»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'perl6: say 3;' or rakudo:, niecza:, std:, or /msg camelia perl6: ... | irclog: irc.perl6.org | UTF-8 is our friend!
Set by sorear on 25 June 2013.
dwarring trying ^add_fallback from advent 2013 day 09 - perl6advent.wordpress.com/2013/12/0...and-pairs/ 00:00
dwarring should work on parrot 00:01
timotimo aye, needs a use Test. 00:03
jnthn There'll be some hardening work to come, but on my box I just got a fully passing NativeCall test suite run on Moar. 00:07
TimToady \o/ 00:07
dwarring About to add day 08 tests. I'll fudge it for now 00:07
dwarring s/08/09/ 00:08
dalek ast: 3182fbd | (David Warring [email@hidden.address] | integration/advent2013-day09.t:
adding advent 2013 day 09
00:17
TimToady dwarring: 3 doesn't look like a caddr to me, it looks like a cddr 00:24
or a cddar 00:25
TimToady or have I forgotten how those guys work? 00:26
dwarring had just blindly cut and pasted 00:27
TimToady in any case, it doesn't have a nice failure mode 00:28
TimToady I guess I'm remembering them backwards 00:36
TimToady the problem is that the match $0 is true even when it matches the null string 00:38
TimToady masak: ^^ with regard to your add_fallback in advent 2013-09 00:39
needs to be $0 ne '' or so
dalek ast: b13f601 | larry++ | integration/advent2013-day09.t:
fix caddr fallback test

  $0 is always true because (<[ad]>*) always matches, needs $0 ne ''
00:41
MadHatter1987 knock knock 00:42
dalek ast: 831ac2a | larry++ | integration/advent2013-day09.t:
remove fudge for caddr test
00:43
dwarring TimToady++
MadHatter1987 anyone ? 01:07
TimToady anyone who? 01:11
PerlJam MadHatter1987: not just *anyone*! 01:12
TimToady anytwo, apparently
MadHatter1987 :D 01:14
what I meant was "is anyone around ?"
so i'm a perl coder 01:15
was snooping around
hoping if anyone around here had implemented AI in perl 01:16
PerlJam MadHatter1987: For Perl 5, there are some AI modules on CPAN. 01:16
MadHatter1987 was more interested in the implementation 01:16
i used to C# 01:17
and it had quite of support for AI modules
but in perl I cant find anything interesting around
AI related
PerlJam MadHatter1987: so ... perhaps I should mention that this channel is for Perl 6 which is a different language than Perl 5. 01:18
MadHatter1987: but in any case, there are modules on CPAN and you can read their code (for Perl 5)
MadHatter1987 well I wasnt seeing any discussions so I thought on asking around 01:19
well thnx anyway
hoelzro I don't mean to hate on the good folks at GH, but I found this PR interesting: github.com/github/linguist/pull/748 02:20
seldom does one find good commentary on the Internet =)
TimToady
.oO(maybe we should use a .m extension for moar... :)
02:40
lue also could use .m for Mu, a.k.a. "every document" :P 02:41
uvtc Has much changed of late in the state of the art of syntax highlighting Perl 6 code? Any new syntax highlighters available recently? 03:40
lue GitHub has a highlighter now, don't know of much else. 03:41
uvtc Using Xfce, I just checked mousepad, and it seems to use libgtksourceview. 03:43
thanks, lue 03:44
uvtc Hm. Looks like there's substantial [docs for creating a Kate syntax highlighting file](docs.kde.org/development/en/applica...ght.html). They even start out with a Perl 5 example. :) 04:10
hoelzro uvtc: we have a start on a Kate file
lue should really get back to his emacs mode for Perl 6, at least to unbreak it :P
hoelzro should work on Perl 6 for Vim =/ 04:11
uvtc I was just installing Kate (er, KWrite) now just to see what the editor looks like. apt-get is grabbing a lot of packages. 04:12
Hi, hoelzro. I remember you were doing something with Perl 6 syntax highlighting. 04:13
hoelzro yeah, I wrote the pygments highlighter
lue uvtc: wild guess: you've not installed anything KDE-based before, have you? :)
uvtc lue, no, never. :) Running Ubuntu + Xfce here, but switching to Debian Testing this weekend, so I thought, "what the heck, pull in the kde hoards and have a look before starting from scratch again." :) 04:14
Nice thing about a Kate highlighter is that it would then trickle down into Pandoc (markdown processor with syntax HL via kate SH definition files). 04:17
hoelzro and Qt creator
my life is just about to stabilize, so I intend to help on that front once it has =) 04:18
uvtc hoelzro, could you point me to the start on a Kate SL def file? I'm not seeing it at github.com/perl6/
.
hoelzro hang on
lemme find it
uvtc: github.com/timo/perl6-kate-syntax
uvtc Oh, wow. It's looking pretty substantial already. Thanks for the link, hoelzro. 04:19
hoelzro np
thank timotimo for the work =)
uvtc ++timotimo 04:21
uvtc Hm. Looks like the rather disused Perl 6 wiki has fallen into disrepair (well, wiki.perl6.org rendering without any styles on my system). 04:29
uvtc moritz, I've made backups of my "mostly harmless docs" that were started there. I have no qualms if you'd like to remove the site until a hero appears who is willing to maintain a wiki. 04:31
lue I know I can use &foo.assuming to prime some parameters, but I don't see a way to re-prime parameters :( [e.g. &foo.assuming(:dbg).assuming(:!dbg), which doesn't DWIM]
uvtc Personally though, in general, of late, I've found wikis to be not as useful as I'd hoped. 04:32
lue (except for keeping a separate hash of unprimed functions, but that feels like something I shouldn't have to do...) 04:33
uvtc Is phenny still around? <perl6.org/community/irc> indicates yes. 04:37
lue perl6.org/community/irc would be wrong then :) 04:38
wow, that list is old, it's talking about nqp-rx
uvtc Oh, replaced by yoleaux?
