github.com/moarvm/moarvm | IRC logs at irclog.perlgeek.de/moarvm/today
Set by moderator on 27 September 2013.
02:01 benabik joined 02:02 FROGGS joined 06:32 diakopter joined 06:38 foo_bar_baz joined
diakopter jnthn: curious... I set the gen2 ratio to 1 like you suggested, and nmake selftest All tests successful. 06:57
jnthn: .. and with --optimize, selftest takes only 16 seconds o_O 07:02
JimmyZ and master? 07:03
diakopter I'll try it.. I doubt it's any different though 07:04
07:04 FROGGS joined
diakopter JimmyZ: does nqp t\\moar\\dirs.t fail on master? 07:05
JimmyZ nope
diakopter hrm
could you try the gcorch branch and help me diagnose it? 07:06
JimmyZ ../moarvm nqp.moarvm t/moar/dirs.t failes
While looking for './MASTTesting.moarvm': no such file or directory
diakopter heh, well, you're the first person to try that :P
FROGGS hmmm, I thought the t/moar tests must be run with ../moarvm directly, no? 07:26
yoleaux 00:05Z <Mouq> FROGGS: lol @ nqp/cursless 927729b :) I like it much better than what I wanted to do; create a simple QLang for NQP with :cclass and reorganize the code in NQP and Rakudo to reflect this :p
diakopter FROGGS: no, the t/moar can only be run with parrot's nqp 07:28
FROGGS ahh, okay 07:30
thanks
JimmyZ diakopter: nqp t/moar/dirs.t pass in gcorch here
t/moar can be run with moar's nqp, someday 07:31
diakopter well, if you can make a version of MASTTesting.nqp that can be compiled to a MASTTesting.moarvm, then yes :)
08:02 foo_bar_baz joined 08:22 odc joined
jnthn diakopter: oddness; I suspect it's a case of hiding a bug, though... 08:27
08:46 lizmat joined 09:25 grondilu joined 11:48 colomon joined 14:21 jnap joined 14:28 cognominal joined 15:21 FROGGS joined 15:26 grondilu joined 16:18 benabik joined 16:23 ssutch joined 17:05 jnap joined 17:26 jnap joined 17:54 jnap joined 18:23 cognominal joined 19:17 grondilu_ joined 19:52 grondilu joined 19:54 japhb_ joined 19:56 foo_bar_baz joined
grondilu $ perl6 -e 'say .WHAT for (^3).list.combinations(2).perl.tree' 20:42
(Array) (Array)...
so I guess that's fine
TimToady well, except maybe for the overhead of creating arrays 20:44
grondilu not sure what you mean
TimToady it would be better, I think if whatever was process the lol could simply bind to a **@lol rather than transforming each parcel to an array, then binding 20:45
*processing 20:46
just seems kinda like wasted motions
grondilu well, it's a first draft, isn't it? Can I send a pull request? 20:53
grondilu does. Won't take offense if rejected :)
github.com/rakudo/rakudo/pull/211 20:55
benabik ww? This is #moarvm, not #perl6 20:56
grondilu oops
sorry
benabik NBD, just thought you might get more response over there. ;-)
21:13 jnap joined 21:41 donaldh joined
diakopter jnthn: "a case of hiding a bug" as opposed to what? 22:15
jnthn diakopter: Who knows, I hadn't drunk any coffee when I wrote that :P 22:21
diakopter ok. I mean, I thought the whole purpose of the experiment was to see if was hiding a bug. :) 22:30
jnthn No 22:41
The purpose of the gen2 ratio change was to *try* and make *more* crash, not hide the bug.
Maybe choose a nubmer like 3 22:42
diakopter hrm
ok
whaaa 22:44
yeah, fail
during parsing
23:57 FROGGS joined