»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'perl6: say 3;' or rakudo:, niecza:, std:, or /msg p6eval perl6: ... | irclog: irc.perl6.org/ | UTF-8 is our friend! | Rakudo Star Released!
Set by diakopter on 6 September 2010.
sorear Should role Foo { } be considered a subtype of Mu? 00:03
I guess it has to bee 00:16
otherwise it's impossible to pass Foo into a function like &infix:<~~> or &infix:<but>
which leaves the question - are roles Mu, or Any? 00:17
if Foo ~~ Any, then, is it also the case that ClassWhichDoesFoo ~~ Any? Subtyping ought to be transitive
Does *that* mean that class Bar is Mu does Foo is illegal? 00:18
TimToady: jnthn: ping
dukeleto sorear: sounds like some interesting edge cases 00:22
dukeleto gets popcorn
masak rakudo: role Foo {}; say Foo ~~ Mu; say Foo ~~ Any 00:50
dalek rixel: e438505 | diakopter++ | sprixel/ (5 files):
added a TAILCALL_NOT_TRAMPOLINE preprocessor option. it seems 1.5x as slow

I'll try it on Mono on Windows and Linux sometime.		 00:51
p6eval rakudo 668719: OUTPUT«(timeout)»
masak locally, that says '1\n1\n'
diakopter the poor p6eval vps is bogged with something. looking. 00:53
oh. swapping horribly due to rebuilding parrot. 00:54
that is getting out of hand.
actually no, gen/perl6.pbc is compiling gen/core.pm 00:55
and using 90% of the system ram
6 hours of cpu?
oh, minutes. :) 00:56
rakudo: role Foo {}; say Foo ~~ Mu; say Foo ~~ Any 00:58
p6eval rakudo 668719: OUTPUT«1␤1␤»
masak diakopter++ 01:06
diakopter I didn't do anything; I just waited for rakudo to be done rebuilding 01:08
masak ah. 01:10
sjohnson "perl6: ready to rock and/or roll" 01:41
sjohnson perl! 02:00
colomon 6!
sjohnson heh
++perl5; ! 02:01
masak sjohnson: "Perl 6: ready to rock, paper or scissors." 02:45
plobsing rakudo: "rock", "paper" or "scissors" 02:53
p6eval rakudo 668719: ( no output )
plobsing rakudo: say "rock", "paper" or "scissors"
p6eval rakudo 668719: OUTPUT«rockpaper␤»
masak ok, then. time to go to Paris. 03:29
dalek ecza/master: a3cf7c1 | sorear++ | / (4 files):
Implement type objects for roles
 06:40
ecza/master: 4d2eb8a | sorear++ | / (6 files):
Implement compilation of parametric roles
nadim on p5 someone said to not use this <<<< say~~grep{$_<0}@array >>>> which may be. some one cares to tell me why? 08:52
moritz_ what does ~~ as a prefix do? 08:53
nadim: but the best chances are to ask the one who told you
nadim in p5 it two unary ~ to force scalar context 08:54
moritz_ you can use $(...) for that in Perl 6 08:55
sorear Why are you writing code that ugly?
moritz_ nadim: there's also the scalar() in p5
nadim I know it all started because I wanted to write >>> say scalar grep {$_ < 0} @array ; <<< shorter. please ddon't ask why, there is no reason. 08:56
masak oh hai from OSDC.fr! \o/ 09:02
nadim you're everywhere K. 09:02
jnthn_fr o/ #perl6 09:06
masak rakudo: my @i = <rock paper scissors>; my %n = %(@i.kv).invert; sub w($a, $b) { (%n{$a} - %n{$b}) % 3 }; my ($p1, $p2) = @i.roll(2); .say for $p1, $p2; say $_ ?? "Player $_ wins" !! "A tie" given w($p1, $p2) 09:15
p6eval rakudo 668719: OUTPUT«paper␤scissors␤Player 2 wins␤»
masak rakudo: my @i = <⽯ ⎘ ✄>; my %n = %(@i.kv).invert; sub w($a, $b) { (%n{$a} - %n{$b}) % 3 }; my ($p1, $p2) = @i.roll(2); .say for $p1, $p2; say $_ ?? "Player $_ wins" !! "A tie" given w($p1, $p2) 09:20
p6eval rakudo 668719: OUTPUT«✄␤✄␤A tie␤»
jnthn_fr That rocks on paper, but srsly, cut it out. :P 09:22
masak :P
masak cuts out of IRC
sorear hello jnthn_fr 09:25
jnthn_fr: niecza's metamodel is not so unlike 6model as I once thought
the mistake I made was to think of DynMetaObject as a bizarre HOW 09:26
it's not, it's much closer to STable
the one thing I'm really missing is reification of representations as objects
sorear blah 09:27
masak rakudo: class A { has $.x }; say A.new(:x<a b c>).x.WHAT 10:12
p6eval rakudo 668719: OUTPUT«Seq()␤»
moritz_ wonders why item context turns a Parcel into a Seq 10:20
sorear because S02 says so. 10:24
TimToady and pmichaud are adamant S02 is right but I haven't grokked their explanations yet
moritz_ hopes that it becomes clearer when pmichaud updates the specs with the new list/iterator API 10:25
masak "specs" in plural makes it sound like there's more than one spec... :/ 10:25
it's enough of a bother with the one we have. 10:26
moritz_ we have the written S\d\d documents, the test suite and #perl6 collective knowledge 10:27
masak I'm pretty sure you only meant the first of those, though. 10:27
moritz_ right
masak which I'd call "one spec" but "several synopses".
