Concerning the bug in the doom library, there will (already is) a workaround for it. It is not that I "don't care", I just prefer to have bugs fixed at their origin. As for this problem, it is too early to say. One thing that happened is that the audio.device was recompiled with SAS/C instead of lattice, and SAS/C does not only have a different register allocation (triggering the bug in the doom sound library), but it also optimizes better - which might be a problem with hardware register access.