»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.
raydiak r: say [1,2,3,4,5][4,3,2,1][1,2,3][2,1][1] 00:01
camelia rakudo-parrot 4a74e7, rakudo-jvm 4a74e7, rakudo-moar 4a74e7: OUTPUT«3␤»
timotimo i made a tiny bit of progress on little_big_int \o/ 00:20
PonderingDonkey can reduction operators be used at the end of a list? 00:28
PonderingDonkey so instead of: [+] 1..100 we have something like (1..10).[+] 00:29
smls p: say (1..10).reduce(&[+]); # you can do this 00:34
camelia rakudo-parrot 4a74e7: OUTPUT«55␤»
smls p: say (1..10).reduce(&infix:<+>); # long form 00:35
camelia rakudo-parrot 4a74e7: OUTPUT«55␤»
PonderingDonkey nice. thanks 00:51
TimToady_ p: say (1..10).reduce(*+*) 01:25
camelia rakudo-parrot 4a74e7: OUTPUT«55␤»
TimToady_ at bit shorter
but &[+] is probably faster
timotimo humans align to words anyway 01:45
clsn__ On my other computer, my \x=20 works but my \φ=20 doesn't and I can't figure out why. Must be something weird in the compiling. 02:01
timotimo libicu? 02:02
clsn__ I dunno. I'm not too worried about it, I guess, since it plainly isn't a bug in rakudo or anything. 02:07
clsn__ r: subset Even of Int where * % 2; my Even $e; $e=2; say $e.succ; 02:14
camelia rakudo-parrot 4a74e7, rakudo-jvm 4a74e7, rakudo-moar 4a74e7: OUTPUT«Type check failed in assignment to '$e'; expected 'Even' but got 'Int'␤ in block at /tmp/tmpfile:1␤␤»
clsn__ r: subset Even of Int where * % %2; my Even $e; $e=2; say $e.succ; 02:15
camelia rakudo-parrot 4a74e7: OUTPUT«use of uninitialized value of type Nil in numeric context in block at /tmp/tmpfile:1␤␤Divide by zero using infix:<%>␤ in method sink at gen/parrot/CORE.setting:12235␤ in method BUILDALL at gen/parrot/CORE.setting:918␤ in method bless at ge…» 02:15
..rakudo-moar 4a74e7: OUTPUT«use of uninitialized value of type Nil in numeric context␤Divide by zero using infix:<%>␤ in block at src/gen/m-CORE.setting:12218␤ in sub infix:<%> at src/gen/m-CORE.setting:4404␤ in sub infix:<%> at src/gen/m-CORE.setting:4012␤ in sub infi…»
..rakudo-jvm 4a74e7: OUTPUT«use of uninitialized value of type Nil in numeric context␤Divide by zero using infix:<%>␤ in method sink at gen/jvm/CORE.setting:12214␤ in method BUILDALL at gen/jvm/CORE.setting:923␤ in method bless at gen/jvm/CORE.setting:842␤ in method new …»
clsn__ argh, irc client eating my percent signs.
r: subset Even of Int where * %%2; my Even $e; $e=2; say $e.succ;
camelia rakudo-parrot 4a74e7, rakudo-jvm 4a74e7, rakudo-moar 4a74e7: OUTPUT«3␤»
clsn__ That's probably right... though I find myself wondering if "4" isn't a better answer. It's probably right as is.
clsn__ Hmm. But tricky to make a sub that will do the other thing. Have to run the parameter's refinement on the local var somehow. 02:25
clsn__ Yay. sub subsucc(\x) { my $n=x; $n=$n.succ; my $w=x.VAR.of.^refinement; $n=$n.succ while !$w($n); $n }. Have to pull the code into a var so it doesn't get an invocant. 02:28
colomon r: subset Even of Int where * %%2; my Even $e; $e=2; say $e.WHAT; say $e.succ.WHAT 03:10
camelia rakudo-parrot 4a74e7, rakudo-jvm 4a74e7, rakudo-moar 4a74e7: OUTPUT«(Int)␤(Int)␤»
colomon r: subset Even of Int where * %%2; my Even $e; $e=2; say $e.WHAT; $e =.succ; say $e.WHAT 03:10
camelia rakudo-parrot 4a74e7, rakudo-jvm 4a74e7, rakudo-moar 4a74e7: OUTPUT«(Int)␤(Even)␤»
colomon r: subset Even of Int where * %%2; my Even $e; $e=2; say $e.WHAT; $e =.succ; say $e.WHAT; say $e 03:11
camelia rakudo-parrot 4a74e7, rakudo-jvm 4a74e7, rakudo-moar 4a74e7: OUTPUT«(Int)␤(Even)␤(Even)␤»
colomon r: subset Even of Int where * %%2; my Even $e; $e=2; say $e.WHAT; $e = $e.succ; say $e.WHAT; say $e
camelia rakudo-parrot 4a74e7, rakudo-jvm 4a74e7, rakudo-moar 4a74e7: OUTPUT«(Int)␤Type check failed in assignment to '$e'; expected 'Even' but got 'Int'␤ in block at /tmp/tmpfile:1␤␤»
clsn__ Yeah, I've been experimenting etc... You can get the subset by looking at the .VAR.of of the variable. I was just tinkering with something like that on parameters, and then the sub would introspect... 03:13
It's kind of dumb though (the latter thing I'm playing with). Subsets in signatures get translated to parameters with constraints, which is okay, but if you know the subset when you're writing the signature you might as well just put the condition inside the code rather than have to introspect to find it. You shold know it already. 03:14
japhb__ .tell pmichaud Congratulations to @child[1], and good luck to @child[0]! 03:23
colomon clsn__: sorry, got distracted there. seems to me the problem is there's no reason the where clause should be simple. 03:27
gh, kebard j sped wrkng agan. :( 03:28
BenGoldberg .ping 03:37
No bot around listening for .tell 03:38
lue oh wow, even preflex is gone. 03:41
colomon clsn__: what I was trying to say... 03:42
r: subset SmallPrime of Int where { $_ < 100 && $_.is-prime }; my SmallPrime $sp = 2; $sp = $sp.succ; say $sp; $sp = sp.succ; 03:44
camelia rakudo-parrot 4a74e7, rakudo-jvm 4a74e7, rakudo-moar 4a74e7: OUTPUT«===SORRY!=== Error while compiling /tmp/tmpfile␤Undeclared routine:␤ sp used at line 1␤␤»
colomon r: subset SmallPrime of Int where { $_ < 100 && $_.is-prime }; my SmallPrime $sp = 2; $sp = $sp.succ; say $sp; $sp = $sp.succ;
camelia rakudo-parrot 4a74e7, rakudo-jvm 4a74e7, rakudo-moar 4a74e7: OUTPUT«3␤Type check failed in assignment to '$sp'; expected 'SmallPrime' but got 'Int'␤ in block at /tmp/tmpfile:1␤␤»
colomon wonders why he doesn't always use his external Mac keyboard, much nicer typing on it than the MBP's keyboard. 03:45
colomon afk # got to try to finish reading Republic of Thieves so I can start reading Steles of the Sky. 03:46
uvtc jnthn , thanks for posting your recent 2014 fosdem slides. :) Aside from the content itself, they were quite entertaining. Great images in there. 04:01
phenny, are you still around? 04:02
bots, roll call! 04:03
uvtc (jnthn, I meant to imply that the content was great and interesting on its own.) 04:06
uvtc Not sure exactly what would make for a suitable GSoC project, but it would be nice to have a Kate syntax description file for Perl 6, such that some more editors could highlight P6, as well as Pandoc. 04:37
japhb__ I added a number of additional GSoC project ideas just now. 04:39
uvtc: You should add yours to the existing doc!
uvtc Would the syntax definition file idea be suitable for a GSoC project though? Not sure of the specific GSoC goals I suppose. 04:40
[Coke] moritz: etoomuchdayjob 04:41
uvtc Hm. Seems like maybe not suitable. That is to say, you probably want coding projects which benefit the student a bit more... 04:42
Is there already a site with some of those nice svg plots showing Perl 6 impl performance data over time? 04:47
uvtc I thought there was one for Rust. Maybe <huonw.github.io/isrustfastyet/>. Could've sworn there was another one for some $other-lang... 04:49
diakopter js
arewefastyet.con
uvtc Ah. Thanks, diakopter . I think that was it. 04:50
diakopter you know, the language that's a few thousand times more used than rust :)
(also from mozilla)
uvtc hehehe. Yes, that obscure language slipped my mind. :)
Well, anyhow, I'd rather see a Perl 6 tutorial than a perf site. But though it might be worth mentioning as a potential idea for GSoC. 04:51
s/though/thought/
uvtc Oh. Right. There's perl6/bench-scripts. Wasn't someone running them at some point and graphing the data? Maybe [Coke] ? 04:57
japhb__ uvtc: I'm editing the ideas list now -- did you want me to put that in the list for you? 05:00
uvtc I don't remember if an "are we fast yet?" site was already done in the past, and then scrapped for whatever reason. 05:01
Also, I think I still have my commit bit, though it's not been sharpened in some time. :) 05:02
japhb__, Thanks for asking, regardless. :)
japhb__ uvtc: timotimo and I recently (and several others in the past) have asked for an arewefastyet or speed.pypy styled site, based on perl6-bench (which may need some minor improvements to support such a thing, and will definitely need more tests). 05:05
japhb__ I think it would be an excellent GSoC project. 05:05
japhb__ OK, I've committed my latest ideas. uvtc, you're up. :-) 05:07
lue "hasn't been implemented in any backend yet ... could implement this in the JVM or the MoarVM backend." well then :) 05:12
uvtc japhb__, Ok, added that GSoC idea. :) 05:15
All the bots are on vacation it seems. Sipping digital martinis on some luxury virtual cruise. 05:16
masak morning, #perl6 05:26
diakopter masak: g'morn 05:28
japhb__ uvtc: expanded. :-) 05:29
uvtc japhb__, Nice. 05:30
japhb__ raydiak: OK, that's a doozy of a pull request. It may take me a day or two to get that reviewed.
