»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'p6: say 3;' or rakudo:, or /msg camelia p6: ... | irclog: irc.perl6.org or colabti.org/irclogger/irclogger_logs/perl6 | UTF-8 is our friend!
Set by moritz on 22 December 2015.
AlexDaniel m: sub foo(@array ($first, @rest)) { say @rest }; foo <1 2 3> 00:01
camelia rakudo-moar 641ee1: OUTPUT«cannot stringify this␤ in sub foo at <tmp> line 1␤ in block <unit> at <tmp> line 1␤␤»
AlexDaniel yes, I know that it is *@rest 00:02
AlexDaniel rakudobugs it
AlexDaniel j: sub foo(@array ($first, @rest)) { say @rest }; foo <1 2 3> 00:07
camelia rakudo-jvm cd19db: OUTPUT«__P6opaque__75@437a1841 in sub-signature of parameter @array␤ in sub foo at <tmp> line 1␤ in block <unit> at <tmp> line 1␤␤»
AlexDaniel /o\
japhb I see there was a big argument earlier about database categories, in which "what Google does" came up a few times. Unless they've actually worked for Google, it's likely that what any given person thinks Google does on the inside is probably at least missing some nuance, if not just plain wrong. There's a LOT more complexity than simple assumptions would indicate. I'd be willing to bet the same is true at the other top-tier tech companies. 01:04
theRegularPreten Good evening 01:24
It has been asked a million times already, I feel, but there I go:
Is there a CodeCademy-like site that teaches Perl?
dalek ateverable: 3cf1e29 | (Aleks-Daniel Jakimenko-Aleksejev)++ | build.p6:
Build all tagged commits

This way it will be possible to dig as far as you'd like, in the expense of building less than a hundred commits.
Fixes issue #30		 01:43
tushar theRegularPreten, for Perl 5 Udemy has one or two courses. In addition, Dave Cross wrote a blog kind of thing for Udemy for Perl 5. For Perl 6, I have not seen anything yet. 01:45
SmokeMachine____ Hi! I am writing a constraint programming problem solver in perl6... if anyone get a time have any hint/comment, please? github.com/FCO/ProblemSolver 02:21
perlawhirl so... 04:37
is irclog.perlgeek.de down for anybody else?
tadzik seems like it: isup.me/http://irclog.perlgeek.de/ 04:38
dalek whateverable: 4d18cb4 | (Aleks-Daniel Jakimenko-Aleksejev)++ | / (2 files): 04:39
whateverable: Add special checks for exit code 125
whateverable:
whateverable: Normally this should not happen, but a user may decide to skip some
whateverable: commits manually by exiting with exit code 125. That's totally OK, but
AlexDaniel perlawhirl: it's not completely down 04:40
just a bit unstable
at least that's how it was a few hours ago
dalek line-Perl5: c630c25 | niner++ | / (2 files):
Pass Map as pure Perl 5 hashes

Maps are immutable, i.e. we don't expect any writes to make it back to the Perl 6 object. So they are the perfect candidate for signalling that it can be safely copied to a Perl 5 hash with no ties to the original.		 06:27
perlawhirl I need to upgrade my Inline::Perl5... nine++ seems to have made a spree of optimisations in the past 24 hours 06:30
perlawhirl can panda update modules... or do i just... 'install --force' it 06:33
<trailing question mark>
moritz perlawhirl: panda --force install The::Module 07:01
(note that the --force must come before the 'install')
nine perlawhirl: panda install The::Module will install a new version. You need --force only if the version number is the same as the installed one's. 07:12
dalek line-Perl5: ebba898 | niner++ | META.info:
Version 0.12
 07:13
perlawhirl cool
nine perlawhirl: while you're at it. Could you please try the following line with the old and new version? 07:14
git clone github.com/Tux/CSV.git && cd CSV && for i in $(seq 1 10000); do echo 'hello,","," ",world,"!"'; done > /tmp/hello.csv && time perl6 csv-ip5xs.pl < /tmp/hello.csv && time perl6 csv-ip5xs.pl < /tmp/hello.csv
perlawhirl I already updated, but I can reinstall an old commit and test... though i warn you my 32-bit laptop ain't the fastest 07:21
nine Just the number for the new version can help, too. 07:23
Trying to find out, why the 40 % improvement I see here does not show um on Tux++'s daily benchmark.
perlawhirl So... I installed version 0.10 ran the test, then version 0.12, but it was more or less the same... ~8.85s. maybe I did it wrong, or maybe my 32-bit machine hit it's speed limit :D I will try on another box later. 07:30
nine perlpilot: thanks 07:34
Oops...wrong tab completion 07:35
nine .tell perlawhirl note that just installing an older version when a newer one was already installed, won't make rakudo use that old version. If no :ver is specified in the use statement, we pick the newest version. 07:57
yoleaux nine: I'll pass your message to perlawhirl.
masak good antenoon, #perl6 08:45
arnsholt o/ masak
DrForr wow, almost afternoon here. I guess I'd better get something accomplished. 08:46
arnsholt has ordered more stuff for his new keyboard
And read about SOAP 08:47
DrForr As keyboards normally *are* accessories, I assume this is of the all-keys-in-a-row variety? :) 08:48
masak arnsholt: my heartfelt sympathy
arnsholt DrForr: Dampener rings to put under the keycaps
It's a mechanical keyboard, and recent feedback from colleagues have indicated that a bit of soundproofing may be in order =) 08:49
DrForr Ah.
masak been there, done that :) 08:50
arnsholt And a few custom keycaps while I'm at it: a red escape key and a pair of Portal keycaps to put on the Windows keys
masak: I've only read a tiny bit, so I'm not scarred for life just yet 08:52
And other people have implemented most of the SOAP stuff. Apparently, I'm supposed to implement MTOM (sending binary stuff in binary next to the XML, basically) in existing code
jnthn so far today fixed 2 bugs and ate a plum 08:53
I liked the plum more than the bugs.
