»ö« | perl6.org/ | nopaste: paste.lisp.org/new/perl6 | evalbot usage: 'perl6: say 3;' or rakudo: / pugs: / std: , or /msg p6eval perl6: ... | irclog: irc.pugscode.org/ | UTF-8 is our friend!
Set by moderator on 20 October 2009.
TimToady std: {{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{say(4)}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}} 00:04
p6eval std 28943: OUTPUT«Deep recursion on anonymous subroutine at Cursor.pm line 677.␤Deep recursion on anonymous subroutine at Cursor.pm line 831.␤Deep recursion on anonymous subroutine at Cursor.pm line 987.␤Deep recursion on anonymous subroutine at Cursor.pm line 987.␤Deep recursion on anonymous
..subrou…
TimToady ENOTYET
00:06 tylerni7 joined
cj wow. hi ;) 00:07
diakopter cj: see my /msg to you 00:08
std: {{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{say(4)}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}} 00:28
p6eval std 28944: OUTPUT«ok 00:02 120m␤»
00:31 cotto_work joined 00:33 orafu joined 00:35 Whiteknight joined 00:36 mrsaturns joined, FCO joined
mrsaturns hey again 00:36
00:46 Exodist joined 00:55 xenoterracide joined 00:58 xenoterracide joined 01:02 nbrown joined
carlin moritz_: (From the backlog) what's the failure in IO-Socket-INET.t? All tests there should pass (and do here). 01:05
moritz_: *doh* ... of course ... nevermind. 01:07
01:11 xenoterracide joined 01:14 brunov joined 01:44 quuxx joined 01:51 KyleHa joined 02:13 agentzh joined
dalek p-rx: 501baec | pmichaud++ | src/NQP/ (2 files):
[nqp]: Add statement_prefix:sym<INIT> .
 02:13
p-rx: 773419c | pmichaud++ | t/nqp/33-init.t:
Add test for INIT statements.
p-rx: d8f4a95 | pmichaud++ | build/PARROT_REVISION:
Bump PARROT_REVISION.
03:25 eternaleye joined 03:26 perlsyntax joined
perlsyntax hi 03:27
i got a question how do i start out in perl 6 with like this #!/usr/bin/perl?
__ash__ perlsyntax: whats the issue now? your not sure how to start perl6? 03:33
perlsyntax right 03:34
__ash__ did you build it?
perlsyntax yes
__ash__ and do make install?
perlsyntax yes
__ash__ perl6 should start it then
perlsyntax just type perl6 03:35
__ash__ or #!/usr/local/bin/perl6 if you didn't change the prefix
yeah, assuming /usr/local/bin is in your path, it should be in there
perlsyntax youok 03:36
ok
03:36 meppl joined
__ash__ if you start it without a file it just starts a little command prompt you can write in, it only lets you do single line functions though 03:37
perlsyntax cool
then how do i compile a perl 6 file
i newbie
__ash__ um, you don't? hehe, its not required, they are run through the interpreter 03:39
perlsyntax __ash__Maybe i make some cooltools for perl 6
__ash__ just do: perl6 /path/to/file.pl
perlsyntax i see
cool
i think i going to like perl 6 03:40
__ash__ or, if your in a unix system, chmod +x the file, then do the shebang to perl6 as the first line of your file
that looks something like: #!/usr/local/bin/perl6
perlsyntax ok i see 03:41
i use fedora 11
__ash__ then you can just run the file itself
perlsyntax cool 03:42
03:42 frew joined
__ash__ have you ever used perl before? 03:43
perlsyntax i use perl 5 sometimes. 03:44
i know how to run a perl 5 program.
i run a perl6 file if i am right. 03:45
__ash__ perlgeek.de/blog-en/perl-5-to-6/ might be useful for you then
perlsyntax ok 03:48
i made a hello script hope i can run it.
lol
__ash__ perlcabal.org/syn/ is a great place to look at the current documentation/spec of the language
03:48 envi^office joined
__ash__ woot, step 1 for any real language, hello world 03:49
perlsyntax yepper
thanks
i can't find perl6 is this odd. 03:54
03:57 perlsyntax left
__ash__ how did you install it? 04:03
oh, nm he left
04:04 eternaleye joined, jaldhar joined
dalek p-rx: acc9532 | pmichaud++ | src/NQP/Grammar.pm:
[nqp]: Allow regexes to take parameters.
 04:22
p-rx: e259ff0 | pmichaud++ | src/stage0/ (3 files):
Update bootstrap to include parameterized regexes.
04:23 xenoterracide joined 04:38 felipe joined 04:44 Schwern joined 04:48 xenoterracide joined 04:54 mepplock joined 04:56 xenoterracide joined
pugs_svn r28945 | pmichaud++ | [misc/pm.txt]: Another query. 05:34
r28945 | Pm-7: S05 says that a match's reduction object is given by the
r28945 | C<:action> parameter, STD.pm's initparse has C<:actions> and
r28945 | $*ACTIONS. Should we pick one and stick with it? (PGE and
r28945 | Rakudo have traditionally used C<:action> -- if there's to
r28945 | be a change, now would be a good time for it. :-)
05:49 rlarson87 joined 05:51 Schwern joined, iblechbot joined
pugs_svn r28946 | pmichaud++ | Add a comment to Pm-7. 05:59
06:05 eternaleye joined
dalek p-rx: 67e897d | pmichaud++ | src/ (2 files):
Use $*ACTION contextual instead of $!action cursor attribute (from STD.pm)
 06:07
p-rx: 414f6a0 | pmichaud++ | src/stage0/ (3 files):
Update bootstrap.
06:10 xenoterracide joined 06:15 justatheory joined 06:18 xenoterracide joined 06:24 NorwayGeek_ joined 06:36 c1sung left 06:42 c1sung joined 06:49 Bzek joined
bpetering PerlJam: happy birthday! :) 06:56
07:40 colomon joined, ejs joined 07:50 ejs1 joined
moritz_ good morning 07:54
07:54 rblasch joined 08:02 mariuz joined
colomon moritz_: Should we have defined checks (like your sign) for all the trig functions, etc? 08:04
moritz_ colomon: not really 08:05
the spec explicitly mentions that sign(undef) is undef 08:06
pugs_svn r28947 | moritz++ | [t/spec] unfudge and correct sign() tests
moritz_ it doesn't for other functions
colomon Huh. That seems like an odd detail... 08:07
moritz_ aye 08:08
my goal yesterday was just to move it to the setting, so that closing RT #60674 would become easy 08:09
colomon Guess it's an odd detail I'm perfectly happy not to worry about. :)
moritz_ and I have a patch for that now
colomon moritz_++ # moving things to the setting 08:11
pugs_svn r28948 | moritz++ | [t/spec] test for RT #60674, sign(Complex) 08:13
dalek kudo: 27e0d69 | moritz++ | src/setting/Complex.pm:
sign(Complex) fails. Closes RT #60674
 08:17
colomon Okay, looks like I have successfully installed Ubuntu in a VM on my MacBook, time to go back to bed.
moritz_ good night colomon 08:18
colomon err.... why did you define sign(Any) and sign(Complex)? Isn't the former enough?
moritz_ no 08:19
colomon (And why aren't Complex undefs special-cased?)
moritz_ sign(Any) calls $x.Num.sign
$x.Num fails
and Failure.sign barfs 08:20
colomon ah, instead of $x.sign. I see.
moritz_ well, I guess it's sub-optimal the way I did it
what do you mean by Complex undefs? Complex.sign? or undef + undef\\i? 08:21
colomon rakudo: my Complex $x; sign($x);
p6eval rakudo 8cf27e: ( no output )
colomon rakudo: my Complex $x; say sign($x)
p6eval rakudo 8cf27e: OUTPUT«Use of uninitialized value␤␤»
colomon rakudo: my Complex $x; say $x.WHAT; say $x
p6eval rakudo 8cf27e: OUTPUT«Complex()␤Complex()␤» 08:22
moritz_ but it seems to dtrt
colomon yeah, interesting.