sorry about being so picky. 10:28
I'm still a bit un-fine with the repo being called "specs" as in "eyeglasses".
not that there aren't more important matters to attend to, but still. renaming it would make one small thing more right. 10:29
jnthn++ says "specs -- how you look at Perl 6" :) 10:43
lunch & 10:44
cosimo any examples of sockets usage in parrot or perl6? 11:16
moritz_ cosimo: github.com/mberends/http-server-simple
cosimo perl6 lwp::simple is working fine for me, while i have really big problems talking to a memcached server 11:17
as in, i can't get all the bytes the server is sending me, or get just some lines, ...
moritz_: thanks, i'll take a look
oyse Is the perl6 executable built with debugging symbols by default or do I have to provide som options when running make? 11:20
moritz_ oyse: you probably need to change that in parrot's configure step
hm, looks like by default is with debugging symbols 11:21
oyse moritz_: ok. thanks 11:22
lucas_ hello developers 11:52
I'm just beginning my learning path, so I have some doubts... 11:53
STD.pm(6) is written in Perl 6, right?
was it written in Perl 5 at any point in time before?
masak lucas_: no, STD.pm6 is written in Perl 6 from the start. there's a tool that knows how to translate (the STD-specific subset of) Perl 6 into Perl 5 so that it can run, though. 12:06
lucas_ masak: hello 12:10
what's is the name of this tool?
I run into this file...
masak lucas_: viv.
lucas_ std/std_hilite/STD_syntax_highlight 12:11
and it was written in p5
masak lucas_: its name means "vi -> v", i.e. Roman numerals for "6 -> 5".
lucas_ but it required STD...
moritz_ lucas_: you can install the compiled-to-p5 version of STD via CPAN
lucas_ oh, cool, I wouldn't think that about viv name
masak lucas_: yes, the std_hilite thing is mostly used by Padre, I think. Padre is written in Perl 5.
moritz_ cpanm STD # and wait a bit
lucas_ the lesson I learned now it that there is this p5-compiled version of STD. I didn't know about that. thanks 12:14
masak well, viv compiles the Perl 6 source code in STD.pm6 to Perl 5. 12:17
for the purpouse of them running it.
lucas_ I'm trying to run std/std_hilite/STD_syntax_highlight 12:39
it dies because CursorBase.pm can't find setting CORE and asks if I runned make
running make in std source tree dies because of missing package VAST::Assignment_operator
even after installing Mildew dist from CPAN
sub load_lex in CursorBase.pm is trying to load some yaml file syml/CORE.syml
what are those syml files? I've never heard of these before. stands for "syntax" yaml?
where are they installed?
x3nU i'm trying to use MiniDBI 13:00
and that's what i get
The native library 'libmysqlclient' required for 'mysql_affected_rows' could not be located
of course i have libmysqlclient
moritz_ x3nU: do you also have the header files? 13:33
x3nU moritz_: yes 13:38
colomon loliblogged: justrakudoit.wordpress.com/2010/10/...alindrome/ 13:55
rjbs (16, 8, 4 ... 0) 14:25
So, this gets all Zeno on me and gets to very very small numbers. Fine, I can dig it.
Is there a way to say, "treat everything as integers"?
moritz_ rakudo: .say for 'abcdedbb'.match(:g, /(\w+)\w?<?{$0.flip}>/) 14:26
p6eval rakudo 668719: OUTPUT«abcdedbb␤»
moritz_ rakudo: say ~(64, 32, *div* ... 0) 14:27
Juerd So that matched what, bb? 14:27
p6eval rakudo 668719: OUTPUT«64 32 2 16 0␤»
rjbs wanted to do it without supplying a closure.
moritz_ Juerd: it matched the whole string, because the <?{...}> seemed to have been ignored
rjbs (Just for exploration, not because it's important.)
moritz_ rjbs: then I don't know how to do it 14:28
rjbs Ok, thanks!
