»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!
Set by sorear on 4 February 2011.
diakopter n: 'alive' 00:27
p6eval niecza v22-7-g7df6137: OUTPUT«(timeout)[auto-compiling setting]␤»
diakopter sorear: niecza rebuild seems stuck
dalek rl6-bench: c11b540 | (Geoffrey Broadwell)++ | / (2 files):
Split bench into timeall and analyze
 01:51
rl6-bench: 1ee014f | (Geoffrey Broadwell)++ | / (2 files):
Update EXPECTED and README to refer to split timeall and analyze scripts; minor additional fixes in each
rl6-bench: 3b96734 | (Geoffrey Broadwell)++ | bench:
Remove no-longer-needed bench script
rl6-bench: 511186f | (Geoffrey Broadwell)++ | .gitignore:
Add bench-*.json to .gitignore
japhb Well, that at least splits the original monolithic Perl 5 script into two somewhat more focused Perl 5 scripts. Conversion of `analyze` to Perl 6 may have to wait for another day. :-) 01:52
masak r: my $m = method ($a) { say $a }; class A {}; A.new.$m(42) 04:30
p6eval rakudo 547fcb: OUTPUT«42␤»
masak r: my $m = method ($a) { say $a }; class A {}; sub bind($obj, $method) { -> *@args { $obj.$method(|@args) } }; my $bound = bind(A.new, $m); $bound(42) 04:32
p6eval rakudo 547fcb: OUTPUT«42␤»
masak \o/
Woodi morning today :) 05:13
masak: how that last example work ? looks like one-line MOP for me...
err, def&implementation of MOP... 05:14
especially: how this: $obj.$method(... can happen ? 05:16
raiph Woodi: i think jnthn just changed rakudo to make the my $m = method bit work. i think $o.$m() has long worked. 05:22
Woodi but some random sub shouldn't be allowed to add method to class... 05:23
or it's kind of wrapper ?
*add meth to obj egzaktly 05:24
raiph Woodi: poked around logs. "recent change" might be my imagination. note: irclog.perlgeek.de/perl6/2012-10-08#i_6044469 05:38
ah. and irclog.perlgeek.de/perl6/2012-10-11#i_6051934 05:41
Woodi: this is what i was thinking was the relevant commit: github.com/rakudo/rakudo/commit/5a92ef1c0e 05:46
Woodi raiph: np. I was probably wrong too, with masak intentions. looks now like find of security breach :) 05:50
masak raiph: everything I used above has worked for a long time. 06:13
Woodi: in no way am I adding a method to a class. you may want to look again. 06:14
raiph masak: thanks. 06:14
masak it would only be a "security breach" (though you probably mean "breach of encapsulation/information hiding") if the anon method could access private state. which it can't. only methods in the class block can do that. 06:15
raiph masak: at first i thought you were adding a method to an object. but it's not even that is it?
masak correct. it's not.
the name of the sub -- "bind" -- is a clue.
raiph so exactly what is it? :) 06:16
masak I'm binding an object (to be the invocant) to an anonymous method.
JavaScript has a method .bind on functions which does this. essentially pre-setting 'this' on the invocation.
sorear raiph: $obj.$method(42) is almost synonymous with $method($obj, 42)
perl 6 does not have a magical this value, it's just the first argument 06:17
masak what sorear said. in Perl 5 thre is an *equivalence* between method calls and sub calls with an extra first parameter. in Perl 6 they are different but still very similar.
another way to say that is that there aren't "sub lambdas" and "method lambdas". there's only lambdas. methods just happen to have an extra (often hidden) first parameter. 06:18
there *is* a way to add methods to objects, using anonymous subclassing. 06:19
rn: my $o = 42 but role { method greet { say "OH HAI" } }; say $o; $o.greet() 06:20
sorear n: class A { method foo(:$foo:) {} }; A.foo
p6eval niecza v22-7-g7df6137: OUTPUT«(timeout)[auto-compiling setting]␤»
..rakudo 547fcb: OUTPUT«42␤OH HAI␤»
niecza v22-7-g7df6137: OUTPUT«(timeout)[auto-compiling setting]␤»
sorear r: class A { method foo(:$foo:) {} }; A.foo
p6eval rakudo 547fcb: OUTPUT«Too many positional parameters passed; got 1 but expected 0␤ in method foo at /tmp/sDHRdlw4hS:1␤ in block at /tmp/sDHRdlw4hS:1␤␤»
masak hm, that signature sholdn't parse, methinks. 06:21
nameds must come after positionals, so the colon after ':$foo' can't appear there.