rakudo: my Complex $x; say $x.WHAT; say defined($x)
p6eval rakudo 8cf27e: OUTPUT«Complex()␤0␤»
colomon needs sleep. 08:23
moritz_ nqp: grammar A { token TOP { (.) $0 } }; say(A.parse('bb'))
p6eval nqp: OUTPUT«bb␤»
moritz_ colomon: good night
nqp: grammar A { token TOP { (.) $1 } }; say(A.parse('bb'))
p6eval nqp: OUTPUT«b␤»
moritz_ pmichaud: that doesn't look right ;-)
nqp: my $x := 'b'; grammar A { token TOP { (.) $b } }; say(A.parse('bb')) 08:24
p6eval nqp: OUTPUT«␤»
moritz_ rakudo: our Int @A1; our @A2; @A2[0] = "Str" 08:25
p6eval rakudo 8cf27e: TIMED_OUT
08:26 Schwern joined
moritz_ rakudo: our $c; class T { $c++ }; T.new; say $c 08:27
p6eval rakudo 8cf27e: TIMED_OUT
moritz_ rakudo: sub foo(Int *@a) {}; my Int @b; foo(|@b) 08:28
p6eval rakudo 8cf27e: TIMED_OUT
moritz_ rakudo: say 5.Num.sin('degrees'); 08:36
p6eval rakudo 27e0d6: OUTPUT«0.0871557427476582␤»
moritz_ rakudo: say 5.0.sin('degrees'); 08:37
p6eval rakudo 27e0d6: OUTPUT«0.0871557427476582␤»
moritz_ rakudo: say (5 * 1.0).sin('degrees');
p6eval rakudo 27e0d6: OUTPUT«0.0871557427476582␤»
moritz_ nqp: say 3 ?? 5 : 6; # intentionally wrong 08:40
p6eval nqp: OUTPUT«Confused at line 1, near "say 3 ?? 5"␤current instr.: 'parrot;HLL;Grammar;panic' pc 11756 (gen/hllgrammar-actions.pir:1131)␤»
moritz_ rakudo: (1..3).map({$_ => $_*$_}).perl.say 08:41
p6eval rakudo 27e0d6: TIMED_OUT 08:42
08:43 SmokeMachine joined 08:55 [1]rlarson87 joined
moritz_ rakudo: my $r = 1..9; $r.pop; say $r.perl 09:00
p6eval rakudo 27e0d6: OUTPUT«1..8␤»
09:01 Juerd joined, pmichaud_ joined 09:02 hatsefla1s joined, dukeleto joined, dalek joined
moritz_ down to 520 tickets, 4 of which (at least) can be closed if the testing status is satisfactory 09:04
09:31 JimmyZ joined
JimmyZ rakudo: my $r = 1..*; $r.pop; say $r.perl; 09:32
p6eval rakudo 27e0d6: OUTPUT«Method 'pop' not found for invocant of class 'WhateverCode'␤in Main (file src/gen_setting.pm, line 295)␤»
JimmyZ rakudo: my $r = 1..*; say $r.perl;
p6eval rakudo 27e0d6: OUTPUT«{ ... }␤»
09:35 synth joined 09:45 [1]rlarson87 joined 09:47 am0c joined 10:02 Whiteknight joined 10:04 Su-Shee joined
Su-Shee hi all. 10:05
moritz_ helo
10:21 rgrau joined 10:29 alfieANDfred joined 10:53 agentzh joined 10:58 buubot joined 10:59 Su-Shee_ joined 11:00 dakkar joined 11:12 jnthn joined
jnthn o/ 11:13
11:18 KatrinaTheLamia joined 11:19 ejs joined
mathw oh hai jnthn 11:21
11:34 NorwayGeek joined 11:55 payload joined 12:07 meppl joined 12:12 payload joined, masak joined
masak good not-even-pretending-it's-early, #perl6. 12:12
jnthn lolitsafternoon
masak orz 12:13
jnthn masak++ # making me feel better about the time I got up
masak jnthn: :)
12:14 payload joined
takadonet morning all :P 12:14
masak takadonet: \\o
takadonet already changed Parrot version! I just got rakudo updated yesterday! 12:15
jnthn takadonet: Sorry, we've been migrating onto the new Parrot calling conventions bit by bit, and I've had to patch Parrot a couple of times along the way too, so we've been chasing those patches. 12:16
takadonet I know jnth, been following commit bits and this channel 12:17
takadonet wonders his 3 week old rakudo bioperl6 will work....
moritz_ parsing got slower, running got faster 12:18
so hopefully an overall win for number and string crunshing
takadonet I just hope it works 12:19
we do not have lazy parsing of files using a grammar yet right?
moritz_ right 12:20
pmichaud_ good morning, #perl6
jnthn lol
moritz_ good morning pmichaud :-)
jnthn Somebody who actually *can* say good morning. ;-)
12:20 masq joined
masq Hello. 12:21
moritz_ hi
Su-Shee hi masq :) 12:23
moritz_ btw I turned my LTM notes into an article: perlgeek.de/en/article/longest-token-matching 12:25
pmichaud *click*
moritz_ thanks to everybody who provided feedback
pmichaud moritz++
takadonet YA! new article from moritz_ :) 12:26
moritz_++
moritz_ pmichaud: if you find anything that you understood differently, please let me know 12:28
pmichaud moritz_: I didn't see anything that I disagreed with
moritz_ great 12:29
diakopter wakes up, sorta 12:35
12:42 payload joined, brunov joined
moritz_ nqp: say('pmichaud++') 12:42
p6eval nqp: OUTPUT«pmichaud++␤» 12:43
pmichaud nqp: pir::say('pmichaud++')
p6eval nqp: OUTPUT«pmichaud++␤»
moritz_ nqp: say(' pmichaud++ ') # I fear it doesn't work without spaces, since lambdabot is too dumb
p6eval nqp: OUTPUT« pmichaud++ ␤»
masak wow, an nqp mode! 12:44
moritz_ and it's nqp-rx under the hood
and fast
pmichaud nqp: say(pir::substr('perl6',4))
p6eval nqp: OUTPUT«6␤»
pmichaud nqp: say(pir::substr('perl6 rocks',6,4))
p6eval nqp: OUTPUT«rock␤»
pmichaud yay.
pugs_svn r28949 | Kodi++ | [perl6.org/source/documentation] Added a link to Rosetta Code. 12:46
moritz_ I love it when people contribute to projects I started :-) 12:47
masak aye, that's a nice feeling. 12:50
12:52 JimmyZ joined, huf_ joined 12:53 ejs joined
masak more rakudo muddling from a fellow use.perl-er: use.perl.org/~lilstevey/journal/39815 12:53
moritz_ masak++ # community radar
masak moritz_: I could say the same about you, and your excellent presence on PerlMonks. :) 12:54
jnthn starts to piece together nested siggies 13:00
masak yay, nested siggies! 13:01
jnthn Was going to postpone them until the new grammar, but the parsing of them looked like an easy thing anyway, and the majority of the effort is in stuff that will stay the same.
diakopter moritz_: "A token that also switches on the :ratchet modifier is called a rule." (from book) 13:02
don't tokens already have the :ratchet modifier turned on?
jnthn Maybe meant :sigspace
pmichaud diakopter++ 13:03
masak aye.
and I wouldn't phrase it like that, because tokens that turn on :sigspace are still tokens, not rules.
jnthn Aye, it's more just twiddling the defualts. 13:05
pmichaud "A rule is a token that has the :sigspace modifier enabled" 13:06
or
"A rule is a regex with both :ratchet and :sigspace enabled"
masak nod.
moritz_ diakopter: that's corrected in the source already
diakopter oh 13:07
moritz_ pmichaud++ caught it, iirc
diakopter was reading the pdf
masak man, it took me ages to learn which one of regex/token/rule was which. :)
moritz_ it's actually quite simple
a regex is just... a regex 13:08
and a token is what you use to match the simplest thing possible
such as \\w+ or so
and don't want to backtrack
which just leaves rule for the third option ;-)
if you wrote a grammar, token is very intuitive 13:09
masak nod.
it was in the process of learning to write grammars that I repeatedly chose the wrong one of those, and had to learn which one to use.
moritz_ thinks the JSON parser never backtracks 13:10
diakopter what generated that .pdf? can it be automated/scheduled?
moritz_ it can
and it's a good idea
13:10 [particle] joined 13:11 huf joined
moritz_ moritz.faui2k3.org/tmp/book.pdf # current version 13:11
13:14 NorwayGeek_ joined 13:16 [synth] joined
diakopter to support 'make install' of multiple perl6 implementations, will rakudo install a 'rakudo' binary... and then perhaps another make target (perhaps named make install_as_default_perl6) that creates a symlink (or on Windows, perl6.cmd containing "rakudo %*") to rakudo? 13:19
pmichaud diakopter: hasn't been completely decided yet 13:20
jnthn Wouldn't you just install them to different directories?
And the one that's first in your path wins as the default?
Or does the *nix "omg no I want all my apps in one folder" mentality prevent this? :-) 13:21
diakopter sure, but for $PATH and %PATH% resolution, there's not currently a way to call 'sprixel' as "perl6" while also having rakudo available on the path
13:22 morphic joined
jnthn I dunno. I guess people manage multiple versions of Perl 5 today somehow. I guess that's one precedent we could follow. 13:22
diakopter I'm thinking for /etc/alternatives
a la java/javac
jnthn I think most people will expect make install to give them a Perl 6 binary.
*perl6 binary
I'd suggest we provide an alternative target for people who want it installed as something different than that. 13:23
And keep what most people want as the Common Case.
(e.g. make install)
Perhaps with a "btw you already have a perl6 binary in place, are you sure" style warning. 13:24
If we're about to scribble over another one.
(e.g. one that isn't an older Rakudo.) 13:25
diakopter ok; I can see that. so a debian packaging of rakudo would install it to /usr/bin, but as /usr/bin/rakudo, and then use alternatives to handle multiple impl of /usr/bin/perl6 ?
13:25 rdice joined
jnthn diakopter: Well, if that works for Debian, they can do that. 13:26
diakopter: My platform doesn't have a /etc/alternatives, so it's kinda irrelevant.
And mostly an implementation detail. We more need a guideline that each platform can apply in whatever way is best for that platform. 13:27
diakopter yes, but your platform also doesn't 'make install' to %SYSTEM_ROOT% :P
jnthn True, though one of them will still be in the path ahead of others. 13:28
diakopter on windows the default "make install" doesn't necessarily (or ever, by default) install perl6.exe to a directory that's on the %PATH% 13:29
but my point is that on the rest of the platforms, it does
13:29 SmokeMachine joined 13:30 cj joined, cj_ joined
jnthn diakopter: for better or worse, yes, true. 13:31
diakopter: I think that most people who are building stuff and doing "make install" will expect to get a perl6 binary at the end of it.
pmichaud ...unless we change their expectations such that they get 'rakudo' 13:32
I wonder what python does with its various implementations
[particle] i imagine the same thing AS perl does 13:33
provide a checkbox that asks if you want to modify path and other env settings
diakopter probably same as discussed; alternatives/symlinks to provide the default on *n*x, and manual %PATH% mgmt on windows..
[particle] (above for windows) 13:34
pmichaud Parent isn't a Class.
(error message from "add_parent".)
That's a LTA error message.
"What parent? What class?"
13:38 barney joined, NorwayGeek joined
jnthn pmichaud: means the argument to add_parent wasn't an instance of the Class PMC. 13:39
pmichaud I know what it means. :)
But I don't know which class I'm attempting to add_parent on.
jnthn Ah.
pmichaud and since there are a bunch of classes floating about, it's not helpful to tell me which one is failing or why.
jnthn aye 13:40
13:40 __ash__ joined
jnthn Probably easily patched. 13:40
I'd offer to do it, but I have Slovak class now. 13:41
pmichaud I'll live with it for now.
