On Wed, Apr 28, 2010 at 8:44 PM, Matthias Kilian <[email protected]> wrote: > On Wed, Apr 28, 2010 at 03:41:00PM +0200, David Coppa wrote: >> I'm trying to reproduce the problem running this trivial script many >> times, but the bastard doesn't want to segfault anymore :( > > Holy shit! How often did you try?
for i in $(jot 50); do ./ghc-segfault-test.sh ; done > If this isn't pure luck, the problem is even more bizarre than I > thought, because it means that if you get the segfaults on building > haddock, your stage2 is broken (and if you don't get the segfaults, > your stage2 is not broken). I will try again by leaving it run over night: maybe I'll be more lucky. > To all the kind testers who already wasted so much CPU time: if you > currently have failed build (segfaulting when building haddock with > stage2) on your disk, could you please waste some more cycles and > try Davids script? Please report back wether it segfaults *always* > or only *sometimes*. > > Ciao, > Kili > > ps: I'll probably not have time to really hunt his bug down (or > just disable threading support) before sunday, so don't despair if > you don't get feedback from me until then. Ok! Cheers, David
