IRCloggy #perl6 2018-07-29

Logs Search ←Prev date Next date→ Channels Documentation

Provider of IRC logs since 2005.
WARNING: As Freenode became unjoinable and lost all warnings in topics, we cannot log channels on Freenode anymore.

2018-07-29

timotimo here in line 456 it does ~~ m:global/<$selector_tree_key_pattern>/00:00
that goes via EVAL00:00
the profiler dying or at least getting confused when confronted with lots of EVAL isn't new :(00:02
MasterDuke and an explicit EVAL here https://github.com/jdv/p6-data-selector/blob/master/t/03-apply_tree.t#L9400:05
fake_space_whale ryn1x, Juerd, thanks for the advice I would look at packaging it up00:13
→ mcmillhj joined00:16
→ fbynite joined00:19
← mcmillhj left00:21
← fbynite left00:21
← epony left00:28
Geth ¦ doc: bddb87747a | Coke++ | doc/Language/5to6-nutshell.pod600:40
¦ doc: pass finicky links test 00:40
¦ doc: review: https://github.com/perl6/doc/commit/bddb87747a00:40
synopsebot Link: https://doc.perl6.org/language/5to6-nutshell00:40
→ epony joined00:42
→ mcmillhj joined00:43
← mcmillhj left00:48
→ kalkin-- joined00:54
← kalkin- left00:57
→ mcmillhj joined01:02
← ryn1x left01:03
← mcmillhj left01:06
← wamba left01:21
→ araraloren joined01:23
→ mcmillhj joined01:24
← mcmillhj left01:29
→ mcmillhj joined01:51
← mcmillhj left01:56
→ mcmillhj joined02:09
← mcmillhj left02:13
→ pmurias joined02:33
→ mcmillhj joined02:35
← mcmillhj left02:40
← diakopter left02:40
→ mcmillhj joined02:48
← mcmillhj left02:52
← epony left02:57
→ epony joined03:02
← Sgeo left03:03
→ Sgeo joined03:05
→ mcmillhj joined03:10
Geth ¦ ecosystem: 3e945a603f | (Fernando Correa de Oliveira)++ (committed using GitHub Web editor) | META.list03:14
¦ ecosystem: Add Injector to the ecosystem 03:14
¦ ecosystem:03:14
¦ ecosystem: Adding https://github.com/FCO/Injector03:14
¦ ecosystem: review: https://github.com/perl6/ecosystem/commit/3e945a603f03:14
← mcmillhj left03:15
→ mcmillhj joined03:27
← vrurg left03:29
← mcmillhj left03:31
avuserow m: sub f (FatRat() \y, FatRat() \z) {108 - ((815 - 1500 / z) / y)}; my @results = 4.0, 4.25; for 2 .. 200 -> $i {@results.push(f(@results[$i-1], @results[$i-2]));}; say @results[200];03:34
camelia rakudo-moar b7379a804: OUTPUT: «4.9999999999999999999999999999999999999999999914634955237594519840605021696245253531402251293461657465278749240871156571863477542965001140165357␤»03:34
avuserow the above is from https://medium.com/@bellmar/is-cobol-holding-you-hostage-with-math-5498c0eb428b03:35
Perl 6 works very well if you ask for FatRats, and pretty well if you just use regular rational numbers03:35
benjikun can also be slow so you have to be careful03:39
avuserow true. but accuracy comes first, and it's great to have the accuracy within such easy reach03:41
m: sub f (Num() \y, Num() \z) {108 - ((815 - 1500 / z) / y)}; my @results = 4.0, 4.25; for 2 .. 20 -> $i {@results.push(f(@results[$i-1], @results[$i-2]));}; say @results[20]; # floating point version, note that it explodes at 20 versus 20003:42
camelia rakudo-moar b7379a804: OUTPUT: «100.00001247862016␤»03:42
benjikun Indeed03:43
Some would criticize p6 for using rationals by default03:43
Easy to write it off as "it's slow by default, I'm done", though being an incorrect synopsis of the language itself03:44
avuserow hmm. I had forgotten that, with my recent adventures using native types almost exclusively due to nativecall and handling binary file formats03:45
benjikun I do think it is SOMEWHAT weird that perl6 uses rats by default, despite it being more accurate, but I've gotten used to it03:46
Hopefully newcomers can too03:46
jdv79 i get the evals in the test file - is there a better way to deep clone a hash like that?03:56
i didn't know about the regex boiling down to an eval.03:56
is there a better way to do that?03:56
plus what's wrong with eval?03:56
→ mcmillhj joined04:02
geekosaur safety04:04
although slowness also comes into it, we precomp for a reason but EVALs can't be precomp-ed because the thing to be compiled isn't known at compile time, only at runtime04:05
← mcmillhj left04:06
jdv79 its unfortunate a regex has to eval04:15
the explicit eval i could get around easily but probably not the regex implied ones04:15
geekosaur regex has to eval if you use <$foo> because it has to compile $foo in04:19
→ mcmillhj joined04:20
← mcmillhj left04:25
[Coke] cd sandbox/perl604:25
04:25
oops04:25
→ vrurg joined04:35
fake_space_whale Am I correct in thinking that one can not call .WHAT on a object of type Distribution::Resource (from the %?RESOURCES object)?04:35
← vrurg left04:36
fake_space_whale The reason I am thinking this is because when I add a line of code `say $filename.WHAT` I get the error "Cannot look up attributes in a Distribution::Resource type object"04:38
geekosaur .WHAT is a compiler macro; resources are runtime04:40
this might qualify as an LTA error04:40
$filename.^name might work better04:40
sjn o/04:42
sjn bumped into interesting the might be worth looking at for speeding up floating point -> string conversion: https://pldi18.sigplan.org/event/pldi-2018-papers-ry-fast-float-to-string-conversion04:44
sjn maybe this belongs in #rakudo?04:46
fake_space_whale geekosaur, thank you04:46
do you know of any docs or hints or even code about what the Distribution::Resource is?04:47
→ mcmillhj joined04:48
geekosaur no, I think that whole area is a bit underspecced at the moment?04:49
benjikun sjn: Moreso perl6-dev imo04:49
You could also start a conversation about it on rakudo's github04:49
← mcmillhj left04:52
fake_space_whale okay, thanks anyway04:52
Geth ¦ doc: coke self-unassigned Sort Table of Contents https://github.com/perl6/doc/issues/39205:01
← aindilis left05:11
→ kaare_ joined05:22
jdv79 which makes regexes less useful for high perf settings:(05:24
at least p5's are pretty fast:)05:24
benjikun There are ways of avoiding regexes for now sometimes05:28
They'll undoubtedly perform better in the future05:28
fake_space_whale it appears, to my noob eyes, that the Distribution::Resource objects are IO::Path objects.05:29
→ molaf joined05:32
Zoffix sjn: we're pretty much already using state of the art for Grisu3 algo for floating point stringification. I flipped through that talk and I see the guy reference Dragon4 algo, but neither Grisu no Errol, which makes me a bit dubious he really came up with something that's faster than Grisu3+Dragon4 fallback that still fullfills the same properties (Str->Num->Str roundtripping without drift and stringifying to05:51
shortest strings). Right now we use sprintf as fallback for Grisu3, and it could be made faster by using Dragon4 instead, but that covers just 0.5% of cases Grisu3 can't handle, so there won't be any major improvents. There are some references and explanations in this commit: https://github.com/MoarVM/MoarVM/commit/067c0594103a025472c39ae7f8670e47211b874b05:51
fake_space_whale: no, but it attempts to fake being one05:51
m: say Distribution::Resource.^mro.map: *.^name05:51
camelia rakudo-moar b7379a804: OUTPUT: «(Distribution::Resource Any Mu)␤»05:51
Zoffix fake_space_whale: and the bug with .Str/.gist/.perl is something you could fix, perhaps? R#214305:55
