github.com/moarvm/moarvm | IRC logs at colabti.org/irclogger/irclogger_logs/moarvm
Set by AlexDaniel on 12 June 2018.
nine .tell jnthn niner.name/files/inline-issue/master.log.gz vs. niner.name/files/inline-issue/patched.log.gz 07:51
yoleaux nine: I'll pass your message to jnthn.
nwc10 good *, #moarvm 07:52
nine .tell jnthn To reproduce run: `time perl6 csv-ip5xs.pl < /tmp/hello.csv` from Tux' CSV repository and compare Inline-Perl5's master branch to the inline-demonstration branch (1 commit difference). I do not think that it's JIT related. Feels more like inlining will push the inliner's size over the inline limit and we'd benefit from that much more. Or so. 07:54
yoleaux nine: I'll pass your message to jnthn.
brrt ohai #moarvm 13:05
dogbert2_ hi brrt 13:08
nwc10 good *, brrt 13:10
dogbert2_ nwc10: did you notice that jnthn added another debug variable? 13:12
nwc10 oooh, now
er, no
naughty fingers 13:13
dogbert2_ last week, see github.com/MoarVM/MoarVM/commit/7f...7043ede44d
MasterDuke this is probably a better spot for this question. anyone have an idea why this patch gist.github.com/MasterDuke17/a618b...d984af2d3e would cause `Unable to allocate an array of 4607182418800017409 elements` when building NQP with my make-int-default branches? 13:39
nwc10 4607182418800017409 is 0x3ff0000000000001 13:42
completely guessing here, but there's an internal representation choice, isn't there, for some sort of ints (but I forget which) to either be a pointer to a bigint, or the actual int stored inline
is it getting confused by that, somehow?
I repeat, "completely guessing here" 13:43
but that bit pattern is strange.
brrt hmmm 13:46
no idea :-(
MasterDuke looks like set_size_internal was called with 4607182418800017408 (and then internally adjusts to the number reported), so 0x3ff0000000000000, but that doesn't help me 13:51
hm, maybe something is getting upgraded from 32-bit to 64-bit? 13:56
nwc10 it's more likely to be 16-bit to 64-bit 13:57
brrt hmm yeah 13:59
MasterDuke if it helps, here's the full error gist.github.com/MasterDuke17/2b956...d07facd838
timotimo nwc10: that representation is internal and should never become visible at all :) 14:00
lizmat and another Perl 6 Weekly hits the Net: p6weekly.wordpress.com/2019/01/21/...-youtubed/ 22:11