Comment 28 for bug 1492037

Good job tracking it further! I kind of gave up, especially after figuring out the workaround by choosing UXA as AccelMethod.

I'm not sure exactly what the proper behavior should be here (should everything check if the mipmap tree is null?), which makes it harder for me to get started on a patch. Trying to trace the failing path in the kernel is a bit beyond my skills and what I have time for at the moment