github.com/moarvm/moarvm | IRC logs at irclog.perlgeek.de/moarvm/today
Set by moderator on 27 September 2013.
diakopter BenGoldberg: heh. 00:29
01:13 benabik joined 01:50 FROGGS_ joined 02:15 foo_bar_baz joined 02:59 krunen joined 03:48 benabik joined 04:44 BinGOs_ joined 04:45 jnthn_ joined
TimToady Ooh, just noticed: All tests successful! Yay for our team! \\o/ 05:17
diakopter ^ insert <but I commented out two tests> wet blanket here 05:21
TimToady is okay, I live for my delusions... :) 05:22
06:54 BinGOs joined 07:02 FROGGS_ joined 07:40 cognominal joined 08:26 grondilu joined
jnthn o/ 08:47
yoleaux 03:26Z <lue> jnthn: I was somehow able to solve that sanity check fail by converting my pointy blocks to inner subs. The pointy blocks that manipulate the DB still work as intended though.
JimmyZ \\o 08:52
10:35 cognominal joined 10:54 dalek joined 11:06 cognominal joined 12:33 colomon joined 14:16 jnap joined 14:23 jnap left
diakopter . 14:54
moritz .. 15:05
FROGGS ... 15:06
jnthn ∓
FROGGS ⁘
masak ::: 15:07
FROGGS ⁘⁚⁘ 15:08
TimToady ???? 15:10
15:25 colomon joined 15:32 sorear joined 15:40 ggoebel2 joined 15:51 benabik joined 16:04 FROGGS joined
grondilu what's up guys? you've been quiet pretty much ever since nqp.moarvm passed all test. I guess jnthn takes some well deserved rest, but still... 16:04
diakopter zzzzzzz
grondilu: you gotta know, the test suite has only like 40% coverage of what rakudo needs to bootstrap 16:05
grondilu oh, that's a turn off indeed.
jnthn It's not *quite* that bad. :) 16:06
If you include t/serialization and we pass all the bigint tests (do we? :))
TimToady grondilu: also, you should know that ports like this always give the appearance of accelerating because bugs tend to fail early but the fixes are cumulative 16:12
16:31 BenGoldberg joined
timotimo huh. the nqp-cc doesn't build any more 16:46
where do i make selftest to look at all the cool progress?
jnthn "doesn't build"? 16:48
timotimo first a gajillion warnings like this one: ../../src/core/ops.c:987:9: warning: initializer-string for array chars is too long for C++ [-Wc++-compat]
then a crapton of messages like this one: error:imcc:syntax error, unexpected DOT ('.') in file 'MASTOps.pir' line 4920
diakopter yeah doens't bjuild ;basform emoem iethere
timotimo ... yeah, that's what i thought
16:57 not_gerd joined
not_gerd o/ 16:57
diakopter o/ 16:58
not_gerd timotimo: the warnings can be safely ignored
you need a fairly recent nqp to build moarvm 16:59
timotimo ah, mhm
jnthn Did you re-Configure?
It shoulda told you that
timotimo yes, i did
jnthn I bumped le NQP_REVISION...
hmm
dalek Heuristic branch merge: pushed 46 commits to MoarVM/ctypes3 by gerdr 17:00
TimToady had to include the --gen-parrot last night 17:01
diakopter too
timotimo so nqp-cc is the right place to go?
timotimo does the --gen-parrot trick
TimToady and then perl Configure.pl --gen-parrot
diakopter nthn: I had to blow away install/, nqp/, parrot/
TimToady didn't 17:02
TimToady hasn't even done a make clean for a week or two 17:03
timotimo no ladies coming over to visit recently?
diakopter not to clean! 17:04
TimToady install any of N different Your Mom jokes here... 17:06
oh wait, there's only 1 different one 17:07
jnthn walk, dinner &
TimToady which is the Meta Your Mom joke I just made
diakopter never meta your mom I didn't like 17:08
TimToady yes, we know
timotimo so, ready to hook up the qregex tests? it seems like it'd require a QRegex.moarvm to be generated 17:15
.o( how i meta your mom )
or is the nfa implementation missing at the moment? :P
dalek arVM/ctypes3: fa66d80 | (Gerhard R)++ | src/native/reprs/C (2 files):
Wire up CStruct and CUnion
17:18
diakopter jnthn: you always wanted your own C interpreter.... ;) 17:19
17:29 ggoebel3 joined
TimToady
.oO( "Anything you can do, I can do meta!" from Annie, Get Your Mom )
17:32
masak .oO( "...I can do mater..." ) 17:39
jnthn If we're parsing NQP without the NFA, I'd like to know how :P
Three's a QRegexMoar 17:40
diakopter no, a crowd
I mean, company.
jnthn :P
diakopter >.<
dalek arVM/ctypes3: bcc4354 | (Gerhard R)++ | / (8 files):
Make CPtr work again

Feature-wise we should now have recovered from the refactor.
17:45
timotimo how come i set ulimit -m to something as low as 6000 and moarvm still happily fills up all my 8 gigs of ram and then a couple of gigs of my swap, making my whole system unusable for a few minutes each time? 18:25
ulimit -m 6 should allow it to use 6 kilobytes, but it just keeps growing and growing 18:26
-v seems to help better
(segfaults oh yeah!)
PerlJam timotimo: From the bash docs: -m The maximum resident set size (many systems do not honor this limit) 18:27
timotimo: that last parenthetical might be important here :)
timotimo ah, great :) 18:28
now i turned moarvm into a D state
diakopter derivative? 18:29
timotimo holy hell, killing that took ages
i think i ought to turn off core dumping
18:34 colomon joined
timotimo it seems to be that match(Str, Regex, :global) is messed up in moarvm. 18:34
at least it doesn't get past my @parts := match($line, /\\T+/, :global); 18:35
when do we get an evalbot for moar-nqp? :P 18:37
nwc10 probably the non-smiley answer is "once it self hosts NQP" 18:39
diakopter well, we could have one now, probably 18:41
ish
19:27 colomon joined 19:52 japhb_ joined 20:18 BenGoldberg joined
not_gerd bye, #moarvm 20:37
FROGGS o/ 20:38
20:39 not_gerd left
timotimo as soon as you fix match :global, you'll probably get a whole bunch more passing tests :P 20:49
(once you copy over the qregex tests and replace use QRegex with QRegexMoar)
21:14 foo_bar_baz joined 22:09 benabik joined 22:58 BenGoldberg joined 23:33 benabik joined