jni/Android.mk file to change the path of required osg libraries for your system. To get the apk for android I do this things. android update project -p . -t 1 # -t is the target defined (android list devices to get the list) ndk-build ant debug If things work inside bin/ you will get the .apk Greets, Rafa. 2011/1/18 Robert Osfield <[email protected]>: > Hi Jorge, > > I know little about the issues of Android platform, so read with > interest in the various developments that yourself and others are > working on, it does mean I can't yet contribute too pro-actively > though. I'll will try and chip in where I can. > > On Mon, Jan 17, 2011 at 10:37 PM, Jorge Izquierdo Ciges > <[email protected]> wrote: >> Pthreads. The Set functions are trivial and i don't think that could cause a >> very big damage. But pthread_cancel is something different, because i've >> seen that it's widely used. And we would need a suitable replacement to >> don't let the thread go on. > > With the OSG usage of threads I have implemented co-operative exit of > the thread loops using done flags, instead of relying upon > Thread::cancel(). Given this, not having cancel() working as before > may well not be that critical. > > Robert. > _______________________________________________ > osg-users mailing list > [email protected] > http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org > -- Rafael Gaitán Linares Instituto de Automática e Informática Industrial http://www.ai2.upv.es http://gvsig3d.blogspot.com Ciudad Politécnica de la Innovación Universidad Politécnica de Valencia
osgviewerandroid.tar.gz
Description: GNU Zip compressed data
_______________________________________________ osg-users mailing list [email protected] http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
