OK, so, one of the original ideas by javispedro was to write our own WSEGL. I decided to follow up a bit on that idea.

A 'open enough' specification of wsegl.h can be found from http://qt.gitorious.org/qt/qt-gberg/...owervr/wsegl.h

An example WSEGL can be found from http://qt.gitorious.com/qt/qt/blobs/.../pvrqwswsegl.c

Now, this file mentions an interesting part of being a WSEGL:

/* Swap the contents of a drawable to the screen */
static WSEGLError wseglSwapDrawable(WSEGLDrawableHandle _drawable, unsigned long data)

So, my idea is to do the following: Have a WSEGL that our powervr.ini points to. This WSEGL is actually just dlopen()'ing the X11 WSEGL and wraps around it, replacing/extending some functions.

The idea is then in static WSEGLError wseglSwapDrawable(WSEGLDrawableHandle _drawable, unsigned long data) to send a OMAPFB_UPDATE_WINDOW initially for the full screen when the procedure has been executed. For optimization (send only parts of screen), it should be possible to retrieve information on the drawable's nature in longer term.

An initial skeleton of this 'wrapperwsegl' can be found at http://stskeeps.subnetmask.net/maemo...wrapperwsegl.c