github.com/moarvm/moarvm | IRC logs at colabti.org/irclogger/irclogger_logs/moarvm
Set by AlexDaniel on 12 June 2018.
02:15 MasterDuke left
Kaiepi so wide strings on moarvm requires setting the locale with newlocale/uselocale/freelocale (and whatever's used on windows) 02:18
should the locale be set each time a thread starts or should it only be set and cleared when the locale's needed?
the functions i mentioned are threadsafe btw, and can set the locale to whatever the system uses 02:20
05:45 domidumont joined 06:07 lizmat left 06:13 domidumont left 06:14 domidumont joined 06:16 domidumont left 06:39 lizmat joined 07:01 lizmat left 07:12 MasterDuke joined 07:18 domidumont joined 07:34 domidumont left 07:35 domidumont joined 07:47 domidumont left
MasterDuke dogbert11: i've been getting a bunch of fails in what appear to be random test files when i set MVM_GC_DEBUG=1 and MVM_NURSERY_SIZE=8192. i haven't been able to repro any particular fail, but i suspect they're just new appearances of whatever causes github.com/rakudo/rakudo/issues/3030 09:29
but maybe you can dig something new up
09:44 Kaiepi left 09:45 Kaiepi joined 09:50 sena_kun joined 10:01 Kaiepi left 10:05 Kaiepi joined 10:21 Kaiepi left, Kaiepi joined 10:38 Kaiepi left 10:39 Kaiepi joined
AlexDaniel squashable6: status 11:15
squashable6 AlexDaniel, 🍕🍕 SQUASHathon is in progress! The end of the event in 1 day and ≈2 hours. See github.com/rakudo/rakudo/wiki/Mont...Squash-Day
dogbert11 . 12:43
MasterDuke: I'm get plenty of failing spectests when running with a 16k nursery, quite many as a matter of fact 13:17
*I'm getting ..
timotimo oof
dogbert11 all of them vanishes if I set MVM_SPESH_DISABLE=1 13:18
MasterDuke i suspect a single root cause (though i have absolutely no idea what that is) 13:20
dogbert11 hmm, I might have to take that back. I'm sloppy it seems, after rebuilding Rakudo the number seems to go down quite a bit 13:21
timotimo ah. it happens :)
dogbert11 grrr :)
I was playing around with jnthn's frame-opts branch yesterday and forgot to rebuild rakudo after going back to master 13:22
dogbert11 changing the nursery to 8k just like MasterDuke 13:23
timotimo better escape analysis could also make spec test runs with small nurseries just a little bit less horrible :) 13:24
dogbert11 btw, I believe that the third gen Ryzen CPU's will be released tomorrow 13:25
in case someone is thinking of building a new system
MasterDuke i am. and luckily ram prices have dropped a bit 13:28
my desktop isn't scheduled to arrive (along with the rest of the household goods) for another couple weeks. maybe i'll just buy the stuff beforehand and only have to set it up once 13:30
dogbert11 what are you going to buy? 13:31
MasterDuke 3700X 13:33
dogbert11 is that the 8-core 65W TDP model? 13:36
MasterDuke yeah 13:37
dogbert11 very cool 13:39
are you getting a X570 MB as well? 13:43
MasterDuke i think that makes sense, since i need a new MB anyway 13:44
there don't appear to be massive difference, but they aren't horribly more expensive either 13:45
dogbert11 you get PCI4 which might possibly come in handy 13:47
MasterDuke yeah. i don't think i would go out and buy an X570 if i was given something else. but since i have to get something new anyway, might as well future-proof as much as i can 13:53
dogbert11 after running a spectest with an 8k nursery I got ~10 spectest failures
one test hung and I had to kill it. t/spec/S07-slip/slip.t 13:55
several sockaet related test files failed as well but I believe that we have reported those already 13:56
ugexe whatever the current flapper on appveyor is (06-struct ?) also goes away wih MVM_SPESH_DISABLE=1
that one seems to have started earlier this year 13:57
dogbert11 ugexe: that test fails om my RPi 13:59
ugexe does it fail with MVM_SPESH_DISABLE=1 ? hopefully they are related 14:00
dogbert11 let me investigate 14:05
sigh, now it doesn't want to fail anymore (grumble) 14:13
AlexDaniel squashable6: status 15:19
squashable6 AlexDaniel, 🍕🍕 SQUASHathon is in progress! The end of the event in ≈22 hours. See github.com/rakudo/rakudo/wiki/Mont...Squash-Day
AlexDaniel Now with an issue tracker! github.com/ajs/perl6-Math-Sequences/issues
18:53 robertle joined 18:59 lizmat joined
AlexDaniel samcv: while we're squashathoning away, any news about the moar release? 19:24
Geth MoarVM/jit-expr-float: 26 commits pushed by (Bart Wiegmans)++
review: github.com/MoarVM/MoarVM/compare/2...12aa478823
19:48
19:49 brrt joined
brrt \o 19:49
timotimo yo 19:52
brrt turns out that type systems are subtle business, who would've guessed 19:55
timotimo ha
ain't that the truth
Kaiepi so i'm on the last step of implementing wide string support on moarvm, which is dealing with locales 20:09
should i set the locale globally with setlocale or individually for each thread with newlocale/uselocale/freelocale? 20:10
the latter is thread-safe
alternatively i can only set the locale while wide strings are being used 20:12
timotimo that sounds much more sensible, even if multiple locales in the same process isn't necessarily common
Kaiepi aight 20:13
20:13 brrt left 20:16 brrt joined 20:45 lizmat_ joined 20:49 lizmat left 20:53 lizmat_ is now known as lizmat 21:20 brrt left 23:48 sena_kun left