DrForr I guess I'm one of the last people to not look at my keyboard when typing :)
arnsholt Yeah, I need to learn to touch type properly, really
But the custom keycaps are mostly just bling =) 08:54
jnthn was lucky enough to get typing lessons when he was young 08:56
DrForr Same here.
arnsholt We were supposed to learn it at school when I Was about 16, but at that point it was too late for me
jast it's never too late! 08:57
jnthn Though I started on a typewriter. After a while of using the home computer's keyboard in a similar manner, I got told off for being too loud and busting the space bar. :P 08:58
DrForr Drape a tea towel over your hands and start making errors. You'll learn very quickly where the backspace key is :)
arnsholt Yeah, it's mostly a question of time and being reluctant to go through the effort of breaking two decades of muscle memory of faking it
jast in my experience, muscle memory isn't that hard to change if you don't make a big deal out of it
arnsholt jnthn: I typed on a Model M clone for many years. I have been told similar things =)
DrForr On occasion I switch to Dvorak just for the hell of it. And pissing off anyone who tries to type on my laptops. 08:59
jast it does take time, though, no argument there
jast DrForr: I think colemak will confuse people much more because it's somewhat closer to QWERTY 08:59
El_Che full disk. Arrg. Linux VM + zillion docker images takes space. Who would have thought it (I need to get rid of this $work Windows installation) 09:11
lambd0x Hi everyone! 10:13
m: say 3//2; 10:14
camelia rakudo-moar 7e3506: OUTPUT«3␤»
lambd0x m: say 1//2;
camelia rakudo-moar 7e3506: OUTPUT«1␤»
lambd0x m: say -1//2;
camelia rakudo-moar 7e3506: OUTPUT«-1␤»
lambd0x how does the // operator work? 10:16
brrt i think it is integer division 10:17
DrForr docs.perl6.org/language/operators#infix_//
lizmat it returns the left side if it is defined, otherwise the right side
brrt oh…
lizmat nothing to do with division
brrt it's the same as perl
hehe
brrt was confused with python
DrForr You're maybe thinking of 'div'?
brrt m: -1 div 2
camelia rakudo-moar 7e3506: OUTPUT«WARNINGS for <tmp>:␤Useless use of "div" in expression "-1 div 2" in sink context (line 1)␤»
DrForr (and I'm probably thinking of TT perl5)
brrt m: say -1 div 2;
camelia rakudo-moar 7e3506: OUTPUT«-1␤»
lambd0x lizmat: hm... when won't it for example be defined? 10:19
lizmat m: my $a; say $a / 42
camelia rakudo-moar 7e3506: OUTPUT«Use of uninitialized value of type Any in numeric context␤ in block <unit> at <tmp> line 1␤0␤»
lizmat m: my $a; say $a // 42 10:20
camelia rakudo-moar 7e3506: OUTPUT«42␤»
DrForr When the left side is a variable that you haven't defined?
lizmat DrForr: s/defined/initialized/ ? 10:21
m: no strict; say $a // 42
camelia rakudo-moar 7e3506: OUTPUT«42␤»
lambd0x DrForr: ahahah, I was just wondering if therefore u have a Int defined variable, by default it has a 0. So it's always defined.
lizmat :-)
DrForr 'declared' was probably a better word to use there :/ 10:22
lizmat m: my Int $a; say $a # still undefined
camelia rakudo-moar 7e3506: OUTPUT«(Int)␤»
lizmat m: my $a; say 42 + $a; # acts as 0 with warnings
camelia rakudo-moar 7e3506: OUTPUT«Use of uninitialized value of type Any in numeric context␤ in block <unit> at <tmp> line 1␤42␤»
lambd0x lizmat: ok, that's nice :) 10:23
DrForr: yeah, sorry...just woke up haha.
DrForr Talking to myself, not you. 10:25
Woodi hi #perl6 :) 10:42
Woodi I keep getting ===SORRY!===\nMissing block for: if $kh<milk> { line... any ideas ? 10:43
"This is perl6 version 2012.10 built on parrot 4.6.0 revision 0" 10:44
moritz Woodi: try using a perl 6 that' isn't nearly 4 yars old 10:45
Woodi moritz: I can't. I'm on my way back in time and this prevents progress... 10:47
moritz Woodi: then seing the code around it might or might not help 10:49
Woodi here: github.com/slunski/caffeeExample/b...e.pl6#L203 10:52
pmurias arnsholt: vim in my experience is very good for learning to touch type, as it makes you keep your hands in proper position ;)
Woodi except that %kh is $kh to reach that point in time :)
moritz Woodi: we didn't have adverbs back than, I believe
Woodi: so %kh.exists('milk') 10:53
or check with .defined, if that's acceptable
Woodi moritz: nothing changes. maybe just line numbersare off... 10:55
moritz Woodi: then I don't know 10:57
Woodi moritz++ thanx anyway :) 10:59
Woodi ok, problem was in 'next;' few lines above... btw. 'unless' still don't allows 'else' ? 11:08
masak Woodi: doesn't allow it by spec. 11:10
Woodi: it's not like it's hard to implement so that it allows it.
S04: "The C<unless> statement does not allow an C<elsif> or C<else> in Perl 6." 11:11
Woodi masak: I wonder why ? natural language betterness ? 11:14
moritz finds unless + else confusing 11:16
masak Woodi: yes, you could call it that.
Woodi masak: oki :) 11:18
masak Woodi: Perl 6 is pretty good at staying out of your preferences and letting you choose.
Woodi: but in this case, the language makes the call that if you do `unless` followed by `else`... you're confused, and you should be told you're confused 11:19
Woodi: maybe Perl 6 is a little bit better than Perl 5 at detecting and preventing "useless double negations". `unless` + `else` is one. `!!expr` is another -- Perl 6 has `?expr` 11:20
(but in that latter case, doesn't enforce) 11:21
m: say !!"OH HAI"
camelia rakudo-moar 4b1864: OUTPUT«True␤»
Woodi masak: realy I do not wanted unless-else, just mechanical try to replace 'next' discovered that :)
masak I still think you got the nudge for the right reasons :) 11:21
Woodi ok, so I will do half city walk :) later 11:25
Ven` o/ 11:25
masak Ven! \o/ 11:27
ShimmerFairy OK, how do I get a C++ style namespace thing going, where I can use abbreviated names inside and not outside? That is, module A { class B { ... }; class C { my B $.d; } }, and then anyone who 'use's the module has to use A::B etc.? 11:38
lizmat ShimmerFairy: -use- is compunit level, so the only thing you can use is the module A 11:40
lizmat if you really want to support A::B, you would have to make that a separate compunit (file) which brings the problem of not being able to refer to B from C without prefixing A:: 11:41
hope I made sense ? 11:42
ShimmerFairy lizmat: huh, I could swear an our-scoped thing in a module got exported so that you could skip the A::, but apparently not? 'class B' apparently forces A::B, where I thought I need to do stuff like 'my class B is export' or such
lizmat you were asking about forcing people to have to say "use A::B" 11:43
you can only -use- a compilation unit (compunit) 11:44
class B inside module A is part of the compunit in which it lives, you cannot -use- it by itself
ShimmerFairy lizmat: no I wasn't, I was asking about forcing people not to do things like 'use A; say B'. I was having that issue when I thought 'my class Foo is export' would do what I wanted.
lizmat aaaahhhhh 11:45
use ne -use- :-)
ShimmerFairy I don't know why, but I really hate and find it hard to wrap my head around how Perl 6 handles names and "namespaces" and such. C++'s 'namespace' approach is so much easier for me. 11:46
masak I tend to not nest things in blocks so much nowadays. 11:47
I do things in separate compunits, like lizmat++ suggested. that seems to work for me.