lue yep, I'll go update that list unless someone else is already on it. 04:39
uvtc .tell moritz <irclog.perlgeek.de/perl6/2014-03-15...439991> 04:40
yoleaux uvtc: I'll pass your message to moritz.
uvtc *hordes ! 04:44
hoelzro, glad to hear about life stabilization. That's a nice thing. :) 04:45
hoelzro yeah, tell me about it! =) 04:46
dalek href="https://perl6.org:">perl6.org: 92b3739 | lue++ | source/community/irc.html:
Update the list of bots on #perl6
04:49
uvtc Built Rakudo+Moar a couple of days ago. Was pleased to find that it starts up and runs little scripts in under a half a second (0.4s)! 04:50
uvtc When trying to run <rosettacode.org/wiki/Colour_bars/Di...erl_6>, Rakudo-Moar tells me about the "my @colors" line: "This type cannot unbox to a native integer". 05:09
moritz .tell uvtc I've disabled wiki.perl6.org 05:36
yoleaux moritz: I'll pass your message to uvtc.
04:40Z <uvtc> moritz: <irclog.perlgeek.de/perl6/2014-03-15...439991>
avuserow is HTTP::Client the current preferred web library for Rakudo? I'm wanting to play around with it, ideally on MoarVM 06:04
vendethiel So, apparently, perl sometimes makes the imgur frontpage ... imgur.com/gallery/yDV6s 08:25
jercos I have it on reasonable authority that PHP makes the imgur frontpage. 08:30
Most of the other pages too.
nwc10 jnthn++ # nativecall 08:47
and who is going to try DBIish?
FROGGS I am still working on CPANish, so... :o9) 09:01
-9
nwc10 FROGGS++ # working on stuff 09:07
masak antenoon, #perl6 09:43
pmurias hoelzro: do you think it would be possible to generate all the different highlighters from a general highlighter description? 10:30
dalek p: f13dca0 | (Arne Skjærholt)++ | tools/build/gen-jvm-properties.pl:
Add some more information about how to compile things for JVM.

Hopefully this'll let NativeCall with the JVM backend compile libs on Windows platforms.
11:09
volaj: 9975032 | (Arne Skjærholt)++ | t/CompileTestLib.pm:
Try to make library compilation more robust on Windows JVM backend.
arnsholt jnthn: I'm hoping that pair of commits lets us compile libs on the JVM-Windows pair 11:11
colomon \o/ 11:28
jnthn arnsholt: Testing. 11:30
dalek kudo/eleven: a0a5990 | lue++ | src/core/ (2 files):
Rename .ast to .made, as per S05.

The .ast method is still around, though now it's deprecated.
11:42
timotimo lue: not deprecated! 11:43
timotimo is opinionated
JimmyZ 3 days ago..
timotimo it has been undeprecated
JimmyZ if you see url ...
jnthn w and t. The keys are like, right next...oh... 11:44
timotimo oh
jnthn Hm, that patch does depreate it though.
OK, I declare we're not deprecating it until after the 2014.03 release. 11:45
dalek ecs: 70f396a | (Elizabeth Mattijsen)++ | S11-modules.pod:
Introduction of "supersedes" and "superseded_by"

And some other minor clarifications.
jnthn Whatever the spec says.
Just out of ecosystem consideration and how close we are to the monthly release.
lizmat +1 11:46
timotimo aye
jnthn Oh, that patch went into a branch, anyways, though :)
timotimo yes
JimmyZ it's a merge ...
jnthn ah... :)
FROGGS soo much confusion... should I feel sorrow? :o) 11:47
JimmyZ and dalek exited
jnthn And indeed, nom has it undeprecated
So, good :)
Keep calm and continue hacking :P
JimmyZ :O
jnthn arnsholt: Doesn't seem to work out :( 11:49
arnsholt: Investigating.
