Hi, I opened MR to fix this issue and started a discussion about the proper solution: https://gitlab.freedesktop.org/mesa/mesa/merge_requests/1258
Thanks, Andrii. On Wed, Jul 24, 2019 at 10:32 AM Olivier Fourdan <four...@gmail.com> wrote: > Hi, > > On Wed, May 22, 2019 at 4:01 AM Ian Romanick <i...@freedesktop.org> wrote: > > > On Thu, Jul 19, 2018 at 12:08 PM andrey simiklit > > > > > > Jason, can we reconsider Andrii's patch? It still applies cleanly > > > (https://patchwork.freedesktop.org/patch/237490/) > > > > Looking at the patch and the "Simple reproducer" in the bug, I think > > this just papers over the issue. It seems like the problem is somewhere > > down inside the driver's handling of glXBindTexImageEXT. My best guess > > is that the texture is GL_TEXTURE_2D but the miptree backing it is > > GL_TEXTURE_RECTANGLE. It seems that the glXBindTexImageEXT handling > > should mark the miptree as GL_TEXTURE_2D when binding the image to a > > texture that is GL_TEXTURE_2D. Or is that not possible for some > > non-obvious reason? > > Sorry to be a pain, but I still get bug reports in xfce for this, for > everyone using a Mesa build with debug enabled (or a pre-release for > that matter) on Intel, the `assert()` are enabled and this bug occurs > with the xfce compositor. > > Maybe Andrii's patch is just hiding the issue, but that's already the > case without the `assert()` enabled (i.e. all stable releases of > Mesa), so I guess it's not such a big deal anyway. > > Can we agree on a fix on this? > > Cheers, > Olivier >
_______________________________________________ mesa-dev mailing list mesa-dev@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/mesa-dev