github.com/moarvm/moarvm | IRC logs at colabti.org/irclogger/irclogger_logs/moarvm
Set by AlexDaniel on 12 June 2018.
Geth MoarVM: MasterDuke17++ created pull request #949:
Cleanup *_I ops		 01:26
brrt \o 11:19
lizmat o/ 11:20
brrt ohai lizmat 11:35
lizmat waves from the middle of the Irish Sea 11:36
(via satellite :-) 11:37
brrt ooh, nice
lizmat land ahoy! :-) 11:38
timotimo crossing the sea on a satellite? that's decadent 11:55
i do hope you have a safe landing 11:57
lizmat only the best will do :-)
only 45 mins to disembarking
japhb Disembarking from a satellite would be quite the event. I suspect you'd need excellent timing ... 15:30
TimToady I had to quickly leave the South Satellite at Seatac earlier this week... 15:35
fortunately my connection was also delayed due to the incredible pall of smoke rolling down Puget Sound from BC 15:36
or we woulda missed it
we could stare at the orange sun comfortably through the muck, and it was coming from a different country 15:40
usually the CA at fault this way is the country of California 15:42
(the other one our president is waging war against...) 15:43
but I digress, please ignore these satellite remarks... 15:44
diakopter O_- 15:53
Geth MoarVM: 9601ac45bc | (Samantha McVey)++ | tools/release.sh
Don't store owner or group in our release's tar archive

Set owner and group to ID 0 and don't store the same either. This is done so the archive will not show the username of the user who created the release.		 19:04
samcv going to do this ^ for nqp and rakudo as well 19:09
maybe should also disable extended attributes too? 19:10
ah looks like that's disabled by default 19:11
[Coke] samcv++ 19:16
travis-ci MoarVM build failed. Samantha McVey 'Don't store owner or group in our release's tar archive 19:25
travis-ci.org/MoarVM/MoarVM/builds/420254216 github.com/MoarVM/MoarVM/compare/e...01ac45bc1e
diakopter jnthn: how much of GC time is spent marking? most? some? 20:14
jnthn Well, marking - at least for nursery objects - isn't really marking, it's copying. But I'd guess from various profiles I've looked at it's about two thirds going on that, and around a third on the sweep step after. 20:23
diakopter I wonder if the dual-nursery benefit has been eclipsed by all the other specializations for generations 20:34
timotimo diakopter: how do you mean? we shouldn't have semispace any more, or we shouldn't split young and old gen or what? 20:49
timotimo yeah, dual nursery sounds a lot more like semispace 20:54
jnthn I don't think anything has changed much yet in terms of the workload we're putting on the GC. EA will probably make the biggest difference. 21:50
diakopter what's EA :D 21:56
timotimo Escape Analysis 21:57
dogbert17 is this a false positive? 22:08
Thread 4 bound to an attribute of an object (ThreadPoolScheduler::GeneralWorker) allocated by thread 1 22:09
at SETTING::src/core/ThreadPoolScheduler.pm6:237 (./CORE.setting.moarvm:run-one)
from SETTING::src/core/ThreadPoolScheduler.pm6:281 (./CORE.setting.moarvm:) 22:10
timotimo yeah, i'd say it is
dogbert17 darn, was hoping it had something to do with 'MoarVM panic: SC index out of range' problem 22:11
timotimo the worker gets created by the thread that owns the TPS, and the worker sets its "working" flag to 1 at the beginning and 0 at the end of every piece of work
dogbert17 aha 22:12
will try to figure out another approach then 22:13

Powered by Cro and the Raku® Programming Language.

Please report any issues / comments / feature requests as an issue on App::Raku::Log.

Thank you!