timotimo m: my @colors = map -> $r, $g, $b { Buf.new: ($r, $g, $b) xx 16 div 8 }, 0, 0, 0, 255, 0, 0; say @colors;
camelia rakudo-moar a2db06: OUTPUT«This type cannot unbox to a native integer␤ in method new at src/gen/m-CORE.setting:5164␤ in method new at src/gen/m-CORE.setting:5154␤ in any at src/gen/m-Metamodel.nqp:1403␤ in block at /tmp/GtIRhXNHHf:1␤␤»
jnthn arnsholt: oh, wait
timotimo m: my @colors = map -> $r, $g, $b { Buf.new: ($r, $g, $b) xx (16 div 8) }, 0, 0, 0, 255, 0, 0; say @colors;
camelia rakudo-moar a2db06: OUTPUT«This type cannot unbox to a native integer␤ in method new at src/gen/m-CORE.setting:5164␤ in method new at src/gen/m-CORE.setting:5154␤ in any at src/gen/m-Metamodel.nqp:1403␤ in block at /tmp/PJaC80S63_:1␤␤»
timotimo m: Buf.new: (1, 2, 3) xx 2 11:49
camelia rakudo-moar a2db06: OUTPUT«This type cannot unbox to a native integer␤ in method new at src/gen/m-CORE.setting:5164␤ in method new at src/gen/m-CORE.setting:5154␤ in any at src/gen/m-Metamodel.nqp:1403␤ in block at /tmp/FSN3HwxKg_:1␤␤»
timotimo m: Buf.new: (1, 2, 3)
camelia ( no output )
timotimo m: Buf.new: 1 xx 10 11:50
camelia ( no output )
timotimo m: Buf.new: (1, 2) xx 10
camelia rakudo-moar a2db06: OUTPUT«This type cannot unbox to a native integer␤ in method new at src/gen/m-CORE.setting:5164␤ in method new at src/gen/m-CORE.setting:5154␤ in any at src/gen/m-Metamodel.nqp:1403␤ in block at /tmp/sD_bs46M9V:1␤␤»
timotimo p: Buf.new: (1, 2) xx 10
camelia rakudo-parrot a2db06: OUTPUT«This type cannot unbox to a native integer␤ in method new at gen/parrot/CORE.setting:5168␤ in method new at gen/parrot/CORE.setting:5158␤ in any at gen/parrot/Metamodel.nqp:1481␤ in block at /tmp/96yeoY8J4d:1␤␤»
timotimo oh, that's not a moar-only bug then
rosettacode.org/wiki/Colour_bars/Display#Perl_6 ← talking about this
jnthn arnsholt: I forgot to pull the nativecall repo. It works \o/
dalek volaj: 115d5ce | jnthn++ | t/01-argless.c:
Emit missing newline after a test.
11:57
dalek nda/depends: 8346347 | (Tobias Leich)++ | lib/Panda/ (2 files):
switch to META6.json, and fix depends/provides
12:39
dalek ecs: f901166 | (Elizabeth Mattijsen)++ | S22-package-format (2 files):
We want to keep S22 in its 5 year old incarnation around for now
12:49
masak timotimo: I think you're expecting too much flattening from Perl 6's signatures. 13:37
timotimo masak: well, the original author of that code seems to expect that :) 13:40
masak :) 13:44
maybe I'm wrong. maybe the thing after the colon isn't actually a signature, but a general expression.
timotimo hmm. the : *does* turn it into a listop precedence thing 13:45
masak right.
timotimo but i don't think it should automatically do something like |(...) for the expression
that could lead to extraordinarily big WATs in my expectation
well, i put a |( ... ) there and now it works, i think i should edit that rc code. 13:52
hm. this could be using a buf8 13:54
interesting. turning that buf into a Buf[int8] makes it neither faster nor save any memory (maxrss of course) on moar 13:55
well, that may be because there's only 8 of them 13:56
pmurias what is rakudo/eleven? 13:58
timotimo it's for packaging things related to the new S11 synopsis 14:02
surely, by now eqat is in the stage0 for all our nqp backends? 14:06
*weird*. our parrot stage0 doesn't have it? 14:08
FROGGS well, was there a reason to bump it lately?
timotimo in the last 4 months? :) 14:10
FROGGS well, yeah :o)
timotimo apparently not!
FROGGS see
timotimo last bump there was 5 months ago
FROGGS :/
timotimo ah well.
i have a quest that says "implement and use eqat on all backends"
but that's not a good reason to introduce another ~megabyte of download to nqp 14:11
retupmoca apparently the new moarvm nativecall blows up when precompiled
all tests pass if I don't precompile, but when I do I get 'Missing deserialize_stable_size'
timotimo ah. jnthn has already suspected that would happen
jnthn Yeah, I was sure a bunch of those were gonna be unported so far. 14:15
timotimo: Buf is a buf8 by default 14:18
timotimo oh! 14:21
arnsholt jnthn: Hooray! 14:23
moritz timotimo: remeber that 2048 game? it lets you continue after you win, and now I managed to get to the 4096 14:33
hiscore 59612
timotimo oh wow.
colomon moritz++
timotimo get back to work ;)))
also, did you see 2048-3d?
colomon covers his ears with his hands and goes "la la la" so he can't hear more 14:34
moritz timotimo: no, I didn't
hoelzro pmurias: I wish =( 14:47
probably not, though
well, we could generate some of the features
like list of builtins, classes, etc
timotimo aye, those are kind of easy
timotimo i even put the generation code into my kate definition files :) 14:48
hoelzro but for example, the pygments highlighter uses actual Python code to parse some more complicated constructions
which, afaict, most highlighters don't even allow
jnthn retupmoca: Latest Moar supports pre-comp 14:49
hoelzro jnthn: you've been working on native call on moar, right? 14:50
jnthn hoelzro: A little...
hoelzro is that support just in Moar? or in NQP as well?
jnthn hoelzro: Code is almost all in Moar, in NQP there's really just op mappings. 14:51
hoelzro but you haven't pushed up the op mappings at all?
jnthn hoelzro: No, they're there. Here, I get a fully passing NativeCall.pm too.
hoelzro hmm
I can't get mine to work =/
jnthn hoelzro: Didn't bump NQP_REVISION or MOAR_REVISION yet.
hoelzro maybe I'm using too old of an NQP
jnthn hoelzro: You need bleeding edge of everything.
hoelzro I'm just trying to play with it =)
oooooh 14:52
that's why
jnthn I'll bump stuff soon. It's just that things went yesterday from passing 2 test files to passing all 8 :) 14:52
hoelzro wooooooooo
jnthn++
jnthn And figured I'd do a final bump when the dust settles :)
colomon \o/
JimmyZ Are you going to rewrite CStruct compose? :) 14:53
jnthn JimmyZ: I maybe "should" but I really can't be bothered right now :)
JimmyZ: If you fancy taking it on, feel free.