13:41 TSa joined 13:42 morphic left
pugs_svn r28950 | moritz++ | [t/spec] test for RT #69424, typed slurpy params should accept typed arrays 13:48
pmichaud is that spec?!?
moritz_ that's common sense ;-) 13:49
pmichaud I'm not so sure it is
but okay
I'm still wondering exactly what it means to have a "typed slurpy".
13:50 payload joined
moritz_ if (Int *@a) doesn't accept a Positional[Int], what would it accept at all? 13:50
pmichaud Having (Int @a) accept a Positional[Int] makes sense to me
but slurpies don't bind to single params
sub foo(Int *@a) { ... } ; foo(1,2,3,'a'); 13:51
moritz_ unless they do ;-)
pmichaud should that fail?
moritz_ probably
pmichaud sub foo(Int *@a) { ... } ; foo(1,2,3); # how about this ? 13:52
moritz_ and if I understood TimToady correctly also lazily
not sure about that
pmichaud i.e., where's the Positional[Int] in that case?
jnthn bbiab - slovak (sorry)
would join slurpy discussion otherwise, since I have thoughts on that :-) 13:53
laters...
moritz_ let's wait from jnthn that, at least I am in no particular hurry
pmichaud okay
I'm fine if we have type constraints on slurpies -- I'm just not at all certain about the mechanics.
And it becomes particularly weird if people start passing junctions into slurpies. 13:54
diakopter or captures
moritz_ I think the decree was that junctions don't autothread over slurpies
but I could be wrong here
pmichaud right
so then
sub foo(Int *@a) { ... }; foo(1,2,3|4); # fails? 13:55
moritz_ resolves RT #69424 anyway ;-)
13:55 colomon joined
moritz_ pmichaud: either fail, or it might just interpolate the eigenstates 13:55
pmichaud interpolate eigenstates would be wrong, certainly
moritz_ speaking of junctions (mostly unrelated)... I found them very useful if used sparingly 13:57
Patterner like salt?
13:57 rgrau joined
moritz_ try to use them as fundamental part of a design seems almost always wrong 13:58
Patterner: exactly
mathw my best use for them is things like if $foo == any(1, 2, 3)
moritz_ my best use for them is in filtering things
mathw so I use them as syntactic sugar, really
moritz_ whenever you you have a matching condition 13:59
suppose you write sub grep($matcher, @things) { gather for @things { take $_ if $_ ~~ $mather } }
then it's a very nice option to pass /foo/ & /bar/ as $matcher
(uhm, make that Object $matcher, maybe) 14:00
and since smart-match is really smart, you can use a closure as the matcher, even if the author never thought of that 14:01
mathw smart match really is incredibly smart 14:02
masak my favourite use of junctions so far is in this post: use.perl.org/~masak/journal/38924
the line `return [@slots] if all @slots;` suggested by moritz_++ 14:03
use.perl.org/~masak/journal/38924 [15:03]
ERC> the line `
urgh.
ERC--
moritz_ detects a fail by a non-irssi user and is happy ;-) 14:04
masak :)
I managed to confuse ERC. it's never happened before. not even sure how I did it.
anyway `return [@slots] if all @slots;` reads extremely well. 14:05
14:05 silug joined
mathw it does 14:05
this is something I've always liked with Perl
14:05 nnunley joined
mathw statement modifier conditionals 14:05
also being able to use 'and' and 'or' as conditional expressions
open my $filehandle, '<', $filename or die "You suck at specifying valid filenames" 14:06
etc.
moritz_ lol
mathw further awesome in Perl 6 14:09
compare this line of C#
DataGridViewComboBoxColumn combo = new DataGridViewComboBoxColumn();
you know that's so much nicer in Perl 6
[particle] .= nicer
mathw my DataGridViewComboBoxColumn $combo .= new()
why *should* you have to write the type name twice??
diakopter well, it's var combo = new DataGridViewComboBoxColumn();
to be fair
mathw diakopter: depends which version of C# you can use 14:10
14:11 iblechbot joined
diakopter ok; same with perl :P 14:11
14:16 envi^home joined 14:20 kidd` joined 14:26 lichtkind joined 14:31 nothingmuch joined 14:40 nihiliad joined 14:46 ejs1 joined 14:48 NorwayGeek_ joined
colomon rakudo: say (2 * (1, 2, 3, 4)).perl 15:03
p6eval rakudo 27e0d6: OUTPUT«8␤»
15:03 justatheory joined
colomon rakudo: say (0 xx 8).perl 15:03
p6eval rakudo 27e0d6: OUTPUT«[0, 0, 0, 0, 0, 0, 0, 0]␤»
15:05 brunov joined 15:08 Psyche^ joined
colomon masak: ping? 15:14
masak colomon: pong. :)
colomon I'm trying to understand your second Langford script. 15:15
I think I understand your approach, but the details are puzzling me.
masak colomon: it'll take me a while to remind myself of what I was doing, but go on.
colomon I think it's the [@slots] and push @found that are confusing me the most. 15:16
If all @slots, you've got a solution, right?
masak the [] around the @slots is just to create an array instead of a list.
colomon: yes.
colomon is chatting while his shiny new Ubuntu VM installs kcachegrind in the background... 15:17
array instead of list: why?
mathw This is something I don't really understand
what is the difference between an array and a list
masak colomon: ISTR because the list would flatten, and then it wouldn't work. 15:18
mathw: a list is immutable, an array is mutable. 15:19
colomon ooo, nice distinction. 15:20
Is @found an array of all possible solutions, then?
lichtkind are there any attempt to unify all the pathtools in one module and egt rid of Cwd File::Spec ...?
masak colomon: ah, there you go. @found needs to be an array of arrays.
colomon: if you were to return a list, it would get mixed with other solutions. 15:21
mathw: also, lists flatten, arrays don't.
well, arrays flatten in certain special circumstances, such as in push @a, @b
colomon So you do a depth first search and return all the solutions you find?
masak colomon: something like that. 15:22
colomon masak++
masak well, both solutions do that, but the second one does it better.
colomon BTW, @new-slots[$k, $k+$c+1] = $c, $c; both delights and terrifies me. 15:23
masak jnm has an even better solution in the comments, IIRC.
colomon masak: the second solution does a smart depth first search, yeah.
jnthn loli'mback
masak jnthn++ # iz back!
jnthn lolmasakblogged! 15:24
masak um. no :) 15:25
jnthn oh wait
I'm sure I saw a link
masak I only thought about it...
that was moritz_++
jnthn aww
masak jnthn: oh, the link to the old blog post.
jnthn yeah
moritz_ didn't blog either
since friday
so it's time agian
jnthn I got excited and thought it was a new masakpost
masak jnthn: :) 15:26
moritz_ so what should I blog about?
masak moritz_: nested signatures!
and why they rock.
moritz_ masak: but... they don't work yet, do they?
jnthn masak: ETOOSOON :-P
moritz_ topic rejected, for now
masak :)
15:29 araujo joined
dalek p-rx: 7d38dff | pmichaud++ | src/ (4 files):
Add <LANG> subrule for handling language braids and actions.
 15:30
p-rx: 3d8df04 | pmichaud++ | build/Makefile.in:
Slight reordering of build sequence.
p-rx: ef8e140 | pmichaud++ | (2 files):
PBC refactor #1 -- create Regex.pbc .
p-rx: ca98719 | pmichaud++ | (7 files):
More .pbc refactoring and bootstrapping.
p-rx: 7515ff5 | pmichaud++ | (3 files):
More library refactoring, final stage of bootstrap.
p-rx: d526e5f | pmichaud++ | build/Makefile.in:
Eliminate HLLGrammar.pbc from build -- it's now HLL.pbc .
colomon Wow, pmichaud++ makes me feel very unproductive... 15:33
jnthn Me too!
Time to get back to hacking on nested siggies.
15:34 drbean joined
colomon Using kcachegrind with parrot... why is perl6;!TOPERL6MULTISUB using up 201% of the time? 15:34
oh, wait, that's not the total time... maybe because it's in a cycle? 15:35
moritz_ colomon: it's confused with CPS control flow
colomon I find I'm generating very pretty graphs with no idea what they mean. :) 15:36
jnthn hehe
But they're *pretty*.
pmichaud right now I'm refactoring so that I can put contextuals in regexes 15:37
but it may also give me a huge number of wins on the regex engine as well 15:38
we shall see :)
moritz_ really misses parameterized regexes in rakudo
pmichaud nqp-rx has them!! :-) 15:39
moritz_ I know
masak \\o/
jnthn moritz_: We'll get 'em for free my stealing the nqp-rx grammar.
pmichaud I may also have variables and closures in regexes working before the day is out.
jnthn wow
pmichaud (very likely -- not hard to do at this point, just so many low-hanging-fruit items to work on)
[particle] how much pain will there be in converting rakudo actions to the new std-based grammar? 15:40
pmichaud there will be pain
I think we'll basically start over.
(actions, anyway)
moritz_ enough for a masochist ;-)
masak where do I sign up?
jnthn Well, in a sense, there won't be pain converting them because we won't be converting them. ;-)
pmichaud many things that we handle bottom-up in rakudo's current grammar/actions are handled top-down in STD.pm
jnthn There'll just be pain building the whole thingy up again.
pmichaud on the plus side, I've found it to be a very enjoyable pain in nqp-rx 15:41
it's like "yes, I already did this once before.... but this is WAY cooler"
[particle] ok, so let me ask a different question, then...
