00:41 sena_kun left
vrurg greppable6: \$!dynamic_context 00:57
greppable6 vrurg, Found nothing!
00:58 sena_kun joined
sjn Hm. faq.raku.org has a cert error (invalid domain) 01:45
02:42 sena_kun left 02:55 sena_kun joined 04:43 sena_kun left 04:56 sena_kun joined
rba sjn: After the server crash faq.perl6.org went down and it‘s not up again. So not migrated to raku domain. If one could point me which repo it‘s come from, or what we should do I will bring it back. Where is it linked? 06:25
06:42 sena_kun left 06:48 MasterDuke left 06:49 domidumont joined 06:54 domidumont left 06:55 sena_kun joined 07:34 domidumont joined 08:08 MasterDuke joined
lizmat Files=1301, Tests=109805, 210 wallclock secs (28.49 usr 8.15 sys + 2935.26 cusr 272.18 csys = 3244.08 CPU) 08:32
MasterDuke lizmat: did you see my question from last night? 08:41
lizmat scrolled back and doesn't have an answer 08:42
08:42 sena_kun left
lizmat appears to have developed a write-only comment mode 08:42
MasterDuke heh, no worries 08:44
then a more open-ended question. think an Int.chars that uses a new nqp::chars_I op that doesn't have to actually create the string just to count the digits would be useful? 08:46
lizmat aaaah...now I remember 08:53
yeah, that could be useful 08:54
it'll come back to me why I thought that'd be useful :-) 08:55
probably after I put down by Rakudo Weekly hat which I've just put on 08:56
08:58 Kaiepi left, sena_kun joined
Geth rakudo: 06ae7d614d | (Elizabeth Mattijsen)++ | src/core.c/Mu.pm6
Make the Mu.perl fallback method a multi

As suggested by Vadim Belman. This fixes
   github.com/rakudo/rakudo/issues/3433
09:51
lizmat sena_kun: ^^^
afk for a few hours&
09:53 Kaiepi joined
sena_kun lizmat++ 09:55
10:19 |Tux| left 10:23 |Tux| joined 10:42 sena_kun left 10:58 sena_kun joined
|Tux| Rakudo version 2019.11-506-g06ae7d614 - MoarVM version 2020.01-17-gefb014f19
csv-ip5xs0.724 - 0.742
csv-ip5xs-206.404 - 6.428
csv-parser22.743 - 23.032
csv-test-xs-200.418 - 0.426
test7.455 - 7.712
test-t1.781 - 1.798
test-t --race0.789 - 0.795
test-t-2029.829 - 29.946
test-t-20 --race8.677 - 8.799
11:14
12:42 sena_kun left 12:57 sena_kun joined 13:30 lucasb joined 14:41 sena_kun left 14:56 sena_kun joined
Geth nqp: cb3c3369df | (Elizabeth Mattijsen)++ | src/how/NQPAttribute.nqp
Add NQPAttribute.is_bound method

Just to be on the safe side, as HLL Attribute class has one as well.
15:55
lizmat notable6: weekly 16:13
notable6 lizmat, No notes for “weekly”
lizmat well, that's easy then
16:19 Kaeipi joined 16:23 Kaiepi left 16:40 domidumont left 16:41 sena_kun left 16:45 Kaeipi left 16:48 Kaiepi joined 16:56 sena_kun joined
AlexDaniel sena_kun: btw sometimes your commits are signed and sometimes they are not :) 17:05
sena_kun AlexDaniel: is there anything wrong with that? :) I think it depends on the repo I commit from... 17:12
AlexDaniel sena_kun: why not set it up system-wise?
I mean, just put it in your global git config or whatever is required to do so 17:14
sena_kun AlexDaniel: by the way, as you're here, I have a question related to whateverable. I see both cro and HTTP::UserAgent are used. And Cro is used in a single place to do a simple query. Was it intentional? I think the footprint can be made much smaller removing one of these.
AlexDaniel sena_kun: new code is using Cro, old code is not touched :D
sena_kun AlexDaniel: so if we're migrating, we're migration to Cro?