ShimmerFairy masak: the problem is that I don't like typing out long names a bunch (e.g. class Really::Long::Name::B { has Really::Long::Name::A $.foo } ), and I'm not always a fan of a Java-esque "one class per file" approach. 11:48
lizmat putting this in Foo.pm: module Foo { class A is export { } } 11:49
ShimmerFairy If it's a big class, sure, but if I'm just writing up a bunch of AST node classes, one per file would be more annoying than helpful.
lizmat and then use Foo; makes Foo::A appear in scope
ShimmerFairy: so adding "is export" to the class seems to do the trick, if I understand what you want 11:50
ShimmerFairy lizmat: except I (somehow) don't need the 'is export' even. I don't even know why.
lizmat ah, indeed... 11:53
I guess that's because classes are by default "our" rather than "my"
if I change it to "my class A { }", it doesn't work, not even with "is export" added... 11:54
which feels like a bug to me
ShimmerFairy lizmat: I could swear that before I discovered our-scoped things in a module would export as its short name, hence why I was trying something like "my class Foo is export" 11:55
lizmat ShimmerFairy: not sure what to make of the current state, hope for some feedback from TimToady jnthn masak moritz 11:59
ShimmerFairy lizmat: because I really, really dislike how names are handled at the moment. Like I said, C++'s ways are much easier and sensible for me. (I also like how 'namespace Foo { }' just means "put these things in a bucket called 'Foo'") 12:01
lizmat well, that's 'package' for you in Perl
ShimmerFairy really, you're allowed to write 'package Foo' multiple times? I thought that was just as impossible as doing the same with 'module' and such. 12:02
moritz our-scoped and "is export" are orthogonal 12:03
moritz and several times "package" works because a package is the stub for more specialized packages, like classes, modules, roles, grammars 12:07
ShimmerFairy I was wondering if I'd have to end up making a NamespaceHOW to have things done like I want :) 12:08
Ven` p6: package a { sub f { 5 } }; package a { sub g { f - 2 } }; a::g 12:15
camelia rakudo-moar 4b1864: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Undeclared routine:␤ f used at line 1␤␤»
dalek line-Perl5: a46e341 | niner++ | lib/Inline/Perl5.pm6:
Add missing Hash and Map methods to Perl5Hash
 12:16
ShimmerFairy Ven`: ah boo, looks like I would need to create a NamespaceHOW type thing then :P
pmurias what's the difference between nqp::fc and nqp::lc? 12:21
masak pmurias: fc is for comparing case-insensitive 12:22
lizmat nine: feels like those changes could do with a loop and some mop magic :-) 12:23
pmurias masak: when do they behave differently?
avar pmurias: p5 docs, but relevant: perldoc.perl.org/functions/fc.html 12:24
masak avar++
I know the differences are minor, and maybe nonexistent for English
nine lizmat: better would be to give them real implementations. Right now each of those methods just create a Perl 6 hash to delegate to. method elems() for example could call hv_scalar to get the number of items. 12:26
lizmat nine: ah, yes :-) 12:27
nine But that's or after my talk at the Dancer conference ;) 12:27
huf how does casefolding deal with the turkish i problem? 12:28
ShimmerFairy huf: the UCD has a couple of special cases for the turkish i, but in general I believe the CLDR is for region-specific stuff.
moritz m: say lc 'İ'
camelia rakudo-moar 4b1864: OUTPUT«i̇␤»
lizmat nine: confirm spectest now clean
moritz m: say fc 'İ'
camelia rakudo-moar 4b1864: OUTPUT«i̇␤»
lizmat afk& 12:29
huf moritz: well that's broken :) 12:30
masak m: say fc("Iİ")
camelia rakudo-moar 4b1864: OUTPUT«ii̇␤»
masak hm.
moritz huf: what do you think it should case-fold to?
nine lizmat: \o/ 12:31
moritz huf: it's not like folding to lower-case i without dot would be a better answer :-)
huf moritz: i dont think a good solution exists
but an i + a combining dot seems double wrong
moritz huf: I find the current answer quite satisfactory
it just doesn't round-trip, but I'm used to that from the ß -> SS case
huf yeah, it's never going to roundtrip 12:32
ShimmerFairy www.unicode.org/Public/UCD/latest/u...olding.txt <-- grep for T; 12:32
huf humans have messed up writing far too much for that :)
moritz so, what's the problem then?
ShimmerFairy Perl 6 doesn't handle anything locale-specific such as those Turkish special cases AFAIK. 12:33
huf the combining dot looks wrong, but maybe that's just me
i guess it makes casefolded dotted-capital-i distinct so it's cool
ShimmerFairy m: say "Unicode sez: \x0130 -> \x0069 and \x0049 -> \x0131 for turkish langs" 12:34
camelia rakudo-moar 4b1864: OUTPUT«Unicode sez: İ -> i and I -> ı for turkish langs␤»
huf yeah, it's easy when you know the locale, but by that point you've already lost :) 12:35
masak not sure "knowing the locale" ought to qualify as losing 12:36
ShimmerFairy m: say "Unicode sez: \x0130 -> \x[0069,0307] and \x0049 -> \x0069 for others"
camelia rakudo-moar 4b1864: OUTPUT«Unicode sez: İ -> i̇ and I -> i for others␤»
ShimmerFairy really what we want are things like a CLDR module :)
huf masak: it means you cant treat strings uniformly and you get into messes like mixed-locale text
how to casefold/capitalize a turkish sentence containing a foreign name or phrase, etc :) 12:37
ShimmerFairy Programs rendering mixed bidi text seem to manage that sort of thing decently enough :P
huf do they?
ShimmerFairy well, the good ones
huf i've always had lots of trouble with them :)
moritz translations face the same problem, mostly when words have different meanings in both languages 12:38
xiaomiao The pressure takes place in Germany.
