Mark H Weaver skribis: > Andreas Enge writes: >> It looks like the patch broke dmd on i686 and x86_64: >> http://hydra.gnu.org/eval/52352 > > Wow, those tests must be very fragile. All I did was increase all of > the sleep times by a factor of about 3. I didn't look closely at the > logic. I think this is fixed by dmd commit 3c45967 (below), which I just pushed. Thanks, Ludo’.