AlexDaniel yea 17:15
definitely
sena_kun I see
ok, will see if I'll be able to get to it somehw
AlexDaniel I'd appreciate that a lot, but it's probably not as easy as it sounds :) 17:18
there's also a dep on HTTP::Server::Async
17:18 domidumont joined
AlexDaniel pretty sure that code was written before Cro existed, or at least before it was awesome 17:18
17:32 Kaiepi left 17:33 Kaiepi joined
lizmat just noted that sena_kun is missing from CREDITS :-) 18:41
18:41 sena_kun left 18:57 sena_kun joined
lizmat and another Rakudo Weekly News hits the Net: rakudoweekly.blog/2020/01/20/2020-...aittraced/ 19:04
19:39 domidumont left 19:46 sena_kun left 19:54 dumarchie joined
dumarchie lizmat: you said you would like to see .List and .Array coercers implemented in the Iterator role 20:01
tellable6 2020-01-19T20:50:18Z #raku-dev <MasterDuke> dumarchie: i don't have an informed opinion on whether it should be hidden (i'd lean towards no absent any other information), but if not, then yes it should be spectested and written up in the docs
2020-01-19T23:02:33Z #raku-dev <lizmat> dumarchie I'd rather see .List and .Array coercers implemented in the Iterator role
dumarchie lizmat: to me it seems weird to coerce an Iterator to an Iterable 20:03
lizmat yet, that is what List.from-iterator does ? 20:04
dumarchie I don't see that as a coercer, more as a constructor. 20:05
(that's why I mistakenly called it List.new-from-iterator)
lizmat in any case, you want to turn an iterator into a List, no ? 20:06
dumarchie Not exactly. I want to create an Iterable (in this case a List) from an Iterator. 20:07
Coercion can make sense if there is some acceptable way to disable other code from using the Iterator afterwards. Seq does this by undefining the underlying iterator. 20:10
lizmat ah, I see, you would not want a reified List, but one that would still need to be reified 20:13
and having the iterator external, runs the risk of being used again 20:14
well, that's the general issue with using iterators directly, I'd say
dumarchie You're right. Maybe List.from-iterator should be hidden after all. 20:15
lizmat yeah, in general, iterators are things that run under the hood, so to speak
dumarchie Does ALLCAPS count as hiding? 20:16
lizmat it effectively does, yes
so maybe we should just ALLCAP it
this would then also apply to List.from-slurpy / from-slurpy-onearg / from-slurpy-flat 20:18
and Array.from-iterator / from-list and reification-target in Array and List 20:19
dumarchie Probably... I just searched a fairly recent checkout of the roast and from-iterator does not appear. 20:20
I'm still wondering if there's no decent way to "claim" an object for another object... 20:23
lizmat not sure what you mean? 20:24
dumarchie Well, if you could somehow tell the Iterator that it should only accept calls from a certain Iterable everything would be fine, wouldn't it? 20:27
lizmat ah, yes... no, unless you want to set up a system like vowed promises
dumarchie Too much overhead, I suppose :) 20:28
lizmat yeah :-( 20:29
dumarchie On the other hand there's a story about rope and a gun. So we could make List.from-iterator public with a stern warning. 20:31
lizmat as I just said... it's one of a set in List 20:33
why do that one, and not other ones?
and I still think it *is* an implementation detail
it would be different if we would disguise them as coercers
currently coercing an iterator to a List is meaningless: it creates a List with the iterators as its only element 20:34
dumarchie The coercers could work, but the argument would have to be in Iterable in my opinion. Every Iterable should have an iterator. 20:36
lizmat m: class A does Iterable { } 20:37
camelia ( no output )
lizmat m: class A does Iterable { }; A.new
camelia ( no output )
lizmat hmmm... that should complain, I guess
my role Iterable {
method iterator() { ... }
dumarchie m: Any.iterator 20:38
camelia ( no output )
lizmat Ah, I guess that's why...
dumarchie multi method iterator(Any:) 20:40
lizmat yeah 20:43
dumarchie There's already Mix/Bag/Set(Hash) coercers in Iterable. These use private ALLCAP subs. It doesn't get more hidden than that.
