timotimo When invoking frame_name_9928, Provided outer frame 0x3890200 (MVMStaticFrame frame_name_9929) does not match expected static frame type 0x3890080 (MVMStaticFrame message) 00:01
er ... oops?
FROGGS_ timotimo: 1 if <[a]> gives that too 00:02
timotimo: 1 if /<[a]>/ <-- this
timotimo maybe it's my debug output 00:03
FROGGS_ inline if? 00:05
timotimo yeah
er, i mean
$left := $other if $other < $left 00:06
my replacement for the min call
FROGGS_ that is likely 00:07
timotimo very weird, it explodes right in the middle of everything
ah, i think i may know what's wrong
grondilu uv__loop_delete: Assertion `!((((*(&(loop)->active_reqs))[0]) == (&(loop)->active_reqs)) == 0)' failed. 00:08
timotimo that always happens if we've outputted something to stderr
FROGGS_ ahh
1 if /<[a]>/ only explodes because $_ is Mu or something
timotimo mhm
FROGGS_ perl6-m --ll-exception -e '$_.match()' 00:11
When invoking frame_name_10012, Provided outer frame 0x4efb4e0 (MVMStaticFrame frame_name_10013) does not match expected static frame type 0x4efb360 (MVMStaticFrame message)
grondilu makes a manual pull in ./nqp/MoarVM 00:12
[Coke] --gen-nqp=master --gen-moar=master 00:16
(at least until we get our version #'s fixed up)
grondilu I tried that, but I got the failed assertion above 00:17
timotimo gets a newer nqp and hopes that changes something 00:32
but not much changed in between :(
grondilu tries 'rm -rf install nqp' and Configure.pl again 00:35
00:36 jnap joined 00:48 jnap joined 00:49 colomon joined
grondilu watches the compilation tackling m-CORE.settings 01:24
lue grondilu: just so you know, if you get a segfault at the end of the m-CORE.setting step, you should try running 'make' one or two more times. 01:27
grondilu ok 01:28
what does the numbers at each stage mean, already? 01:29
for instance: Stage parse : 455.780
lue How long it took that stage to happen, in seconds. 01:30
grondilu oh yeah I could have guessed that 01:31
wow, compilation has ended with no error 01:37
./perl6-m -e 'say "hi";'
hi
\o/
well done, guys!
JimmyZ \o 01:38
grondilu runs make spectest 01:51
grondilu gladly notices that rakudo counts from 1 to 100 on moar faster than on parrot 01:58
it's funny though, when I run: 01:59
time ./perl6-m -e '.say for 1 .. 100;'
I got a segfault at the end
but not when I run: 02:00
time ./perl6-m -e '.say for 1 .. 99;'
neither when I run:
./perl6-m -e '.say for 1 .. 100;'
hum, it's actually not faster, I spoke too fast. 02:01
diakopter is sad that it's even measurable in any backend 02:02
lue Running perl6-m -v gets me the version info, but perl6-m -e 'say 42' gets me "Missing or wrong version of dependency 'gen/moar/stage2/QRegex.nqp'" :| 02:23
03:17 jnap joined 03:36 jnap joined 04:37 jnap joined 07:39 jnap joined 08:40 jnap joined 08:52 odc joined 09:00 eternaleye joined 09:01 eternaleye joined 09:02 eternaleye joined 09:41 jnap joined 10:41 jnap joined 11:42 jnap joined 12:32 cognominal joined 12:43 jnap joined 13:24 colomon joined 13:44 jnap joined 14:06 jnap joined 15:11 benabik joined 15:38 cognominal joined 15:39 sorear joined 17:54 japhb_ joined 18:00 colomon joined 19:27 dalek joined 19:28 masak joined 19:29 dalek joined 19:32 [Coke] joined
FROGGS perl6-m --ll-exception -e 'Nil.match()' 19:33
When invoking frame_name_10012, Provided outer frame 0x40ebb70 (MVMStaticFrame frame_name_10013) does not match expected static frame type 0x40eb9f0 (MVMStaticFrame message)
what is wrong with Nil?
ahh 19:34
this should happen:
Cannot call 'match'; none of these signatures match:
:(Cool:D : $target, *%adverbs)
nwc10 jnthn: sorry for delay - 80a4be01d5 cherry picked atop my local torture and debugging hacks "passes" 19:50
by inspection, it doesn't seem to add or remove any allocations, so that implies that the GC runs will happen at exactly the same places, which in turn implies that it does fix the mark issue
likely to be some delay in rebasing my torture code onto current HEAD, and having a re-run 19:52
dalek arVM: 57de493 | (Tobias Leich)++ | src/io/fileops.c:
die when attempting to print a NULL string
20:36
[Coke] FROGGS: should that have been inside a conditional? 20:39
FROGGS hehe 20:41
>.<
[Coke]++
dalek arVM: 0b69366 | (Tobias Leich)++ | src/io/fileops.c:
actually check for string being NULL, [Coke]++
20:42
FROGGS [Coke]: but my test worked :o)
nwc10 timotimo: Perl 5's sort may actually be fractionally better than timsort. See www.nntp.perl.org/group/perl.perl5....07355.html 21:01
lue I assume "Missing or wrong version of dependency 'gen/moar/stage2/QRegex.nqp'" is a bug I'll need to wait for being fixed, right? 21:04
[Coke] lue: I'm not seeing that on my daily builds. What's the scenario to reproduce? 21:05
FROGGS I don't have that either
[Coke] do we want moarvm tickets on RT or on github?
lue Install to /usr/local, try running outside the git repo :) [this was from last night; maybe it changed?]
FROGGS lue: that might be a problem 21:06
can you run from inside the repo?
I had to revert a path that made it wortk outside the repo for me, but there was a pitfall I seemed to jump over some how 21:07
lue FROGGS: yes, it does work inside. perl6-m -v works outside the repo however.
FROGGS the -v does almomst nothing, it does not load any module for example 21:08
lue FROGGS: I could imagine, just thought I'd mention that something on perl6-m works outside the repo :) 21:16
FROGGS yeah... I will make that work some day but this will take a few hours 21:17
lue Don't worry too much about it, I'm not desperate. 21:23
22:18 jnap joined 23:12 jnap joined 23:32 jnap joined