TimToady Heinlein: out of the gravity well is halfway to anywhere
[particle] how much pain will there be to implement STD (parser only) based on nqp-rx?
pmichaud nqp-rx is based on STD, and it's pretty straightforward. 15:42
[particle] straightforward is good.
pmichaud as before, our goal isn't to run STD exactly, but to get a close enough approximation that we can start to converge
jnthn For the usual Perl 6 definition of straightforward. ;-)
moritz_ ;-)
jnthn Oh hey, that looks like I've got the code gen for nested signatures almost right... 15:43
TimToady
.oO(forward == all directions at once)
moritz_ not sure how good our test coverage is for nested signatures...
probably rather bad
jnthn Probably.
Given I don't think any implementation ever did 'em before. 15:44
pmichaud (straightforward) -- for example, I've just been able to eliminate a few more PIR-based "cheats" and replace them with methods written in NQP.
[particle] pmichaud: yes, that's a good goal. i'm thinking of tuits i might have, and ways to apply them
but, that said, i've just failed to build rakudo with msvc AND strawberry. sigh. 15:45
pmichaud [particle]: well, a couple of weeks ago I was starting on a command line argument processor....
I started to do it in PIR, then said "ugh. I want to write this in NQP."
jnthn [particle]: With msvc++ you failed?
[particle] jnthn: yes
jnthn [particle]: Try upgrading to XP, it works great here.
;-)
[particle] mt.exe : general error c101008d: Failed to write the updated manifest to the resource of file "perl6.exe.manifest". The binary is not a valid Windows image.manifest embedding failed
jnthn Why are we even invoking mt.exe? :S 15:46
[particle] why is it trying to embed a manifest?
jnthn I don't know. Does it not build an EXE?
Or does it want that before building the EXE?
pmichaud I suspect something to do with pbc_to_exe
jnthn Ah, maybe.
pmichaud I've noticed that it builds something called a "manifest" on my system.
jnthn Odd.
:-/
pmichaud and I have no clue what it is.
[particle] yes, pbc_to_exe seems the culprit 15:47
pmichaud pmichaud@orange:~/nqp-rx$ make nqp
/home/pmichaud/nqp-rx/parrot_install/bin/pbc_to_exe nqp.pbc
...
Linked: nqp.manifest
dalek p-rx: 3428caa | pmichaud++ | (4 files):
[p6regex]: Eliminate src/cheats/p6regex-grammar.pir, they're now P6 source.
 15:48
jnthn huh, isn't \\(1,2,3) meant to create a capture?
Oh, wait, that's not prefix:\\ 15:49
[particle] so, i need to get back into perl6 development somehow
moritz_ [particle]: look for RT tickets with the LHF (low-hanging fruit) marker on them ;-)
jnthn rakudo: say (\\(1,2,3)).WHAT 15:50
p6eval rakudo 27e0d6: OUTPUT«List()␤»
jnthn pmichaud: ^^ That's wrong, no?
pmichaud jnthn: what should it be?
[particle] moritz_: i'm wondering if converting STD.pm to nqp-rx is LHF
jnthn I was expecting a Capture?
pmichaud I'm not sure how visible Captures are 15:51
jnthn Are my expectations wrong?
pmichaud they tend to hide themselves for the item
jnthn Hm
pmichaud for example, if I have
my $a = \\%hash;
and I ask for the type of $a, I somewhat expect 'Hash'
jnthn Ah
'cus in item context they turn into a single item...
moritz_ jnthn: my experience is that \\(...) doesn't do really much useful in rakudo
jnthn OK, how do I construct a Capture with a bunch of items?
pmichaud in current Rakudo, or in Perl 6? 15:52
jnthn Perl 6
Well, and in current Rakudo.
pmichaud keep in mind that they're different
and the spec is incomplete
iiuc, (1,2,3) now constructs a Parcel
and there's likely some way to convert a Parcel into a Capture
jnthn .Capture I guess. 15:53
pmichaud I dunno
anyway, Rakudo doesn't presently have a way of constructing an object that is truly of type "Capture"
jnthn OK.
pmichaud because the spec and tests have been ambiguous on that particular point.
jnthn I probably need to change that now.
e.g. by implementing .Capture for a few types.
pmichaud I'd prefer to see a clear spec before we do too much.
somehow I doubt that the answer is .Capture 15:54
(not only "doubt", but "highly doubt")
jnthn OK, what I'm *actually* wanting to implement is coercion of something else into a Capture.
Given .Int gives us an Int, .Str gives us a Str, etc, it follows that for that, .Capture gives a capture. 15:55
pmichaud Yes, I follow. I just don't know the answer from a "real P6" perspective.
why wouldn't .Capture simply be the same as \\(...) though ?
[particle] wait for a 'real p6' to come out, and see.
TimToady because that's just a non-flattening parcel
not a cpature 15:56
*capture
jnthn I'm thinking about this in the context of nested signatures.
pmichaud so how does one "make" a capture ?
TimToady typically, by using a parcel as an argument list, but that doesn't rule out an explicit conversion
jnthn Last time this was discussed, the answer was that we took the thingy to bind to a nested signature and coerced it to a Capture.
pmichaud if we ask a capture for its type, what do we get? 15:57
TimToady yes, binding turns parcels/whatevers into captures
jnthn e.g. sub foo([$head, *@tail]) { ... }; my @a = 1,2,3; foo(@a); # the binder does something like @a.Capture, and binds the resulting capture agaisnt the nested signature
pmichaud not @a.Capture
jnthn Why?
pmichaud it would be the parcel containing @a
not @a itself 15:58
jnthn By the time the binder sees this, we've already capturized the original parcel.
pmichaud because foo(@a, 3) constructs a capture from (@a, 3), not from @a and 3 separately
jnthn IIUC
OK, putting it another way:
sub foo([$head, *@tail]) { ... } is like sub foo(@temp) { my :($head, *@tail) := @temp.Capture; } 15:59
pmichaud I think that's wrong
jnthn Well, not .Capture there since binding against a signature implies the coercion of the RHS to capture.
pmichaud it's like
moritz_ thought jnthn was doing nested signatures, not signature unpacking 16:00
pmichaud sub foo(|$capt) { ... }
TimToady same thing, really
jnthn moritz_: Waht's the difference?
pmichaud jnthn: okay, I see what you're getting at
moritz_ jnthn: in my mind nested signatures are when you have a Callable in a signature, and that thing has a signature again, and you do multi dispatch on that
jnthn pmichaud: No, that'd capture *all* the arguments. 16:01
pmichaud: We want to unpack a single positional.
pmichaud jnthn: still, .Capture seems wrongish to me, at least until TimToady++ declares it to be right-ish
TimToady pmichaud: that grabs the supercapture, but [] grabs the first arg of that
pmichaud right, I get it.
I see the difference.
jnthn Sorry, an example with multiple parameters woulda been clearer.
TimToady and [] is really just sugar for :() that emphasizes the positionalness 16:02
jnthn TimToady: I'm assuming that we put the thing passed into list context before binding the nested signature too?
pmichaud I'm just wondering how visible Captures tend to be. i.e., what would ~(@temp.Capture).WHAT produce?
TimToady jnthn: that always depends on the signature in question 16:03
jnthn TimToady: In the case of :([$head, *@tail])
16:03 drbean_ joined
pmichaud most of the test suite tends to want captures to be highly invisible creatures 16:03
TimToady that implies list to @tail, but to nothing else 16:04
jnthn TimToady: ah, OK.
TimToady: That makes sense.
TimToady if you bind @foo it's an object
16:05 desertm4x joined
TimToady basically, only * does list context 16:05
jnthn TimToady: So are you saying :([$head, *@tail]) is really like :(\\$temp) and then :($head, *@tail) := $temp ?
dalek p-rx: 3a49ad1 | pmichaud++ | src/stage0/ (3 files):
Update bootstrap files.
pmichaud I need lunch here... bbiaw
jnthn kk 16:06
TimToady that seems one level off somewhere
maybe lose the :() around \\$temp
or use :(|$temp) 16:07
jnthn TimToady: Ah, I was meaning the whole signature, not that we nest the \\$temp in one.
TimToady okay
jnthn That is, sub foo([$head, *@tail]) { ... } would be more like sub foo(\\$temp) { ...unpack... }
TimToady then yes
jnthn OK.
So that bit is easy enough. :-) 16:08
TimToady in fact, you don't even need the \\
(I think)
since the later binding capturizes
jnthn Well, we'd enforce item context without it, but yes, maybe that isn't a problem actually.
pmichaud in some sense it feels like "capturize" is "binding"
i.e., it's binding that converts a parcel into a capture 16:09
jnthn pmichaud: Not really. The binder needs a capture to work on.
pmichaud: I mean, "obtaining a capture" is kinda a pre-requisite to the binding.
TimToady it's the phase of binding where we find the named args
pmichaud right
jnthn You could see it that way I guess.
TimToady if they aren't handy, we don't have a capture yet
pmichaud it's a phase of binding, pre-requisite to completing the bind, but not a pre-requisite to starting it
jnthn Well, depends what you are counting in as "the bind". 16:10
OK, let me put it another way.
pmichaud I'm trying to follow the meaning that TimToady++ is using here :)
16:08 <TimToady> since the later binding capturizes
16:09 <TimToady> it's the phase of binding where we find the named args
TimToady I basically mean everything that := does 16:11
pmichaud right
TimToady which includes binding :)
pmichaud I tend to think of "binding" as :=
even for signature binding
TimToady but I can see the other view
jnthn Where does the logic to know "I have a List and want it to use its elements as the positionals" or "I have a Hash and wnat to use it as my nameds" or "I have some Object and want to use its attributes as nameds"
TimToady in which case := is the coerce to capture and bind operator
jnthn belong?
TimToady which is a mouthful
pmichaud isn't it really "I have a parcel..." ? 16:12
as opposed to "I have a List" or "I have a Hash" or ....