Woodi hmm, why add $obj as "invocant" for subs/methods with syntax exactly like method on object call ? 06:22
raiph r: my \m = method () { say 1 }; sub bind (\o, \m) { -> { $o.$m }; my $b = bind (1, m); $b(); 06:33
p6eval rakudo 547fcb: OUTPUT«===SORRY!===␤Variable $o is not declared␤at /tmp/cRNu4bJ4jc:1␤»
TimToady r: my $m = method ($self: $a) { say $a }; class A {}; my $bound = $m.assuming(A.new); $bound(42); 06:37
p6eval rakudo 547fcb: OUTPUT«42␤»
TimToady r: my $m = method ($a) { say $a }; class A {}; my $bound = $m.assuming(A.new:); $bound(42);
p6eval rakudo 547fcb: OUTPUT«===SORRY!===␤Unable to parse expression in argument list; couldn't find final ')' at line 2, near ":); $bound"␤»
TimToady n: my $m = method ($a) { say $a }; class A {}; my $bound = $m.assuming(A.new:); $bound(42);
p6eval niecza v22-7-g7df6137: OUTPUT«(timeout)[auto-compiling setting]␤»
TimToady heh
n: my $m = method ($a) { say $a }; class A {}; my $bound = $m.assuming(A.new:); $bound(42); 06:38
p6eval niecza v22-7-g7df6137: OUTPUT«(timeout)[auto-compiling setting]␤»
TimToady um...
n: say 42
p6eval niecza v22-7-g7df6137: OUTPUT«(timeout)[auto-compiling setting]␤» 06:39
TimToady this is...suboptimal...
raiph nr: sub s (\a, \b) { }; s (1, 2); 06:40
p6eval niecza v22-7-g7df6137: OUTPUT«(timeout)[auto-compiling setting]␤»
..rakudo 547fcb: OUTPUT«Not enough positional parameters passed; got 1 but expected 2␤ in sub s at /tmp/iPzMmiWHGe:1␤ in block at /tmp/iPzMmiWHGe:1␤␤»
raiph r: sub s (\a, \b) { }; s (1, 2); # what am i doing wrong? 06:42
p6eval rakudo 547fcb: OUTPUT«Not enough positional parameters passed; got 1 but expected 2␤ in sub s at /tmp/n2XtQv4sC8:1␤ in block at /tmp/n2XtQv4sC8:1␤␤»
TimToady sorear: see timeouts above 06:45
sorear yeah I'm aware 06:54
TimToady r: my $m = method ($a) { say $a }; class A {}; my $bound = $m.assuming(A.new); $bound(42); 07:04
p6eval rakudo 547fcb: OUTPUT«42␤»
TimToady I guess that works too
r: my $m = method ($a) { say self.WHAT, ' ', $a }; class A {}; my $bound = $m.assuming(A.new); $bound(42); 07:05
p6eval rakudo 547fcb: OUTPUT«A() 42␤»
TimToady even works right :)
masak: so I think your bind() is a bit of overkill 07:07
zzz & 07:08
sjohnson bed! 07:12
masak TimToady: heh, I had .assuming on my list of alternative ways to do .bind, and then promptly forgot about it ;) 07:16
jnthn morning o/ 07:18
brrt hi \o
jnthn raiph: (doing wrong) passing a single parcel argument 07:19
r: sub s (\a, \b) { }; s(1, 2); 07:20
p6eval rakudo 547fcb: ( no output )
raiph jnthn: ah. simple enough. thanks. and good morning. :) 07:23
masak jnthn! \o/ 07:24
sorear don't count on a prompt resolution of the p6evalissue. 07:32
raiph jnthn: does the Rakudo source tree contain any lines of pure PIR, or is all PIR code embedded in NQP? How many lines of NQP contain PIR?
