Hi Rui,

just a note about QTimer:

> It's great to see the Qt integration could run in a much beter shape.
> I'd like to have a try these two days and try to make use of a
> separated thread to advance frames instead of the timer.

Using the QTimer with time zero actually triggers the timer on any cpu idle 
time. This is the way to mimics osgViewer::run behaviour. osgViewer spawns 
main loop that checks events, updates scene and makes a frame. We are not 
owning the main loop in Qt. Actually, we can create our own and check for 
events manually, but it brings some problems (as Qt doc says). Thus, it may be 
a better solution to let the Qt keep their main loop and use, for instance 
QTimer with zero time, an user event that is circularly sent from its handler, 
or thread, as you suggest.

Maybe, if there is high GUI load, or high main thread load, separate thread 
may bring some benefits. If this turns to be the case, it is probably the 
problem of osgViewer also as it is also advances frames from main thread.

If you try the separate thread, keep ON_DEMAND run scheme working. I spend 
noticeable time on it ;-) . ON_DEMAND scheme requires osgQt::setViewer() for 
setting the viewer.

John

_______________________________________________
osg-submissions mailing list
[email protected]
http://lists.openscenegraph.org/listinfo.cgi/osg-submissions-openscenegraph.org

Reply via email to