I'm going to start looking at Star. 14:54
I'll make a branch to work on getting Star multi-backend aware.
JimmyZ jnthn: I just want to Apply gist.github.com/zhuomingliang/9541251
:P
jnthn JimmyZ: Oh, that looks like a simple cleanup 14:55
JimmyZ yeah
jnthn There's still a nasty bug though.
attr_iter = (MVMIter *)MVM_iter(tc, attributes);
That allocates
There's no reason to use it afaict
Just grab elems and then do a normal for loop 14:56
JimmyZ oh, I was porting the orig code... 14:57
jnthn I suspect we get away with more on Parrot 'cus it never moves objects. 14:58
JimmyZ My patch just re-orders the code to avoid MVMROOT current_class var
jnthn Yeah. But the rooting is missing that MVM_iter requires 15:02
hoelzro anyone mind +1'ing this PR to linguist that improves Perl 6 detection? 15:05
github.com/github/linguist/pull/900
masak wow, the first comment ever that perfectly straddles spam and non-spam such that I can't immediately tell. 15:06
it's a comment for perl6advent.wordpress.com/2013/12/0...pe-object/
the text is "Hi all just started programming with the ruby programming language so I'm quite a bit of a newcommer!
But yet have found your websiye very motivating and instructive.
Many thanks!"
and the URL is learn-programming-quickly.blogspot....y-way.html 15:07
masak the text makes me think it's genuine; the URL makes me think it's spam. 15:07
skids I'm lost among all the github icons. How do I +1 a PR?
masak oh and the URL is linked up with the text "Kids programming". I don't know enough about Wordpress to tell whether that's actually the user name or something.
moritz skids: you write +1, I'd guess :-) 15:08
skids Oh, I thought there was actually a voting system :-) 15:09
timotimo skids: i think it's :+1 15:11
at least all the icons start with : and you get a little "autocompleter"
:+1:
skids Got it eventually. I feel so old. 15:15
hoelzro thanks JimmyZ, skids 15:17
masak release week next week. 15:18
++tadzik
anyone want to nab the release manager role for May? there are still openings.
hoelzro tadzik++
dalek ar/tristar: da40b10 | jnthn++ | tools/star/Makefile:
Initial Makefile updates to include MoarVM.

Untested, may need tweaks, but should be about right.
15:19
ar/tristar: 361d740 | jnthn++ | tools/star/release-guide.pod:
Some release guide updates.
isBEKaml :wq 15:20
hugme hugs isBEKaml, good vi(m) user!
jnthn will be submitting talks for Dutch/Polish/Czech Perl workshops soon. If there's any topics anyone attending would especially like to hear me talk about, feel free to say so :) 15:22
dalek ecs: 3d55709 | (Elizabeth Mattijsen)++ | S (2 files):
Restart S22, move appropriate stuff in from S11
15:37
masak jnthn: I'd like to hear more about category theory... 15:51
:P
random insight from www.youtube.com/watch?v=jScLjUlLTLI -- technologies that don't restrict you too much (i.e. more Perl than Python) tend to spawn "Best Practices"-style cottage industries 15:52
Juerd The great thing about best practices is that they evolved from a lot of experience 15:53
masak putting on my Python hat, though, I could make the same case for the restrictions/opinions built into the language. 15:54
I'm by no means a Python expert. but half of the time I feel hemmed in by some "we know better" decisions; the other half of the time I kind of admire it. 15:55
TimToady hems and haws 15:57
FROGGS TimToady: howdy!
jnthn masak: I wsa hoping folks would pick topics I knew about :P 15:57
masak jnthn: just introduce yourself as a co-expert on category theory! :P 15:58
jnthn :D
masak .oO( In Soviet Russia, category theory reverses YOU! )
dalek ecs: cc3524b | (Elizabeth Mattijsen)++ | S22-package-format.pod:
Tweaks, make examples more in line with each other
15:59
lizmat hopes that TimToady and masak have time to look at the new S22 spec 16:00
(or anybody else, really, specifically people with packaging backgrounds) 16:01
to make sure that I'm on the right track
jnthn
.oO( In Putin theory, YOU reverse to Soviet Russia... )
lizmat
.oO( stop Putin fear in our hearts )
16:02
masak next random thought, also based on the same video: if you're more opinionated, that's a kind of giant's shoulders that the community can stand on. contrariwise, flexibility is "broad" but not "tall". 16:05
also, interestingly, in that talk the (Ember) speakers keep defining themselves in terms of Angular. a bit like Python people have defined themselves in terms of and distanced themselves from Perl. 16:06
"So if you're ready to stop going it alone, and joining a community that values making the right way feel right -- use Ember. You won't regret it." 16:07
yep. Ember is now officially the Python of JavaScript web frameworks.
jnthn Hmm...what... 16:11
So the Star build script claims it will download an NQP/Parrot, but also bundles them? 16:12
masak sounds a bit redundant. 16:14
TimToady you can say that again 16:15
jnthn is very confused 16:17
dalek ecs: c7aab52 | (Elizabeth Mattijsen)++ | S22-package-format.pod:
Add optional "authors" tag to META6.json
sjn lizmat: pad.hackeriet.no/p/p6-deploy 16:19
dalek ar/tristar: 31977c5 | Jonathan++ | Configure.pl:
Add --gen-moar option to Configure.pl.