jnthn raiph: iirc, there's two (non-generated) .pir files in NQP, in src/cheats/ 07:33
And I think we've none at all for Rakudo
moritz nr: say 42 07:34
p6eval niecza v22-7-g7df6137: OUTPUT«(timeout)[auto-compiling setting]␤»
..rakudo 547fcb: OUTPUT«42␤»
masak n: say '(timeout)[auto-compiling setting]' 07:37
p6eval niecza v22-7-g7df6137: OUTPUT«(timeout)[auto-compiling setting]␤» 07:37
tadzik yeah, it finally compiled the setting and now reacts correctly :) 07:38
masak that's the only reasonable explanation I can see, anyway. 07:43
moritz n: say 1 07:45
p6eval niecza v22-7-g7df6137: OUTPUT«(timeout)[auto-compiling setting]␤»
masak dang, hypothesis falsified. :P 07:47
tadzik quick, do something before it kills feather again :)
moritz huh, on feather3, the process 'perl poll' takes 100% CPU 07:52
GlitchMr n: say [*] 1.. 6 #`?? 08:01
p6eval niecza v22-7-g7df6137: OUTPUT«(timeout)[auto-compiling setting]␤»
masak if a routine defines itself as returning a Str, what does a no-argument 'return' return? 09:01
jnthn r: sub foo() returns Str { return } 09:02
p6eval rakudo 547fcb: ( no output ) 09:03
hoelzro guesses Str
jnthn ...apparently, not an exception...
jnthn kinda expected that to complain
jnthn Oh, I wonder if it returns Nil, and Nil is accepted in any context? 09:04
jnthn r: sub foo() returns Str { return }; say foo.WHAT 09:04
p6eval rakudo 547fcb: OUTPUT«Type check failed for return value; expected 'Str' but got 'Nil'␤ in sub foo at /tmp/3JL_UDkW2p:1␤ in block at /tmp/3JL_UDkW2p:1␤␤»
jnthn Oh!
I didn't call it above
duh.
OK, then it is doing what I expect
masak ah. 09:05
jnthn I think that's sane
masak yes, that was one of two possibilities that I expected.
jnthn If you said you're gonna return a Str and then you don't... 09:06
masak the other one being that 'return' is context-sensitive enough to know that it should return Str.
moritz that'd sound like a piece of context flowing in the wrong direction 09:19
though maybe necessary for returns of native type 09:20
(or maybe not)
Caldrin Hi! Is there a new POD for perl6 or do I use the old POD for documenting my modules? 09:24
tadzik there's new Pod, yes
Caldrin Can you link me to documentation, please? 09:25
Caldrin Ah, I found this: zag.ru/perl6-pod/S26.html 09:28
tadzik: Thanks for your answer.
masak Caldrin: welcome! there's indeed a new Pod, but I should warn you that rendering support is not yet up to the standard of the old POD. 09:30
oh and by the way, tadzik++ is the guy who implemented almost all of the Pod parsing in Rakudo :) 09:31
Caldrin masak: Well, as far as I see there are other places too, where Perl6 is not yet up to the standard of Perl5. Thats fine for me. I want to be an early adopter of "The Next Big Thing". 09:33
masak then double welcome.
doubly*
Caldrin Thank you 09:37
jnthn heads to Stockholm for NPW & 11:08
dalek rlito: b50d8dc | (Flavio S. Glock)++ | / (3 files):
Perlito5 - js - new structure p5ArrayOfAlias
 11:47
masak jnthn: good luck in Stockholm! say hello from me! :) 12:14
[Coke] realizes he did not re-check the daily run builds to make sure today's daily run was ok. whoops. 12:37
"wait and see", I guess.
[Coke] drinks a coffee and hopes for the best. 12:40
rindolf [Coke]: shouldn't you drink coke? ;-) 12:51
grondilu What's the equivalent of perl5's "SUPER" pseudo-class? 12:58
masak closest equivalent is nextsame, callwith, et al. 12:59
but "SUPER" doesn't make much sense in a MI language. 13:00
it's all about method resolution orders instead. hence nextsame and gang.
grondilu ok 13:01
moritz the question about SUPER is a FAQ now 13:05
somebody[tm] should really add it (and the answer) to the perl6/faq repo
grondilu nr: class Foo { method talk { say "foo" } }; class Bar is Foo { nextsame; say "bar" }; .talk given my Bar $x .= new;
p6eval niecza v22-7-g7df6137: OUTPUT«(timeout)[auto-compiling setting]␤»
..rakudo 547fcb: OUTPUT«nextsame is not in the dynamic scope of a dispatcher␤ in block at src/gen/CORE.setting:485␤ in block at /tmp/EkcVMqf5Sv:1␤␤»
grondilu callsame and co seem to deal with multi dispatch, not inheritance :/ 13:07
moritz so does SUPER 13:08
what else do you need it for, except calling methods from superclasses?
