Le mercredi 22 juillet 2009 à 18:39 +0200, Lionel Landwerlin a écrit : > Hi ! > > I started playing with DirectFB at the beginning of the month. I now > have a gfxdriver for the proprietary drivers I'm working with. I can run > almost all directFB examples (df_andi, df_porter, df_neo) without any > problem. But I'm getting problems with directfb windows. > > I had developed some kind of simulator for my proprietary driver layer > and I'm able to see directfb buffers (and so surfaces, layers, ...) > while the directfb application is running. > What I see is that all the GTK+ windows are correctly rendered > offscreen, but they never get blitted on the layer they belong to (I'm > using only one layer, it's an hardware limitation). > > Is that a window manager problem ? I tested with both default and sawman > window managers and I had the same problem in both case. > > I actived debug traces and here is the traces I've got after the last > drawing operation on a directfb window : > > (-) [Main Thread 3.629] (29528) IDirectFBSurfaceW: > IDirectFBSurface_Window_Flip( 0x488700, 0x4007d514, 0x00000000 ) > (-) [Main Thread 3.629] (29528) IDirectFBSurfaceW: -> FLIP 4, > 26- 575x 372 > (-) [Main Thread 3.629] (29528) Core/LayerContext: > dfb_layer_context_lock( 0x42f958 ) > (-) [Main Thread 3.629] (29528) Core/LayerContext: -> locked 1x now > (-) [Main Thread 3.629] (29528) Core/WM: > dfb_wm_update_window( 0x479808 [0,0-600x400], [4,26-575x372], 0x0 ) > (-) [Main Thread 3.629] (29528) Core/LayerContext: > dfb_layer_context_unlock( 0x42f958 ) > > As you can see Window_Flip does not trigger any blit or whatever. So the > primary (and only one) layer stays empty. > > I'm looking for help on that point. Any idea is welcome.
Recompiling GTK+ fixed the problem... Sorry for annoyance. -- Lionel Landwerlin <lionel.landwer...@openwide.fr> _______________________________________________ directfb-dev mailing list directfb-dev@directfb.org http://mail.directfb.org/cgi-bin/mailman/listinfo/directfb-dev