synopsebot R#2143 [open] : https://github.com/rakudo/rakudo/issues/2143 [easy to resolve][good first issue] Can't .gist/.Str/.perl Distribution::Resource:U 05:55
fake_space_whale oh, I am not making any promises but I will definitely consider doing that.05:57
Zoffix sweet \o/05:57
→ rindolf joined06:03
Zoffix Talk Slides and Recording: "Intro Into Perl 6 Regexes and Grammars Perl 6": https://rakudo.party/post/Intro-Into-Perl6-Regexes-and-Grammars-Perl-6-Talk--Slides-Recording06:03
sjn Zoffix: he references Grisu3 several times, and compares it to his proposed algorithm06:04
I'd say have another look :)06:04
→ psychoslave joined06:05
Zoffix sjn: ah, cool06:05
|6d f.p. stringification talk: https://pldi18.sigplan.org/event/pldi-2018-papers-ry-fast-float-to-string-conversion06:05
ZofBot Zoffix, Will remind you on 2018-08-04T02:05:39.259931-04:00 about f.p. stringification talk: https://pldi18.sigplan.org/event/pldi-2018-papers-ry-fast-float-to-string-conversion06:05
← Zoffix left06:17
buggable New CPAN upload: Text-Names-0.0.1.tar.gz by RIBNOTTER https://cpan.metacpan.org/authors/id/R/RI/RIBNOTTER/Perl6/Text-Names-0.0.1.tar.gz06:21
→ domidumont joined06:24
→ SteffanW joined06:25
← domidumont left06:29
→ domidumont joined06:30
← reportable6 left06:31
→ reportable6 joined06:32
→ wamba joined06:33
← domidumont left06:34
→ stmuk joined06:35
← stmuk_ left06:37
→ domidumont joined06:37
← kaare_ left06:43
→ kaare_ joined06:43
← psychoslave left06:50
← [particle]1 left06:51
→ jmerelo joined06:55
← Actualeyes left06:56
→ [particle] joined07:09
← fake_space_whale left07:09
← kaare_ left07:26
→ kaare_ joined07:26
→ xq joined07:27
← kaare_ left07:34
← spycrab0 left07:35
→ vrurg joined07:36
← CurtisPoe left07:37
→ kaare_ joined07:43
→ eponym joined07:51
← eponym left07:51
← vrurg left07:51
← araraloren left07:53
→ araraloren joined07:53
← epony left07:54
← SteffanW left08:04
← [particle] left08:22
→ [particle] joined08:38
← lizmat left08:55
→ lizmat joined08:58
→ tokomer joined09:09
→ sena_kun joined09:10
→ araraloren_ joined09:10
← araraloren_ left09:13
→ araraloren_ joined09:13
← araraloren left09:13
→ imcsk8_ joined09:15
← imcsk8 left09:18
→ epony joined09:18
→ spycrab0 joined09:24
Geth ¦ doc: 54200d05f0 | (Tom Browder)++ (committed using GitHub Web editor) | htmlify.p609:39
¦ doc: tweak grammar 09:39
¦ doc: review: https://github.com/perl6/doc/commit/54200d05f009:39
¦ doc: 9ca6a1740b | (JJ Merelo)++ | doc/Language/variables.pod609:42
¦ doc: Changes to sentence case refs #2223 and eliminates non-current comment 09:42
¦ doc: review: https://github.com/perl6/doc/commit/9ca6a1740b09:42
¦ doc: 737c6eb645 | (JJ Merelo)++ | 2 files09:42
¦ doc: Adds constant prefix to the Variables page 09:42
synopsebot Link: https://doc.perl6.org/language/variables09:42
Geth ¦ doc:09:42
¦ doc: This closes #1566 although there was a pretty good section on the09:42
¦ doc: Terms page, for some reason. I have added just a simple example and09:42
¦ doc: linked the previous page, there's no need for duplicates.09:42
¦ doc: review: https://github.com/perl6/doc/commit/737c6eb64509:42
→ timeless13 joined09:46
timeless13 Interested in reasonably priced GLOBAL IRC ADVERTISING? Contact me on twitter https://twitter.com/nenolod or linkedin https://www.linkedin.com/in/nenolod09:46
← timeless13 left09:46
lizmat not here also? :-(09:46
benjikun :/09:49
stmuk looks like that twitter user is being attacked .. ironically it's drawn my attention to his anti-spam IRC bot :)09:50
https://github.com/kaniini/antissh09:50
"In 2018, there was a resurgence of IRC spam attacks that were undetected by traditional proxy scanning methods. This is because the attackers were using vulnerable SSH daemons running on routers, IPMI devices and other embedded devices to proxy the connections, using the direct-tcpip subsystem.09:52
jmerelo stmuk: what's an IPMI device?09:53
stmuk: ah, OK, intelligent platform management interface. Never heard of it before. https://www.itworld.com/article/2708437/security/ipmi--the-most-dangerous-protocol-you-ve-never-heard-of.html09:56
→ isBEKaml joined09:56
Geth ¦ doc: 9b57216da4 | (Tom Browder)++ (committed using GitHub Web editor) | lib/Perl6/Documentable.pm610:04
¦ doc: add attribute for Language sections 10:04
¦ doc: review: https://github.com/perl6/doc/commit/9b57216da410:04
→ HaraldJoerg joined10:05
benjikun Yeah, there have been huge port scans for several devices like this for years upon years :/10:05
→ parv joined10:17
stmuk SSH-2.0-dropbear_2013.6210:21
Geth ¦ doc: f62168e44d | (Tom Browder)++ (committed using GitHub Web editor) | htmlify.p610:29
¦ doc: prepare for future menu autogen 10:29
¦ doc: review: https://github.com/perl6/doc/commit/f62168e44d10:29
→ sarna joined10:32
sarna o/10:32
stmuk "Format string vulnerability in Dropbear SSH before 2016.74 allows remote attackers to execute arbitrary code via format string specifiers in the (1) username or (2) host argument.10:32
← pmurias left10:32
→ pmurias joined10:33
stmuk I don't think antissh would help us with CVE-2016-7406 .. maybe just block based on sshd banner version?10:33
or just block all matches to dropbear10:34
benjikun There are tons of holes like this in tons of things10:34
I doubt every spammer that has spammed #perl6 uses this vulnerability10:34
stmuk I think there is probably just one spammer using one (or a small number) of vulnerabilities10:35
Geth ¦ doc: 27dc7cfa7e | (Tom Browder)++ (committed using GitHub Web editor) | htmlify.p610:35
¦ doc: clean up comment 10:35
¦ doc: review: https://github.com/perl6/doc/commit/27dc7cfa7e10:35
benjikun I've seen a few different messages10:36
maybe 3-410:36
← lizmat left10:36
stmuk but mostly attacking freenode ops (even on other IRC networks)10:36
tbrowder_ sarna: \o10:36
benjikun stmuk: Yeah, I'm not sure how to feel about that10:37
stmuk I wish I had a record of the other ips to check banner versions.. I bet its all dropbear10:38
benjikun stmuk: Look through logs10:39
just grep for words you remember10:39
tbrowder_ [Coke]: any objections to moving some of the htmlify.p6 subs to Htmlify.pm6?10:40
← isBEKaml left10:40
tbrowder_ jmerelo: ^^^?10:40
stmuk benjikun: that gives me IRC nick only .. I can use /WHOWAS for a short period of time but failed to look up more10:40
benjikun I see10:43
jmerelo tbrowder_: not really10:46
tbrowder_: the main problem is that since htmlify.p6 is not tested, we might now know what's broken until it shows. But moving stuff to a module will really help test it; you can't simply test a script except by running it10:48
tbrowder_ ok, gotcha.10:50
Geth ¦ doc: 0bd4fcc53e | (JJ Merelo)++ | 4 files10:53
¦ doc: Moves to 2018.08 references to 2018.07 10:53
¦ doc:10:53
¦ doc: Since 2018.07 is not really going to happen https://twitter.com/zoffix/status/102329404826896793710:53
¦ doc: review: https://github.com/perl6/doc/commit/0bd4fcc53e10:53
jmerelo tbrowder_: it's a catch-22, really.10:53
tbrowder_ yep10:53
jmerelo tbrowder_: but I guess we'll have to start somewhere... so by all means do whatever you had in mind if everyone agrees with it.10:56
tbrowder_ i just saw your comment on site generation speed. until sake buid stuff is ready, there are some makefile dependencies that maybe could be tweaked to help. for instance, i’m not sure the type graph and type doc info should have to rebuild every time, cache or no cache, if the source hasn’t changed.11:01