PonderingDonkey why is this producing different results: 05:53
r: my %hash = (3=>"C",1=>"A",2=>"Z") ; .say for %hash.sort
camelia rakudo-parrot 4a74e7, rakudo-jvm 4a74e7, rakudo-moar 4a74e7: OUTPUT«"1" => "A"␤"2" => "Z"␤"3" => "C"␤»
PonderingDonkey r: my %hash = (3=>"C",1=>"A",2=>"Z") ; %hash.sort>>.say
camelia rakudo-parrot 4a74e7, rakudo-jvm 4a74e7, rakudo-moar 4a74e7: OUTPUT«"3" => "C"␤"1" => "A"␤"2" => "Z"␤»
PonderingDonkey does >> still operating on the "unsorted" list 05:54
simarly 05:56
r: my @=(5,3,4); @a.sort.say
camelia rakudo-parrot 4a74e7, rakudo-jvm 4a74e7, rakudo-moar 4a74e7: OUTPUT«===SORRY!=== Error while compiling /tmp/tmpfile␤Variable '@a' is not declared␤at /tmp/tmpfile:1␤------> my @=(5,3,4); @a.sort.say⏏<EOL>␤ expecting any of:␤ …»
PonderingDonkey r: my @a=(5,3,4); @a.sort.say 05:57
camelia rakudo-parrot 4a74e7, rakudo-jvm 4a74e7, rakudo-moar 4a74e7: OUTPUT«3 4 5␤»
PonderingDonkey r: my @a=(5,3,4); @a.sort>>.say
camelia rakudo-parrot 4a74e7, rakudo-jvm 4a74e7, rakudo-moar 4a74e7: OUTPUT«5␤3␤4␤»
PonderingDonkey concurrency?
different results in term of ordering 05:59
TimToady_ correct, you may rely only on the ordering of values returned--you may not rely on the order of side effects 06:05
and IO is a side effect
PonderingDonkey thanks 06:10
masak "First off, you're in for a lot of work…years of work…most of which will be wandering in the desert." -- www.drdobbs.com/architecture-and-de.../240165488 06:15
TimToady if you ain't Moses, don't start :)
masak .oO( let's not split hairs^Wseas here ) 06:19
lue masak: "The first tool that beginning compiler writers often reach for is regex. Regex is just the wrong tool for lexing and parsing" obviously never seen Perl 6 :P 06:21
masak lue: well, there *is* a distinction between "regex" and "parser" ;) 06:23
I had a talk about that, but it seems I didn't put it online... :/ 06:24
well, not as a PDF, at least. 06:26
lue: see here: github.com/masak/talks/blob/master...ges/slides
lue masak: to be fair, I've only ever seriously used regexes in Perl 6, so that may affect my personal connotations of what regexes are :)
masak hehe. 06:27
lue: I have a prototypical anti-example for you: daringfireball.net/projects/markdown/ 06:28
lue: go ahead and have a peek at the Perl 5 code that drives that parser/emitter.
lue To the point where one of my "eventually" projects is a libp6regex for C++. I just like P6 regexes (and grammars) too much I guess :) 06:28
masak that's exactly how *not* to use regexes for parsing.
tadzik good morning!
masak tadzik! \o/
lue: don't get me wrong -- that .pl file is very successful, and heavily used. I use it for my blog. but it's not the way to parse a language. Perl 6 grammars would have been much better. 06:29
masak lue: this is a much better way to parse it: github.com/gildor478/ocaml-markdow...arkdown.ml 06:30
moritz \o 06:40
tadzik hello
tadzik moar, moar ideas to GSoC page! 06:40
is there a plan to start as a Rakudo org? 06:41
masak lue: I maded you a talk PDF: masak.org/carl/lpw-2011-small-languages/talk.pdf :) 06:51
there's still a lot of context missing -- things I said out loud that are not in the slides -- but I think the basic idea of "don't use regexes, write real parsers" comes through. 06:52
masak guten morgen, moritz 06:59
masak is on a train 07:40
diakopter is not on a train 07:41
masak the train got stuck after another (slower) train, and the service manager on board was being very passive-aggressive about it on the PA system.
saying things like "apparently they don't feel like letting us past" and "we called traffic control, and they were spectacularly unhelpful about the whole thing"
diakopter "maybe we should get out and push" 07:42
raydiak japhb__: yeah, sorry if I made work for you; hadn't seen you chat for a while, and didn't want to keep bugging with questions, so I just kinda went with it; there are rough spots and plenty to discuss at some point, but I figured doing it would be better than not :) 08:36
ribasushi what the hell... 08:47
can someone see what was my quit message?
no idea how I fell off...
raydiak last one I see for you is a split 08:48
it's been raining DDoSes on freenode lately
ribasushi ah, that may have been why
Ven r: say [1, 2].WHICH 08:56
camelia rakudo-moar 4a74e7: OUTPUT«Array|139746001520328␤»
..rakudo-jvm 4a74e7: OUTPUT«Array|1668028044␤»
..rakudo-parrot 4a74e7: OUTPUT«Array|7540192417656216563␤»
diakopter p: say 7540192417656216563.is-prime 08:57
camelia rakudo-parrot 4a74e7: OUTPUT«False␤»
moritz masak++ # slides 08:59
woolfy Wow, Damian Conway does a course on Perl 6 in Switzerland on April 9: it.oetiker.ch/consulting/courses/2014.en.html 09:07
tadzik nice!
Ven Is there a feature perl6 doesn't have you wish it had ? ;D 09:15
moritz does "speed" count as a feature?
mature IO system 09:16
Ven moritz: "speed" is implementation-dependant 09:18
Timbus i was going to answer but moritz cut me off :<
ill add 'lower memory requirements' then 09:19
Ven then again, that doesn't seem to be specifically tied to the language
Timbus or language features eh
Timbus oh* 09:19
tadzik IO, IO 09:20
Ven yeah, else it's not specifically related to the language.
moritz and honestly, I'm not entirely happy with our current object initialization system
Ven moritz++ # I can agree some parts seem tricky for not much reason 09:21
p: say +^1 09:22
camelia rakudo-parrot 4a74e7: OUTPUT«-2␤»
Timbus uhm... i would have liked less strange coroutines i guess, but the current concurrency (woah) spec really fills in that hole
Ven Timbus++ # I'd like to see different keywords, need to finish reading the syn about concurrency 09:26
tadzik looks like twitter is pretending to be iOS7 now 10:27
FROGGS yoleaux: \o/ 10:27
tadzik fsvo of leaux :P 10:28
FROGGS hehe 10:29
tadzik it's a funny world we live in, where I say "fsvo of leaux" and it's perfectly understandable in some circles :) 10:30
timotimo o/ 10:34
tadzik \o 10:47
timotimo added a peephole optimizer to the GSoC ideas 10:58
FROGGS O.o
timotimo the last time i looked, i didn't see very obvious optimization opportunities, but i was looking at the output of some nqp or rakudo source file 11:03
it'd be better to look at some custom code instead
tadzik . o O ( Python for MoarVM ) 11:07
well, domain-specific hilighting is on my "somebody do this" list, but hardly Perl6 related 11:09
timotimo what does that imply? 11:10
tadzik well, I'm not particularly fond of syntax hilighting. But it'd we awesome to have vim hilighting based on profiling, or test coverage
so you don't need an additional viewer for those
timotimo i have a key binding + function that highlights lines based on their indentation level with grayscale 11:30
also, vim has support for icons to be placed on the left side in the gutter 11:31
Ven r: 0, 1, *+* ... * > 100 12:48
camelia ( no output )
Ven r: say 0, 1, *+* ... * > 100 12:49
camelia rakudo-parrot 4a74e7, rakudo-jvm 4a74e7, rakudo-moar 4a74e7: OUTPUT«0 1 1 2 3 5 8 13 21 34 55 89 144␤»
Ven not sure I understand that. I thought `$^a ... $^b > $^c` means "execute $^a for $^b..$^c" 12:50
timotimo er .. wha? 12:51
Ven *s03*
timotimo how did you get to $^a from that?