(direct translation of "Der Druck findet in Deutschland statt" - printing, not pressure, same word in german)
timotimo :D
huf yeah, translation is 12:39
xiaomiao seen in a contract for a multimillion-$ project
huf much worse
timotimo fortunately perl6 don't promises a vm-builtin for translating human language ...
huf :)))
praise the lord
timotimo will we be offering nqp::turkishfc() ?
ShimmerFairy
.oO("Oder wie wär's mit chicken ?" "Ich will nichts schicken! Ich will was essen!")
moritz xiaomiao: I mostly meant mixing in words from other languages that also have a meaning in your own
timotimo what's wrong, DHL? schicken? - nobody calls me schicken! 12:40
masak yes, it's a pity that the world is too messy to fit in the neat little labeled boxes we've designed for it :P
timotimo the world needs a stern talking to by a group of comp sci nerds
ShimmerFairy How to be more like FORTRAN: A world that isn't so compli 12:41
xiaomiao www.youtube.com/watch?v=yR0lWICH3rY 12:42
masak even translating from Chinese (simp) to Chinese (trad) is a hard problem 12:43
xiaomiao even translating german to german can be difficult ;) 12:45
smls masak: Doesn't Wikipedia do that automatically though? 12:47
masak perhaps they do. doesn't mean it's not hard, or that when they do it they don't make lots of mistakes. 12:49
I mean, Google Translate is at the same time fantastic and incredibly naïve when it translates 12:50
moritz I remember using it once to translate some poetry, and it translated a weird word order into a weird word order in the other language. Bug or feature? How to tell, even? :-) 12:51
xiaomiao masak: google translate is roughly good enough to figure out the topic and intent of a foreign-language text 12:52
xiaomiao it's like, say, reading dutch as a native english speaker 12:53
huf only if the language is close to english
or has received a lot of attention from someone
i've never found the *<>hu translations any good
tony-o *<>hu ?
huf anything to hungarian 12:54
tony-o ah
timotimo translate anything to huf
huf and i think gtranslate does generic a->b translations through english
which introduces interesting artifacts
timotimo like "translation party" :) 12:55
huf oh well, it's slave labor, we shouldnt expect good quality 12:57
moritz the French <-> German <-> English triangle seems to work rather well
huf the "3 big ones" work better than the rest? shock! 12:58
Ven` yeah, gtranslate uses english as a middleman
ShimmerFairy which is fair, considering 2n is usually much less than n! :) 13:00
SmokeMachine____ Hi there! I am writing a constraint programming problem solver in perl6... does anyone get a time to give me any hint/comment, please? github.com/FCO/ProblemSolver 13:02
SmokeMachine____ There's a EVAL on that project that I'd like to remove... but I can't find how... 13:03
tony-o SmokeMachine____: what is the eval doing? 13:04
SmokeMachine____ tony-o: creating a sub 13:04
No, a block
SmokeMachine____ With a variable number of input parameters... 13:05
Here's the EVAL: github.com/FCO/ProblemSolver/blob/...em.pm6#L67 13:07
timotimo well, if it's not important that the arity of the sub is correct, you can just build the block with a *@a parameter 13:08
SmokeMachine____ timotimo: the problem is that it's important on that case... 13:09
tony-o m: my $r = { say $_; }; $r(123);
camelia rakudo-moar 4b1864: OUTPUT«123␤»
timotimo hehe.
in that case, here's a trick for you
SmokeMachine____ I put that on a array and grep by the signature... 13:10
timotimo m: sub test(*@a) { say @a.elems } but role { method arity { 5 }; method count { 5 } }; (1..100).map(&test)
camelia rakudo-moar 4b1864: OUTPUT«1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤…»
timotimo er, oops?
oh, by the signature
masak m: my &foo = { say "OH HAI" }; foo
camelia rakudo-moar 4b1864: OUTPUT«OH HAI␤»
timotimo m: sub test(*@a) { say @a.elems } but role { method arity { 5 }; method count { 5 } }; &test.signature does role { method arity { 5 }; method count { 5 } }; (1..100).map(&test)
camelia rakudo-moar 4b1864: OUTPUT«1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤…»
timotimo :(
SmokeMachine____ Yes, by the signature...
:(
And if some one have any comment about the code... I'd love to hear! 13:12
Not only about the code... 13:13
timotimo: I do: @funcs.grep: {%values ~~ .signature} 13:15
timotimo i see, ok 13:16
well, you can still create the signature declaratively
installing it is possible to by mixing in a role that overrides the signature method 13:17
here be dragons :)
SmokeMachine____ timotimo: interesting! 13:17
SmokeMachine____ m: my $a = {say "ok"}; say {:bla} ~~ $a.signature; $a.signature = :(:$bla!); say {:bla} ~~ $a.signature 13:21
camelia rakudo-moar 4b1864: OUTPUT«False␤Cannot modify an immutable Signature␤ in block <unit> at <tmp> line 1␤␤»
SmokeMachine____ say {:bla} ~~ $a.signature
Sorry
mst timotimo: NuSQL?
timotimo yup
like Nu Metal
mst timotimo: o/ since you never did a join in the first place / maybe it's time to have the tables turned / in the interests of all involved, we've got the scaling solved / and the verdict is postgres o/ 13:22
timotimo i don't recognize those lyrics
SmokeMachine____ m: my $a = {say "ok"}; say {:bla} ~~ $a.signature; $a.signature does role {method signature{ :($bla!)}; say {:bla} ~~ $a.signature
camelia rakudo-moar 4b1864: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Missing block␤at <tmp>:1␤------> 3e{ :($bla!)}; say {:bla} ~~ $a.signature7⏏5<EOL>␤ expecting any of:␤ postfix␤ statement end␤ statement modifier␤ statement…»
timotimo oh, no, not like that
now you're installing a signature method on the signature you get from .signature on the block 13:23
mst timotimo: then you don't know nu metal worth shit, that's from Spit It Out by Slipknot :D
timotimo so you'd get that from $a.signature.signature
ah, sorry. i didn't slipknot at all during that time
mst slipknot, deftones, korn, systemofadown were basically -the- bands associated with the nu metal term first 13:24
timotimo i was an ignorant little kid back then :) 13:24
of those i only latched on to SoaD 13:25
mst google the spit it out lyrics and you'll find my filk fits pretty well 13:26
mst can rap out that entire song from memory in time to the track still
I can sort of do the same for a few SoaD bits but they generally have tunes so I don't do it as often for the safety of other people's ears 13:27
nine didn't know there was "nu metal" when listening to those bands 13:28
timotimo hehehe 13:30
RabidGravy things I didn't expect to be doing on a Monday morning: upgrading a "live" MySQL installation :-\ 13:33
timotimo oh wow
mullagainn is there plans for a p6 equivalent to pythons asyncore 13:45
timotimo i don't remember what asyncore is about 13:46
we do have asynchronous I/O, though
jnthn That looks much like that the built-in IO::Socket::Async is for 13:50
jnthn (if the first google hit is anything to go by... :)) 13:50
timotimo "Network servers are rarely processor bound, however." - unless your business logic is written in a slow language :P 13:51
timotimo in principle a module could be written that gives you the same API as asyncore 13:53
except we can't "handle_expt" because moarvm doesn't support OOB data for TCP sockets
mullagainn yeah, that is what I am seeing. Also looking at the doc I cant see anything related to raw sockets. 13:54
timotimo correct, we don't have raw socket support at the moment
of course, if you really need it, you can NativeCall into libc for that
mullagainn at all or just within the Async module?