moritz_ hm, <?{ is a assertion
rakudo: .say for 'abcdedbb'.match(:g, /(\w+)\w?<{$0.flip}>/)
p6eval rakudo 668719: OUTPUT«ded␤bb␤»
moritz_ colomon: that's a simpler solution... just look for the match with the highest length 14:29
colomon moritz_: isn't that grotesquely inefficient? 14:30
moritz_ colomon: it probably is 14:31
colomon: but it's in the regex engine, which is a bit faster than regular Perl 6 code
colomon I mean, with a thousand character string, there will be 1,000,000 possible matches to test, no? 14:34
moritz_ correct 14:35
colomon I was figuring a regex solution would look like /(\w)(\w)(\w)$3$2$1/
moritz_ that's only for length 6
moritz_ but it'd be probably more efficient to to look for length 2 and 3 palindromes, and then go to the left and right for there on 14:36
colomon moritz_: right, but you can pretty easily generate the equivalent matches for longer lengths, and as soon as you hit one that does match (if you're counting up in length) you know your last match was the correct answer. 14:39
moritz_ colomon: nopaste.snit.ch/24122 is my current approach 14:42
still a bit buggy, it seems
moritz_ but takes only 2.3s on the example string from the challenge 14:43
colomon ooooooooo
zippy
that only gets the length 3 palindromes, right? 14:44
well, and length 2
moritz_ length 2 and 3, and then the inner while-loop should expand them
going to left and right simultaneously as long as left and right are the same
colomon ah, I see. nice. 14:45
moritz_ but there's a bug
at walks over the end
that's a bad description 14:46
moritz_ it produces 'allh', which is not a palindrome at all 14:46
mberends x3nU: in a debian-ish system, the module required for MiniDBI is libmysqlclient-dev. It's a bit like using libicu-dev when building Parrot. 14:47
x3nU i'm using CentOS 14:49
and i have mysql-dev
devel*
/usr/include/mysql.h
is present
/usr/include/mysql/mysql.h *
mberends x3nU: no other ideas right now, but I'll try installing on CentOS next week. Thanks for pointing this out :) 14:50
x3nU hope it will eventually work for me :) 14:53
colomon moritz_: think you need "my $right = .to;" 14:59
and then kill the $right--; at the end.
that said, it's generating correct results for me, but not THE correct result, which is weird.
never mind, the correct result is indeed in there. 15:01
and this version is screamingly fast. moritz_++
colomon moritz_: now you need to blog your solution. :) 15:11
moritz_ colomon: I will 15:29
colomon I've updated my post pointing out your brilliance. ;) 15:30
moritz_ lolnowibloggedtoo: perlgeek.de/blog-en/perl-6/longest-...regex.html 15:46
moritz_ anyone got a not-so-slow prime tester in perl 6 at hand? 15:51
moritz_ rakudo: foo($x) 15:55
p6eval rakudo 668719: OUTPUT«===SORRY!===␤too few positional arguments: 2 passed, 3 (or more) expected␤»
moritz_ now that's really LTA
rakudo: sub foo($) { }; foo($x)
p6eval rakudo 668719: OUTPUT«===SORRY!===␤too few positional arguments: 2 passed, 3 (or more) expected␤»
moritz_ rakudo: say (1, 2, 2).uniq 16:00
p6eval rakudo 668719: OUTPUT«12␤»
x3nU win 30 16:26
oops
in 30
TimToady rakudo: ~(64,32,16 ...^ Rat) 16:28
that ought work 16:29
p6eval rakudo 668719: OUTPUT«(timeout)»
TimToady but it doesn't
rakudo: ~(64,32,16 ...^ Num)
p6eval rakudo 668719: OUTPUT«(timeout)»
TimToady that's not it...
moritz_ rakudo: say ~(64,32,16 ... *)[^10]>>.WHAT
p6eval rakudo 668719: OUTPUT«Int() Int() Int() Rat() Rat() Rat() Rat() Rat() Rat() Rat()␤»
TimToady is is not really doing a smartmatch? 16:30
colomon moritz_: I used the old 2, 3, *+2 ... * > $possible-prime.sqrt method, which was plenty fast for the second challenge. 16:31
TimToady rakudo: say ~(64,32,16 ...^ 0..^1) 16:31
p6eval rakudo 668719: OUTPUT«Method 'shift' not found for invocant of class 'Range'␤ in 'infix:<...^>' at line 838:CORE.setting␤ in main program body at line 22:/tmp/UGYc0ZNN_f␤»
TimToady rakudo: say ~(64,32,16 ...^ (0..^1)) 16:31
p6eval rakudo 668719: OUTPUT«Method 'shift' not found for invocant of class 'Range'␤ in 'infix:<...^>' at line 838:CORE.setting␤ in main program body at line 22:/tmp/7LgcdWL3o9␤» 16:32
TimToady rakudo: say ~(64,32,16 ...^ ^1.item)
p6eval rakudo 668719: OUTPUT«Method 'shift' not found for invocant of class 'Range'␤ in 'infix:<...^>' at line 838:CORE.setting␤ in main program body at line 22:/tmp/WWDCUhFtqJ␤»
moritz_ colomon: I've done the same now; fast enough indeed
TimToady rakudo: say ~(64,32,16 ...^ (^1).item)
p6eval rakudo 668719: OUTPUT«Method 'shift' not found for invocant of class 'Range'␤ in 'infix:<...^>' at line 838:CORE.setting␤ in main program body at line 22:/tmp/SFdSJozP11␤»
moritz_ colomon: but I don't have a good idea on how to approach the third 16:32
TimToady rakudo: my $r = (^1); say ~(64,32,16 ...^ $r) 16:33
p6eval rakudo 668719: OUTPUT«Method 'shift' not found for invocant of class 'Range'␤ in 'infix:<...^>' at line 838:CORE.setting␤ in main program body at line 22:/tmp/3UG0Zw9Glp␤»
colomon no approach at all, or just no fast approach?