jmerelo tbrowder_: sometimes it's rebuilt even if non-doc is changed.11:01
I mean, every little time.11:01
I proposed this some time ago https://github.com/perl6/doc/issues/193711:02
tbrowder_ some of those steps, if they could be extracted into separate progs, could be more easily be controlled by the makefile.11:03
jmerelo tbrowder_: and then there's this thing https://github.com/perl6/doc/issues/195211:03
tbrowder_: in fact, [Coke] created that here https://github.com/perl6/doc/commit/e9a057962d09ceeb815624b5acc6649d571fba26 but it's using text, not the Pod structure that pod::to::Bigpage uses11:04
tbrowder_: that makes it really useful for stuff like testing text, but not really for generation11:05
tbrowder_: if you're finally attending TPC in Glasgow, that could be something to be discussed in the Doc BoF11:06
tbrowder_ unfortunately i won’t be able to go. i sure would like to attend some specific p6 hacking gathering, though. where do you think a good international gathering place would be? iceland?11:09
estonia?11:09
jmerelo tbrowder_: I love Iceland.11:10
tbrowder_: never been to Estonia, though.11:10
tbrowder_ any p6 people in iceland?11:10
El_Che isn't iceland expensive as hell?11:10
jmerelo tbrowder_: I know some Perl mongers. El_Che yep, that's right.11:11
tbrowder_: Belgium during FOSDEM would be a nice alternative. But that's next february. I'll guess meanwhile we'll have to make do with IRC and issues.11:11
stmuk belgium/holland is probably easier for many than iceland11:11
tbrowder_ how about spain?11:12
→ koto joined11:16
→ ignoreme joined11:18
← ignoreme left11:18
← sena_kun left11:18
jmerelo tbrowder_: wonderful place. I can organize it in Granada if there's enough interest.11:22
tbrowder_: maybe something like perl 6 toolchain summit11:23
Would everyone want to do something like this?11:23
I got just the place: http://carmendelavictoria.ugr.es/ whose rooms are free if booked from the uni11:24
also you can stay there if booked with enough time, it's usually booked solid with university visiting people11:24
There are direct Delta flights from NY to Málaga, 1.5 hours away by bus11:25
El_Che wow11:25
jmerelo If there's enough interest, just let me know and I can start moving it, maybe for november or early December11:26
(above I mean meeting rooms, not bedrooms)11:27
→ raiph joined11:27
El_Che hehe11:28
good clarification :)11:28
jmerelo I have added discussion on that topic to the Perl 6 documentation BoF in Glasgow http://act.perlconference.org/tpc-2018-glasgow/wiki?node=Perl6%20documentation%20BoF11:34
stmuk: since we are both around here, can we talk a bit about this? https://github.com/perl6/ecosystem/issues/404#issuecomment-40867042111:42
stmuk: quite clearly, you don't agree with me, as repeated thumbs-down show. But I'm trying to find common ground here. Can you please ellaborate a bit?11:43
stmuk jmerelo: I don't think there is any common ground. I regard the ticket as a waste of time and probably any further discussion.11:44
jmerelo: if I mark a module as DEPRECATED it means exactly that (see wikipedia definition) I wouldn't expect to be contacted by an open source volunteer asking me if I really mean I wanted it deleting11:46
jmerelo stmuk: why do you think there's no common ground?11:47
stmuk I've just explained that11:48
jmerelo stmuk: you keep discussing about that particular aspect of what I propose, but that's not the main point. The main point is to have some clear procedure for removing stuff from ecosystem (not CPAN)11:50
→ lizmat joined11:50
stmuk there is no problem11:50
jmerelo stmuk: for instance, see this for NPM https://docs.npmjs.com/cli/unpublish11:50
It clearly specifies things like: if you've deleted something, you can't add it back.11:50
stmuk: there's all kind of stuff that might happen. For instance, you might want to delete some "latest" version but leave some versioned thing. That's also specified there.11:51
stmuk: plus we might want to eventually have everything in CPAN6.11:51
stmuk I don't see the perl community about written procedures and rules in the sense you do11:52
jmerelo stmuk: there are two options. Enforce it in code (see CPAN) or if you don't, enforce it as a rule. I'm all for enforcing it in code.11:53
stmuk I don't want to see the github based ecosystem replaced by CPAN611:53
anyway I can't see this discussion going anywhere11:53
→ psychoslave joined11:53
jmerelo stmuk: well, it should go towards trying and find common ground. If we can't we could maybe do a poll in the issue itself.11:54
stmuk: but if we can't even agree on that, I can close the issue, no problem. I don't want anyone unhappy and/or wasting time.11:55
Of course, as the proverbial grumpy old man, I reserve the right to say "I told you so" if this causes problems down the line, but really, I don't want to waste anyone's time.11:56
← kaare_ left11:58
El_Che jmerelo: lol, non registered users can not send you a message.12:12
you got too much spam?12:12
jmerelo El_Che: not really. use email12:14
El_Che let me look for a client :)12:14
less hot now in Granada?12:14
← domidumont left12:24
tbrowder_ jmerelo: ref probs building doc website and publishing. would sme kind of mutual lock work? i’m fuzzy on details but with all the asynchronicty (?) between to12:27
the two processes something needs to intervene and be a traffic cop.12:28
ref ecosystem and cpan: Zoffix and AlexDaniel and lizmat and others i thnk would vote for dumping ecosystem.12:30
tyil I personally prefer cpan for hosting modules, but some people prefer the simplicity of the ecosystem repo on github iirc12:35
the docs did mention the ecosystem repo to be deprecated "soon", but that was removed as nobody seemed to actually know about any real plans to deprecate it12:37
jmerelo El_Che: it's OK.12:38
tbrowder_ tyil: i have temporarily stopped working on =defn, but still on my plate12:39
tyil tbrowder_: ok, thanks for the headsup :>12:39
tbrowder_ first attempts haven’t worked, got to dig deeper in grammar12:39
tyil I've since been working on a module to ease creating new Pod formatters with as well12:39
jmerelo tyil, tbrowder_ the ecosystem is OK as long as there are 1 PRs every other week. If it becomes really popular, we're in trouble.12:39
tyil which I hope to use to make a Pod::To::Man module soom(tm)12:40
and I have some lispy thing to finish and write an article about ;~;12:40
AlexDaniel tbrowder_: you mean leave cpan but get rid of github thingie?12:40
tyil jmerelo: I know, I am personally in favour in CPAN because it scales better12:40
I also dont like depending on a proprietary platform like github12:40
jmerelo Just closed the issue.12:41
Never mind.12:41
tyil I think we can keep the ecosystem repo for now, but teach new contributors to use CPAN as the default12:43
El_Che tyil: I don't think it's clear at all12:44
therefor jmerelo's discusion was relevant12:44
do we want to push people to cpan?12:44
do we want to base the ecosystem on a Microsoft company?12:44
tyil what was clear?12:45
jmerelo El_Che: you can use also gitlab adn bitbucket12:45
El_Che how an econsystem should look like12:45