TimToady wonders if "mitsuboshi" is a word in japanese... 16:21
isBEKaml what's tristar?
TimToady * with three backends, presumably
isBEKaml Ah. 16:22
dalek ar/tristar: 07094e4 | jnthn++ | README.star:
Update an outdated URL.
16:28
ar/tristar: 79a6cbd | jnthn++ | .gitignore:
Update .gitignore.
jnthn My attempt to get us a Star to have 3 backends, which will probably need to be re-done afterwards, because I'm awful at making build systems, but that's still less awful than it not being worked on.
TimToady is reminded of himself + rakudo for some reason 16:30
jnthn Well, I didn't work on the Rakudo one, which you can mostly be glad of. 16:30
TimToady see the comments in metaops/flatmap for an example of my nqp coding skilz 16:31
jnthn nqp::if(Mu 16:32
Taht's basically "if false" 16:33
Meaning it'll only ever execute the first branch...
TimToady yes, but it blue op if I tried to simplify it
so obviously my mental model was wrong sumhau 16:34
jnthn hm, odd
TimToady there's some "Oh obviously you have to change that X to a Y in that case" that I'm not cognizent of yet 16:35
dalek ecs: e9913fd | (Elizabeth Mattijsen)++ | S22-package-format.pod:
Introduce "license" tag in META6.json
jnthn I'm also not entirely sure how to try out the Star module build without having releases cut, which is a pain in so far as I wanted to fix blockers... 16:37
pmurias masak: looking at the ember website aren't they more ruby on rails then python? 16:40
lizmat this just on perl-qa: [17:36:25] GumbyNET7 CPAN Upload: Test-Ignore-Me-0.001 by FROGGS metacpan.org/release/FROGGS/Test-Ignore-Me-0.001 16:43
the first (unsuccessful) Perl 6 module distribution upload to CPAN 16:44
TimToady EYOUDIDN'TIGNOREME!!!!
lizmat actually, the upload should be ok
the indexing was still attempted, when it shouldn't have been :-(
TimToady hopefully it really gummed up the works :D 16:45
lizmat rjbs and Andreas are looking at it :-)
FROGGS no, the indexing bit is not yet patched, but the upload itself should go into a different directory, and that failed
lizmat
.oO( version 0.01, there is room for improvement :-)
16:46
TimToady FROGGS++ for being an Agent of Change
FROGGS hehe
masak pmurias: I think arguably Ruby is opinionated in sort of the same way as Python. they're just... less explicit about it, perhaps? 16:48
I could very well be wrong.
TimToady they just have a different one way to do it :)
isBEKaml masak: Ruby is hipster crowd, they SHOULDN'T be opinionated. :P 16:49
masak isBEKaml! \o/
isBEKaml hello, masak
dalek ar/tristar: 7e41622 | jnthn++ | tools/build/Makefile.in:
The Rakudo debugger binary is built by Rakudo now.

Don't need to handle it specially in Star any more.
16:51
japhb__ jnthn: Have you bumped the version files for Moar and NQP? dalek gets knocked off so often, it's hard to tell -- and I was about to kick off one of my Build All The Things scripts. (Wouldn't want to get to the end and find out it was pre-NativeCall-awesomeness.)
TimToady Bump All the Things! 16:53
jnthn japhb__: No, I didn't yet. 16:56
japhb__: Though I probably could now.
japhb__: Given I fixed the pre-comp thingy...
japhb__ jnthn: Ah, very good. 16:57
jnthn Lemme do it now... 16:58
dalek ecs: c7c81c9 | (Elizabeth Mattijsen)++ | S22-package-format.pod:
Elaborate in "support" tag in META6.json
16:59
TimToady thinks Buf.new should be flattening its args like any self-respecting slurpy
it's a naturally flat structure
TimToady has been watching too much anime lately, and thinks the N in N'th is crying because of the spurious apostrophe 17:04
dalek p: d5ec544 | (Elizabeth Mattijsen)++ | README.pod:
Mention --gen-moar as a parameter
17:06
p: 04ffa12 | (Elizabeth Mattijsen)++ | README.pod:
Change copyright to 2014
japhb__ TimToady: Or maybe just embarrassed? 17:07
TimToady it's embarrassin', I'm opinin'
dalek p: 83c0d27 | jnthn++ | tools/build/MOAR_REVISION:
Bump MOAR_REVISION for native call support.
17:10
kudo/nom: a24091e | jnthn++ | tools/build/NQP_REVISION:
Bump to an NQP with native call improvements.

Both MoarVM support, and JVM improvements.
17:12
isBEKaml we should have a bum pall 17:13
lizmat pulls, builds and spectests
isBEKaml to bump 'em off one by one.
TimToady masak: Best Practices cultures fighting with You Can't Even Think Wrong Thoughts cultures have been with us at least from the time of Athens and Sparta :) 17:15
arnsholt TimToady: "This! Is! Python!"? =) 17:16
isBEKaml Or "Tell me about it." (I'm Socrates and my wife just poured bucketful of water on me)
masak TimToady: now there's a YAPC talk right there. 17:20
TimToady: anyway, I'm not taking sides. much. I still like Perl a lot, even when I'm letting the Python-think get to me. :) 17:21
TimToady: also, it's interesting to note that we are sorta-kinda saying, with Perl 6, "we're *still* leaning towards the flexible end of the spectrum... but some things are worth being opinionated about, at least by default"
japhb__ TimToady: Not knowing my ancient Athenian history that well -- am I correct in assuming BP was Sparta and YCETWT was Athens?