TimToady usually
jnthn pmichaud: If I've passed in a Hash and expect to unpack it, then I have a Hash. :-) 16:13
moritz_ or "I have a dream" # SCNR
TimToady I don't have enough coffee (in me)
pmichaud jnthn: in that case, I'd tend to thinkof the Hash as being a parcel with one element
jnthn pmichaud: That's fine, but doesn't help me.
pmichaud: I still need to capturize it. 16:14
So I've got somewhere to pull the named arguments from.
pmichaud can't just pull the named arguments from the hash directly?
jnthn Well sure, if I am going to teach the binder to know to do that.
pmichaud I mean, Hash tends to support a .hash interface :-)
jnthn You're missing my point.
pmichaud Agreed.
jnthn Do I bake the logic to handle these things into the signature binder?
pmichaud oh. I think "yes". 16:15
Because I think of it as "binding", not just "signature binding"
TimToady hmm
pmichaud (perhaps that's inaccurate/naive, but that's my current mental model)
jnthn pmichaud: You see, I'm a tad uncomfortable doing that.
quuxx masak blogged "I'm in for an interesting month": use.perl.org/~masak/journal/39817?from=rss
masak quuxx: lolyounoticed! 16:16
TimToady if that's the primitive, then .Capture would have to be written my (|$cap) := self
seems a bit inside out
jnthn I mean, essentially baking a bunch of coercions into the binder feels odd.
pmichaud I tend to not think of the binder as doing coercions, but I see your point (and the self-contradiction) 16:17
what I do know is that every time I try to figure out the meaning of the ":=" operator, it ends up looking a lot like a signature bind.
jnthn Basically, I see two directions. 16:18
1) The binder looks at something it's got to bind against a signature and has a set of rules built in to it to know what to do. (e.g., it's is a Capture, use it, or this does Associative so use this to obtain nameds, etc) 16:19
2) The binder says "Ok, is this thing a Capture already? if not, do XXX to the thing I have to coerce it to a Capture" 16:20
I'd been under the impression that direction 2 was more the Right Way.
pmichaud I think that "coerce to a Capture" depends on context.
i.e., I can't decide if something goes in as named or position arguments until I know what I'm binding to
either that or there has to be something that explicitly capturizes things before sending them to the binder 16:21
TimToady I think you're getting confused between the supercapture and the subcaptures (if any)
pmichaud very likely.
anyway, I'm low-blood-sugar, so I'll get lunch, let you two work it out, and find out the result :)
jnthn In sub foo($n, [$head, *@tail]) { ... } 16:22
16:22 ejs joined
jnthn foo(42, @thingies) 16:22
Then the binder initially has the capture we got from the parcel (42, @thingies) to operate on. All well and good.
But then it sees it has a sub-signature, or nested signature, or whatever you like to call it.
TimToady that's the supercapture
there are no subcaptures (yet) 16:23
jnthn It needs to take @tail, which it got out of the supercapture, and bind it against the nested signature.
moritz_
.oO( too many positional arguments: 3 passed, 3 expected )
what an awesome error message :)
TimToady jnthn: I think you meant @thingies
16:24 cdarroch joined
jnthn TimToady: Yes, sorry, I did. 16:24
TimToady at that binding point, it needs to turn @thingies into a capture
jnthn Right!
That's what I've been trying to argue. :-)
pmichaud question:
jnthn So what's the mechanism by which @thengies gets turned into a capture? :-)
pmichaud what if it was called as foo(42, $thingies)
and $thingies is an Array 16:25
TimToady whether that is done inside of "binding" outside is immaterial, except to the implementors )
*or outside
pmichaud would having $thingies change the bind ?
jnthn TimToady: It's material in the sense of, should we be providing people a way to specify how their type coerces to a Capture?
TimToady no, in either case the second arg is not *, so in item context
jnthn pmichaud: There may be an argument that [...] implies a type check for Positional. 16:26
pmichaud jnthn: in my case, $thingies does Positional, yes.
because $thingies is an Array.
jnthn pmichaud: OK, then I don't think it matters.
pmichaud my question is whether it still unpacks
I'm hearing "yes"
jnthn I belive so.
TimToady unpacks?
what does that mean?
pmichaud i.e., if my $thingies = [1,2,3,4] 16:27
and we do foo(42, $thingies)
is $head an Array, or is it 1 ?
TimToady it's 1
pmichaud okay.
TimToady there is *no* difference between $thingies and @thingies, because they're both items
pmichaud understood.
TimToady the @ only makes a difference if bound to *
[particle] goes karmic 16:28
pmichaud and that's what you meant by *... right
thanks.
afk, lunch
(for real this time)
jnthn pmichaud: yeah right
TimToady because it has the "flatten me baby" marker
jnthn ;-)
TimToady but item ignores it
jnthn TimToady: OK, so we seem to agree @thingies needs to be turned into a Capture. 16:29
japhb I read that as the "flatten my baby" marker ... and wondered exactly when Perl 6 got the Halloween spirit.
TimToady more precisely, the second positional item, whever it is
*whatever
jnthn Right.
So my question is, how is that done?
TimToady I don't see why .Capture can't do it, though it might want inlining for performance 16:30
jnthn Sure, I'm wanting to get us semantically right before making us OMG FAST 16:31
:-)
OK, so supposing this, I have just one more question before I go and write code. ;-)
TimToady but if we have a proto, and we enforce [] subsigs as part of that, then we could do capturizing at compile time
jnthn True. Again, optimization, but there'll be ways to make it faster.
sub traverse ( BinTree $top ( $left, $right ) ) { 16:32
}
In this example, I have some type BinTree. I'm supposing left and right are attribute names?
Further, I'm assuming they must be public attributes?
TimToady or a hash with those keys
jnthn e.g. class BinTree { has $.left; has $.right; } 16:33
OK.
TimToady we can probably handle either
jnthn Yes, I figure me can.
Essentially I was seeing this as building a capture with left and right as named paramters.
Thing is, what if we have a bunch of other attributes?
TimToady that sig is more or less (|$top, :$left, :$right) 16:34
jnthn *nod*
But we had to turn BinTree into a Capture.
TimToady though | ... yes
jnthn Suppose we had something with like 10 attributes, and we only really wanted to pull out 2 of them. 16:35
Building a capture out of them only to consume two feels slightly wasteful.
TimToady but | just makes it raw bind to original BinTree
the Capture is a temporary
maybe
anyway, can probably optimize it out somehow
jnthn I mean, as a first cut I can do it that way. In the end, we may want something capture-ish that actual proxies. 16:36
So only obtains the thingies we need.
TimToady Capture as a role, perhaps
jnthn Yeah.
Anyway, at this point I'm guessing best is I go away and implement something along these lines.
And then we see how it works and iterate on it. 16:37
TimToady sounds like we're in the ballpark
masak should &open die in Perl 6 if it fails?
jnthn OK, great. Thanks. :-)
masak: only under "use fatal"
TimToady which fights against the homerun metaphor, I guess
jnthn :-)
TimToady masak: or if you use the failure value
jnthn *nod* 16:38
16:38 Chillance joined
TimToady and void context counts as "using" :) 16:41
(void context is also eager, interestingly, or for loops wouldn't run to completion) 16:42
moritz_ void context needs spec
or did I just miss it? 16:43
TimToady the eager part is specced
moritz_ but not how it's determined, iirc
Tene pmichaud: will nqp-rx be able to use the STD.pm grammar? 16:44
jnthn Tene: Not directly, but something way closer to it than we have now.
16:45 pyrimidine joined
TimToady writing an nqp-rx emitter for viv might be fairly easy 16:45
it would still have to emulate laziness like gimme5 does, alas 16:46
16:46 lichtkind joined
TimToady well, maybe not for the regexes :) 16:46
16:50 mj41 joined 16:56 desertm4x joined
diakopter I wonder how big the nqp-rx version of STD.pm would be 17:01
TimToady depends on how easy/hard it is to interweave the fate structure with nqp-rx 17:02
at worst, it would expand out to the same size-ish
slightly better insofar as we use . instead of -> everywhere :) 17:03
but slightly worse insofar as nqp is less expressive than Perl 5 (debatable)
we can probably more aggressively combine tests into PATTERN checks 17:04
but until nqp does LTM/fates, it won't shrink a lot, I'd guess 17:05
jnthn whee. It works. \\o/ 17:11
TimToady \\o/ # what works?
jnthn sub foo($n, [$head, *@tail]) { say $n; say $head; say @tail.perl; } 17:12
my @x = 1,2,3,4; foo(42, @x)
Gives output
42
1
[2, 3, 4]
moritz_ w\\oo/t
TimToady yeaye
jnthn When I first saw that in the spec I was like "way cool, can't wait until somebody implements that" :-)
17:13 astrojp left
TimToady well, somebody did :) 17:13
jnthn Yeah, I kinda didn't see me being the who at that point. ;-)
lichtkind TimToady: are there any plans to put Cwd, File::Spec and so on into a well defines path::tools module?