moritz Ven: the series operator is a bit more involved than that 12:52
r: say 1, 2, 4 ... * > 1000
camelia rakudo-parrot 4a74e7, rakudo-jvm 4a74e7, rakudo-moar 4a74e7: OUTPUT«1 2 4 8 16 32 64 128 256 512 1024␤»
Ven it can ... "guess" the difference ? 12:53
moritz yes
m: say 1, 3, 5 ... 11
camelia rakudo-moar 4a74e7: OUTPUT«1 3 5 7 9 11␤»
Ven m: say 1, 5 ... 100 12:54
camelia rakudo-moar 4a74e7: OUTPUT«(timeout)»
Ven m: say 1, 5 ... 30
camelia rakudo-moar 4a74e7: OUTPUT«(timeout)» 12:55
timotimo m: say (1, 5 ... 100)[^20]
camelia rakudo-moar 4a74e7: OUTPUT«1 5 9 13 17 21 25 29 33 37 41 45 49 53 57 61 65 69 73 77␤»
timotimo it never hits 100, or 30
Ven 1, 5 ... > 100 12:55
timotimo needs a * to create a closure 12:55
Ven m: 1, 5 ... * > 100
camelia ( no output )
Ven m: say 1, 5 ... * > 100 12:56
camelia rakudo-moar 4a74e7: OUTPUT«1 5 9 13 17 21 25 29 33 37 41 45 49 53 57 61 65 69 73 77 81 85 89 93 97 101␤»
Ven Damn I'm used to my REPL
m: say 1, 5 ...^ * > 100
timotimo :)
camelia rakudo-moar 4a74e7: OUTPUT«1 5 9 13 17 21 25 29 33 37 41 45 49 53 57 61 65 69 73 77 81 85 89 93 97␤»
Ven timotimo++ # ^ is used to discard the first non-matching value
m: 1, 2, 5 ... * > 30
camelia rakudo-moar 4a74e7: OUTPUT«Unhandled exception: Unable to deduce sequence␤ at src/gen/m-CORE.setting:10811 (/home/p6eval/rakudo-inst-2/languages/perl6/runtime/CORE.setting.moarvm:throw:129)␤ from src/gen/m-CORE.setting:12212 (/home/p6eval/rakudo-inst-2/languages/perl6/runtim…»
timotimo the first matching value, actually 12:57
Ven uh, yeah.
did not negate the correct part of the sentence 12:58
Ven m: [1] + (0, 5 ... 100)[1..*] 12:59
camelia ( no output )
Ven m: say [1] + (0, 5 ... 100)[1..*]
camelia rakudo-moar 4a74e7: OUTPUT«21␤»
Ven m: say [1] Z (0, 5 ... 100)[1..*] 13:00
camelia rakudo-moar 4a74e7: OUTPUT«1 5␤»
Ven m: say (0, 5 ... 100)[1..*].perl
camelia rakudo-moar 4a74e7: OUTPUT«(5, 10, 15, 20, 25, 30, 35, 40, 45, 50, 55, 60, 65, 70, 75, 80, 85, 90, 95, 100)␤»
Ven m: say [1].append (0, 5 ... 100)[1..*] 13:01
camelia rakudo-moar 4a74e7: OUTPUT«===SORRY!=== Error while compiling /tmp/3JHA2k_Cxx␤Two terms in a row␤at /tmp/3JHA2k_Cxx:1␤------> say [1].append ⏏(0, 5 ... 100)[1..*]␤ expecting any of:␤ method arguments␤ postfix␤ …»
timotimo you want the , operator
r: say (1, (0, 5 ... 100))
camelia rakudo-parrot 4a74e7, rakudo-jvm 4a74e7, rakudo-moar 4a74e7: OUTPUT«1 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100␤»
timotimo r: say (1, (0, 5 ... 100)).perl 13:02
camelia rakudo-parrot 4a74e7, rakudo-jvm 4a74e7, rakudo-moar 4a74e7: OUTPUT«(1, (0, 5, 10, 15, 20, 25, 30, 35, 40, 45, 50, 55, 60, 65, 70, 75, 80, 85, 90, 95, 100).list)␤»
timotimo r: say (1, (0, 5 ... 100)).list.perl
camelia rakudo-parrot 4a74e7, rakudo-jvm 4a74e7, rakudo-moar 4a74e7: OUTPUT«(1, 0, 5, 10, 15, 20, 25, 30, 35, 40, 45, 50, 55, 60, 65, 70, 75, 80, 85, 90, 95, 100).list␤»
Ven that's nested, isn't it ?
timotimo you don't need .list if you assign it to an @ variable or iterate over it
because those operations impose a flattening context
for (1, (0, 5 ... 20)) { say "v: $_" }
m: for (1, (0, 5 ... 20)) { say "v: $_" }
camelia rakudo-moar 4a74e7: OUTPUT«v: 1␤v: 0␤v: 5␤v: 10␤v: 15␤v: 20␤»
Ven Uhm ... I'm not sure I like automatic flattening 13:03
timotimo it's just that the , operator in this case doesn't impose any sort of context and you'll have to decide for yourself 13:04
.tree and .lol and .item are your friend
(not your food)
Ven . 13:05
m: (1,).^name.say; [1].^name.say;
camelia rakudo-moar 4a74e7: OUTPUT«Parcel␤Array␤»
Ven I don't get the need for both "lists" and "arrays" 13:06
timotimo the big difference is containers
i constantly confuse which one is which, though :P
m: my $foo = [1, 2, 3]; my $bar := $foo[1]; $bar = 100; say $foo
camelia rakudo-moar 4a74e7: OUTPUT«1 100 3␤»
timotimo m: my $foo = (1, 2, 3).list; my $bar := $foo[1]; $bar = 100; say $foo 13:07
camelia rakudo-moar 4a74e7: OUTPUT«No such method 'STORE' for invocant of type 'Int'␤ in block at /tmp/LTZ8bhLV69:1␤␤»
timotimo as you can see, the list does not have a container we can use for binding
timotimo does that help you any? 13:09
Ven timotimo: Yeah, that does. I really dislike the idea that everything is a ref, though.
moritz native types aren't refs
Ven yeah but that's an Int here, right 13:10
or is that because of := ?
timotimo no, it does create an Int
tadzik I think you'll find that it doesn't bother you as much as you think it would :)
timotimo unless you specifically ask for an int, you get an Int
Ven well, the thing I really dislike is that, if 2 different lists share an element at pos. 0, if you modify $list[0], $list2 would be different too 13:12
timotimo you only get that if you :=, though
tadzik r: my @a = (1, 2, 3); my @b = @a; @b[0] = 5; say @a.perl 13:13
camelia rakudo-parrot 4a74e7, rakudo-jvm 4a74e7, rakudo-moar 4a74e7: OUTPUT«Array.new(1, 2, 3)␤»
tadzik r: my @a = (1, 2, 3); my @b = @a; @b[0]++; say @a.perl
camelia rakudo-parrot 4a74e7, rakudo-jvm 4a74e7, rakudo-moar 4a74e7: OUTPUT«Array.new(1, 2, 3)␤»
Ven m: my $val := 1; $l = [$val, 2, 3]; my $l2 = [$val, 5, 6]; $l[0] = 100; say $l2.perl
camelia rakudo-moar 4a74e7: OUTPUT«===SORRY!=== Error while compiling /tmp/m39ZzIq_f6␤Variable '$l' is not declared␤at /tmp/m39ZzIq_f6:1␤------> my $val := 1; $l⏏ = [$val, 2, 3]; my $l2 = [$val, 5, 6]; ␤ expecting any of:␤ postfix␤»
tadzik now you're forcing it
Ven m: my $val := 1; my $l = [$val, 2, 3]; my $l2 = [$val, 5, 6]; $l[0] = 100; say $l2.perl 13:13
camelia rakudo-moar 4a74e7: OUTPUT«[1, 5, 6]␤»
tadzik why would you do binding in the first place? 13:14
Ven better keep away from := then
timotimo the assignment you did to $l and $l2 copied the target of the container into a new container
(that's what arrays do)
you have to ask for binding much more explicitly for it to happen to you 13:15
and when it happens, you will very rarely be surprised
Ven I don't get the distinction between list and array nor the need for it in the first place. It seems weird to have several types of sets
timotimo at least you're not asking for WTH there is a Parcel on top of that as well :P 13:16
Ven I'll add that question then
timotimo haha :D
well, one thing i have to say is that there's a rewrite of the synopsis for lists and friends coming up 13:17
that's the reason why Nil still disappears in flattening context on rakudo; we're waiting until all changes to the list synopsis are done before changing big things around in the implementation
moritz ... assuming that such a mythical condition is every fulfilled 13:18
Ven But is there a "better" container ? Why does list require "@" while arrays dont ?