timotimo actually ... i don't know
mst surely you'd be better NCing into libuv 13:55
timotimo we currently use libuv for sockets of all kinds (even synchronous) so maybe all it needs is to set a flag to some value and we'd get a raw socket instead of a tcp socket?
mst ... which I'm now sorely tempted to try at some point to see if I can get at unix sockets
timotimo :) 13:56
jnthn ...alternatively, just patch IO::Socket::Async given it's based on libuv anyway :)
timotimo right
i'm not sure what parts are needed to support unix domain sockets or raw sockets
jnthn The only reason it doesn't expose those things is nobody got around to it yet.
timotimo maybe it's super easy and doesn't even require code in moarvm
jnthn timotimo: Probably needs a bit
timotimo it'd likely require someone to write some failing tests first :) 13:57
mst jnthn: well, yes, that's why I was suggesting NCing into libuv
since hopefully I can manage to re-use at least -some- of the existing stuff that way
timotimo mst: do note that moarvm will not give you access to its _uv structs
you also don't get access to MVMThreadContext structs from user code
unless, you know, you dig through process memory for something that looks like it could be a TC 13:58
like, if you read out the stack pointer you can surely find the TC
mst and then NativeCall into MoarVM
it'll be fine
jnthn Is this your way to make me drop other stuff to do it properly? :P
.oO( It worked for Devel::Declare... :P )
Well, not on me, but :P 13:59
mst GRINS
the parallels were not lost on me
brrt mst: if you make sure it is JITted, you can access the threadcontext pointer from r14 14:01
brrt libuv is kind of aggressive about 'owning' sockets though 14:05
timotimo a bit, yeah
you'll probably want to create your own uv_loop struct and create the sockets on your own ;) 14:06
pmurias NativeCalling into MoarVM dependencies seems like inventing a more evil way to do XS ;) 14:07
timotimo we can have "use VISHNUS-MULTI-ARMED-FORM;" and then "nqp::becomedeathdestroyerofworlds()" to get the current MVMThreadContext
brrt that seems like it'd satisfy mst just fine 14:09
pmurias refuses to implement getting the current MVMThreadContext on the js backend ;)
timotimo :D 14:11
DrForr Oh, package names in kebab-case work? I thought I had a problme with doing that. 14:12
timotimo i think package names are CamelCase
DrForr (need to rewrite some things tonight then.) 14:13
pmurias could look into adding unix sockets support to the MoarVM if it means stopping people from poking into VM internals 14:14
mst argh, you're not in #perl6-dev 14:15
timotimo cool :)
mst which I just moved the conversation to as more appropriate
timotimo right
without really announcing it, though?
mst I *thought* everybody that was likely to care was already in there 14:16
had not noticed the lack of pmurias
timotimo :)
pmurias mst: joined #perl6-dev 14:27
mst \o/ 14:28
RabidGravy mst, I'd love you even more than I do now if you could do unix domain sockets :) 14:36
masak .oO( roses are seven, violets are twelve, my love is isomorphic to a subset of itself ) 14:47
timotimo :D 14:48
Ven`
.oO( Wherefore art thou Russel ? )
mst masak++
Jolust Is it correct to say that when @a».bar will be made auto-threaded that the compiler will figure out the best number of threads to use for the operation? 15:07
Or will it always use a specific number of threads
timotimo there's no guarantees about this :) 15:08
perlpilot Jolust: it's correct to say "*magic happens*" :)
Jolust perlpilot: but what does it mean? What does it mean to "auto-thread"? Why is it useful? 15:09
what sort of magic 15:10
jnthn It'll use whatever logic .hyper and .race use by default. At the moment that's a fixed default value, but it'll likely move to be based on the number of cores in the shorter term, and to do some kind of hill-climbing algo or similar in the longer term
brrt it means that the compiler has the liberty to schedule it to run on multiple threads
timotimo rakudo may change how it decides how to automatically allocate threads for your task
jnthn Quite possibly with some per-callsite stats
brrt doesn't mean that it does, but it can, and it makes sense depending on how much individual time each of the operations take
jnthn So I'd expect it'll, on average, get smarter with thime. :) 15:11
*time
Jolust Alright. Thanks.
timotimo thime makes you smarter? damn
timotimo raids kitchen drawers
perlpilot jnthn: but will it be manually tunable for those rare cases where you want to force a certain number of threads?
brrt i don't expect so perlpilot
Jolust perlpilot: you'd use a hyper/race for those cases, I'd think 15:12
timotimo you can use .race and .hyper instead of .>>
brrt (question wasn't directed at me, i know)
timotimo yeah, those take two arguments that let you tune exactly how things go down
jnthn perlpilot: I think the >>.foo syntax will just pick the defaults, and what others have said for configuring it.