jmerelo El_Che: I really think so, but it seems to be impossible to find some common ground, so I guess it's better to close it now.12:46
stmuk you could host ecosystem/META.list somewhere other than github12:46
El_Che there is the go moduel of pushing stuff to github or gitlab and that's it12:46
tyil I didn't go into clarifying what it should look like12:46
I just think a git repo doesn't scale well into the future12:46
El_Che there is the centralised cpan or JS model12:46
tyil and I dont want to depend on a proprietary platform such as github12:46
El_Che there the catastrophy of a ecosystem called java12:46
tyil which are my two main reasons for going CPAN12:46
stmuk and people could use whatever git repos they like inside META.list12:46
El_Che tyil: yes, I am arguing that is something that needs to be reflected upon12:47
cpan has also downsides12:47
the interface is *very* oldschool12:47
tyil surely, if only because its harder to get started with for new people12:47
El_Che administration is needed, etc12:47
tyil which is something we shouldn't ant12:47
want*12:47
stmuk the main win of CPAN is the mirroring12:47
tyil but CPAN is stable, and proven12:47
El_Che distributed, indeed12:47
tyil and with modules, easy to work with12:47
jmerelo El_Che: I think so, but that does not seem to be the consensus. So let's just leave it at that and move on to some other thing.12:47
El_Che for perl5 the testing framework somewhat related to cpan12:48
tyil I'd like to see a friendlier UI to PAUSE for module creators, but I don't think it's worth the resources at this point in time to make it12:50
but since I use assixt for uploading them, I don't really have to work with the interface anyway12:51
→ andrzejku joined13:09
andrzejku araraloren_: hey :)13:09
araraloren_ andrzejku13:10
hey13:10
andrzejku araraloren_: I saw you was working hard13:10
araraloren_ :) haha13:11
← sarna left13:12
← jmerelo left13:20
← rgrau left13:20
AlexDaniel “ecosystem police” xD13:24
what's next, ecosystem moral police? :)13:25
tbrowder_: well, there is at least one p6 person in Estonia :) Not sure if it's a good place for an international gathering though13:34
tbrowder_: although I'm all for it, but yeah, I'm biased :)13:34
stmuk imagines the ecosystem police attempting to herd cats13:37
MasterDuke are many people planning to be at the london perl workshop?13:39
→ rgrau joined13:41
stmuk "we have to revoke your perl 6 licence because your META6.json isn't compliant with the changes made last week and you had trailing whitespace"13:41
tyil stmuk: I've already mentioned we should have a meta-version key in META6.json to deal with such breaking changes ;)13:42
← mephinet left14:02
← reu left14:02
→ vrurg joined14:04
stmuk AlexDaniel: don't you have internet millions of times faster than everyone else though?14:05
AlexDaniel stmuk: IIRC it's even faster in Japan14:06
stmuk: but I don't have a good connection14:06
AlexDaniel looks at https://www.fastmetrics.com/internet-connection-speed-by-country.php14:08
← AlexDaniel left14:08
→ AlexDaniel joined14:09
← vrurg left14:09
→ vrurg joined14:09
xiaomiao such things are not very useful - Japan has stupidly fast local connections but is badly connected to the rest of the world, for example14:10
→ aindilis joined14:10
xiaomiao a 10mbit connection in germany might be more pleasant to use than a gbit connection in japan14:10
Geth ¦ ecosystem: kalkin++ created pull request #405: Fix "Add update for Ddt to v0.5.3" 14:11
¦ ecosystem: review: https://github.com/perl6/ecosystem/pull/40514:11
kalkin-- → kalkin-14:13
← vrurg left14:14
← molaf left14:14
→ mephinet joined14:25
← parv left14:33
→ zakharyas joined14:43
→ fake_space_whale joined14:46
→ reu joined14:50
← lizmat left14:55
→ kerframil joined14:59
→ HaraldJoerg1 joined15:03
← HaraldJoerg left15:03
→ jmerelo joined15:05
jmerelo squashable6: status15:06
squashable6 jmerelo, ⚠🍕 Next SQUASHathon in 4 days and ≈18 hours (2018-08-04 UTC-12⌁UTC+14). See https://github.com/rakudo/rakudo/wiki/Monthly-Bug-Squash-Day15:06
tbrowder_ tyil: actually i have solved first part of =defn which uses text on same line as term and following para as def15:07
second use in speculations is first line of para is term, following lines are def15:08
BUT neither is specced as far as i can tell.15:09
imo, term on =def line makes the most sense, and forget other way15:10
thoughts?15:11
jmerelo tbrowder_: you get to spec it, I guess. term on =defn line makes sense to me too.15:15
HaraldJoerg1 → HaraldJoerg15:19
lucs jmerelo: Do you happen to see private /msg that are addressed to you?15:22
jmerelo just seen one15:23
lucs :)15:23
→ Sgeo_ joined15:23
← Sgeo left15:24
kalkin- Can some one please merge this https://github.com/perl6/ecosystem/pull/405 ?15:26
← psychoslave left15:28
jmerelo kalkin-: not me, sorry. You can ask for specific reviewers if you want when you do the PR.15:30
lucs The documentation sometimes uses the terms "longname" and "shortname" to designate those concepts (function name with parameters added or not).15:31
Is such usage a good idea, or should plain "long name" be sufficient?15:32
jmerelo lucs: where does it do that?15:32
lucs: do you mean using long name as opposed to longname?15:33
lucs I've seen it in S12 I think, so I'm wondering if it should be used in the regular docs.15:33
jmerelo: Yes, that's what I mean.15:33
jmerelo lucs: it's probably better if you use long name.15:34
lucs: git grep on the docs returns longname only in code15:35
lucs I kind of agree, but "longname" does imply that there is a technical meaning to it, which "long name" does not make obvious.15:35
jmerelo lucs: and there's a single reference to long name, and it uses a split term.15:36
lucs: hum.15:36
lucs: I kinda fail to see the relationship between those words "long name", and what they design.15:36
lucs See what I mean :)15:37
jmerelo lucs: so another option is just go for "double semicolon"15:37
lucs Oh, that's something else entirely.15:37
jmerelo lucs: ok, let me check.15:37
lucs This is what I'm talking about (let me find it...):15:38
jmerelo lucs: also, the example is wrong, right?15:38
lucs https://design.perl6.org/S12.html#Multisubs_and_Multimethods , right under that blue box.15:39
The example, like Zoffix said, is LTA.15:40
jmerelo lucs: I still think long name is the way to go. It's used that way in S12, so...15:40
lucs You'll also see 'longname' in S12... :/15:41
jmerelo lucs: 2 vs 1815:41
lucs Not sure what you mean...15:42
Geth ¦ ecosystem: 21496997af | (Bahtiar `kalkin-` Gadimov)++ | META.list15:42
¦ ecosystem: Fix "Add update for Ddt to v0.5.3" 15:42
¦ ecosystem:15:42
¦ ecosystem: This fixes the commit 518af6f212e900bfcdb794282695b8a973051d32.15:42
¦ ecosystem:15:42
¦ ecosystem: - included back all the previous versions of Ddt.15:42
¦ ecosystem: - latest Ddt version uses also a tagged version.15:42
¦ ecosystem: review: https://github.com/perl6/ecosystem/commit/21496997af15:42
¦ ecosystem: f9402863df | (Steve Mynott)++ (committed using GitHub Web editor) | META.list15:42
¦ ecosystem: Merge pull request #405 from kalkin/master 15:42
¦ ecosystem:15:42
lucs Oh, number of occurences.15:42
Geth ¦ ecosystem: Fix "Add update for Ddt to v0.5.3"15:42
