On Thu, 2016-03-24 at 11:59 +0100, Marc-André Lureau wrote: > On Thu, Mar 24, 2016 at 12:39 AM, Marc-André Lureau <mlureau@redhat.c > om> wrote: > > > > > > > > When / how it can happen ? Looking at gl_scanout() there is > > > g_return if > > > scanout == NULL, if NULL is a valid/expected return value then > > > that > > > g_return should be changed. > > Good point. AI think I wrote this initially thinking that no one > > should ever call get_gl_scanout() without having first received the > > scanout notification, but that assumption is wrong. It should be > > fine to call get_gl_scanout() and it's easier to deal with a NULL > > return value rather than checking the returned scanout.fd == > > -1. I'll change the assert. (btw, the set_static_boxed for the > > property should be fine) > > Actually, gl_scanout() should only be called with a valid scanout, > the > patch doesn't need to be modified then. If you prefer, we can add a > comment above the g_return_if_fail(scanout). > Yes please Pavel _______________________________________________ Spice-devel mailing list Spice-devel@xxxxxxxxxxxxxxxxxxxxx https://lists.freedesktop.org/mailman/listinfo/spice-devel