timotimo list does not require @ 13:19
moritz Ven: the sigils are about what flattens
so you can put anything into a $ that can be put into a @
it just won't flatten when you iterate over it
Ven okay -- need some paracetamol ...
timotimo r: my %foo := (1, 2, 3).list; say %foo # what the hell are you doing?!
camelia rakudo-moar 4a74e7: OUTPUT«Type check failed in binding; expected 'List' but got 'Associative'␤ in any bind_error at src/vm/moar/Perl6/Ops.nqp:216␤ in block at /tmp/tmpfile:1␤␤»
..rakudo-parrot 4a74e7, rakudo-jvm 4a74e7: OUTPUT«Type check failed in binding; expected 'Associative' but got 'List'␤ in block at /tmp/tmpfile:1␤␤»
moritz doc.perl6.org/language/containers
timotimo haha 13:20
wat :D
moritz that's my attempt to provide a cohesive picture
timotimo that's funny
moritz it's also sad :(
timotimo no, i meant the output of camelia there
moritz yes
timotimo i have an idea why it might happen, though
Ven OKAY - so @ means that everything will be flattened. At least for lists. 13:21
Ven m: my $a = (1, (2, (3, 4))); for $a { .perl.say } 13:21
camelia rakudo-moar 4a74e7: OUTPUT«$(1, (2, (3, 4)))␤»
Ven .
moritz m: my @a = (1, (2, (3, 4))); say .perl for @a 13:22
camelia rakudo-moar 4a74e7: OUTPUT«1␤2␤3␤4␤»
Ven yeah, cause @ is flattening. I understand that now. But what if I *don't* want to flatten ?
moritz then you could use .tree, for example
r: say (1, (2, (3, 4))).tree.perl 13:23
camelia rakudo-parrot 4a74e7, rakudo-jvm 4a74e7, rakudo-moar 4a74e7: OUTPUT«LoL.new(1, $(2, (3, 4)))␤»
moritz that $(...) will prevent the list from flattening when you iterate over it
timotimo there we go.
Ven okay. So that's just on one level, right ?
moritz m: say .perl for (1, (2, (3, 4))).tree
camelia rakudo-moar 4a74e7: OUTPUT«1␤$(2, (3, 4))␤»
timotimo oh, i need to create a meme picture with a cat now.
moritz Ven: right
Ven if I want a structure to act as I nested it, is there an alternative to .tree that is recursive ? 13:24
timotimo takelessons.com/blog/wp-content/upl...py-cat.gif "invisible rakudo commit"
or rather "invizibl rakduo comit"
[Coke] moritz: mark keating just sent out a message about TPF submitting as a member org for GSOC 13:25
moritz [Coke]: thanks, I've seen it
Ven .perl.say for [1, [2, [3, 4]]];
m: .perl.say for [1, [2, [3, 4]]];
camelia rakudo-moar 4a74e7: OUTPUT«[1, [2, [3, 4]]]␤»
FROGGS [Coke]: and we are welcome to add our P6 tasks
Ven m: .perl.say for [1, [2, [3, 4]]].tree; 13:26
camelia rakudo-moar 4a74e7: OUTPUT«1␤[2, [3, 4]]␤»
Ven m: [1, [2, [3, 4]]].tree.perl.say;
camelia rakudo-moar 4a74e7: OUTPUT«(1, [2, [3, 4]]).list␤»
[Coke] I imagine we should decide if we want to try going our own way.
timotimo yeah, tree listifies the outermost thingie and keeps all other things deeply-nested
[Coke] but if all we have to do is come up with ideas and let mark k. be the admin, that sounds vaguely winny.
moritz [Coke]: it does, if TPF actually gets in 13:27
timotimo m: [1, [2, [3, 4]]].lol.perl.say;
camelia rakudo-moar 4a74e7: OUTPUT«LoL.new(1, [2, [3, 4]])␤»
timotimo wow, that took a while.
moritz does anybody remember on which host dalek usually runs, and as which user? 13:28
timotimo shouldn't your irclog tell you?
timotimo from the joins/parts/quits? 13:28
moritz well, not the ones on the web 13:29
those don't track anything but nicks
FROGGS my brane says host07, but do not trust my brane
moritz and my irssi doesn't log joins and leaves 13:30
timotimo fixed the order of got/expected
Ven timotimo: I probably need to write some perl to get the gist of it. At first sight it really seems confusing and unnatural
timotimo it confused me a whole lot at the beginning, too 13:31
Ven thanks for bearing with me, timotimo++ && moritz+=
moritz timotimo: shouldn't it simply throw the right exception type, and then methodd message takes care of
FROGGS FreeNode-#perl6.log:Sep 30 12:54:54 *dalek (~dalekbot@feather3.perl6.nl) hat #perl6 betreten
moritz: ^^
moritz FROGGS: do you have anything from this year too? :-)
FROGGS hold on 13:32
timotimo moritz: it does throw the typed exception
but it goes via the registration mechanism
Ven Looking forward to a possible rewrite about LoLs etc
timotimo and *that* registers a sub with *positional* arguments
FROGGS irc.perl.org-#parrot.log:Feb 01 13:12:34 *d4l3k_ (~dalekbot@feather3.perl6.nl) hat #parrot betreten
irc.perl.org-#parrot.log:Feb 01 13:18:23 *d4l3k_ (~dalekbot@feather3.perl6.nl) hat #parrot betreten
moritz FROGGS: thanks
FROGGS yw
Ulti Ven what languages do you know other than Perl6? 13:35
Ven Ulti: asm, c, c++, c#, java, scala, tcl, ruby, js, php; for the most part 13:36
Ulti right so how do you deal with int[] Vector<Integer> ArrayList<Integer> LinkedList<Integer> in Java? 13:37
Ven yea
Ulti int[] is super different to all the others 13:38
Ven I think arrays in java are a mistake 13:38
Ulti but even C right, you have a linked list of structs or you can have an array of them 13:40
Ven but yes, every container has different implementations. They however have the same API and don't have a backed-in syntax
Ulti and by array I dont necessarily mean something statically used and on the stack, you can malloc enough room for N structs in a row, or you can wrap a struct in some other struct with a null pointer and some pointers for previous and next in the list etc. 13:41
Ven yes, you have linked lists, double linked lists, etc, but the only one who gets its own syntax is the array
Ulti Perl6 just has a different set of oddball things which are more homogenous in syntax
Ven you don't a[1] on double linked list, for example
Ulti which is both confusing, but most of the time you dont notice because the syntax works for you rather than against you 13:42
Ven you could if you could overload the [] operator ;)
Ven not in C
Ulti sure
Ven and then again, that's library-wise, not a special *syntax* 13:43
there's not <[]> to create a linked list and >[]< to create a vector
I thought I'd look at a simple example like conway's game of life. It seems to use "Array is dim (5, 5)" 13:44
m: my @life = Array of bit is dim (20, 20) is default(0); @life.perl.say 13:46
camelia rakudo-moar 4a74e7: OUTPUT«===SORRY!=== Error while compiling /tmp/PuLNQTZxcQ␤Two terms in a row␤at /tmp/PuLNQTZxcQ:1␤------> my @life = Array ⏏of bit is dim (20, 20) is default(0); @l␤ expecting any of:␤ postfix␤ i…»
Ulti also the word list and array is fairly well defined in general, list implies something containers/pointers-in-a-line like rather than the actual bits of data being in a line... usually to get around wanting different types of things in a line for the most part
parcels are what get me in Perl6
Ven m: my Array of bit @life is dim (20, 20) is default(0); @life.perl.say
camelia rakudo-moar 4a74e7: OUTPUT«===SORRY!===␤Type 'Array' is not declared␤at /tmp/Y7zHUY2WfL:1␤------> my Array of ⏏bit @life is dim (20, 20) is default(0);␤Malformed my␤at /tmp/Y7zHUY2WfL:1␤------> my Array of ⏏bit @lif…»
Ven wut 13:47
moritz the error message is wrong 13:48
it probably actually stubmles over 'bit' not being a known type
m: bit
camelia rakudo-moar 4a74e7: OUTPUT«===SORRY!=== Error while compiling /tmp/2oyyhTMsAm␤Undeclared routine:␤ bit used at line 1␤␤»
diakopter m: my Array[bit] @life is dim (20, 20) is default(0); @life.perl.say
Ven m: my Array of Int @life is dim (20, 20) is default(0); @life.perl.say
camelia rakudo-moar 4a74e7: OUTPUT«===SORRY!===␤Cannot invoke null object␤»
rakudo-moar 4a74e7: OUTPUT«===SORRY!=== Error while compiling /tmp/DmSPwlU61F␤Can't use unknown trait 'is dim' in a variable declaration.␤at /tmp/DmSPwlU61F:1␤------> my Array of Int @life is dim (20, 20) ⏏is default(0); @life.perl.say␤…»
Ulti also sized arrays aren't a thing yet either right?