¦ ecosystem: review: https://github.com/perl6/ecosystem/commit/f9402863df15:42
jmerelo lucs: control-F and search. "long name" appears 18 times, "longname" only 2.15:43
lucs Yep.15:43
→ molaf joined15:43
stmuk does ugexe still use IRC much? He seems to be on freenode but not on channels?15:43
lucs So, not worth introducing 'longname' "officially" -- sticking to "long name".15:44
jmerelo lucs: I mean, it's your call. You want to introduce it, let's stick to that then.15:44
kalkin- stmuk: thanks15:45
lucs "long name" feels lighter -- I'll just make sure it gets into the glossary too :)15:46
→ vrurg joined15:46
← andrzejku left15:47
lucs & # Moyse Sonorité time!15:47
→ ofperfection_ joined15:48
→ damaxi joined15:49
→ thisamu joined15:49
← drudgesentinel left15:50
← thisamu left15:51
→ lizmat joined16:09
→ raynold joined16:14
← araraloren_ left16:17
kalkin- How can I find all the classes implementing a specific role?16:19
jmerelo kalkin-: in the ecosystem, in a program, in a module?16:29
kalkin-: also, why would you want to do that?16:31
kalkin- in all installed modules16:32
jmerelo kalkin-: in all the modules that you have installed locally? Again, why would you want to do that?16:32
kalkin- jmerelo: reasons16:32
jmerelo kalkin-: let's break down the problem then. First, you need to know all modules that are installed, right?16:33
kalkin- right16:33
then i need to iterate over each module and lookup for each class and check if it implements a role?16:34
koto sounds so.16:34
jmerelo kalkin-: zef --installed list16:34
kalkin-: then you need to access the module itself, with all the stuff it provides. Again, zef.16:35
kalkin-: let me check16:35
kalkin-: that would be zef locate16:36
← vrurg left16:36
jmerelo kalkin-: but you could also just "use" it and access its stuff using the meta-object protocol16:36
kalkin- I can't just use classes which are not known at program creation time16:37
jmerelo kalkin-: you do one program to generate a list of "uses", and then add stuff to that program16:38
kalkin- basically i wanted to do plugins. The current straight forward solution would be to use Pluggable, but this forces all plugins to have the same namespace16:38
jmerelo kalkin-: those are the reasons?16:38
kalkin- hmm, but maay be it's not that a bad of idea if they use same namespace16:38
→ pecastro joined16:38
jmerelo anyway.16:38
m: say Rat.^roles16:38
kalkin- jmerelo: Yeah partly, partly curiosity16:38
camelia rakudo-moar 62f7c3784: OUTPUT: «((Rational[Int,Int]) (Real) (Numeric))␤»16:38
jmerelo ^roles will give you all the roles a particular class implements.16:39
kalkin- yeah MOP is definitely the way to go16:39
I just thought may be there is a shortcut, without iteration over all modules and over all classes in modules.16:39
koto you can EVAL `use $foo`, then in this scope do `MY::.keys`, iterate over needed stuff, do not forget to do it recursively going deeper with `.WHO.keys`. You check if it's `thing.HOW.WHAT ~~ Metamodel::ClassHOW` and do what you want, e.g. check roles.16:40
jmerelo so if zef --installed list returns Foo Bar Baz you do "use Foo; use Bar; use Baz;" and then for <Foo Bar Baz> -> $class { say ::$class.^roles }16:40
Although what koto says seems reasonable too :-=16:41
buggable New CPAN upload: Perl6-Ecosystem-0.0.2.tar.gz by JMERELO http://modules.perl6.org/dist/Perl6::Ecosystem:cpan:JMERELO16:41
kalkin- yes thank you koto & jmerelo. I probably just use Pluggable16:41
koto I am not an expert though, but I think there is no shortcut for this. I mean, in this case it'd be motion to children from parents, and type system is not really a graph DB or something like that.16:41
→ sarna joined16:41
jmerelo koto: the MOP _is_ the shortcut...16:42
sarna hey, why do my methods return Mu instead of Nil if I don't return anything?16:42
koto kalkin-, I bet you want to use Pluggable instead of black magic I've described above. : )16:42
jmerelo sarna: maybe it's returning a class and it's inialized to that by default? Mu is the nil for classes.16:43
koto jmerelo, kind of. I meant, there is no call like `^children` or something like that. Also, `say ::$class.^roles` approach will omit basically everything inside of package, so recursion is necessary.16:44
sarna jmerelo: makes sense. how can I avoid type errors then?16:44
jmerelo sarna: you can try exceptions, but it's difficult to say without looking at the code.16:44
sarna because if I returned Nil it wouldn't complain, now it throws a type error because Mu isn't MyClass16:44
jmerelo sarna: why don't you show us a bit of code? Difficult to say without the rest of the context16:46
sarna jmerelo: well alright, hold on a second16:46
jmerelo Here's the "river" score for all perl6 modules, including whether they fail or not. The "score" includes how far down they are in the chain of dependencies. High score and true in the fail column makes them a candidate for bitrot squashaton16:48
jmerelo clickbaits https://github.com/JJ/p6-river/blob/master/data/river-scores.csv16:48
sarna jmerelo: https://gist.github.com/sarna/70eeeead6f3f488501407a061a12f47016:49
jmerelo: winner is the problematic method16:49
jmerelo sarna: (also, if you've got the time and the will, please golf it and post in StackOverflow. I'll check this now anyway)16:49
sarna: winner is returning a Coin, and if Coin is nill, it's initialized to Mu, is that correct?16:50
sarna jmerelo: alright, I'll try :) (I've tried 5 minutes ago but failed)16:50
jmerelo: yes16:50
lizmat will be afk for most of the evening&16:50
jmerelo sarna: check-rows and columns do not have a default value for what they return. It can happen both alternatives are false16:51
sarna: also, returning from inside a loop, well...16:51
sarna jmerelo: what's wrong with returning from inside a loop"16:52
?16:52
jmerelo: and I thought Nil || Nil would return Nil, but16:52
jmerelo sarna: it's better to end the loop and then return. And you're not guaranteed check-rows and check-columns are returning Nil. If they don't return, they will return the last value in the sub16:53
sarna jmerelo: what do you mean by "end the loop and then return"? I don't need to continue if I found a match16:55
jmerelo: oh, so adding Nil at the end should fix it. I thought it was there implicitly16:55
jmerelo m: sub foo() returns Any { loop( my $i =0; $i < 3; $i++ ) { say "Hello" }}; say foo;16:55
camelia rakudo-moar 62f7c3784: OUTPUT: «=== SORRY!=== ␤Word 'loop' interpreted as 'loop()' function call; please use whitespace around the parens␤at <tmp>:1␤------> sub foo() returns Any { loop ⏏ ( my $i =0; $i < 3; $i++ ) { say "Hello" ␤Unexpected block in infix position …»16:55
jmerelo sarna: never mind the loop and the returning.16:56
sarna: not sure what is there implicitly. But returning Nil might give you what you want.16:56
sarna jmerelo: yeah, no side effects which won't be run in my loop :)16:56
jmerelo: thanks for taking your time!16:57
jmerelo sarna: good luck :-)16:57
sarna m: class Foo {}; sub bar(--> Foo) {if False {return Foo}}; bar()17:04
camelia rakudo-moar 62f7c3784: OUTPUT: «Type check failed for return value; expected Foo but got Slip (Empty)␤ in sub bar at <tmp> line 1␤ in block <unit> at <tmp> line 1␤␤»17:04
sarna jmerelo: the shortest I could get :)17:04
