Welcome to the main channel on the development of MoarVM, a virtual machine for NQP and Rakudo (moarvm.org). This channel is being logged for historical purposes.
Set by lizmat on 24 May 2021.
Nicholas good *, * 07:00
MasterDuke exciting-not-in-the-best-of-ways *, * 08:10
but now that we've had a release, and it seems all known mimalloc issues are fixed, anybody have any thoughts about github.com/MoarVM/MoarVM/pull/1645 ? 08:14
nine I for one welcome our new mimalloc overlords! 08:19
MasterDuke and if we want to switch to another microsoft-written allocator, there's github.com/microsoft/snmalloc ! 08:22
some commentary about the two here lobste.rs/s/kfwxvu/microsoft_mimalloc#c_cvykcm 08:23
Geth MoarVM: dogbert17++ created pull request #1689:
Update libuv to version 1.44.1
11:55
lizmat and yet another Rakudo Weekly News hits the Net: rakudoweekly.blog/2022/03/21/2022-...easomatic/ 14:56
Voldenet I've noticed something… relatively strange 15:07
raku.land/cpan:BDUGGAN/Dawa
last paragraph's header looks like this i.imgur.com/PoxH3cW.png 15:08
I'm pretty sure it's caused by readme.md being rendered like this, so `#dists`, `#info` and similar colliding things would break the display on that page 15:10
s/like this/with the name of the header as the id/ 15:14
Oops, I should've posted that on #raku 15:15
MasterDuke i figured out the bug in my implementation of pre-sizing $!mbc when writing frames. now to try and remember why i was doing that... 16:07
i think at least part of it was assuming that when the backing VMArray got very big, it would be more efficient to only increase by the known desired amount, instead of doubling 16:10
nine I just realized: the day we switch to RakuAST for everything will be The Day The World Went Away 17:16
lizmat indeed... World.nqp will be no more 17:24
jnthnwrthngtn Clearly the World is not enough. 18:07
drakonis when is that day :V? 19:37
MasterDuke hopefully tomorrow never dies 19:41
drakonis perhaps raku will finally have excellent editor extensions once rakuast lands 20:24
comma is nice but i'm partial to emacs 20:25
does moarvm have TCO? 20:40
MasterDuke not really 20:44
drakonis i see, it should still be possible to introduce TCO into it, right? 20:46
or in rakudo, i'm not sure where it has to be implemented to allow it
jnthnwrthngtn I suspect at the language level tail calls should be explicitly requested rather than introduced as an optimization. And yeah, it'd need some MoarVM support. 21:30
MasterDuke jnthnwrthngtn: since you're around, do you happen to have an opinion on github.com/rakudo/rakudo/pull/4807 ? 21:32
jnthnwrthngtn MasterDuke: About the last one, did you try doing a substitution in an action method? 21:56
(One that takes the edited code path, that is) 21:57
The first two look reasonable to me
*two commits
MasterDuke jnthnwrthngtn: not sure what you mean re doing a substitution in an action method. to test that the change is correct? or as a better way of implementing that change? 22:05
jnthnwrthngtn To test the change is correct; $/ is bound readonly in that context usually 22:38
MasterDuke m: grammar G { token TOP { <thingy>* }; token thingy { "thingy" } }; class A { method thingy($/) { my $p = "page"; $p ~~ s:g[ \w ] = "a"; $p } }; my $g = G.parse("thingythingy", actions => A.new); say $g 22:55
camelia Cannot assign to a readonly variable or a value
in method thingy at <tmp> line 1
in regex thingy at <tmp> line 1
in regex TOP at <tmp> line 1
in block <unit> at <tmp> line 1
MasterDuke something like ^^^ ? i get the same error on my branch 22:56
jnthnwrthngtn OK, I guess it already trips over the issue before it gets to that point, then 23:21
MasterDuke (and that example does in fact go through the new fast path here github.com/rakudo/rakudo/pull/4807...4804R1813) 23:24
jnthnwrthngtn Alright 23:25
Then I've nothing to add :)
MasterDuke cool. i'll go ahead and merge. lots of time to revert before the next release if need be 23:26
jnthnwrthngtn :) 23:32
MasterDuke++