grondilu r: class Foo { method talk { say "foo" } }; class Bar is Foo { callwith; say "bar" }; Bar.new.talk;
p6eval rakudo 547fcb: OUTPUT«callwith is not in the dynamic scope of a dispatcher␤ in block at src/gen/CORE.setting:463␤ in block at /tmp/uqHSYuAAcL:1␤␤»
moritz well, you need to call it inside a method 13:09
huf where's method talk in Bar?
moritz as a first step at least :-)
grondilu oops
moritz r: class Foo { method talk { say "foo" } }; class Bar is Foo { method talk { callsame; say 'bar' } }; Bar.new.talk 13:09
p6eval rakudo 547fcb: OUTPUT«foo␤bar␤»
grondilu my bad
indeed.
grondilu Thanks :-) 13:10
moritz np & yw :-) 13:11
[Coke] rindolf: it's a little early for the cold stuff. I switch to Coke Zero about 10am local time. :) 13:14
rindolf [Coke]: ah. 13:16
grondilu rn: say 3.infix:[my $ = "*"](4); 13:22
p6eval rakudo 547fcb: OUTPUT«No such method 'infix:<*>' for invocant of type 'Int'␤ in block at /tmp/SlLgrXlTKT:1␤␤»
..niecza v22-7-g7df6137: OUTPUT«(timeout)[auto-compiling setting]␤»
grondilu Is this kind of indirection NYI? 13:24
moritz r: say 3.&infix:[my $ = "*"](4); 13:24
p6eval rakudo 547fcb: OUTPUT«12␤»
moritz grondilu: if you want to use a subroutine as a method, you must include the & in the name 13:25
grondilu ok, but I guess S12 needs to be corrected then. 13:26
around line 379
moritz I don't see why 13:29
in that case $methodobj contains a callable
moritz but you were using a method *name* 13:30
grondilu it took some time but yeah I think I get it. 13:42
r: class Foo is Str { method infix:<++>($s) { self ~ $s } }; say Foo.new.infix:<++>("bar");
p6eval rakudo 547fcb: OUTPUT«bar␤»
grondilu here we don't need '&' indeed.
moritz right, because infix:<++> already is a method, not a sub 13:43
dalek rlito: 1a8d3e4 | (Flavio S. Glock)++ | / (3 files):
Perlito5 - js - p5ArrayOfAlias - shift, pop
 13:46
moritz css.dzone.com/articles/anti-javascript-perl-6 13:50
(sorry if that's already been posted, I haven't backlogged thorougly for a time)
www.johndcook.com/blog/2012/10/09/t...avascript/ is the original URL 13:52
[Coke] wow, that's a surprisingly postive take on the situation. :) 13:55
moritz aye :-) 13:59
r: class A { method !foo($x) { say "private $x" }; method public { $!foo("stuff") } }; A.new.public 14:01
p6eval rakudo 547fcb: OUTPUT«===SORRY!===␤Attribute $!foo not declared in class A␤at /tmp/FnPF5WO3bX:1␤»
moritz according to S12:471, that should work 14:02
as a short way to say $( self!foo('stuff') )
though personally I'm not sure the conflict with attribute names is worth it
supernovus r: my $m = method ($a) { say "$.name says $a"; }; class A { has $.name; }; my $foo = A.new(:name<Bob>); $foo.$m("hello"); 16:15
phenny supernovus: 10 Oct 02:01Z <[Coke]> tell supernovus I tried to write a test that this passed and rakudo failed, and failed. Can you add a test to the ticket or to roast?
p6eval rakudo 547fcb: OUTPUT«Bob says hello␤»
Ulti r: my @list = <hello>; @list.unshift(1..Inf); say @list[*-1]; 16:22
p6eval rakudo 547fcb: OUTPUT«(timeout)»
Ulti I think it starts to reify Inf
yeah it does looking at my RAM usage :D 16:23
if you go the other way around you can't push on the end of an Inf list it raises an exception 16:24
n: my @list = <hello>; @list.unshift(1..Inf); say @list[*-1];
p6eval niecza v22-7-g7df6137: OUTPUT«(timeout)[auto-compiling setting]␤»
colomon n: my @list = <hello>; @list.unshift(1..Inf); say @list[*-1]; 16:25
p6eval niecza v22-7-g7df6137: OUTPUT«(timeout)[auto-compiling setting]␤»
colomon huh. why still compiling the setting? 16:26
Ulti would be kind of neat if it just noticed the end was already reified and /after/ the inf list.... like recursive laziness
colomon n: say "Hello"
p6eval niecza v22-7-g7df6137: OUTPUT«(timeout)[auto-compiling setting]␤» 16:26
Ulti but yeah thats a nasty edge case for the spec I guess 16:27
colomon seems like something is wrong with p6eval here....
supernovus [Coke]: Okay. I'll check out roast and look at adding a test.