without `if False` it just returns Nil17:06
→ psychoslave joined17:10
jmerelo sarna: it's returning whatever "if" is returning.17:12
m: my $what's-this = if False { say "Hello" }; say $what's-this17:13
camelia rakudo-moar 62f7c3784: OUTPUT: «=== SORRY!=== ␤Word 'if' interpreted as a listop; please use 'do if' to introduce the statement control word␤at <tmp>:1␤------> my $what's-this = if ⏏ False { say "Hello" }; say $what's-this ␤Unexpected block in infix position (two …»17:13
jmerelo m: my $what's-this = do if False { say "Hello" }; say $what's-this17:14
camelia rakudo-moar 62f7c3784: OUTPUT: «()␤»17:14
jmerelo m: my $what's-this = (if False { say "Hello" }); say $what's-this17:14
camelia rakudo-moar 62f7c3784: OUTPUT: «()␤»17:14
jmerelo m: my $what's-this = say "Hello" if false; say $what's-this17:14
camelia rakudo-moar 62f7c3784: OUTPUT: «=== SORRY!=== Error while compiling <tmp>␤Undeclared routine:␤ false used at line 1␤␤»17:14
jmerelo m: my $what's-this = say "Hello" if False; say $what's-this17:14
camelia rakudo-moar 62f7c3784: OUTPUT: «(Any)␤»17:14
sarna jmerelo: oh! makes sense :)17:14
← benjikun left17:19
→ domidumont joined17:21
→ benjikun joined17:21
→ vrurg joined17:41
← vrurg left17:44
damaxi Hey17:49
sarna m: say <1 2 3> + <4 5 6>17:51
camelia rakudo-moar d77a51de2: OUTPUT: «6␤»17:51
sarna why doesn't it throw a type error :D17:51
tobs sarna: because it can coerce17:52
m: <1 2 3>.Int.say17:52
camelia rakudo-moar d77a51de2: OUTPUT: «3␤»17:52
sarna tobs: ohh, I understand now17:53
thanks17:53
→ DrForr joined17:59
jmerelo Hey, DrForr . How are you doing?18:00
DrForr Reasonably well given the circumstances, right now waiting for a friend to come online so I can ask a few questions. Noticed this box had been bounced so I thought I'd reset things.18:01
El_Che hi drforr18:01
DrForr Howdy.18:02
Other than that chilling with the dogs who are *completely* asleep on the job as they haven't noticed this dee..spoke too soon, there they go.18:02
jmerelo DrForr: good luck with everything...18:02
DrForr Well, it *might* work out better than I'd initially thought, because of an offer I got on Friday.18:03
El_Che the twitter one?18:03
sarna hey, can I make a thing that's not known at compile-time immutable? like, assign once?18:05
→ raschipi joined18:05
← El_Che left18:05
DrForr No, something else. I need to learn about some logistics from a friend, but if I can get insurance straightened out I might just go with the deal I've been offered for 3-6 months.18:05
jmerelo sarna: constants are decided at compile time... You can still bind, which is like declaring a literal, which is constant-ish18:05
m: my $constant-ish := 33; $constant-ish = 34;18:06
camelia rakudo-moar d77a51de2: OUTPUT: «Cannot assign to an immutable value␤ in block <unit> at <tmp> line 1␤␤»18:06
sarna m: my $x = 3; my $constant-ish := $x; $x = 4; say $constant-ish18:07
camelia rakudo-moar d77a51de2: OUTPUT: «4␤»18:07
jmerelo m: my $þ = 7; my $constant-ish := 33+ $þ; say $constant-ish18:07
camelia rakudo-moar d77a51de2: OUTPUT: «40␤»18:07
jmerelo m: my $þ = 7; constant $constant = 33+ $þ18:07
camelia rakudo-moar d77a51de2: OUTPUT: «Use of uninitialized value of type Any in numeric context␤ in block at <tmp> line 1␤»18:07
Geth ¦ doc: f94b0a38e8 | (Zoffix Znet)++ | doc/Programs/03-environment-variables.pod618:07
¦ doc: Rename RAKUDO_EXCEPTIONS_HANDLER 18:07
¦ doc:18:07
¦ doc: Procspec: https://github.com/perl6/roast/commit/7d2432590e18:07
¦ doc: Rakudo fix: https://github.com/rakudo/rakudo/commit/5e1df41b3418:07
¦ doc: review: https://github.com/perl6/doc/commit/f94b0a38e818:07
[Coke] tbrowder_: in general, no objections. practically, it's going to make it harder for me to keep coke/build going due to merge stuff, but that's on me, not you18:07
jmerelo sarna: in that case, you're binding the container, not the value.18:08
[Coke] (extract to separate programs in build) that's the opposite of what I'm trying to do.18:08
jmerelo m: my $x = 3; my $constant-ish := $x; say $constant-ish; $x = 4; say $constant-ish18:08
camelia rakudo-moar d77a51de2: OUTPUT: «3␤4␤»18:08
DrForr stitches together a Spreadsheet API.18:08
[Coke] (I'm trying to move things *into* perl6, but keep them better segregated).18:09
But given my lack of time to commit to the work, I'm not going to fight.18:09
jmerelo [Coke]: you mean the Makefile and the web app, right?18:09
tbrowder_ [Coke]: I may be going down a rabbit hole anyway. Can we see your version?18:10
← ofperfection_ left18:12
[Coke] tbrowder_: f62168e44d2d441a8851b2340ca2fc6512fa6cf5 - you added a comment that's basically the same as the comment that's there. (and are failing the whitespace tests now)18:13
sarna jmerelo: ah yeah, I remember now. thanks18:13
→ El_Che joined18:14
El_Che (shell server upgrades it seemsà18:15
)18:15
tyil tbrowder_: I think "=defn <term>", with the following Para as its definition is the most sane to read/write18:16
sarna is there a shortcut for this? foo(x => $x) (x is a named argument)18:16
tyil sarna: foo(:$x)18:16
El_Che :$x18:16
sarna thanks!!18:16
it's not in the docs :(18:17
jmerelo sarna: foo( :$x )18:17
tbrowder_ tyil: thanks, i agree. i’m thinking of asking Damian’s opinion since he wrote most of S26 i believe.18:18
sarna jmerelo: yep, just complaining about the docs now :)18:18
← lizmat left18:18
jmerelo sarna: don't just complain, create an issue :-) (Although I'm pretty sure that is in the docs)18:18
tyil might be an issue of being too hard to find, which we can still look into18:18
tbrowder_ [Coke]: i’m sorry about ws but i didn’t see any evidence of it, but i think web edit is conducive to that18:19
→ Zoffix joined18:19
masak TimToady: is COMPILING always reaching into a static lexpad? if yes, could there be something like COMPILING which reaches into a runtime lexpad?18:19
sarna jmerelo: well, it's not here https://docs.perl6.org/type/Signature#index-entry-named_argument18:19
Zoffix sarna: most of those shortcuts: https://rakudo.party/post/Perl-6-Colonpairoscopy18:19
yoleaux 14:10Z <samcv> Zoffix: well that would seem to make sense18:19
jmerelo sarna: it is: $ = :(:$a);18:20
sarna Zoffix: thank you :)18:20
jmerelo: oh. it's a bit cryptic though, I don't know if you agree18:20
Zoffix sarna: and `else`-less `if`/`with`/`without` return Empty when condition is false, to support constructs like (1, 2, 3, (42 if $whatever), 45, 6)18:20
jmerelo sarna: yep, it kinda is.18:20
sarna: so, issue that :-)18:21
sarna jmerelo: ok :)18:21
Zoffix: makes sense now!18:22
← Zoffix left18:22
Geth ¦ doc: 5d247af6c8 | (JJ Merelo)++ | doc/Type/Signature.pod618:23
¦ doc: Just fixing some typographic error 18:23
¦ doc: review: https://github.com/perl6/doc/commit/5d247af6c818:23
synopsebot Link: https://doc.perl6.org/type/Signature18:23
← jmerelo left18:26
← raiph left18:28
sarna m: sub foo($x) {x++}; foo(2)18:31
camelia rakudo-moar d2116efd5: OUTPUT: «=== SORRY!=== Error while compiling <tmp>␤Undeclared routine:␤ x used at line 1␤␤»18:31
sarna m: sub foo($x) {$x++}; foo(2)18:32
camelia rakudo-moar d2116efd5: OUTPUT: «Cannot resolve caller postfix:<++>(Int); the following candidates␤match the type but require mutable arguments:␤ (Mu:D $a is rw)␤ (Int:D $a is rw)␤␤The following do not match for other reasons:␤ (Bool:D $a is rw)␤ (Bool:U $a …»18:32