timotimo well, not *exactly*, it's really just two knobs, but it lets you decide enough, IMO
you can potentially override the HYPER_* subs in your program scope? 15:13
though probably not
we might enable that, though
perlpilot Jolust, timotimo: I was thinking more along the lines of forcing some number of threads for all >>. operations. 15:14
in any case, my imaginary scenario is sated ;) 15:15
jnthn perlpilot: @a>>.foo is jsut short for @a.hyper.map(*.foo)
jnthn uh, .hypermap I guess 15:15
But anyway, it desugars, and you'd just call the thing it desugars to with options. 15:16
perlpilot jnthn: right ... in my imagination I'd already have a bunch of @a>>.foo and didn't want to rewrite them as .hypermap 15:17
timotimo someone might introduce a slang that gives us >>o(10, 10)o>> syntax, or something :P
perlpilot But ... whatever, it wasn't a *real* problem (yet) :-)
harmil_wk Damn, I keep trying to write "say for blah". It's my Perl 5 muscle memory... 15:45
timotimo sorry about that :(
harmil_wk It's good. I'll get used to it eventually. But I just need to learn to double-check my code before sending out advice. 15:46
timotimo oh, it was you on that mailing list 15:47
o/
Jolust m: put for "blah" 15:56
camelia rakudo-moar 4b1864: OUTPUT«␤»
timotimo oh, we probably want to add put to the list of things that will warn 15:57
Jolust m: print for "blahg" 15:58
camelia rakudo-moar 4b1864: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Unsupported use of bare "print"; in Perl 6 please use .print if you meant $_, or use an explicit invocant or argument, or use &print to refer to the function as a noun␤at <tmp>:1␤------> 3print7⏏5 f…»
Jolust aye
timotimo if you want to, you can send that as [email@hidden.address] or what the address is
Jolust I may just fix it :) 15:59
timotimo that'd be cool, too
[Coke] ignores backscroll. Hallo, everyone. 16:04
moritz \o [Coke]
timotimo o/
dalek osystem: 31eee22 | (Lee Johnson)++ | META.list:
Add Geo::IP2Location::Lite to ecosystem

see github.com/leejo/geo-ip2location-lite-p6		 16:10
moritz leejo++ 16:14
timotimo cool 16:15
but the reverse, you know, only works if you write a visual basic gui to trace an ip.
Jolust is having second thoughts about `put for "blah"` 16:24
It's not Perl 5, so throwing X::Obsolete is a bit iffy, and it also disables put; with X::Comp::Group throwing 16:25
timotimo oh, we have a list of things that you could just add it to
it'll do the error throwing properly for you
it's inside either the Grammar.nqp or Actions.nqp 16:26
Jolust Yeah, that's what I'm talking about.
timotimo i don't know if put; is supposed to work, though?
Jolust This one: github.com/rakudo/rakudo/blob/nom/....nqp#L3224
Jolust Yeah, prints nl 16:26
timotimo oh, you mean because it's not "from perl5", it's fine to not explode? 16:27
Jolust s: &put, \()
SourceBaby Jolust, Sauce is at github.com/rakudo/rakudo/blob/4b18...ors.pm#L41
timotimo OK
Jolust timotimo: more like it's not "obsolete" so throwing X::Obsolete doesn't make sense entirely.
timotimo yeah
Jolust And there's an option to treat it as perl6-ism, but then the error message doesn't include the "use .put to mean $_" 16:28
m: bag for 1
camelia rakudo-moar 4b1864: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Function "bag" may not be called without arguments (please use () or whitespace to denote arguments, or &bag to refer to the function as a noun)␤at <tmp>:1␤------> 3bag7⏏5 for 1␤»
Jolust ^ you get that
timotimo oh
Jolust I'll just leave it as is. :)
TimToady we could add .bag to the message 16:31
timotimo .bag it and .ship it 16:32
Jolust m: "blah".tclc 16:33
camelia ( no output )
Jolust m: "blah".wordcase
camelia ( no output )
Jolust OK, I'll do that then
[Coke] anyone want to try to help the user who asked about pack on Jul 29 on perl6-users? 17:14
Jolust m: CATCH {default {}}; Perl ~~ Cool… take note; 17:20
camelia rakudo-moar 77a7a4: OUTPUT«Noted␤»
Jolust ^_^
timotimo m) 17:21
Jolust m: say WHICH new Perl: is => Cool: 17:32
camelia rakudo-moar 77a7a4: OUTPUT«Perl|58183408␤»
Jolust downloads it
poohman hello all, i have one question in Grammars, what will happen if I have multiple "makes" in a action method for a token?? 17:40
will token.made have the last make value??
smls poohman: As far as I know, yes. 17:44
harmil_wk poohman: say grammar { rule TOP {.} }.parse("a", :actions(class { method TOP($/) {make "first"; make "second";} } )).made.perl
m: say grammar { rule TOP {.} }.parse("a", :actions(class { method TOP($/) {make "first"; make "second";} } )).made.perl
camelia rakudo-moar 77a7a4: OUTPUT«"second"␤»
harmil_wk There you go
poohman thanks 17:45
dalek line-Perl5: cb1f745 | niner++ | / (2 files):
Fix possible segfault when returning a single coderef from P5

retval was uninitialized and used as a pointer for some random corruption Version 0.13		 18:06
lizmat gfldex tbrowder are there things to mention for the P6W regarding documentation ? 18:20
m: my %h; sub a($a) { %h{$a}++ }; a(1|2|3|4|5); dd %h # how will this fail when autothreading becomes really multi-threaded? 18:22
camelia rakudo-moar 77a7a4: OUTPUT«Hash %h = {"1" => 1, "2" => 1, "3" => 1, "4" => 1, "5" => 1}␤»
tushar I asked question yesterday about "grep-index" method for searching an index for a specific element of an array. Does that method exist? 18:48
timotimo m: say <a b c d e f g>.grep({ Bool.pick }, :k).say 18:49
camelia rakudo-moar 77a7a4: OUTPUT«(3 6)␤True␤»
timotimo oops, one .say too many 18:50
but this is what you want
smls m: say <a b c d e>.first("c", :k); # Or this
camelia rakudo-moar 77a7a4: OUTPUT«2␤»
timotimo well, .grep isn't the same as .first
and grep-index is more like grep with :k than first with :k 18:51
smls timotimo: But they said "index for a specific element"
timotimo oh
i skipped everything after "grep-index" ;)
or maybe i read up to "searching an index" 18:52
tushar timotimo and smls, thanks for your input.
masak tushar: .grep-index used to exist, but got folded into .grep via the :k option
it's nicely consistent with :k also working on .[] and .{} 18:53
tushar I am in love with chain operations of Perl 6.. You can many stuff in one code line.. At the same time it's very readable..
do*
timotimo masak: and also :kv and such :) 18:54
masak m: my @a = "A".."Z"; say @a[14, 7, 0, 8]; say @a[14, 6, 0, 8] :k
camelia rakudo-moar 77a7a4: OUTPUT«(O H A I)␤(14 6 0 8)␤»
masak tushar: it's only readable because you're weird, like we are :P 18:55
*and now it's too late to turn back*
mwhahaha
TimToady thinks everyone is weird 18:55
optikalmouse ^ 1000x
masak weird is already here, it's just unevenly distributed
tushar masak: hahaha.. may be.. That's why I never lost my faith from Perl.. And never moved to Python.. ;) 18:56
masak tushar: the Python people are far too normal, agree
pathologically normal
TimToady yeah, it's really weird
tushar masak: hahha..
timotimo i was never a perl person. i came straight from python to perl6 18:56
masak timotimo: it's because we're weird but classy 18:57
tushar timotimo: great to hear that
masak: true..
timotimo masak: that, we are *adjusts monocle*
masak adjusts the sea lion on his shoulder
timotimo :D 18:58
raydiak wonders what the dress code is in this sort of establishment...
optikalmouse my only issue with perl6 so far is that there seems to be a lag in the startup time?