TimToady yes, that's because I want you to treat the Range as a scalar!!!
maybe Any.shift should be an identity 16:34
moritz_ colomon: there's always brute forcing, but that's ugly, and too slow in Perl 6
(2**22 would be fine in C, for example) 16:35
iterations, that is
colomon moritz_: okay, yes, I have verify experimentally that brute force is too slow in Rakudo. :)
*verified
TimToady rakudo: say ~(64,32,16 ... 1); # course, this works fine too
p6eval rakudo 668719: OUTPUT«64 32 16 8 4 2 1␤»
moritz_ rakudo: .say for [X~] (0, 1) xx 3 16:38
p6eval rakudo 668719: ( no output )
colomon TimToady: Range does Positional, so it will try to match it to the @rhs version of sequence op
moritz_ rakudo: .say for [X~] [0, 1] xx 3
p6eval rakudo 668719: OUTPUT«000 001 010 011 100 101 110 111␤»
moritz_ \o/
colomon I don't know why the rate one doesn't work.
s/rate/Rat/ 16:39
moritz_ the series code uses $limit.defined in a lot of places
colomon moritz_++ 16:40
hmmm. wonder if 16, 8, 4 ...^ Rat is a known bug? 16:44
TimToady well, the $r should have shielded the Positionalness in any case 16:48
TimToady lunch & 16:50
colomon moritz_: think I have a workable solution for the third problem (currently executing) 17:32
moritz_ colomon: what approach do you take? 17:33
colomon recursion
moritz_ I considered that too
colomon my first full run just ended, but I don't know if the result is correct yet or not.
moritz_ probably a good idea
colomon it's a lot faster because it is easy to knock out dead ends 17:34
moritz_ right, that was the reason not to brute force it with a [X~] [0, 1] xx 21 bitmaks 17:34
colomon moritz_: my answer was indeed right. 17:39
btw, [X~] [0, 1] xx 21 never occurred to me. I just counted from 1 to 2**22. ;) 17:40
moritz_ is 2236 correct? 17:41
nope
colomon nope
that's much too high
moritz_ sees that he counted lots of cases like 78 + 0 == 78 17:43
moritz_ when I remove those I get 200, but that's still wrong :( 17:47
much better now... 17:50
colomon I think you must have another instance of the same sort of error. ;) 17:54
moritz_ I've got the right solution now :-)
and I have a complete list of fulfilled equations along the way 17:55
colomon oooo, nice, I didn't try to make a list of equations. 17:57
moritz_ very helpful for debugging :-) 17:58
nopaste.snit.ch/24126 if you're interested in the code
colomon justrakudoit.wordpress.com/2010/10/...g-subsets/ 17:59
how fast is yours?
moritz_ checks
23s 18:00
probably faster if I skip debugging output, and passing along the already selected values as an array
sorear Good * #perl6!
colomon \o 18:01
moritz_ o/
colomon moritz_: interesting, yours seems to be a bit faster than mine. (at a guess, I've not run it locally to get a real comparison.) 18:01
moritz_ colomon: I put in some prior knowledge. Like that the last two numbers can never appear on the LHS of a sum 18:03
16s without the debugging code
but probabaly more important, I don't look for a specific target
but rather at each subsum I look if the sum appears in the array 18:04
rakudo: say 40.16 / 15.95 # time ratio on my machine 18:07
p6eval rakudo 668719: OUTPUT«2.51786833855799␤»
sorear seen tylercurtis 19:54
aloha tylercurtis was last seen in #perl6 1 days 6 hours ago joining the channel.
dalek ast: d8e4ca9 | moritz++ | S02-builtin_data_types/array.t:
[array.t] unfudge test for RT #63350, also add a similar one for $()

Also covers RT #78284, which seems to be mostly a duplicate of 63350		 20:40

Powered by Cro and the Raku® Programming Language.

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

Thank you!