sarna why doesn't it copy the parameter by default?18:32
timotimo i think here explicit is better than implicit18:33
i.e. you can immediately see if the value of your parameter is going to change in the function or not18:33
geekosaur I think because most people don't intend to mutate parameters and this allows ffaster code18:33
and yes, seeing it explicitly is good IMO18:33
sarna can I somehow guarantee that this function won't mutate the parameter?18:33
Geth ¦ doc: 342466f69e | (Zoffix Znet)++ | assets/sass/style.scss18:33
¦ doc: Don't show BG camelia on 404s 18:33
¦ doc:18:33
¦ doc: Fixes https://github.com/perl6/doc/issues/222818:33
¦ doc: review: https://github.com/perl6/doc/commit/342466f69e18:33
sarna besides is copy?18:33
timotimo how do you mean?18:35
oh, also, this is only surface-level, not "deep"18:35
sarna like `foo(const int a)` in C18:35
timotimo that's close enough to the default you have in perl618:36
sarna oh, I'm sorry, it's quite late. I got it now :D18:37
thanks for explaining18:37
timotimo NP18:37
← psychoslave left18:37
kalkin- For some reason my $name = 'Foo'; try require ::($name) returns a (Foo), but Foo::.keys is empty18:45
if I do try require Foo, then Foo::.keys is filled up18:45
I also tried ::($name)::.keys18:45
DrForr Huh. Not every afternoon that your hacking is interrupted by a black bear walking through your property.18:51
timotimo kalkin-: i cannot explain why it works differently, but try ::($name).WHO.keys18:52
kalkin- timotimo: thanks! that does the trick18:53
tbrowder_ tyil: instead of waiting till i could implement both types of =defn, i could just do the first type soon and wait to do the second type later. thoughts?do you see any reason i coul19:01
disregard text after question mark19:02
← sarna left19:12
→ vrurg joined19:14
← vrurg left19:15
← molaf left19:17
xinming I got a perl6 program segfault.19:20
I'll try to narrow down.19:20
→ sarna joined19:23
← sivoais left19:26
→ sivoais joined19:26
DrForr I got a lot of those when I was first starting out.19:29
Juerd Back then they were much more common :)19:34
DrForr Doubtless.19:36
xinming finally, narrowed down19:37
← fake_space_whale left19:38
← zakharyas left19:40
← lookatme left19:40
→ lookatme joined19:41
← tokomer left19:46
→ mcmillhj joined19:48
timotimo .tell jjmerelo can you explain what makes your data different from finanalyst's ModuleCitation thing? http://finanalyst.github.io/ModuleCitation/19:52
yoleaux timotimo: I'll pass your message to jjmerelo.19:52
← kybr left19:52
kalkin- as far as I am reading the docs correctly I can't import symbols at runtime19:54
xinming http://singularity.esky.cloud:2080/pub/p6-segfault.tgz <---- This is the segfault bundle where I can narrow down with my best effort. :-)19:54
raschipi kalkin-: Why not?19:54
timotimo right, because the compiler has to know what is available and what isn't19:54
kalkin- So require Foo; is not the same as: my $name = 'Foo'; require ::($name);19:54
xinming anyone here can try this19:54
to confirm what I tried.19:54
timotimo even with require Foo; it won't import the symbols, IIUC only the package itself gets made available19:55
→ vrurg joined19:55
timotimo you can, however, pass a list fo symbols that you want to have available19:55
kalkin- Currently my plugin system consists of multiple module which export MAIN subs. I find out which modules I need to load via Plugabble at run time19:55
timotimo: about require Foo; I see you have to be explicit: require Foo <&something>19:56
timotimo: but the list you are mentioning, need to be available at compile time, right?19:57
xinming I think it's because of multi with is export19:57
timotimo yes19:57
kalkin- so there is NO METHOD for importing it at runtime?19:58
timotimo yes, it's simply not possible without time travel19:58
kalkin- k19:59
← lookatme left19:59
→ lookatme joined19:59
kalkin- K, at least I get the modules loaded. Now I could iterate over them collect all the exported MAIN functions, and then forward @*ARGS some how to the right one20:00
does this make sense?20:00
timotimo doesn't immediately sound wrong20:00
kalkin- given I have a list of MAIN Subs and an array of arguments, how can I do the same as perl6 does when it resolves a multi?20:02
timotimo hm, that's more difficult20:02
since every one of them already has their own proto sub20:02
i wonder if you can use the MOP to get this20:02
kalkin- currently this gives me all the multi subs in a unit → say ::($name).WHO<EXPORT>.WHO<DEFAULT>.WHO.values[0].candidates[0]20:03
m: use JSON::Fast20:03
camelia rakudo-moar d2116efd5: OUTPUT: «===SORRY!===␤Could not find JSON::Fast at line 1 in:␤ /home/camelia/.perl6␤ /home/camelia/rakudo-m-inst-2/share/perl6/site␤ /home/camelia/rakudo-m-inst-2/share/perl6/vendor␤ /home/camelia/rakudo-m-inst-2/share/perl6␤ CompUni…»20:03
← mcmillhj left20:04
kalkin- timotimo: what exactly need I get?20:04
timotimo i haven't worked with this part of the mop a lot20:05
kalkin- I assummed I there will be some function EXECUTE(@subs, |args) which does the magic20:05
timotimo not like that, no20:06
i'd think you'd build "your own" multi sub and just shove all the candidates into it20:06
Geth ¦ doc: eb0c969e57 | (Zoffix Znet)++ | 2 files20:06
¦ doc: [6.d] Document $*ARGFILES being fed by $*IN inside MAIN 20:06
¦ doc:20:06
¦ doc: Even when @*ARGS has stuff up in it.20:06
¦ doc: Rakudo impl: https://github.com/rakudo/rakudo/commit/84d45f07fb20:06
¦ doc: Propspec: https://github.com/perl6/roast/commit/88499ac36620:06
¦ doc: review: https://github.com/perl6/doc/commit/eb0c969e5720:06
xinming seems no one interested in rakudo segfault issue. I'll trouble the channel later. :-)20:07
timotimo sorry, xinming20:07
my brain is dramatically slowed down by the heat20:07
El_Che xinming: raise an issue so there is a trace (for people not in the channel atm)20:08
raschipi .cando -> Returns a list of candidates that can be called with the given Capture20:08
MasterDuke xinming: i can repro it20:08
timotimo yeah, me too20:09
it explodes while compiling a regex or something?20:09
MasterDuke xinming: btw, `\o where { o ~~ Int }` could just be `Int \o`20:09
still happens with the jit disabled20:09
and with spesh disabled20:10
xinming: are you creating an issue?20:10
xinming Not yet20:11
MasterDuke huh, no segv under valgrind20:11
xinming what does valgrind mean?20:11
MasterDuke it's a debugging tool20:12
timotimo it's a pretty cool tool that can figure out a lot of different kinds of memory usage errors20:12
like using memory that has recently been freed20:12
→ mcmillhj joined20:12
timotimo or running out of bounds on a buffer or array or something20:12
xinming Ok, got it, thanks20:12
timotimo MasterDuke: it looks like perhaps GC_DEBUG will tell us something20:13
→ psychoslave joined20:15
MasterDuke hm, doesn't seem to have shown anything new yet20:15
→ dct joined20:16
← mcmillhj left20:16
→ uzl joined20:16
MasterDuke oh, i need to create a log file...20:17
← Kaiepi left20:17
→ Kaiepi joined20:18
→ Zoffix joined20:18
Zoffix It bisects to https://github.com/rakudo/rakudo/commit/722d8711b359a1e5b9f21c3d2a954f18074bffd820:18
Geth: ver https://github.com/rakudo/rakudo/commit/722d8711b359a1e5b9f21c3d2a954f18074bffd820:18