Ulti well this fails on my local install of Rakudo too
though not sure if its this months Rakudo* 16:28
colomon Ulti: I would never expect unshift an infinite list to work. :) 16:28
but say "Hello" ought to work. 16:29
Ulti heh yeah
the real problem is unshift of an infinite /does/ work 16:30
Ulti hmm weird 16:32
Ulti the latest Rakudo* flips out on the act of unshifting the infinite range, Rakudo* from August flips out when you ask for the last element 16:33
^ scratch that they both dont like the initial unshift 16:35
GlitchMr perl6: @(1, 2, 3).perl.say 16:39
p6eval niecza v22-7-g7df6137: OUTPUT«(timeout)[auto-compiling setting]␤»
..rakudo 547fcb: OUTPUT«(1, 2, 3).list␤»
GlitchMr r: @(1, 2, 3).unshift.perl.say
p6eval rakudo 547fcb: OUTPUT«(1, 2, 3).list␤»
supernovus Hmm, if I commit this change to roast, adding the test for my patch, then anyone running rakudo without the patch will now get failures in S32-str/pack.t until the patch is applied. 16:40
grondilu rn: role Foo { method postfix:<!>() { "Foo!" } }; my $x; say ($x but Foo)!; 16:45
p6eval niecza v22-7-g7df6137: OUTPUT«(timeout)[auto-compiling setting]␤»
..rakudo 547fcb: OUTPUT«===SORRY!===␤Confused␤at /tmp/1cC6VROLGs:1␤»
diakopter std: say "h"~~/<<<<<<:m/>>>>>>/>>>>>>/a/ # TimToady what does this mean 16:52
p6eval std 77327a4: OUTPUT«ok 00:00 44m␤»
dalek ast: f244f31 | (Timothy Totten)++ | S32-str/pack.t:
Added test for 'x' in pack()

rakudo patch: rt.perl.org/rt3/Public/Bug/Display...?id=115120
diakopter TimToady: I can't tell what's terminating what 16:54
TimToady that's what viv is for
supernovus [Coke]: Tests for the 'x' flags added to S32-str/pack.t 16:55
Ulti I think I just discovered my favorite bit of perl six "my @list = (); say if @list eq Nil;" rather than in perl5 where you do a check for undef and then scalar @list == 0 16:56
TimToady why do you use eq? 16:57
and @list == 0 is already in scalar context 16:58
diakopter r: my @list = [Nil, Nil]; say 1 if @list eq Nil;
p6eval rakudo 547fcb: OUTPUT«1␤»
diakopter too bad niecza p6eval is broken 16:59
TimToady Ulti: I guess I don't know what you're really trying to accomplish there 17:00
PerlJam Is feather really really slow i^Hor is it just me? 17:01
jnthn o/ from Stockholm 17:02
PerlJam jnthn: o/ 17:03
Ulti TimToady: in perl5 you can't do something like if @list eq []
or () 17:04
at least if you can I am doing it wrong
TimToady why do you want to use 'eq' for that? that's a string coercion
Ulti I don't its just in rakudo I got a load of errors about Nil and numerical comparisson
"use of uninitialized value of type Nil in numeric context" 17:05
PerlJam Ulti: Are you just trying to see if the array is emtpy? 17:05
TimToady Nil is not really a value, it's the absence of a value
Ulti PerlJam: yeah
TimToady 'unless @list'
of 'if @list'
just use it in Boolean context 17:06
r: my @list = (); say so @list;
p6eval rakudo 547fcb: OUTPUT«False␤»
Ulti specifically arrayrefs though in perl5, basically every line of selectall_arrayref in DBI has a messed up if afterwards
TimToady well, you just have to deref it in P5 to get array behavior 17:07
Ulti yeah, but you first check for undef too otherwise you are deref'ing undef
TimToady okay, I see what you're saing now 17:08
*y
Ulti sorry I was really not clear
but yeah just double checked, perl6 just magics the arrayrefness for you 17:09
so the logic of testing an array or arrayref is identical
which is really nice
TimToady bows, then remembers who wrote Perl 5... :) 17:10
Ulti heh 17:12
meh perl5 just /always/ use refs then you dont get into any trouble 17:13
PerlJam TimToady: You can strut for Perl 6 without being humble about Perl 5. It's okay. :-) 17:14
Ulti :) 17:16
whenever I get frustrated over anything datastructure related in perl5 I flick back to some of my past Java code with huge nested generic declarations for stuff like HashTable<String,HashTable<String,ArrayList<String>>> 17:18
its all good
hashes and lists should just work and they should be primitives of all languages 17:19
jnthn The thing is that Java also doesn't have type inference. So you often have to write those things *twice*, once on the thing that'll hold it and once in the new statement.