Ven Ulti: that's from www.perlmonks.org/?node_id=253366 13:49
moritz Ven: don't trust Perl 6 advice (and especially syntax) from 2003 13:51
Ven oh whoops. Should check date first ! 13:51
Ulti heh
moritz back then, there was basically nothing implemented of Perl 6, and the syntax was just a though in the wind
Ulti yeah the spec changes more often than once every 11 years atm :3 13:52
Ven I guess I just got used to advent gifts :p
Ulti hmmm 13:53
m: my Array of Int @life; @life = ((0) x 2 ).item x 2; say @life.perl; 13:54
camelia rakudo-moar 4a74e7: OUTPUT«Type check failed in assignment to '@life'; expected 'Array[Int]' but got 'Str'␤ in method REIFY at src/gen/m-CORE.setting:8363␤ in method reify at src/gen/m-CORE.setting:7252␤ in method reify at src/gen/m-CORE.setting:7235␤ in method reify at …»
Ulti m: my Array[Int] @life; @life = ((0) x 2 ).item x 2; say @life.perl;
camelia rakudo-moar 4a74e7: OUTPUT«Type check failed in assignment to '@life'; expected 'Array[Int]' but got 'Str'␤ in method REIFY at src/gen/m-CORE.setting:8363␤ in method reify at src/gen/m-CORE.setting:7252␤ in method reify at src/gen/m-CORE.setting:7235␤ in method reify at …»
moritz Ulti: this can never work, because (0 x 2).item isn't declare as Array[Int] 13:55
Ulti ahh ok
also its stringafying too
which I didnt expect
moritz oh right, use xx instead
x is string repetition
xx is list repetition
Ulti m: my Array @life; @life = ((0) xx 2 ).item x 2; say @life.perl;
camelia rakudo-moar 4a74e7: OUTPUT«Type check failed in assignment to '@life'; expected 'Array' but got 'Str'␤ in method REIFY at src/gen/m-CORE.setting:8363␤ in method reify at src/gen/m-CORE.setting:7252␤ in method reify at src/gen/m-CORE.setting:7235␤ in method reify at src/g…»
Ulti doh two x! 13:56
m: my Array @life; @life = ((0) xx 2 ).item xx 2; say @life.perl;
camelia rakudo-moar 4a74e7: OUTPUT«Type check failed in assignment to '@life'; expected 'Array' but got 'List'␤ in method REIFY at src/gen/m-CORE.setting:8363␤ in method reify at src/gen/m-CORE.setting:7252␤ in method reify at src/gen/m-CORE.setting:7235␤ in method reify at src/…»
diakopter now that just kills me
Ulti heh and need to arrayafy
o___O
Ven m: say ((0 xx 5).item xx 5).perl
camelia rakudo-moar 4a74e7: OUTPUT«((0, 0, 0, 0, 0).list.item, (0, 0, 0, 0, 0).list.item, (0, 0, 0, 0, 0).list.item, (0, 0, 0, 0, 0).list.item, (0, 0, 0, 0, 0).list.item).list␤»
Ulti m: my @life; @life = ((0) xx 2 ).item xx 2; say @life.perl;
camelia rakudo-moar 4a74e7: OUTPUT«Array.new((0, 0).list.item, (0, 0).list.item)␤»
Ulti \o/
sort of
lol
moritz currently I recommand against using typed arrays
diakopter lol indeed
Ulti LoL even 13:57
moritz r: my @live = [0 xx 5] xx 3; @live[1, 2] = 1; say @live.perl
camelia rakudo-parrot 4a74e7, rakudo-jvm 4a74e7, rakudo-moar 4a74e7: OUTPUT«Array.new([0, 0, 0, 0, 0], 1, Any)␤»
moritz r: my @live = [0 xx 5] xx 3; @live[1][2] = 1; say @live.perl
diakopter wait, why are those .item calls deferred in the .perl output?
camelia rakudo-parrot 4a74e7, rakudo-jvm 4a74e7, rakudo-moar 4a74e7: OUTPUT«Array.new([0, 0, 0, 0, 0], [0, 0, 1, 0, 0], [0, 0, 0, 0, 0])␤»
moritz deferred? 13:58
Ven moritz++ # cleanest !
Ulti also the arrayest 14:00
Ven I like arrays 14:01
Not to write my own Game of life
diakopter I like trains
Ven I like turtles 14:02
moritz turtles like you!
Ven and they say "hello !" 14:03
Ulti dont the more efficeint game of life implementations avoid allocating the whole world as a 2D array though, we need some fresh syntax in Perl6 for Octree or something more 2Deee
is there a unicode character for an octopus leg? 14:04
FROGGS moritz: wiki.enlightenedperl.org/gsoc2014/ideas
I put our stuff there too
moritz we don't need syntax for everything :-) Sometimes plain old method calls are fine :-) 14:04
FROGGS++
Ulti U+1F419
diakopter .u 1F419
yoleaux U+1F419 OCTOPUS [So] (🐙)
Ulti emoji win on OSX :3 14:05
diakopter fail on my OSX 14:05
Ven $*OS eq 'MSWin32' ?? 'cls' !! 'clear' ==> shell; 14:07
@.grid.map(*.join).join("\n") # that's how I perl 14:08
Ven Okay, that's enough perl for now, headaches. I'll read your article later, moritz 14:09
moritz :-) 14:10
Ven Thanks all, really ! :)
masak r: my @live = [0 xx 5] xx 3; @live[1, 2] = 1; say @live.perl 14:34
moritz: did you mean @live[1; 2] ?
camelia rakudo-parrot 44ab3c, rakudo-jvm 44ab3c, rakudo-moar 44ab3c: OUTPUT«Array.new([0, 0, 0, 0, 0], 1, Any)␤»
moritz masak: not sure what I meant, besides "thinko" 14:40
masak ;) 14:42
Timbus <Ven> if I want a structure to act as I nested it, is there an alternative to .tree that is recursive 14:48
isnt that... 14:49
flat?
colomon oooo, Octree module. Hmmm....
Timbus my mind
Ven Timbus: not sure what you mean
TimToady1 .tree is *supposed* to be recursive, but it currently has the desired semantics of lol 14:50
Ven oh.
Ven so it's just a "not-currently-working-as-indented" case? 14:50
TimToady1 nodnod 14:51
FROGGS Ven: that sounds like an invitation to rakudo core hacking :o)
Ven FROGGS: do-does it ?
FROGGS hehe
yeah :o)
Ven o-oh, I see 14:52
lot of stuff I don't know in the compiler :D 14:53
FROGGS ohh, rakudo is not that hard to grok 14:53
it is written in "our" language, you see 14:54
Ven I see. I've worked with self-hosted compilers before, so that doesn't surprise me that much :)
FROGGS cool! 14:55
Ven I get "nom" is the dev branch ?
FROGGS no, it is the master branch
Ven: what is your compiler experience ooc?
Ven FROGGS: mainly coffeescript/livescript
also hacked a bit on MRI (ruby) 14:56
FROGGS cool
TimToady1 oh, I guess .tree is specced to default to .lol
that's silly
Ven there's one on List and one on Any 14:57
(what's that :D:)
TimToady1 :D allows only happy values, : marks invocant 14:58
self
Ven uuu
TimToady1 where happy really means "definite", or instantiated, or some such
anyway, leaves out the type objects 14:59
Timbus type objects werent happy about not being allowed in 15:00
FROGGS there is alse :U for undefined, and in theory :T for type objects
Ven ah yes, nqp doesn't allow for non-primitive type checks right ? 15:01
masak Ven: "nom" is the integration branch. dunno if that's what you mean by "dev branch".
Ven masak: I mean "where you work"
masak Ven: basically, if it passes all the old and new spectests, you can put the commit on nom.
Ven: if it doesn't, you normally shouldn't.