Geth Zoffix, version bump brought in these changes: https://github.com/perl6/nqp/compare/2017.09-40-ga6a1aa0...2017.09-59-g60f79d320:18
Zoffix Geth: ver https://github.com/perl6/nqp/commit/60f79d38e99985c207a2b9a91afda44a6d5e8aa520:19
Geth Zoffix, version bump brought in these changes: https://github.com/MoarVM/MoarVM/compare/2017.09.1-553-ga4fef0b...2017.09.1-575-gd4e230a20:19
xinming issue submitted.20:20
uzl I just watched your 'Wow, Perl 6' presentation, Zoffix. Good stuff!20:20
xinming https://github.com/rakudo/rakudo/issues/214720:21
← HaraldJoerg left20:22
Zoffix Thanks.20:23
uzl You mentions the use of ++$ as a counter. I know it's somewhat simple but is there any documentation on it?20:23
→ HaraldJoerg joined20:23
← raschipi left20:24
Zoffix uzl: https://docs.perl6.org/language/variables#index-entry-%24_%28variable%2920:24
Geth ¦ doc: 1b9e1e5877 | (Zoffix Znet)++ (committed using GitHub Web editor) | doc/Language/variables.pod620:27
¦ doc: Improve example with explicit `state` 20:27
¦ doc:20:27
¦ doc: Using state in `where` clause is UB and should not be shown in the docs.20:27
¦ doc: review: https://github.com/perl6/doc/commit/1b9e1e587720:27
synopsebot Link: https://doc.perl6.org/language/variables20:27
uzl "In general, it is better style and p6ier to declare a named state variable in case you have to refer to it several times."20:27
← Kaiepi left20:28
uzl what does "p6ier" mean in this context?20:28
timotimo i guess "idiomatic perl6"?20:28
→ mcmillhj joined20:28
Zoffix Also.. says who?20:29
ah, it's about using $ multiple times. OK20:29
Geth ¦ doc: 4ca4b13a54 | (Zoffix Znet)++ (committed using GitHub Web editor) | doc/Language/variables.pod620:30
¦ doc: Remove confusing adjective 20:30
¦ doc:20:30
¦ doc: http://colabti.org/irclogger/irclogger_log/perl6?date=2018-07-29#l81220:30
¦ doc: review: https://github.com/perl6/doc/commit/4ca4b13a5420:30
synopsebot Link: https://doc.perl6.org/language/variables20:30
← Zoffix left20:30
→ Kaiepi joined20:31
← mcmillhj left20:33
→ buggable joined20:35
ChanServ set mode: +v20:35
Geth ¦ doc: 7adf41fe82 | (Zoffix Znet)++ (committed using GitHub Web editor) | doc/Language/glossary.pod620:37
¦ doc: Add "UB" to glossary 20:37
¦ doc: review: https://github.com/perl6/doc/commit/7adf41fe8220:37
synopsebot Link: https://doc.perl6.org/language/glossary20:37
← domidumont left20:37
→ mcmillhj joined20:39
← mcmillhj left20:43
← damaxi left20:45
→ damaxi joined20:53
→ mcmillhj joined21:08
← psychoslave left21:09
← mcmillhj left21:12
← pmurias left21:15
← rindolf left21:17
← HaraldJoerg left21:18
→ pmurias joined21:20
← damaxi left21:23
SmokeMachine Isn’t http://modules.perl6.org/ getting the new modules?21:34
timotimo check the far bottom21:35
This page was generated from the files in the modules.perl6.org repository on Sun Jul 29 20:39:36 2018. view build log21:35
SmokeMachine timotimo: thanks21:38
Something wrong on my META6.json21:40
← sarna left21:42
→ mcmillhj joined21:47
← mcmillhj left21:52
→ ofperfection_ joined21:53
SmokeMachine This line is breaking the indexation of my module (Injector), but my .travis.yml isn’t using panda...21:55
https://github.com/FCO/Injector/blob/master/.travis.yml21:56
→ stee joined21:56
SmokeMachine fatal error while building https://raw.githubusercontent.com/FCO/Injector/master/META6.json: Use of uninitialized value in pattern match (m//) at lib/ModulesPerl6/DbBuilder/Dist/PostProcessor/p30METAChecker.pm line 123.21:58
Looks something wrong on the regex...22:00
Or the .travis.yml’s content couldn’t be loaded22:05
→ Zoffix joined22:07
Zoffix This is just ridiculous.22:07
The PR had a comment, saying the code was broken.22:07
The user who merged the PR was told the code was broken.22:07
And yet still no one did anything and now the site is broken.22:07
← ofperfection_ left22:09
El_Che why was it merged then?22:16
did it fail travis?22:16
→ _uzl joined22:16
← uzl left22:17
← pmurias left22:18
← dct left22:21
Zoffix "why merge a PR that's being claimed to be broken?" "I don't think it is."22:21
And then they fucked off.22:21
SmokeMachine Should it be reverted22:30
Zoffix I'm trying to fix it right now, but Types::Standard seems to be glitching and doesn't respect Maybe[] constraints on 5.2822:31
→ mcmillhj joined22:33
→ lizmat joined22:35
← mcmillhj left22:37
Geth ¦ modules.perl6.org: 866b0af6c9 | (Zoffix Znet)++ | lib/ModulesPerl6/DbBuilder/Dist/PostProcessor/p05DataFetcher.pm22:41
¦ modules.perl6.org: Fetch travis file contents 22:41
¦ modules.perl6.org: review: https://github.com/perl6/modules.perl6.org/commit/866b0af6c922:41
¦ modules.perl6.org: 9e1db1d97e | (Zoffix Znet)++ | use-me-for-commit-triggers22:41
¦ modules.perl6.org: [NEWDB] rebuilt stuff 22:41
¦ modules.perl6.org: review: https://github.com/perl6/modules.perl6.org/commit/9e1db1d97e22:41
Zoffix "Types::Standard seems to be glitching" was actually changes in newer Mew: https://github.com/zoffixznet/Mew/pull/422:42
← Zoffix left22:43
← koto left22:44
Geth ¦ modules.perl6.org: fd967f934e | (Zoffix Znet)++ | lib/ModulesPerl6/DbBuilder/Dist/PostProcessor/p30METAChecker.pm22:50
¦ modules.perl6.org: Harden meta checker against failed content fetches 22:50
¦ modules.perl6.org: review: https://github.com/perl6/modules.perl6.org/commit/fd967f934e22:50
¦ modules.perl6.org: 4c03426399 | (Zoffix Znet)++ | lib/ModulesPerl6/DbBuilder/Dist/PostProcessor/p05DataFetcher.pm22:52
¦ modules.perl6.org: Fix copy-pasta in error message 22:52
¦ modules.perl6.org: review: https://github.com/perl6/modules.perl6.org/commit/4c0342639922:52
← pecastro left22:53
→ Ven`` joined22:56
← Ven`` left22:58
Geth ¦ modules.perl6.org: c22c885f33 | (Zoffix Znet)++ | 11 files23:00
¦ modules.perl6.org: Fix @INC for newer perls 23:00
¦ modules.perl6.org: review: https://github.com/perl6/modules.perl6.org/commit/c22c885f3323:00
→ Ven`` joined23:04
← wamba left23:04
← _uzl left23:05
→ spider-mario joined23:19
→ mcmillhj joined23:20
← mcmillhj left23:25
→ mcmillhj joined23:33
← mcmillhj left23:37
vrurg Are there a way to get a custom DESTROY submethod (installed with .^add_method) called?23:37
Wrapping an existing DESTROY works, but installing a new one – doesn't.23:38
timotimo it's important to call .^compose for such changes, but it may still not be enough23:48
→ limarfrek joined23:50
← kerframil left23:51
limarfrek → kerframil23:51
vrurg timotimo: Thanks, but doesn't work for me either. I'm trying to create an attribute trait and all the work happens in Attribute's compose. Calling class .^compose causes a error.23:53
timotimo ah, ok, i thought you were monkey-patching an existing class23:53
maybe try mixing in a role that has the DESTROY submethod23:55
i wonder if ^add_submethod is a thing that exists23:55
vrurg Worse. Trying to re-implement Moo/Moose lazy attributes in a smarter way than the two existing modules.23:55
add_submethod is not necessary because add_method determines what type of routine it is given.23:56
so, .^add_method( method ) and .^add_method( submethod {} ) are different things.23:56
Mixin doesn't work either, BTW. Only if the class has its own DESTROY.23:57
timotimo ah23:59
well, it can have only one23:59

Logs Search ←Prev date Next date→ Channels Documentation