17:14 astrojp joined
TimToady we would like Perl 6 to be useful, if that's what you're asking :) 17:15
moritz_ I think nobody would object to that notion if somebody came up with such a thing
but that's very non-trivial
17:15 TopoMorto joined
moritz_ and all attempts so far where either either over-engineered or did not work :/ 17:16
TimToady the question of what "well defined" means is, of course, subject to a great deal of p6l negotiation
17:16 stephenlb joined
TimToady there are many dimensions to the problem, and many different opinions about what should be easy, or merely possible 17:17
moritz_ "if somebody shoots a silver bullet news at us, we'll catch it" 17:18
TimToady I can tell you that I don't want p{} for a path literal, though qp{} would be okay
17:19 ejs joined
TimToady I don't think Perl should be in the business of grabbing all the single-letter keywords... :/ 17:19
Q, q, and s are bad enough 17:20
moritz_ m
TimToady that too :)
but we've abandoned y
moritz_ at least one ;-)
17:21 ruoso joined
TimToady anyhoo, not terribly interested in grabbing p 17:21
cj TimToady: my 2 year old is 17:22
TimToady is it blue?
lichtkind TimToady: by first question is if id there an basic agreement toward less modules in this realm :)
17:22 colomon joined
pmichaud Tene: nqp-rx somewhat expects to influence the STD.pm grammar, as opposed to "run it directly" 17:23
cj TimToady: not so far...
moritz_ lichtkind: it's hard to haev a general agreement on something that doesn't exist yet
pmichaud Tene: the general approach at this point is for nqp-rx and rakudo to use something that is a lot closer to STD.pm as it exists today 17:24
lichtkind moritz_: i have tuits to make an API proposal but would like to see in which direction the agreement goes
TimToady there is agreement on the need for better design, but to some people that means more namespaces rather than less, somehow
pmichaud then we'll see about getting the two to converge
TimToady anyway, ack out all the spec references to 'path' to see the current mess 17:26
and bear in mind that p will die
pmichaud there's too many p's floating around anyway
TimToady and it's hard enough to mind your q's without having to mind your p's and q's 17:27
besides p vs q will drive the dyslexics nuts
pmichaud qrobably so.
TimToady no puodt it will 17:28
pmichaud (nqp-rx and fate structure) -- nqp-rx as it is set up now will likely be able to handle fate quite well. What it lacks at the moment is a good DFA for fast prefix matching. 17:29
(or something else that can handle fast prefix matching)
TimToady yes, well, you could just borrow mine...oh wait...that's the part I cheated on...
pmichaud :-) 17:30
anyway, that comes quite a bit later in the process. First step is to get Rakudo on the new grammar, and handling constants correctly, and correctly working with captures, and ...
TimToady are we both subconsciously waiting for the other person to do it? :)
jnthn Don't forget the parcels too :-)
pmichaud TimToady: I'm not particularly waiting for you to do it :) 17:31
17:31 mariuz joined
TimToady not consciously :) 17:31
pmichaud I'm not rushing to do it myself, but I do expect I'll end up writing one before all is said and done
I just want to get other stuff out of the way first, and I don't yet see it as being critical path for Rakudo Star 17:32
as long as we can chea.... er, make a close enough approximation, I think we'll be fine for a while.
17:33 ejs joined, mariuz joined
pmichaud besides, I'm progressively removing lots of cheats out of nqp-rx as we go :) 17:33
TimToady I think I might need to think about merely to figure out how to get derived grammars to share nodes with parent grammars to the extent possible, so as not to blow out memory on gazillions of macros
but that implies splitting my token strings up into nodes 17:34
which is halfway to real DFA
pmichaud oh, I have another question. So far in nqp-rx I haven't had a need for many "stopper" rules. Are they in STD.pm primarily for pruning/optimization, or are they really needed?
or are they needed for things that I'm not encountering yet (e.g., metaops)
17:35 hanekomu joined
pmichaud or perhaps to limit runaways 17:35
or all of the above :)
TimToady they're more about getting good error messages, for specific stoppers
pmichaud okay
I'll undoubtedly be adding them. I like STD.pm's error messaging framework a lot 17:36
TimToady and the general terminators are just to avoid deriving a new grammar every time we nest in different brackets and such
jnthn pmichaud: Can we has pretty colors too? ;-)
pmichaud jnthn: I suppose we could.
just need to rewrite the panic method. Since we can now put methods into grammars, that's pretty easy to do. 17:37
TimToady std: colors are good
p6eval std 28950: OUTPUT«Undeclared routines:␤ are used at line 1␤ colors used at line 1␤ good used at line 1␤ok 00:01 105m␤»
TimToady std: @colors are good
pmichaud it's official, "good" is undeclared.
p6eval std 28950: OUTPUT«[31m===[0mSORRY![31m===[0m␤Two terms in a row at /tmp/beSIdvkOHW line 1:␤------> [32m@colors [33m⏏[31mare good[0m␤ expecting any of:␤ bracketed infix␤ infix stopper␤ standard stopper␤ statement modifier loop␤ terminator␤Other potential difficulties:␤
..Variable @colors is…
pmichaud okay, I need to do less chatting and more coding. 17:38
pmichaud hides window.
TimToady hugme where are you? 17:54
moritz_ goes searching
17:55 hugme joined
jnthn Man we have a lot of bots. 17:56
hugme is my favorite though :-) 17:57
hugme: hug me
hugme hugs jnthn
jnthn \\o/
Juerd Why those +v's? 17:58
moritz_ typical mark for bots
Juerd I see
18:03 Chillance joined, pyrimidine joined 18:06 pmichaud joined
jnthn I've wanted to be able to do this one for ages: gist.github.com/221659 18:15
moritz_ how sick. 18:17
but it's cool ;-)
pmichaud gist.github.com/221659 # we now handle :my declarations and contextuals in regexes 18:19
oops
wrong
gist.github.com/221666
( I got distracted by jnthn++'s shiny)
TimToady heh 18:20
jnthn pmichaud++ has a shiny too :-D
TimToady a dollar short and a day late
/me, not pmichaud++
amazing how well pmichaud can type in a window that's hidden, too!! 18:21
cognominal jnthn, I guess I have to read (and understand the backlog) but I ask anyway: why the square brackets are necessary in the quicksort signature?
jnthn cognominal: You're passing one array.
cognominal: And then unpacking it.
pmichaud what's scary is that adding contextuals to regexes was... 12 lines of code.
jnthn wow 18:22
TimToady that many? tsk, tsk...
pmichaud I counted blank lines.
moritz_ you should clearly get a more expressive programming language.
pmichaud see the commit for yourselves :-)
(when dalek pops it up here)
cognominal jnthn, thx, got it.
diakopter flood death? 18:23
18:23 Tene joined
TimToady alias bots 18:23
okay, 12 bots currently, same as my alias 18:24
are any of the rest of you bots? 18:25
dalek p-rx: 7c48517 | pmichaud++ | src/NQP/ (2 files):
[nqp]: Enable :my declarations in regexes.
diakopter wonders sometimes
jnthn BTW, patch for unpacking coming once spectests finish.
18:26 fax joined, xenoterracide joined
jnthn pmichaud: So, does this mean tomorrow we start doing the Rakudo grammar from this? :-) 18:28
jnthn is has a Rakudo Day.
moritz_ is it installable? 18:30
jnthn Also, will be have to check out and build nqp-rx as well as Parrot? :-) 18:31
__ash__ anyone gotten perl6_ops.c:30:31: error: pmc_callsignature.h: No such file or directory when they try to build rakudo's most recent git commit?
pmichaud jnthn: I think we can start tomorrow, yes.
we'll see how things go for the rest of the evening, but it's looking very good. 18:32
I think I have another shiny coming up...
moritz_ __ash__: is 'git status' clean? did you realclean before re-configuring? 18:33
jnthn __ash__: Suspect your Parrot install is too old.
pmichaud: Excellent.
__ash__ moritz_&jnthn: i did a make realclean before re-configuring and my status is empty, i am re-building parrot to be sure at the revision in build/PARROT_REVISION 18:35
jnthn __ash__: That file was not on Parrot's installed headers list, so I had to add it a couple of days ago. 18:36
Maybe yesterday even
I dunno...I don't remember when I did things. :-)
[particle] that's what logs are for 18:37
__ash__ should i be using parrots trunk's latest revision or the one in PARROT_REVISION?
Tene the one in PARROT_REVISION is more-likely to work. 18:40
that's what the rakudo folks recommend.
__ash__ hmm
weird
i did parrot make install, and it didn't install that .h file 18:41
jnthn Odd. :-/ 18:43
moritz_ maybe PARROT_REVISION lags behind?
pmichaud perhaps the .h isn't in the list of installed files
not all of them are. 18:44
jnthn pmichaud: I added it to that list.
__ash__ where is that list? just cirous
curious*
jnthn Moment though...
pmichaud gist.github.com/221691 # another shiny feature
moritz_ wooot 18:45
pmichaud++
dalek kudo: 4320479 | jonathan++ | (10 files):
First cut of implementing unpacking of arguments with nested signatures. Supports arrays, hashes and objects.
pmichaud TimToady: do you see { ... } closures within regexes as enforcing a new scope, or simply sharing the same scope as the regex block itself? (I've implemented it as a new scope for consistency with other uses of {...} ) 18:46
jnthn Yeah. I put it in at trac.parrot.org/parrot/changeset/42134
PARROT_REVISION has 42158
pmichaud++ # nice!! 18:47
__ash__ hmm, did they change the thing you need to use to install the devel stuff?
i'll ask in the parrot channel
pmichaud __ash__: it was "make install-dev", now either "make install" or "make install-dev" should work.
jnthn pmichaud: I guess we'll be able to parse Perl 6 rather than NQP in there for Rakudo too? 18:48
pmichaud: Or will that be trickier?
pmichaud no difference 18:50
it's parameterized on the outer language
jnthn Nice!
:-D
jnthn is very impressed
pmichaud what TimToady refers to as "language braids" :-)
moritz_ oh wow
pmichaud so if someone wants their regexes to have PHP code, it can be done. 18:51
jnthn Ooh.
moritz_
.oO( is that a reason not to have that feature? ;-) )
pmichaud s/PHP/DSL/ # better? 18:52
jnthn s/PHP/LOLCODE/ # funnier :-)
pmichaud oooh! ooh!
I *gotta* do that.
jnthn will remember to post his Lolsql -> sql translator tonight too :-)
pmichaud guess which language besides Rakudo is going to be (re)implemented first using the new tools
jnthn LOL! 18:53
pmichaud good guess!