Ven: even if it does, sometimes you want to work in a (shared or private) branch.
and later rebase/merge with nom. 15:02
Timbus m: (1, (2, 3, (4, 5))).tree[1][2].perl.say 15:03
camelia rakudo-moar 44ab3c: OUTPUT«4␤»
Timbus so, that should be 4, 5 eh 15:04
Ven that'd be sweet :)
TimToady1 is respeccing .tree to make .tree default to .tree(*)
Ven m: (1, (2, 3, (4, 5))).tree(*)[1][2].perl.say
camelia rakudo-moar 44ab3c: OUTPUT«Cannot call 'tree'; none of these signatures match:␤:(Any:U: *%_)␤:(Any:D: *%_)␤:(Any:D: Cool $count, *%_)␤:(Any:D: &c, *%_)␤ in method tree at src/gen/m-CORE.setting:1569␤ in block at /tmp/auh4NFzXBQ:1␤␤»
TimToady1 which in turn means apply .item everywhere 15:05
Ven TimToady1: if you happen to fix that rakudo-side, mind linking the commit ?
dalek ecs: 977e82f | larry++ | S (2 files):
respec tree to default to all levels

also some initial Seq extirpation		 15:12
Timbus (1, (2, 3, (4, 5))).tree(100).perl # well that was scary to type into the repl 15:13
Ven well. 15:14
m: (1, (2, 3, (4, 5))).tree(*).perl.say
camelia rakudo-moar 44ab3c: OUTPUT«Cannot call 'tree'; none of these signatures match:␤:(Any:U: *%_)␤:(Any:D: *%_)␤:(Any:D: Cool $count, *%_)␤:(Any:D: &c, *%_)␤ in method tree at src/gen/m-CORE.setting:1569␤ in block at /tmp/yNblzQr8eU:1␤␤»
Timbus m: (1, (2, 3, (4, 5))).tree(2).perl.say 15:16
camelia rakudo-moar 44ab3c: OUTPUT«((1,).list.item, LoL.new(2, 3, $(4, 5)).item).list␤»
FROGGS .oO( ar ee es pee ee see RESPEC! )
Ven oh but the 1 gets nested too 15:19
Ven m: .perl.say for (1, (2, 3, (4, 5))).tree(2) 15:21
camelia rakudo-moar 44ab3c: OUTPUT«(1,).list.item␤LoL.new(2, 3, $(4, 5)).item␤»
Ven m: .^name.perl.say for (1, (2, 3, (4, 5))).tree(2)
camelia rakudo-moar 44ab3c: OUTPUT«"List"␤"LoL"␤»
Timbus m: (1, (2, 3, (4, 5))).tree(2)[0].say
camelia rakudo-moar 44ab3c: OUTPUT«1␤»
Ven I'd expect that to be Int[nl]LoL
m: (1, (2, 3, (4, 5))).tree(2)[0].^name.say 15:22
camelia rakudo-moar 44ab3c: OUTPUT«List␤»
Timbus not dwimming. demand a refund 15:23
i want my fun back
Timbus but yeah, .tree needs some sprucing up 15:26
TimToady1 hmm, maybe I'm using * backwards 15:27
Timbus its symmetrical tho ;o 15:28
TimToady1 maybe it should mean I don't care what's below this
so stop at this level
TimToady1 otherwise repeat the final mapper 15:29
Timbus no i like it as is. whatever is down there, tree it'
Ven m: (1)succ 15:32
camelia rakudo-moar 44ab3c: OUTPUT«===SORRY!=== Error while compiling /tmp/nhkjRYbnGD␤Two terms in a row␤at /tmp/nhkjRYbnGD:1␤------> (1)⏏succ␤ expecting any of:␤ postfix␤ infix stopper␤ infix or meta-infix␤ …»
TimToady1 m: 1.succ 15:33
camelia ( no output )
TimToady1 m: 1.succ.say
camelia rakudo-moar 44ab3c: OUTPUT«2␤»
TimToady1 we only have 1 alpha postfix 15:34
m: say 1i
camelia rakudo-moar 44ab3c: OUTPUT«0+1i␤»
masak m: sub postfix:<j>($n) { $n\i }; say 1j 15:35
camelia rakudo-moar 44ab3c: OUTPUT«0+1i␤»
masak \o/
m: sub postfix:<j>($n) { $n\i but role { method gist { "0+1j" } } }; say 1j 15:36
camelia rakudo-moar 44ab3c: OUTPUT«Not enough positional parameters passed; got 1 but expected 3␤ in method BUILD at src/gen/m-CORE.setting:10182␤ in method BUILD_LEAST_DERIVED at src/gen/m-CORE.setting:941␤ in sub infix:<but> at src/gen/m-CORE.setting:16131␤ in sub infix:<but> …» 15:36
masak pff. 15:37
Ven Need to read about `but`
Ven I've read it was supposed to be a contrary to `else`, IIRc 15:37
masak Ven: using it for drive-by mixing-in here.
Timbus not quite..
Ven looks like scala's with 15:38
masak m: say (5 but False); say ?(5 but False)
camelia rakudo-moar 44ab3c: OUTPUT«5␤False␤»
Timbus yeah
masak m: my $special-int = 42 but role { method greet { say "OH HAI" } }; say $special-int; $special-int.greet 15:39
camelia rakudo-moar 44ab3c: OUTPUT«42␤OH HAI␤»
masak m: my $special-int = 42 but role { method greet { say "OH HAI" } }; say $special-int.mro
camelia rakudo-moar 44ab3c: OUTPUT«No such method 'mro' for invocant of type 'Int+{<anon>}'␤ in block at /tmp/QKe1JqyzKl:1␤␤»
Ven r: my @a = 42; @a.perl.say
masak m: my $special-int = 42 but role { method greet { say "OH HAI" } }; say $special-int.^mro 15:39
camelia rakudo-parrot 44ab3c, rakudo-jvm 44ab3c, rakudo-moar 44ab3c: OUTPUT«Array.new(42)␤»
rakudo-moar 44ab3c: OUTPUT«(Int+{<anon>}) (Int) (Cool) (Any) (Mu)␤»
masak that last example shows how it's implemented. (anon subclass) 15:40
Ven masak++ # really cool, especially with subset. I like structural typing !
timotimo Ven: did you see my ADT module yet? :3 15:40
masak :)
timotimo++ # ADT module 15:41
Ven timotimo: told you I used it to bait friends onto perl 6 :D
timotimo oh! you did!
Ven timotimo: you should put up a repo description, btw 15:42
masak timotimo: it'd be fun to work on that missing EXPORT feature together. 15:43
masak timotimo: I also looked at that and thought "ooh, that'd be a nice use for a DSL instead of an intert specification string". 15:43
Ven masak++ # agreed about the DSL ! ; timotimo++ # watched your repo ! 15:49
masak Ven: my brain is currently in those parts of spec-land, pondering. 15:50
basically, I think I discern three levels of language extension/modification:
(a) only need subroutines (so-called "internal DSL" by some)
Ven needs reader macros HA
masak (b) need macros (maybe to change program flow, or insert checking/analysis) 15:51
(c) need a whole different sublanguage (because the syntax is different enough that macros aren't enough)
it'd be nice to collect a few examples on each level. I'm gearing up to write blog posts about that, too.
moritz m: my $x; 'abc' ~~ /$x=(ab)/; say $x 15:53
camelia rakudo-moar 44ab3c: OUTPUT«===SORRY!===␤Error while compiling op bind: QAST::Block with cuid cuid_1_1391529188.76986 has not appeared␤»
moritz m: my $x; 'abc' ~~ /$x=(ab)/; say $x
camelia rakudo-moar 44ab3c: OUTPUT«===SORRY!===␤Error while compiling op bind: QAST::Block with cuid cuid_1_1391529195.38929 has not appeared␤»
moritz m: 'abc' ~~ /$<x>=(ab)/; say $<x>
camelia rakudo-moar 44ab3c: OUTPUT«｢ab｣␤␤»
TimToady1 std: my $x; 'abc' ~~ /$x=(ab)/; say $x 15:54
camelia std 09dda5b: OUTPUT«ok 00:01 127m␤»
TimToady1 n: my $x; 'abc' ~~ /$x=(ab)/; say $x
camelia niecza v24-109-g48a8de3: OUTPUT«===SORRY!===␤␤Non-Match bindings NYI at /tmp/2rh2e31Bkj line 1:␤------> my $x; 'abc' ~~ /$x=(ab)⏏/; say $x␤␤Unhandled exception: Check failed␤␤ at /home/p6eval/niecza/boot/lib/CORE.setting line 150…»
TimToady1 r: my $x; 'abc' ~~ /$x=(ab)/; say $x 15:55
camelia rakudo-moar 44ab3c: OUTPUT«===SORRY!===␤Error while compiling op bind: QAST::Block with cuid cuid_1_1391529311.00125 has not appeared␤»
..rakudo-jvm 44ab3c: OUTPUT«===SORRY!===␤setcodeobj can only be used with a CodeRef␤»
..rakudo-parrot 44ab3c: OUTPUT«===SORRY!===␤Could not find sub cuid_1_1391529308.45272␤»
masak time for a LTA rakudobug?
