Keith Whitwell wrote: >> - Commits ----------------------------------------------- >> commit 8dd90ee19d97c4b032c2b057d96b3e674be3e1fd >> Author: Brian >> Date: Thu Mar 20 09:15:24 2008 -0600 >> >> gallium: temporarily disable the memcmp() in cso_set_framebuffer() >> >> The memcmp() fails to detect buffer size changes... > > Would it be sensible to include the window dimensions in the frambuffer > struct? I know this would be useful in the driver code.
Yes, I was going to do that, actually. > I'm also a bit concerned about other aspects of this -- it shouldn't be > possible to resize a buffer, so I guess what is happening is that a buffer is > being destroyed and recreated with a new size, and just happening to end up > with the same pointer. > > But what worries me about that is why is it possible to destroy a buffer that > is bound as a render target? Should the driver hold a refcount to prevent > this? In any case, something should prevent this from happining... We're discarding the pipe_buffer and creating a new one, but the pipe_surface object which contains it persists. -Brian ------------------------------------------------------------------------- This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ _______________________________________________ Mesa3d-dev mailing list Mesa3d-dev@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/mesa3d-dev