TimToady quite the opposite, in my estimation 17:22
it was the athenians that were always arguing with each other :)
jnthn So somehow the Star build thing magically finds the bundled Parrot and NQP. 17:23
TimToady so maybe Perl 6 is more like Rome, where there's a great deal of flexibility within the constraints of the law :) 17:24
but it's a long learning curve to citizenship
TimToady and if your province gets too uppity, we'll just absorb you :) 17:27
rjbs Today was a great day with a high-stress ending! PAUSE function should return shortly… 17:29
TimToady obviously the name PAUSE is appropriate here :) 17:30
dalek ar/tristar: e17b584 | jnthn++ | tools/lib/NQP/Configure.pm:
Some Configure.pm updates for Moar.
17:35
ar/tristar: ee1d52e | jnthn++ | tools/lib/NQP/Configure.pm:
Fix syntax error.
17:38
lizmat teardown of day 3 of QA Hackathon& 17:42
dalek ar/tristar: db73a21 | jnthn++ | / (2 files):
Try to get --gen-moar working.
17:48
dalek ar/tristar: 1525fbe | jnthn++ | tools/lib/NQP/Configure.pm:
Another try at Moar, to not fetch from Git.
17:56
timotimo so, who would like to do "spec work" for the "variables directly accessible in NativeCall" bit? 18:06
ISTR jnthn said something about the "native var is rw" thing being closely related?
dalek ar/tristar: d6042a6 | jnthn++ | tools/lib/NQP/Configure.pm:
Shut up an error and get a working --gen-moar.
18:15
jnthn It's built a Moar and is now working on NQP... 18:16
timotimo \o/ 18:17
jnthn But that's just the first bit :/
jnthn Makefile.in is all a bit Parrot specific. 18:18
arnsholt timotimo: Yeah, it's essentially the same problem, from memory: containers on natively typed variables 18:19
dalek ar/tristar: 16df539 | jnthn++ | tools/build/Makefile.in:
Remove unused symbols.
18:21
lue 18:28
timotimo fair enough on that case. 18:34
after the tristar, do we make moar's spectests clean? :P 18:36
so that i can compare changes against a known state instead of running the thing twice?
jnthn timotimo: Anybody can work on doing that, provided they add RTs as they fudge 18:40
timotimo ah 18:43
maybe i've found my thing for the evening
masak hey #perl6, can I discuss a talk idea with you? Czech Perl Workshop, which I *may* be going to, has this theme "Future Perl". I want to do something like "I code Perl 5 and Perl 6, and the way I code in one language affects the way I code in the other". 18:44
would you come to such a talk? what, in particular, would you like to hear about?
also, what should the talk title be? :)
timotimo that seems like an interesting idea 18:45
something with entanglement
masak "The future is entangled" 18:50
it's like, I'm so used to p6 that when I code in p5, I tend to use a lot of p6 patterns. 18:51
almost coding in a kind of wishful thinking mode, or coding towards a p6 ideal.
masak sometimes that makes inherently p5 solutions take me completely off guard, like with strangelyconsistent.org/blog/speed-...-6-million 18:52
wow. both of those RT tickets are still open. 18:55
r: say "foooo" ~~ /^ f o ** 4 $/
camelia rakudo-parrot a24091, rakudo-jvm a24091, rakudo-moar a24091: OUTPUT«「foooo」␤␤»
masak r: my $l = 4; say "foooo" ~~ /^ f o ** $l $/
camelia rakudo-parrot a24091, rakudo-jvm a24091, rakudo-moar a24091: OUTPUT«===SORRY!=== Error while compiling /tmp/tmpfile␤Quantifier quantifies nothing␤at /tmp/tmpfile:1␤------> my $l = 4; say "foooo" ~~ /^ f o ** ⏏$l $/␤ expecting any of:…»
timotimo yes, we still can't do ** $foo :(
masak :/ 18:56
r: my $l = 3; say "x" x 10 ~~ / x ** {$l} /
camelia rakudo-parrot a24091, rakudo-jvm a24091, rakudo-moar a24091: OUTPUT«===SORRY!=== Error while compiling /tmp/tmpfile␤Block case of ** quantifier not yet implemented␤at /tmp/tmpfile:1␤------> my $l = 3; say "x" x 10 ~~ / x ** ⏏{$l} /␤ …»
masak :/ ** {$l}
lue Is there some sort of .reassuming thing I can use for functions? I specifically would like to toggle a binary named parameter. (There's always keeping an un-primed version of the function around, but I'd rather not)
masak lue: would recomment keeping the un-primed version around. 18:57
Mouq std: my $l = 4; say "foooo" ~~ /^ f o ** $l $/
camelia std 09dda5b: OUTPUT«Potential difficulties:␤ Unsupported use of atom ** $l as separator; nowadays please use atom +% $l at /tmp/ay4dEwlNuz line 1:␤------> my $l = 4; say "foooo" ~~ /^ f o ** $l ⏏$/␤ok 00:01 127m␤»
masak lue: before I had finished reading your question, that's what came to mind.