I can try to add a List coercer to Iterable that replaces List.from-iterator. 20:45
lizmat a coercer should live in the Iterator role
and that means it will have its method copied to all iterators out there 20:46
so I'm a bit reluctant to do that without approval of jnthn
dumarchie To me Seq.List still makes more sense than Seq.iterator.List 20:48
because (1) it hides the iterator and (2) will be easier to document without delving into the details 20:50
lizmat m: dd (1,2,3,4,5).Seq.List # that already works ?
camelia (1, 2, 3, 4, 5)
lizmat m: dd (1,2,3,4,5).Seq.Array # that already works ?
camelia Array element = [1, 2, 3, 4, 5]
dumarchie (Any) method List 20:52
Same for Array, without mentioning the list method. 20:53
The list method uses List.from-iterator(self.iterator) 20:55
in case of Seq invocant
or more precise, in case of PositionalBindFailover invocant 20:57
20:57 patrickb joined
dumarchie I checked the code base and in most cases from-iterator could be replaced by from-iterable, but List.STORE and sub SEQUENCE in operators are notable exceptions, 21:27
So maybe converting from-iterator & co. to ALLCAPS is the best approach after all. 21:34
patrickb The Google Summer of Code deadline is in two and a half weeks. May I ask everyone to think about whether you are willing to mentor *now*. When people start thinking about whether they'd like to mentor three days before the deadline, this will not work out. 21:59
tellable6 2020-01-17T08:21:08Z #raku-dev <nine> patrickb: you have never answered my question: colabti.org/irclogger/irclogger_lo...-12-10#l39
2020-01-17T16:39:39Z #raku-dev <tony-o> patrickb i have most of the infrastructure for an ecosystem fleshed out, i need some help testing it - i'm writing on a blog post on the current state rn
patrickb Also we not only need mentors, but also project ideas. Experience from last year shows that students do not come up with their own ideas, but only apply for ideas we proposed. So it's good to have a concrete project idea you are willing to mentor for.
jnthn: Last year, you showed willingness to mentor. Would you be willing again this year? Can you come up with a project idea?
brrt: Would you be willing to mentor again? Might a follow up project of last years GSoC be a good project idea?
tellable6 patrickb, I'll pass your message to brrt
patrickb samcv, jmerelo: You mentored last year. Are you willing to do it again? Do you have a project idea?
timotimo: You pondered mentoring last year. Maybe this year?
vrurg, nine, AlexDaniel, sena_kun, lizmat: All of you are very active committers. Can you come up with a project idea that could be suitable for a student? Would you be willing to also mentor such a project?
perl-foundation-outreach.github.io...020-ideas/ 22:01
Geth rakudo: 91a324a066 | (Elizabeth Mattijsen)++ | src/core.c/StrDistance.pm6
Make StrDistance.new about 30% faster

And save 2 Scalar allocations per StrDistance object
22:02
rakudo: fc8bfc2e82 | (Elizabeth Mattijsen)++ | src/core.c/Hyper.pm6
Save one Scalar allocation for each Hyper object
lizmat patrickb I will think about it 22:03
22:07 dumarchie left
samcv patrickb, well. i have no ideas. but if somebody else has an idea i could maybe do one 22:10
hmm... 22:12
do we have any cassandra bindings or kafka? 22:13
patrickb pretty sure cassandra not, kafka I do not know. 22:14
samcv what could be fun is improving the speed of marshalling from raku to perl 5. but idk how much that can be improved or not since i haven't looked into it 22:15
i'm just throwing out ideas based on technology frameworks i use at work with perl 5 that may/may not have raku verisons since i am running out of ideas
MasterDuke i think the idea i proposed last year is still valid, but unfortunately i also still won't be able to mentor 22:20
samcv MasterDuke, what was it again? 22:33
MasterDuke add performance benchmarks/regression tests to our ci toolchain 22:34
i wrote a 2-3 sentence description in some repo jmerelo created, but it wasn't adopted as one of the ideas the perl foundation proposed 22:36
23:18 lucasb left
MasterDuke samcv: github.com/perl-foundation-outreac...vements.md 23:40