>I'm starting to work through your patches, finally :) Thank you Denis.
>Have you seen that CoreSurfaceBuffer is a FusionObject now. It >can be decoupled from the surface meaning the back pointer >becomes NULL, any Deallocate or Unlock may never rely on a >valid 'surface' pointer in a buffer. I had not noticed this yet. Having their own destructors will be helpful. Still makes sense to have a buffer specific notifications (at least for destruction) I believe. >I have done the same with the allocations, so >CoreSurfaceAllocation is a FusionObject as well, but I haven't >committed that yet. Nice. With these additional FusionObjects, it seem to me that it may be helpful to create a consistent approach/way for local/process-specific resources associated with a FusionObject to be destroyed before its shared resources when a) there are no more outstanding references to it (of course) or b) the slave process that owns the local resources exits/releases the super interface. Cheers, Timothy -- Timothy Strelchun Software Engineering NTG Service Provider Division (formally DHG) Intel Corporation _______________________________________________ directfb-dev mailing list directfb-dev@directfb.org http://mail.directfb.org/cgi-bin/mailman/listinfo/directfb-dev