TimToady1 well, I believe it's specced to Just Work
moritz m: 'abc' ~~ /^ @<x>=.+ $ /; say $/
masak submits rakudobug
camelia rakudo-moar 44ab3c: OUTPUT«Nil␤»
masak std: 'abc' ~~ /$x=(ab)/; say $x 15:56
camelia std 09dda5b: OUTPUT«===SORRY!===␤Variable $x is not predeclared at /tmp/BMQCFR0dVk line 1:␤------> 'abc' ~~ /⏏$x=(ab)/; say $x␤Variable $x is not predeclared at /tmp/BMQCFR0dVk line 1:␤------> 'abc' ~~ /$x=(ab)/; say ⏏$…»
masak std: 'abc' ~~ /$x=(ab)/
camelia std 09dda5b: OUTPUT«===SORRY!===␤Variable $x is not predeclared at /tmp/KrehdSr98u line 1:␤------> 'abc' ~~ /⏏$x=(ab)/␤Check failed␤FAILED 00:01 126m␤»
timotimo moritz: does that compile error also appear with --optimize=0?
or =off
moritz m: 'abc' ~~ /^ $<x>=(.)+ $ /; say $<x>
TimToady1 r: my $x; 'abc' ~~ /(ab) { $x = $0 }/; say $x
camelia rakudo-moar 44ab3c: OUTPUT«｢a｣␤ ｢b｣␤ ｢c｣␤␤»
rakudo-parrot 44ab3c, rakudo-jvm 44ab3c, rakudo-moar 44ab3c: OUTPUT«｢ab｣␤␤»
moritz m: 'abc' ~~ /^ $<x> = (.)+ $ /; say $<x> 15:57
camelia rakudo-moar 44ab3c: OUTPUT«｢a｣␤ ｢b｣␤ ｢c｣␤␤»
TimToady1 m: 'abc' ~~ /^ $<x>=.+ $ /; say $<x>
camelia rakudo-moar 44ab3c: OUTPUT«｢abc｣␤␤»
moritz m: 'abc' ~~ /^ $<x> = (.)+ $ /; say $<x>.elems
camelia rakudo-moar 44ab3c: OUTPUT«3␤»
TimToady1 usually uses [] instead of () to avoid giving the impression that $0 is set 15:58
TimToady1 m: 'abc' ~~ /^ <x=:Any>+ $ /; say $<x> 15:59
camelia rakudo-moar 44ab3c: OUTPUT«｢a｣␤ ｢b｣␤ ｢c｣␤␤»
TimToady1 there's a use for :Any :)
timotimo i have no idea what that does
FROGGS :Any is a unicode property lookup 16:00
PerlJam random thought: Just now it strikes me as weird that [] doesn't capture while () does capture. [] are used in arrays and capturing is just like indexing into an array.
timotimo oh!
of course
not the perl6 type
FROGGS no
colomon What does the Unicode property mean?
FROGGS Anything
TimToady1 we decided we could fight the history of () in regexland
*n't 16:01
colomon like . ?
TimToady1 yes
TimToady1 <:Any> is just an explicit way to write . 16:01
moritz m: 'abbbb' ~~ / (b ** 0..3)/; say ~$0; 16:02
camelia rakudo-moar 44ab3c: OUTPUT«␤»
masak r: say ?( 'aaa' ~~ m:g/a/ )
camelia rakudo-parrot 44ab3c, rakudo-jvm 44ab3c, rakudo-moar 44ab3c: OUTPUT«False␤»
TimToady1 moritz: that's correct
FROGGS moritz: position wins 16:03
TimToady1 r: say ('aaa' ~~ m:g/a/).WHAT
camelia rakudo-parrot 44ab3c, rakudo-jvm 44ab3c, rakudo-moar 44ab3c: OUTPUT«(Bool)␤»
FROGGS masak's example is weird though
masak lizmat++ # .made 16:04
FROGGS I've seen it on p6u a minute ago too
masak FROGGS: think it might be doing smartmatching twice, or something.
TimToady1 r: $_ = 'aaa'; say (m:g/a/).WHAT
camelia rakudo-parrot 44ab3c, rakudo-jvm 44ab3c, rakudo-moar 44ab3c: OUTPUT«(List)␤» 16:04
masak right.
TimToady1 that wants to return a pseudo Match
FROGGS :/
TimToady1 so that the smartmatcher knows it's done
masak basically, Perl 6 straightens out a waterbed wrinkle there, which surprises people. 16:05
TimToady1++'s solution might work.
TimToady1 it's really OKness in disguise
masak as long as it's in disguise, it's probably fine :P
TimToady1 but it's specced by merely listing the types that behave like Bool
without making it generalizable :( 16:06
TimToady1 masak: it's not clear where the boundary between your (b) and (c) is. Where does a reader macro (is parsed) fit? What is the scope of the sublanguage replacement? 16:14
can a macro do what 'use Python;' does? 16:15
and we need to detangle the 'immediate' semantics from the language change semantics 16:16
my \term = 42; is language change without an immediate sub call
(well, as currently implemented--we could run all identifiers through a common path)
raiph Pop quiz for #perl6: who recently said that the best code is poetry, that “you’re trying to take something that can be described in many, many sentences and pages of prose, but you can convert it into a couple lines of poetry and you still get the essence, so it’s that compression.” 16:18
masak TimToady1: I wish I had more firm answers. but I figure there are some things that the community will consider as being trivially doable by macros, and others where a new braid/bundle of grammar/actions/world is needed. 16:19
masak TimToady1: kind of like the tipping point where one shouldn't do individual regexes anymore, due to diminishing returns. 16:20
so in my analogy, regexes : grammars :: macros : slangs.
diakopter raiph: I give up 16:24
raiph diakopter: nirvana arguably translates to "i give up", so hopefully u feel great :) 16:25
raiph anyhoo, the new ceo of microsoft 16:25
diakopter figures; not even sentences 16:26
masak raiph: I find the best two ways to make that happen are (a) test-first development, and (b) having already implemented the thing 49 times before. 16:29
(fsvo 49)
smls agrees with PerlJam ("Just now it strikes me as weird that [] doesn't capture while () does capture") 16:35
moritz isn't that how other professions do it? You've done 35 pancakes already, you can be pretty confident that the 36th pancake also will be just fine 16:36
smls () is used for simple grouping outside of regexes, so it would be nice to have similar semantics inside regexes
FROGGS ohh, please don't change that ó.ò 16:36
smls and [...] sort of looks like a pair of grappling hooks, so perfect for "capture this part"
moritz don't the {...} look more like grappling hooks? 16:37
smls FROGGS: Would it really be so bad? I mean, there are lots of major changes between traditional regexes and P6 regexes/grammars already. 16:38
So you can't copy/paste nbetween the two anyways
moritz we have an IMMENSE body of Perl 6 regexes written already
FROGGS smls: do we have to break every bit of P6 code that is "out there" atm?
moritz mostly notably in a bootstrapped compiler
and doing such an invase change in a bootstrapped compiler is a real PITA 16:39
masak besides, I don't think the argument for such a change is a very strong one.
FROGGS exactly, I'd say there must be a very strong reason for changing fundamental things like this 16:40
masak yes, we "break all the [back-compat] that needs breaking", but there's something to be said for (...) or some variant of it meaning "capture" in most dialects of regex syntax. 16:40
moritz and we want the specs to solidfy, not to vaporizes
smls moritz: Sorry, grappling hook was the wrong word, I meant more like a gripper as an inductrial robotic might have as its "hand" 16:41
FROGGS smls: I understand what you say but IMO having cute parens is not worth it 16:42
masak use Regex::BracketsAndParensExchanged;
smls lol
masak it's 2014 and I still don't have my language-modifying modules. 16:43
timotimo masak: well, 2014 has only just begun 16:59
smls In an ideal world, any non-negligible benefit could justify a syntax change before the first officially stable release (rather than only a "very strong" benefit), but yeah I understand the pragmatic argument re. existing compilers and modules.
