00:48 oetiker joined 02:16 jimmy_ joined 03:50 prontotest joined, prontotest left 05:02 kjs_ joined 07:22 ggoebel111111117 joined 07:44 FROGGS joined 07:57 zakharyas joined 09:40 kjs_ joined 10:20 kjs_ joined 12:39 brrt` joined 13:42 tgt joined 13:48 brrt joined 14:20 ggoebel111111114 joined 15:23 ggoebel111111115 joined 18:30 njmurphy joined 18:54 FROGGS joined 19:11 FROGGS joined 20:41 kjs_ joined 21:05 ggoebel111111114 joined 21:34 colomon joined 21:40 colomon joined 21:42 kjs_ joined 21:55 carlin joined 22:38 colomon joined 22:50 brrt joined
brrt \o 22:50
jnthn o/ brrt 22:51
brrt i have flappynqp tests (fail under prove, run directly) on windows with master nqp and master moarvm
jnthn On windows? Not seen that, ever. 22:52
brrt hmm 22:53
well, neither have i, but there it is
jnthn Oddness. I'm wrkin gon an advent post and then I should rest because I slept awfully last night. 22:54
But will look tomorrow to see if I can recreate it.
TimToady if it were just random memory differences, you wouldn't get the skew toward failing under prove, but there does seem to be such a skew, which is maybe a big fat clue, if only we were Sherlock Holmes 22:55
brrt install.linux.ncsu.edu/pub/yum/itec.../bumblebee 22:56
random memory differences? 22:57
TimToady a larger or small env, or argv, for instance
anything that moves allocations onto or off of boundaries
but I'm sayin' it doesn't seem to be random, or sometimes it would work under prove and not directly 22:58
but that seems to rarely be the case
brrt hmmm 23:00
it is the case here though
i don't get it at all
TimToady you said it fails under prove, which is the normal case
brrt could it have something todo with the expired windows license the vm is running on? some features disabled? 23:01
it does not fail when running directly
TimToady some kind of ulimit thing?
brrt possible 23:03
but to be honest i have no idea
TimToady I guess there's options to prove one could play with 23:07
to show failure output and such
23:43 oetiker joined