__ash__ doh, i was at the right revision of parrot, but i was in a tag/RELEASE_1_7_0 thats why i wasn't getting any of the parrot changes
jnthn __ash__: That explains it. :-)
pmichaud: Does this mean we'll be able to easily intersperse LOLCODE and Perl 6? :-) 18:54
my $x = 42; Q:LOLCODE { VIZIBLE $x }
pmichaud "There will be modules. 18:55
"
jnthn \\o/
18:56 am0c joined
cj pmichaud: I met pmichaud but not-you the other day. ;) 18:58
19:02 ejs joined 19:03 ejs left
Tene jnthn: lolcode variables aren't named like that. You'll need some pretty fance language braiding features to get one language to use the identifier rules of what it's embedded in. 19:07
jnthn: more likely, make a closure in lolcode and invoke it with $x 19:08
jnthn: I've been planning to add Q:lang modules for a long time.
jnthn Tene: Yes, true... :-) 19:10
Tene jnthn: I could imagine a lolcode module for rakudo that derives a modified lolcode grammar, though. 19:11
jnthn LOL6! 19:15
19:17 pnu_ joined
japhb fears reading the spec for that 19:20
19:32 ejs joined 19:40 SmokeMachine joined
colomon sorts through backlog, and is excited at masak's new mystery project, and delighted by quicksort. 19:49
[particle] the Exporter pmc can be subclassed to deal with variable naming between namespaces/languages 19:53
moritz_ if it's not there already, somebody could make a test case out of the quicksort example
19:54 xenoterracide joined
jnthn moritz_: I was gonna look a bit at test coverage yeah. 19:55
Of course, to really know how good it is we need to set masak++ on it. ;-)
moritz_ S06-signature/unpack-array.t
S06-signature/tree-node-parameters.t
jnthn Thanks 19:56
Feel free to fudge, or I can later.
moritz_ phone& 19:57
japhb Is there a link for the quicksort example? 19:59
20:00 xenoterracide joined
japhb Not in my backlog, guess I'll try perlgeek 20:01
jnthn japhb: gist.github.com/221659 20:02
japhb Ah!, excellent, thank you 20:03
Man, other languages have to cry when they see that sort of simplicity. 20:04
Of course, median-of-three pivot would probably be a good deal uglier.
20:04 brunov joined
jnthn heh 20:04
Yes, my quicksort is a tad vulnerable. :-) 20:05
pugs_svn r28951 | pmichaud++ | [pm.txt]: 20:09
r28951 | Pm-8: Are closures embedded in regexes creating a new lexical scope,
r28951 | or do they share the same scope as the regex block itself?
r28951 | (Currently I'm assuming they create a new scope, to be consistent
r28951 | with other uses of curlies. --Pm
20:10 pyrimidine left
dalek ok: d7fbc55 | moritz++ | (2 files):
example for chapter 1

back-and-forth reference, and it's already warning		 20:12
20:12 dduncan joined
pugs_svn r28952 | lwall++ | s/action/actions/ as suggested by pmichaud++ 20:22
r28953 | lwall++ | [S02] closure in string makes lexical scope 20:23
r28953 | [S05] closure in regex makes lexical scope
TimToady japhb: S06:880 might help with your median-of-three, if we allow [*@front, *$mid, *@back] 20:25
jnthn TimToady: heh, what? 20:26
How would it decide where to split?
TimToady 'course, that doesn't guarantee median, really
jnthn Yes, and if the list is lazy... ;-)
TimToady just "somewhere"
japhb looking
TimToady it's just a conjecture
jnthn Phew. :-)
TimToady but one resulting from reading Guy Steele's slides on folds considered slightly harmful 20:27
20:27 samlh joined
jnthn Linky? 20:27
japhb Wow. Um. That's powerful. In the "loaded gun initially pointed at foot" way.
But boy howdy would that be nice for some algorithms. 20:28
pmichaud gist.github.com/221783 # more shiny stuff
TimToady research.sun.com/projects/plrg/Publ...Steele.pdf 20:29
20:29 Su-Shee left
japhb [$head, *@front, $mid, *@back, $tail] to get the three possible choices for median-of-three pivot? 20:29
TimToady it's sorta baking mapreduce into the language
moritz_ we all have to cover our eyes to avert harm through increased shinyness 20:31
japhb very nice, pmichaud
jnthn pmichaud++ # single-handedly tripling sunglasses sales as a result of massive increase in shininess 20:32
japhb heh
TimToady pmichaud: does it work with Bengali digits? :) 20:33
(I'll wager it matches, but .Num doesn't work) 20:34
moritz_ rakudo: say +'৩৪' 20:35
p6eval rakudo 432047: OUTPUT«0␤»
moritz_ rakudo: say '৩৪' 20:36
p6eval rakudo 432047: OUTPUT«৩৪␤»
moritz_ rakudo: say '৩৪' ~~ /^\\d+$/
p6eval rakudo 432047: OUTPUT«৩৪␤»
moritz_ TimToady is spot-on
jnthn But how hard can it be? I mean come on, there's a unicode file with the data in it!!
japhb How do you get the Code for a particular multi variant, given name and signature? 20:40
I want to delegate from one multi to another, less specific one.
I guess just being able to re-run the multi dispatch with the current function not included would solve this particular use case (though my larger question still stands) 20:41
[particle] pmichaud: will that work with chaining? e.g. <?{0 <= +$<num> <= 255 }>
pmichaud bengali digits: 20:42
20:42 pnu joined
pmichaud $ cat x 20:42
grammar ABC {
token TOP { $<num>=[\\d+] <?{ +$<num> <= 255 }> }
}
say( ABC.parse("৩৪") );
$ ./nqp x
৩৪
[particle]: nqp doesn't implement chaining yet
jnthn japhb: I think it looks something like &foo:(Int, Str) or so but it may not promise one candidate.
pmichaud (might not ever)
(the bengali digit match is less impressive than it first looks, because it numifies to 0)
japhb jnthn, but even having that might do the trick.
jnthn, so could you do: &quicksort:([$, @]) to get the main multi in your quicksort example? 20:44
jnthn japhb: eww
Not that I know of.
20:45 M_o_C joined
jnthn japhb: If you wanna do crazy stuff like that, just iterate over the .candidates and itnrospect their sigs. 20:45
*introspect
20:45 M_o_C joined
japhb jnthn, here's what I'm playing at: gist.github.com/221787 20:47
typo fixed
jnthn Hmm...I suspect those might be ambiguous candidates in the first place. 20:48
Why not write the second as just 20:49
moritz_ just name it differently
jnthn multi quicksort($pivot, @rest) {
And it'll dispatch by being arity 2.
japhb jnthn, but it would still split the initial list if the @rest was slurpified ...
hmmm
jnthn oh, I see 20:50
you only want to do it on the first time? 20:51
japhb No, every split, until there are <3 left. But I want to reuse the implementation the non-median-of-three quicksort to implement the divide and conquer after I've determined the correct pivot. 20:52
Basically, trying to see if there was a way not to write the body of the original non-MOT quicksort twice 20:53
Because technically MOT quicksort is the same as: find MOT, move MOT to front, perform pivot-on-first quicksort 20:54
jnthn ah, hmm 20:55
But doesn't computing the MOT give you your pivot? 20:56
oh, but you need at lesat 3 elements the last time...
japhb right, so we still need to handle the 0,1,2 element cases
I guess I could handle them explicitly ... 20:57
*editeditedit* 20:58
dalek p-rx: 88ab936 | pmichaud++ | src/ (6 files):
[nqp]: Add code assertions to nqp.
japhb jnthn, well, I guess that's a little better 21:00
Still feels like extra work 21:01
jnthn Yeah, hmm.
japhb Can you have both a packed and unpacked form? As in 'multi quicksort(@list([$head, *@tail]))' or somesuch? 21:08
Er, I meant, can you name both the packed and unpacked form?
jnthn japhb: Not yet in Rakudo, but the spec seems to allow for that. STD may too. 21:11
japhb: The actual binder guts already know how to handle it too.
We just ain't parsing it yet (if STD does it, should fall out naturally in the new grammar.) 21:12
japhb What is the correct syntax?
21:13 pmurias joined 21:17 xenoterracide joined 21:18 nnunley joined 21:20 b_jonas joined
jnthn std: sub traverse ( BinTree $top ( $left, $right ) ) { 21:21
std: sub traverse ( BinTree $top ( $left, $right ) ) { }
p6eval std 28953: OUTPUT«[31m===[0mSORRY![31m===[0m␤Invalid typename BinTree at /tmp/IyyeBbV38C line 1:␤------> [32msub traverse ( BinTree[33m⏏[31m $top ( $left, $right ) ) {[0m␤FAILED 00:02 110m␤»
std 28953: OUTPUT«[31m===[0mSORRY![31m===[0m␤Invalid typename BinTree at /tmp/E7lF7ezKCB line 1:␤------> [32msub traverse ( BinTree[33m⏏[31m $top ( $left, $right ) ) { }[0m␤FAILED 00:01 107m␤»
jnthn std: class BinTree { }; sub traverse ( BinTree $top ( $left, $right ) ) { }
p6eval std 28953: OUTPUT«ok 00:01 110m␤»
jnthn japhb: That is spec'd, so I wonder if it's just...
std: class BinTree { }; sub traverse ( @orig [ $head, *@tail ] ) { } 21:22
p6eval std 28953: OUTPUT«ok 00:03 107m␤»
jnthn Yup
oh, ignore class decl at the start.
Anyway, I hope to copy-paste STD's signature parsing as far as possible into the new grammar.
21:23 astrojp joined
pmichaud should be possible to go far 21:23
jnthn pmichaud: Yeah, I think so do, especially given the stuff you've been putting in of late. 21:24
21:24 quietfanatic joined
jnthn *too 21:24
b_jonas is how $*ARGFILES (the magical variable that replaces *ARGV{IO}) documented in S yet? 21:26
um, verb missing. works.
dalek p-rx: 719cd0e | pmichaud++ | t/nqp/34-rxcodeblock.t:
Add tests for rx codeblocks.
