<![CDATA[QThread stack-address]]>Greetings,

I use Qt4.8 on Windows and I´m searching a way of getting the stack-address of a QThread or
set the stack myself.

With pthreads it´s possible to get/set the stack-pointer by pthread_attr_getstack() and
pthread_attr_setstack().
Since QThreads are built upon pthreads there should be a solution but I only found
QThread::setStackSize and I don´t like the idea of running a pthread in my QtApp only
because of getting the address.

The reason I need this is preventing the garbagecollection of my embedded interpreter
which lives in a QThread from running amok ( http://bugs.ruby-lang.org/issues/2294 )

I would be really thankful for getting some help/advice (also if it´s just telling me that it´s not possible -.-)

]]>https://forum.qt.io/topic/21421/qthread-stack-addressRSS for NodeWed, 16 Sep 2026 18:25:06 GMTWed, 14 Nov 2012 21:05:40 GMT60<![CDATA[Reply to QThread stack-address on Fri, 16 Nov 2012 09:30:00 GMT]]>Helps a lot. Thank you very much for explaining rcari.

]]>https://forum.qt.io/post/157181https://forum.qt.io/post/157181Fri, 16 Nov 2012 09:30:00 GMT<![CDATA[Reply to QThread stack-address on Thu, 15 Nov 2012 13:20:48 GMT]]>Well it is based on pthread on UNIX systems but on native Windows threads on Windows. Qt is a cross-platform API and provides access to what can be used in a cross-platform way. The address of the stack must therefore not be portable, otherwise it would be part of the API.
QThread was not really intended to be subclassed with @QThread::run()@ overloaded and used to fire up custom threads anyway. The initial design goal was to create a secondary event-loop in which QObjects could live and handle signals/events without blocking the main GUI thread. @QObject::moveToThread()@ is used for just that.
So for your specific platform-related need, I think you'll have to go the platform-specific way and create your own pthread/native thread and use the platform API to achieve what you want.
Hope this helps!

]]>https://forum.qt.io/post/157111https://forum.qt.io/post/157111Thu, 15 Nov 2012 13:20:48 GMT