Mouq: that's a "wound" due to spec shifting over time. 18:58
timotimo how about a role for that method that keeps the original around and replaces .assuming to call .assuming on the original instead?
Mouq n: my $l = 4; say "foooo" ~~ /^ f o ** $l $/
camelia niecza v24-109-g48a8de3: OUTPUT«Potential difficulties:␤ Unsupported use of atom ** $l as separator; nowadays please use atom+ % $l at /tmp/wQSWbMGRfw line 1:␤------> my $l = 4; say "foooo" ~~ /^ f o ** $l⏏ $/␤␤Nil␤»
timotimo role UnmodifiedVersion[&foo] { ... } ?
lue masak: yeah, I was kinda hoping I wouldn't need to keep two copies of each function around :) [though the original list can probably just bind to the functions instead of assigning]
Mouq Someone else should really fix that :P 18:59
** {$l}
masak I understand why it's hard.
it's where (static) state machines meet (runtime) variables.
Mouq masak: I've been doing <{"thing**$l"}> when I need it 19:00
masak Mouq: hm. 19:01
Mouq: do you mean "thing" xx $l ?
Mouq is very very happy to see Star and and packaging getting love, though 19:02
s/ / NativeCall/
lue masak: <{}> turns its contents into a regex 19:03
Mouq masak: I don't remember why I used it before, but I believe it was for something like <{"thing**3..{$some calc}"}>
Mouq [jnthn, lizmat, FROGGS, JimmyZ, arnsholt]>>++ 19:05
Mouq hopes he didn't miss anyone, probably did
Mouq must go, though. o/ 19:06
timotimo masak: are you going to revisit your t4 code for performance? 19:18
and, i suppose, idiomaticity? 19:20
moritz and idiocy? :-) 19:32
lue This talk of t4 reminds of a blog post I need to get around to putting up :) 19:33
jnthn moritz: I've got a little further with multi-backend Star. I've reached Makefile.in. Had you pondered what to do with it at all? 19:38
moritz: Would a Makefile-JVM.in etc approach make sense here, as in Rakudo?
masak today's autopun spotting: twitter.com/thereaIbanksy/status/4...6545026048 20:09
masak lue: oh! right. Mouq++ 20:10
timotimo: yes, I imagine I will.
still looking for someone to do an ironman thing with, by the way.
that would help me get back to blogging again, I think.
jnthn masak: Just make a rule that you can only drink Yeti after blogging :P 20:12
masak heh. 20:13
no, I need something more competitive.
timotimo nqp: my $foo = "hello there"; my $bar = "hello there"; say($foo =:= $bar); $bar := ~$bar; say($foo =:= $bar); 20:23
camelia nqp-parrot: OUTPUT«Assignment ("=") not supported in NQP, use ":=" instead at line 2, near " \"hello th"␤current instr.: 'panic' pc 15934 (gen/parrot/stage2/NQPHLL.pir:5941) (gen/parrot/stage2/NQPHLL.nqp:425)␤»
..nqp-moarvm: OUTPUT«Assignment ("=") not supported in NQP, use ":=" instead at line 2, near " \"hello th"␤ at gen/moar/stage2/NQPHLL.nqp:369 (/home/p6eval/rakudo-inst-1/languages/nqp/lib/NQPHLL.moarvm:panic:120)␤ from <unknown>:1 (/home/p6eval/rakudo-inst-1/languages/nqp/lib/…»
..nqp-jvm: OUTPUT«Assignment ("=") not supported in NQP, use ":=" instead at line 2, near " \"hello th"␤ in panic (gen/jvm/stage2/NQPHLL.nqp:378)␤ in infix:sym<=> (gen/jvm/stage2/NQP.nqp)␤ in !protoregex (gen/jvm/stage2/QRegex.nqp:784)␤ in infix (gen/jvm/stage2/NQPHLL.nqp)…»
timotimo nqp: my $foo := "hello there"; my $bar := "hello there"; say($foo =:= $bar); $bar := ~$bar; say($foo =:= $bar);
camelia nqp-moarvm, nqp-jvm, nqp-parrot: OUTPUT«0␤0␤»
timotimo nqp: my $foo := "hello there"; my $bar := $foo; say($foo =:= $bar); $bar := ~$bar; say($foo =:= $bar); 20:24
camelia nqp-moarvm, nqp-jvm, nqp-parrot: OUTPUT«1␤0␤»
timotimo aaw. 20:25
arnsholt jnthn: JAST::Method, many method, very code, wow. 20:31
timotimo yay, i segfault nqp 20:37
the empty string is a poor replacement for every string in the whole program. 20:39
arnsholt Much impress, very achieve!
jnthn arnsholt: Yes, that's the worst one in a sense :) 20:41
arnsholt A lot of attributes, at any rate 20:42
I'm reconstituting (or maybe deconstituting is a better term =) the JAST nodes into Java classes for the direct compilation of JAST
jnthn: Incidentally, Ops.getlex("JAST::Thingy", tc) is TRT to do to get at the SixModelObject for the JAST::Thingy, right? 20:44
timotimo: I'd've never guessed! =D
jnthn arnsholt: Um...not really 20:46
arnsholt: On Moar we pass the node types in.
arnsholt: Into the tree to bytecode thing I mean
Hash of name to node type
arnsholt Ah, right 20:47
So the compile ops have to take an argument with the name-to-type lookup hash
jnthn yeah
arnsholt adds a TODO to the code 20:48
timotimo is there a reason to believe that $*W may not be set to something at some point? 20:51
jnthn Well, before the point in...hm...TOP? Or comp_unit? 20:52
It's declared in one of the two
Pretty early
timotimo mhm 20:53
timotimo it seems like i'm being called from the term rule's action 20:54
jnthn During stage parse?