I don't agreee with the compatibility argument though. I think consistency within Perl 6 is better than consistency between parts of Perl 6 and the analogous parts of other languages.
smls I mean, I myself am a Perl 5 developer with somewhat of a "resistant-to-change" personality, yet I was surprised that this was *not* changed in Perl 6 regexes (considering all the other invasive changes/innovations in that area). 17:02
PerlJam um ... wrt the capturing brackets, I concluded internally right after I had my random thought that we could always put a module out there for people to try that would swap their meanings, so ... what masak said :)
smls I'm not a fan of "cosmetic" syntax-changing modules. 17:05
smls They fracture the language's ecosystem and make it sifficult to merge code from different sources together 17:05
moritz but the same argument doesn't apply to the fractur in time that a syntax change causes? 17:06
timotimo frakturschrift in perl6? :) 17:07
smls moritz: Well, P6's future lifetime (and amount of usage), will hopefully vastly exceed that of the past 17:09
so that the latter would become negligible in comparison
moritz smls: though history shows that old documents are still easily found with google&co, and such cause confusion
smls: just today we had a case here in #perl6 where somebody accidentally quoted Perl 6 syntax from 2003 17:10
smls but I do concede that it is not negligible now for the small P6 team to handle, i.e. thats the pragmatic argument
smls s/handle/migrate/ 17:12
timotimo ==> Successfully installed ADT
no test problems :\
tadzik oh no
timotimo oh, yikes
with ufo and make loudtest i *do* get errors 17:13
tadzik === SORRY === :(
timotimo i don't understand why the tests don't fail with panda 17:16
tadzik $ perl6 bootstrap.pl --prefix HEREPANDA 17:21
Usage: bootstrap.pl [--prefix=<Any>]
ಠ_ಠ
moritz ok, straw poll time. I'm sure many of you have read books or tutorials that explained closures. How many of those actually explained at which point a time a closure is instantiated (or "cloned" in p6 speak)? 17:22
tadzik: use --prefix=HEREPANDA instead? :-) 17:23
tadzik I know, it's just not DWIM :)
moritz patches well done. 17:24
tadzik troo 17:26
tadzik hmm, strange 17:29
smls moritz: Well in P5 one would only think of *anonymous* subroutines as closures, so they would always appear in an expression that is reached during normal control flow. And that's when I always imagined them to be instantiated.
tadzik gist.github.com/tadzik/8808416
in line 4, it shows that it loaded a precompiled File::Find
smls Thouch I don't think the P5 tutorials I read on the subject went into further details on that. 17:30
*though
tadzik but then at the end, it complains that another File/Find.pm is not precompiled
jnthn I agree changing the meaning of (...) now in regex is way too late 17:31
moritz smls: ok, thanks for the data point 17:32
jnthn I also thought we were gonna *add* .made, not remove .ast also... 17:32
tadzik argh 17:36
I hate precompiling stuff
vendethiel smls, moritz: I was the one with that 2003 link. I often get perlmonks results from googling and sometimes it's really outdated, I'm not sure to what extent I can use it or not
moritz vendethiel: that's fine; I just used you as a data point :-)
vendethiel moritz: WRT closures, not sure what you mean. do you mean when values are captured ? 17:37
moritz vendethiel: correct 17:37
for example, javascript also has named closures 17:38
vendethiel moritz: rarely explained. Usually it just says it captures "variables".
We get a lot of those questions in JS/coffee
moritz vendethiel: ok, that matches my own experience
vendethiel "why when I do a loop to bind buttons, my click always gives me the last number?" but that's mainly because JS doesn't (didn't) have block scope 17:39
moritz vendethiel: does coffee also use the "when the outer scope is entered" mechanism?
jnthn moritz: I explain closures relatively often. I tend to not talk precisely about when the clone or instantiation happens, but certainly talk about lifetime and the exact scope that is captured. 17:40
vendethiel moritz: we have JS's semantics, with 1 or 2 exceptions (because we don't have explicit `var`, a variable is declared in the highest scope it's defined) 17:40
s/defined/assigned/
jnthn moritz: C# has an over-sharing issue when you write for (var i = 0; i < 10; i++) things.push(x => x + i); 17:41
jnthn moritz: So I certainly have to give people a good enough mental modle to understand *why* that doesn't do that they want. 17:42
*model
vendethiel jnthn++ # same thing here
jnthn (Worth noting you'll get the same problem in Perl 6 if you use loop, fwiw...) 17:43
moritz has a distaste for implicitly declared variables
timotimo that gives 10 closures that all add 10 to the argument?
vendethiel moritz: oh, I agree
jnthn timotimo: yes
dalek Heuristic branch merge: pushed 40 commits to panda/prefix by tadzik 17:45
jnthn dinner & 17:45
timotimo oh, dalek is back \o/ 17:50
colomon woah, big changes to panda? 18:20
timotimo no, just a merge of master into a branch i believe 18:21
FROGGS break all the things! /o/
FROGGS (joke) 18:22
tadzik yeah, I tried to resume work on prefix 18:38
Util #ps in 4m 19:26
TimToady yay, I'm me again 19:34
FROGGS \o/ 19:35
TimToady had to quit irssi, wouldn't let the nick timeout
nwc10 in real life too, or are you still on the cough syrup etc? 19:39
TimToady I think I've managed to get off the cough syrup with the aid of a humidifier, but I'm also working on a case of shingles, which I'm supposed to wait for blisters to come up before I go back in 19:42
and given that I have cephalic shingles, it can get (back into) my eye (for which reason I was blind in one eye for 14 years)
but now it's just neuralgia on the top of my head mostly, but it makes it hard to think some of the time 19:43
FROGGS O.O
TimToady trigeminal nerve, ophthalmic branch
the other irritation is that when I cough (and I still do a lot of that), it feels as though the top of my head is going to fly off 19:44
but I'm hopeful this may be the first day without fever from the flu 19:45
that's the thing about shingles, it always kicks you when you're already down
nwc10 E has bronchitis, or keeps getting bronchitis again every time she gets a cold 19:47
this is tiring for all of us
TimToady that's why we do this when we're young :) 19:48
timotimo and then the shingles kick you in the shins :( 20:12
dalek kudo/wip-openpipe: d62a7b2 | (Tobias Leich)++ | src/core/IO.pm:
use openpipe on moar
 20:16
TimToady well, on balance I think I would rather have shingles kick me in the shins than in the head, but I don't get to choose 20:21
colomon TimToady: :( 20:40
btyler hi all. just putzing around trying to get a handle on grammars, and decided to make a parser for nginx/apache access logs. can anyone help me figure out why the grammar for 'date' fails but the regex at the bottom matches? gist.github.com/kanatohodets/7b3d3...se-pm6-L11 21:01
timotimo before i look, did you have a look at the Grammar::Debugger and Rakudo::Debugger packages? 21:04
ah, hehe 21:05
you do $<date>=\[ (.+) \]
which means ... assign what matched '[' into $<date> and the stuff after the
sjn btyler: check out perlgeek.de/en/article/debug-a-perl-6-grammar :)
timotimo '[' goes into $0 and then it matches a ']'
btyler sjn/timotimo: thanks! I guess I was confused about how the $<foo> capture things work. do I stick the 'date' capture inside \[\], then? trying to 'grab all the stuff inside a pair of square braces' 21:11
FROGGS p: say '137.22.189.191 - - [02/Feb/2014:23:37:36 -0600] "GET' ~~ /$<date>=[ '[' .+ ']' ]/ 21:13
camelia rakudo-parrot 44ab3c: OUTPUT«｢[02/Feb/2014:23:37:36 -0600]｣␤ date => ｢[02/Feb/2014:23:37:36 -0600]｣␤␤»
timotimo i would suggest .+? instead of .+
that should be much, much faster
actually, use <-[\]]> instead perhaps
FROGGS true
timotimo instead of .
FROGGS p: say '137.22.189.191 - - [02/Feb/2014:23:37:36 -0600] "GET' ~~ /<date=-[\]]>/
camelia rakudo-parrot 44ab3c: OUTPUT«｢1｣␤ date => ｢1｣␤␤»
btyler timotimo: ok, moment..think baby perl6 :P what does those variations mean? 21:14
FROGGS p: say '137.22.189.191 - - [02/Feb/2014:23:37:36 -0600] "GET' ~~ /$<date>=[ <-[\]]>+ ]/
camelia rakudo-parrot 44ab3c: OUTPUT«｢137.22.189.191 - - [02/Feb/2014:23:37:36 -0600｣␤ date => ｢137.22.189.191 - - [02/Feb/2014:23:37:36 -0600｣␤␤»
btyler *do
FROGGS .+ is greedy, .+? is not 21:15
and <-[ ... ]> is a negated charater class
where ... is the actual content
timotimo yeah, your .* would match until the end of the whole file and then backtrack until it finds a ] 21:16
btyler oh, jeez, ok.
timotimo meaning you'll get a few thousand lines as the date of your first entry, in the worst case :P 21:17
skids One doesn't generally parse log files as one string. That's a recipe for high memory footprint. Best to load and discard lines one by one. 21:31
btyler skids: of course, just messing around :) figured that doing a moving window thing was putting the cart before the horse if I didn't have any notion of how grammars worked 21:34
skids ponders the concept of lazy matching on a stream. 21:37
colomon skids: once we have lazy strings... 21:44
timotimo moritz++ # answering a barrage of questions on the mailing list 21:46
skids colomon: right, but what does $/ look like during the match... 21:49
colomon skids: hmmm
lue masak: at parts that slideshow reads like someone trying too hard to impress with buzzwords :P "strictness!" "symbol tables!" "lexical scoping!" 22:38
ooh, didn't know .ast became .made recently 22:42
timotimo lue: that was literally today 23:07
oh, actually, the day just changed a few minutes ago on my timezone 23:12

Powered by Cro and the Raku® Programming Language.

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

Thank you!