Created attachment 106221 [details] proposed patch while running DISPLAY=:2 make test TARGETS=image,xlib with Xvfb on display :2, to check a fix for bug #74779, occasionally device-offset.xlib-render-0_0.rgb24 was reported PASS, even if the code modified was not executed at all by that test. The problem is that in that test (and few others) a temporary Pixmap is created on the server and because render-0_0 does not have FillRectangles an image surface is used to clear the Pixmap, but because the image surface is already 'clear' it is left untouched and in _cairo_surface_unmap the 'clear' image is not unmapped because untouched (image->base.serial == 0) http://cgit.freedesktop.org/cairo/tree/src/cairo-surface.c#n688 this means X's Pixmap is left in its original undefined state and only occasionally it is clear. I suggest to increment the serial member of the image_surface when the 'is_clear' state of the image_surface does not correspond to a clear content of the corresponding xlib_surface, as in the attached patch.
-- GitLab Migration Automatic Message -- This bug has been migrated to freedesktop.org's GitLab instance and has been closed from further activity. You can subscribe and participate further through the new bug through this link to our GitLab instance: https://gitlab.freedesktop.org/cairo/cairo/issues/67.
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.