 21:28
quietfanatic It probably works just like Perl 5's <>
in that it concatenates the files given by arguments, or stdin if there are none 21:29
I'm not certain though
b_jonas yeah, I'm more interested in the details. 21:30
quietfanatic As far as I can tell, it's not specced in detail yet. Too many things in the spec still rely on knowledge of Perl 5. 21:32
japhb jnthn, does the latest version of gist.github.com/221787 look like it would work? 21:41
moritz_ pmurias: I'm writing a Tidings post for my blog right now... is there anything from smop or mildew that I should report? (from the last 2 months) 21:45
jnthn japhb: heh multi quicksort(@a [$a?]) { @a } is cute :-) 21:46
japhb: Probably ;-)
21:46 tak11 joined
moritz_ the perlmmonks are discussing a nice idea right now 21:47
jnthn japhb: It hinges on the [$head, *@front, $mid, *@back, $tail] big being made to work out.
*bit
moritz_ an ARGV function which is called automatically when the file changes over which <> (magic ARGV) iterates
do we want such a thing in Perl 6? 21:48
pmurias moritz_: mildew-js is the new thing in mildew land, but things have slowed down recently as i haven't adjusted to uni fully yet 21:49
21:49 dduncan left
japhb jnthn, ah, cool, I'm glad (@a [$a?]) would work. :-) 21:50
And if only the double-slurpy is a problem, then I think I'm OK with what I have now. Perhaps not perfect, but pretty decent.
Any ideas from the peanut gallery on improving gist.github.com/221787 further?
jnthn japhb: Yeah, I'm quite happy that bit will work. 21:51
japhb afk for a bit, but will backlog 21:52
thanks jnthn++, TimToady++ 21:53
21:54 quietfanatic joined 21:55 jan joined
b_jonas moritz_: yes, that's why I've asked 21:56
actually I'm not sure I want that feature,
but what I would need is a sane bugless ARGV
moritz_ I'm pretty sure that Perl 6's ARGV is not magic
b_jonas 5.10 is better in that than 5.8
moritz_ at least not in the sense that it will execute rm -rf | as a pipe 21:57
21:57 nnunley joined
b_jonas moritz_: sure 21:57
moritz_ kinda likes perl -pe 1 'fortune |'
b_jonas I just want those two long-standing bugs of 5.8 (and I think 5.6) fixed
moritz_ which two? 21:58
b_jonas but perl 6 should have a better exception system so there while(<>) should die on read errors anyway (or at least if you use the right pragma)
moritz_ right
b_jonas moritz_: firstly, readline on any filehandle (not just ARGV) sets $! to EBADF if the last line is incomplete
in the next call when it returns undef that is 21:59
Juerd while (empty list) ? ;)
moritz_ std: while (<>) { }
p6eval std 28953: OUTPUT«[31m===[0mSORRY![31m===[0m␤Obsolete use of <>; in Perl 6 please use lines() or () instead at /tmp/ENx3MrgMWj line 1:␤------> [32mwhile (<[33m⏏[31m>) { }[0m␤FAILED 00:01 107m␤» 22:00
b_jonas secondly, ARGV also used to set the ferror indicator (ARGV->error) if the last line is incomplete
Juerd or ()?
b_jonas the two together means that there's no easy way to reliably detect read errors on ARGV
moritz_ Juerd: as empty list
Juerd Hm
b_jonas Juerd: I did ask whether $*ARGFILES was documented properly
and it's not so I don't want to guess about the proper incantation
22:00 brunov joined
moritz_ b_jonas: but I'm sure abigail has scripts which rely on both bugs, so they can't be fixed in perl 5 *SCNR* 22:01
b_jonas moritz_: 5.10 broke three of my obfus. sure, they're not as great as abigail's, but I don't think even abigail's scripts should stop p5p in this case.
quuxx moritz blogged "Perl 6 Tidings from October 2009": perlgeek.de/blog-en/perl-6/tidings-2009-10.html
jnthn moritz++ # I like the tidings posts 22:02
moritz_ b_jonas: IMHO neither should anything have stopped p5p from making <> less dangerous
jnthn: thanks 22:03
22:03 xabbu42 joined
moritz_ jnthn: though it feels kind of impossible to summarize two months of Perl 6 in less than an hour of writing 22:04
jnthn moritz_:
Mildew
emitter.
Mildew now has an experimental javascript emitter.
is that "emitter." line a fail?
moritz_ it is 22:05
22:06 brunov joined, tak11 joined, fax joined, hugme joined, cdarroch joined, Patterner joined, zaphar_ps joined, slavik joined, japhb joined, charsbar joined, arthur-_ joined, broquaint joined, szabgab joined, c9s joined, breinbaas joined, spinclad joined, pugs_svn joined, BinGOs joined, shachaf joined, sjohnson joined, Grrrr joined, cosimo joined 22:07 pmichaud joined, astrojp joined, pmurias joined, samlh joined, stephenlb joined, lichtkind joined, araujo joined, NorwayGeek_ joined, nihiliad joined, nothingmuch joined, silug joined, __ash__ joined, cj joined, rdice joined, [synth] joined, huf joined, [particle] joined, KatrinaTheLamia joined, Juerd joined, Schwern joined, nbrown joined, tylerni7 joined, mtve joined, dmpk2k joined, Infinoid joined, buu joined, clkao joined, Caelum joined, literal joined, jjore joined, REPLeffect_ joined, cottoo joined, simcop2387 joined, awwaiid joined, hcchien joined, solarion joined, Helios- joined, frew_ joined, cxreg joined, tkr joined, Eevee joined, arnsholt joined, akl joined, blaze-x joined, carlin joined, Bucciarati joined, Trinity94 joined, avuserow joined, estrai joined, allbery_b joined, andreasg_ joined, TimToady joined, diakopter joined, ezra joined
jnthn moritz_: Yes, you can't cover everything, but it's nice to have the salient points picked out. 22:09
22:09 colomon joined
moritz_ fixed. 22:09
does quuxx follow the plantesix feed? or just individual feeds? 22:10
quuxx: owner?
quuxx: help
quuxx: be silent!
thanks for at least obeying one command ;-)
22:11 tak__ joined
__ash__ is there a list of the order of ops for casting? i mean by doing like "40" + 5, does that always try to make them both into a Num? or ...? 22:11
moritz_ by new spec it will probably coerce both to Numeric 22:13
22:13 patspam joined 22:14 Whiteknight joined
__ash__ so, its going to do a multi-method lookup on + for the given types of X and Y 22:14
jnthn Yes, all operators go by multi-dispatch.
22:14 justatheory joined
moritz_ there's an infix:<+>(Any, Any) which does the coercion 22:15
__ash__ kk, so... perl6 isn't weakly typed now
jnthn But you'll probably hit the :(Any,Any) variant.
moritz_ depends on what you mean by "weakly typed" really 22:16
jnthn Well, that or we're just very strong about our weakness. :-)
__ash__ didn't perl 5 do a type coercion when you did "40" + 5, perl6 isn't implicitly converting, its just dispatching to a method that explicitly converts 22:31
jnthn Well, the + operator only works on numerics anyway in Perl. 22:32
cognominal Perl is not Basic! 22:33
jnthn Well, until you do
rakudo: multi infix:<+>(Str $a, Str $b) { $a ~ $b }; say "40" + "5"; # ;-)
p6eval rakudo 432047: OUTPUT«405␤»
jnthn But don't do that. :-)
cognominal I am sure may will do...
*many 22:34
22:34 xenoterracide joined, ruoso joined 22:38 xenoterracide joined, astrojp joined
sjohnson howdy 22:39
jnthn hi sjohnson
sjohnson everyday i use javascript, i come to appreciate Perl 10x as much as the previous day 22:46
</rant>
pmurias ruoso: hi
sjohnson or i should say, window.top.window.endRant();
ruoso hi pmurias
pmurias ruoso: can PurePrototypeHOW be p6opaque specific (in the js implementation) 22:47
?
ruoso you mean by accessing p6opaque internally instead of using the REPR API?
pmurias yes 22:48
ruoso well, you certainly can, but try to make it clearly as a shortcut
so you can later implement the full version 22:49
22:50 nnunley joined
pmurias PurePrototypeHOW is only used for knowhow right? 22:50
ruoso at this point, yes
22:57 payload joined 23:01 pmurias_ joined 23:06 mrsaturns joined
mrsaturns good afternoon everyone 23:06
23:10 rdice joined
am0c rakudo: say '৩৪' ~~ /^.$/ 23:11
p6eval rakudo 432047: OUTPUT«␤»
am0c rakudo: say '৩৪' ~~ /^.+$/ 23:12
p6eval rakudo 432047: OUTPUT«৩৪␤»
pugs_svn r28954 | pmichaud++ | [pm.txt]: Add Pm-9. 23:34
r28954 | Pm-9: Inside of a regex, what happens with C<< <[z..a]> >> ? Is it
r28954 | a compile-time error, an empty range, or ... ?
dalek p-rx: 063d6a1 | pmichaud++ | src/ (6 files):
Remove some files we're no longer using.
 23:41
p-rx: d49ead9 | pmichaud++ | (4 files):
[p6regex]: Add \\xNN, \\XNN, and \\x[NN,NN,...] to p6regexes.
p-rx: c48afee | pmichaud++ | src/Regex/P6Regex/ (2 files):
[p6regex]: Add \\oNN, \\ONN, and \\o[nn,nn,...] escapes.
23:44 kidd joined
sjohnson mrsaturns: hi 23:45
23:46 nnunley joined
quietfanatic pugs: macro term:<.> { '$_' }; for <a b c> {say . ~ 'x'} 23:47
p6eval pugs: OUTPUT«ax␤bx␤cx␤»
23:48 nihiliad joined 23:55 synth joined 23:59 frew joined

Powered by Cro and the Raku® Programming Language.

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

Thank you!