timotimo yes, i think so
aye. 20:55
the only way i can imagine $*W.intern($foo) instead of $foo to fail is that $*W is null or something
oh, duh
jnthn Could fail on the inside... :)
timotimo er, i don't actually have to set the has %!interned_strings to anything, right?
not if you replace the whole body with return $str
jnthn timotimo: Yes, you need to initialize it in BUILD 20:56
NQP's initialization stuff is not so clever as Rakudo's...
timotimo that may make a difference then :)
jnthn chuckles at the bottom variable name currently initialized in there :) 20:57
timotimo code_ref_blocks? 20:58
jnthn oh are you in HLL::World or Perl6::World? :) 21:01
jnthn arnsholt: ohbtw, myself and masak will be about in Oslo on Monday evening, if you're free and want to take a jar... 21:05
arnsholt jnthn: Oh, that sounds great! 21:07
masak arnsholt: what jnthn said. :)
jnthn We can have a "NativeCall runs on all the things" celebratory beer! 21:08
arnsholt Oh, indeed!
What're your plans for dinner, BTW? 21:09
I think I promised jnthn that I'd show him one of the better Indian places, last time he was here
jnthn arnsholt: We'll need to eat. 21:10
arnsholt That specific, eh? 21:11
Anyways, gimme a prod if you wanna grab something together. I can probably think of a place or two =)
jnthn arnsholt: That was a "we didn't make plans and so are open to having dinner" :)
jnthn An Indian works very well for me. But that's always the case :) 21:12
I suspect masak would like it too.
arnsholt Spiffy 21:15
dalek Heuristic branch merge: pushed 20 commits to nqp/moar-conc by jnthn 22:20
kudo/moar-conc: a0a5990 | lue++ | src/core/ (2 files):
Rename .ast to .made, as per S05.

The .ast method is still around, though now it's deprecated.
dalek osystem: f652b00 | (Andrew Egeler)++ | META.list:
Add Compress::Zlib::Raw
22:42
kania Hello, I need some help, I made a program for a class exercise ... 23:11
timotimo oooh, Perl 6 class? 23:13
kania of course timotimo 23:14
timotimo okay
elaborate please?
kania ok... exercise is to fill political Car with 4 each, there are two different political parties, can get 2 of each party, or 4 only one 23:15
I've done it and run, it works, but not always 23:16
also when it ends (I have since finished the 10 cars) shows the threads that have been active :/ 23:17
and do not know what I'm doing wrong, I guess it will close some light, but I can not find is wrong 23:18
timotimo huh? what's a political car?
kania mmm sorry... one car with four politics
yes?
timotimo ah, ok
without seeing any code, i cannot know what's wrong 23:19
kania wait please one moment I up the code 23:19
pastebin.com/3wbyu1h3 23:21
timotimo oh, but that's perl 5 and not perl 6 :(
kania :s
timotimo i'm one of the few who can't do any perl 5, only perl 6, sorry
kania thank you for read me 23:22
lue
.oO(Nothing like something breaking that's apparently ungolfable. Grrrr)
23:38
lue Well, turns out it was just a typo in the test file :/ 23:48
timotimo ouch 23:49
lue though I did encounter a real problem that was masked by both the typo and my use of submethod BUILD: 23:50
r: class C { has $.a = False; method new(:$a) { self.bless(:$a) } }; say C.new.a;
camelia rakudo-parrot a24091, rakudo-jvm a24091, rakudo-moar a24091: OUTPUT«(Any)␤»
jnthn You provided a value for it, so the default won't apply. 23:51
lue the only way to make that work is to say new(:$a = False), afaik
jnthn: yeah, I realize why it doesn't work, I just don't agree with it :) binary adverbs are binary enough I feel... 23:52
*aren't binary enough
jnthn Any way I can think of right off to change that feel likely to cause other kinds of fun... 23:53
I mean, the check could be "is the attribute defined"
But then you can't construct an object and deliberately put a type object in an attr... 23:54
lue tries to find that "coercion" typing thing
jnthn figures he'll get some rest :)
lue r: class C { has $.a = False; method new(Any(Bool) :$a) { self.bless(:$a) } }; say C.new.a;
jnthn 'night, #perl6
camelia rakudo-parrot a24091, rakudo-jvm a24091, rakudo-moar a24091: OUTPUT«===SORRY!=== Error while compiling /tmp/tmpfile␤coercive type declarations not yet implemented. Sorry. ␤at /tmp/tmpfile:1␤------> { has $.a = False; method new(Any(Bool)⏏ :$…»
lue jnthn o/
n: class C { has $.a = False; method new(Any(Bool) :$a) { self.bless(:$a) } }; say C.new.a;
camelia niecza v24-109-g48a8de3: OUTPUT«(Any)␤»
lue n: class C { has $.a = False; method new(Bool(Any) :$a) { self.bless(:$a) } }; say C.new.a; 23:55
camelia niecza v24-109-g48a8de3: OUTPUT«(Bool)␤»
lue nope, guess you have to = False (or some MMD with methods lacking the adverb) 23:56
jnthn: I can imagine there'd be problems by changing things around. This is just one of those corner cases where flags shouldn't be quite the same as adverbs, is all :) 23:58