I should time it but is that something usually happens or is it a problem specific to my crappy laptop?
timotimo raydiak: only dress in a way that makes you comfortable and doesn't make others uncomfortable on purpose? :)
masak optikalmouse: it probably gets exacerbated with a crappy laptop
timotimo optikalmouse: it starts in 0.1s
masak optikalmouse: but yes, it's there. it's not nearly as fast as Perl 5
optikalmouse k will time when I get home 18:59
Xliff_ \o #perl6
Still having issues with gist.github.com/Xliff/0f537c12b415...9484a25b1a
timotimo starting it multiple times in a row to make sure caches are warm is important
Xliff_ At this point, I do not know what to do about that. I don't even know what that error message means.
lizmat and another Perl 6 Weekly hits the Net: p6weekly.wordpress.com/2016/09/19/...he-robots/ 19:30
harmil_wk Saw the note about unicode radixes and so immediately tried: 19:39
ugexe link for Algorithm::LBFGS is missing the username in the url
harmil_wk m: say :⒗<a>
camelia rakudo-moar 77a7a4: OUTPUT«5===SORRY!5===␤Argument to "say" seems to be malformed␤at <tmp>:1␤------> 3say7⏏5 :⒗<a>␤Confused␤at <tmp>:1␤------> 3say :7⏏5⒗<a>␤ expecting any of:␤ colon pair␤Other potential difficulties:␤ Unsupported…»
harmil_wk So, only Unicode characters that represent decimal digits in radixes? 19:41
optikalmouse this looks like a cool project to attempt in perl6: github.com/semantic-release/releas...-generator 19:44
lizmat: awesome :o 19:45
Ivanushka harmil_wk: I see no reason why that shouldn't work. Would you please rakudobug it and maybe ping MasterDuke about it, since they might be more familiar with that code path 19:46
lizmat ugexe: fixed and thanks!
Ivanushka lizmat++ good weekly 19:47
Ivanushka optikalmouse: yup. You should try to make it :) 19:49
optikalmouse welcome to javascript -_- github.com/conventional-changelog that one tool imports this one but that one imports other deps :/ 19:50
bet I can use perl6 to whip something that's smaller + better.
Ivanushka Do it. Then add it to modules.perl6.org
optikalmouse++
robertle hi folks, I am trying to understand how perl6 finds stuff that i "use". is my undertsanding correct that a .pm6 that gets properly installed (CompUnit::Repository::Installation), all contents are inspected and indexed, so it could contain two classes with different names and I could "use" them and they would be found. whereas if the .pm6 is "only" in a CompUnit::Repository::FileSystem, then the path and 19:56
filename are used to look up what I get? so if the contents of the file do not match the filename, then I can't "use" it? seems at odds...
timotimo robertle: installed modules require a meta.json that has a "provides" section 20:25
timotimo the CU::R::FileSystem stuff is for when you use something like -Ilib or "use lib './foo/'" 20:25
(usually you'd just spell that "use lib 'foo'", but i thought i added path markers for clarity) 20:26
robertle timotimo: right, I got that. what I thought was surprising that a .pm6 that gets put into ::Installation has "provides" for anything in the file. e.g. 20:28
"provides" : { ... "SVG::Plot::AxisPosition" : { "lib/SVG/Plot.pm" : { ...} }, ... "SVG::Plot" : { "lib/SVG/Plot.pm" : { ...} ...
timotimo oh? 20:29
that's weird?
robertle so the file "lib/SVG/Plot.pm" provides both SVG::Plot::AxisPosition and SVG::Plot
I tried, I can "use" both
robertle if I put a similar .pm6, with two types in it, into a "lib" irectory, then I can only "use" the filename, not what's inside 20:30
ZoffixMobile Yeah, I noticed that with IRC::Client too 20:30
timotimo well, what you "use" refers to a Compilation Unit, which can be a class (you'd usually spell that "unit class") or a module or a grammar or whatever
robertle so as long as each .pm6 only contains a single type, and that type matches the filename/path the stwo behave the same
timotimo yeah, using stuff isn't supposed to be about types in files, it's about the compunit you want that has what you need in it
that's how i understand it anyway
robertle I am trying to do some sort of "pluggable" thing...
timotimo but this is absolutely not my area of expertise
timotimo ah 20:31
robertle yeah, compunit would make sense...
timotimo github.com/tony-o/perl6-pluggable - like this
robertle timotimo: I am working of just that, but it doesn't work (uses @INC and does not deal with CompUnit::Repository::Installation 20:32
trying to improve it
timotimo oh, damn
that's very good! :)
robertle I totally need Pluggable, I abuse it in perl5 so bad it's not funny anymore
tadzik :) 20:33
ZoffixMobile I think CURI lets you query installed stuff
robertle what's CURI? never heard of it... 20:34
timotimo yup
that's short for CompUnit::Repository::Installation
robertle ah, *that*!
ok, I did see the .installed() method, but can't get it to work and can't understand it from looking at the code either 20:35
timotimo i think you pass it a matcher or something? 20:36
nine no arguments
timotimo ah 20:39
nine oh, it's b0rked
robertle I tried that, but never get anything back. either Nil or "Method 'basename' not found for invocant of class 'Any'"
nine Classic Perl bug, too: method distribution($id) { InstalledDistribution.new(self!read-dist($_.basename), :prefix(self.prefix)) } 20:40
There's no $_ there
timotimo aaaah
robertle with that fixed Pluggable would et much shorter, and I would not have to do my own json file parsing, which is nice because I am only second-guessing the format of that file 20:41
but still need to search filesystem repo's, they don't have a "installed"
nine robertle: filesystem repos are also much simpler
robertle yeah, I do File::Find and that's it... 20:42
nine robertle: fix pushed 20:44
robertle awesome
I need to switch to a HEAD perl6... 20:45
timotimo github.com/timo/cellular-automata-tests - anybody want to look into the performance of this script?
nine Yeah, too bad it just missed a release :/ 20:49
timotimo i think i could turn the inner arrays into native int arrays actually 20:50
hm, but the check for the values isn't a simple min any more 20:51
hm, probably all(@neighbours) == any(...|...)
lizmat timotimo: in a hot loop, I wouldn't use given/when 20:52
at least not at this point in time
timotimo oh? i didn't realize 20:53
lizmat well, for one you're invoking smart matching
timotimo and given without when?
lizmat that I'm not so sure about
robertle I also have a syntax question, one that is kinda hard to google for: what's "$.method()" mean? I am confused about the "naked" dollar sigil... is it another way to say "self" ?
timotimo the only when i have is for SDL event, which ought to be very rare
unless you wiggle your mouse about a whole lot
lizmat robertle: yes 20:54
timotimo yes, it's a way to spell "self"
it's mostly used for accessors
lizmat $.foo is the same as self.foo
timotimo because you write "has $.foo", so you can access $.foo, too
(because perl6 generates an accessor method with the name of the attribute)
robertle great! looks more magic than it is
timotimo right
a profile of that script i posted shows, that enqueue is the most expensive routine 20:57
23.28% with the second place being the inner of the two dimensional loops with 12.93%
timotimo i might get away with running the enqueue thing on a separate thread 21:25
timotimo wow, holy hell, this is a biiiig bunch slower 21:49
timotimo it now calculates frames 2x faster than it renders 'em :\ 21:51
timotimo whoa. removing the whenever to just for-loop over the channel made it much, much, much faster 21:53
timotimo very interesting. when the calculation thread finishes (after frame 500), the render thread becomes a *lot* faster 21:57
timotimo huh, it can apparently deadlock?! 21:59
ugh, 14.5% time in run_gc 22:04
Xliff_ m: my @foo := CArray[int].new; @foo = Bool.roll(100_000) 22:35
camelia rakudo-moar 8be36b: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Undeclared name:␤ CArray used at line 1. Did you mean 'Array', 'array'?␤␤»
Xliff_ m: use NativeCall; my @foo := CArray[int].new; @foo = Bool.roll(100,000) 22:36
camelia rakudo-moar 8be36b: OUTPUT«Potential difficulties:␤ Leading 0 does not indicate octal in Perl 6. Please use 0o00 if you mean that.␤ at <tmp>:1␤ ------> 3Array[int].new; @foo = Bool.roll(100,0007⏏5)␤Cannot resolve caller roll(Bool: Int, Int); none of these sign…»
Xliff_ m: use NativeCall; my @foo := CArray[int].new; @foo = Bool.roll(100000)
camelia rakudo-moar 8be36b: OUTPUT«Cannot modify an immutable NativeCall::Types::CArray[int]␤ in block <unit> at <tmp> line 1␤␤»
Xliff_ s: use NativeCall; my @foo := CArray[int].new; @foo = Bool.roll(100000)
SourceBaby Xliff_, Ehhh... I'm too scared to run that code.
Xliff_ :P
timotimo Xliff_: apparently STORE isn't implemented for CArray yet? 22:38
i think i suggested to someone to implement that some time ago
m: use NativeCall; my @foo := CArray[int].new but role { method STORE(|) { say "hi!" } }; @foo = 1, 2, 3, 4 22:41
camelia rakudo-moar 8be36b: OUTPUT«CArray representation requires a typed array␤ in any compose_repr at gen/moar/m-Metamodel.nqp line 1572␤ in any compose at gen/moar/m-Metamodel.nqp line 3104␤ in any generate_mixin at gen/moar/m-Metamodel.nqp line 1346␤ in any at gen/moar/m-Me…»
timotimo er, oops?
m: use NativeCall; class StorableCIntArray is CArray[int] { method STORE(|) { say "hi!" } }; my @foo := StorableCIntArray.new; @foo = 1, 2, 3, 4 22:42
camelia rakudo-moar 8be36b: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Cannot resolve caller trait_mod:<is>(StorableCIntArray, NativeCall::Types::CArray, Array); none of these signatures match:␤ (Mu:U $child, Mu:U $parent)␤ (Mu:U $child, :$DEPRECATED!)␤ (Mu:U $ty…»
timotimo m: use NativeCall; class StorableCIntArray does CArray[int] { method STORE(|) { say "hi!" } }; my @foo := StorableCIntArray.new; @foo = 1, 2, 3, 4
camelia rakudo-moar 8be36b: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤NativeCall::Types::CArray[int] is not composable, so StorableCIntArray cannot compose it␤at <tmp>:1␤»
timotimo nope.
timotimo the way that is done makes it a bit interesting 22:44
Xliff_ timotimo: Yes. Me. And I've been looking into it, but I am still a bit sketchy on what I've implemented it. 22:45
Was reading blog and thought I could get SourceBaby to tell me if I was in the right place. :P
timotimo ah :)
Xliff_ Of course, I've also been more busy with trying to figure out gist.github.com/Xliff/0f537c12b415...9484a25b1a 22:46
However I haven't been able to get much help.
timotimo yeah, i have no idea what's wrong there :(
Xliff_ I'm at wits end over that.
timotimo have you already written a bit of code for CArray's STORE?
Xliff_ Yes. Nothing for public consumption. 22:47
Not even close.
timotimo oh
any questions i can help with?
Xliff_ Mainly because I am not sure if I should be linking into the typed roles, or into the class itself. 22:48
After several false starts with a role implementation (I used IntTypedCArray for starters) I think I might need to be doing it in the class. 22:49
I haven't had near enough time recently to give it the attention it deserves.
timotimo it's good enough to put it into the main, but putting special methods into the roles might be an optimization for the future 22:50
Xliff_ timotimo: I will get there. Just need a bit of patience. 22:50
timotimo OK
Xliff_ (and I need a can of Wooo Saaaaa)
timotimo: Ah! Well then that confirms I've been barking up the wrong codepath.
Alright. Have to check on dinner. Have a good night. 22:51
o7
timotimo not necessarily the wrong one
timotimo off to bed 23:08
skids anyone got any experience as to typical causes behind "zeroed thread in work pass" problems? 23:15
timotimo memory corruption, i think? 23:24
valgrind may help

Powered by Cro and the Raku® Programming Language.

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

Thank you!