Ulti yup
Ulti I have lots of declarations for class variables that cover like 6 wrapped lines in my Java.... when I complaing to other people who love Java they just ask why I would want a data structure like that... and not a nest of 10 classes 17:20
HashStringListHashOfString datastructure = new HashStringListHashOfString <--- so much better! 17:21
jnthn :/
jnthn is glad he doesn't have to deal with Java 17:22
Ulti I had to deal with it to complete my degree which kind of sucks
GlitchMr HashStringListHashOfString? This is too descriptive name...
Ulti not for Java, the more verbose you are the better the code 17:23
the best is the javadoc comments that just repeat the class names with spaces added :D 17:24
but hey Java vs Perl isn't a fair race
GlitchMr What sort of data structure is that anyways?
HashString?
OfString?
Ulti a great one for holding a graph of word use in a passage of text
in perl you would just use a hash of hashes or something 17:25
GlitchMr Is structure like this: {a: [{b: 'abc'}]}
(this is JavaScript, by the way)
Ulti checks his specific example
GlitchMr I've no idea what this class name actually means 17:26
GlitchMr Probably because I haven't programmed in Java, but not sure 17:26
Ulti HashTable<String,HashTable<String,ArrayList<String>>> === {'a'=>{'b'=>['c','d','e']}}
GlitchMr uhmmm... fun
Ulti I use that quite often
weirdly
GlitchMr What is === anyways? 17:27
Ulti where b is a word following a and then you just have a list of stuff associated with that connection
GlitchMr The part on the left is type, the part on the right is instanceof that type
Ulti GlitchMr just pseudo english/code for are equivalent
GlitchMr: thats me saying the Java declaration of a type is equivalent to the perl literal form
[Coke] I think I (nigh) killed feather again. looks like it's actually one of the pugs tests chewing all memory. 17:28
Ulti vanishes into the ether bbl
GlitchMr So HashTable<> is two arguments type - the first one is type used for keys, the second is type used for values. Am I right? 17:29
[Coke] PerlJam: it wasn't just you
GlitchMr That makes sense
[Coke] all ok now. Disabling cron job until that test gets fudged for pugs.
GlitchMr As for Java, I don't like it because it's language intended to be used with IDE 17:30
Try using it without IDE
[Coke] tempting to add the "perl6-limited.pl" trick so that any one test that may or not be fudged doesn't take down the host system. 17:31
GlitchMr Probably type declaration would be smaller if HashTable would be Hash and ArrayList would be Array, but it's Java... 17:32
jnthn .Net called it Dictionary, which is terribly huffmanized.
At least List is reasonably short 17:33
GlitchMr Isn't .Net Microsoft's response for Java?
GlitchMr So, identifiers length is comparable with Java 17:34
jnthn GlitchMr: Yeah, but C# is a rather better language than Java. By quite some way.
[Coke] supernovus++ 17:36
PerlJam jnthn: that's okay, Java is way better than C++ ;) 17:41
jnthn It's almost as if languages get better with time by learning form the mistakes of others! 17:42
PerlJam jnthn: then how do you explain PHP? They learned the wrong things? 17:43
[Coke] php has been out there for a while. 17:44
jnthn PerlJam: "mistake" is subjective? ;-)
flussence PHP got the ordering of "make it {work,correct,fast}" wrong
PerlJam Coke: php is contemporaneous with Perl 5. I think they missed a few lessons. 17:45
flussence they went for working then fast, and then people started using it and they realised "fast" for plain CGI isn't neccessarily fast for other uses
(in particular, adding new language features is an utter nightmare because their parser does a direct translation between source tokens and VM opcodes instead of building a tree) 17:47
doy that sounds remarkably like perl 5 17:48
PerlJam I had a conversation with a CS prof earlier today whose wife is teaching an "internet programming" course (basically writing web apps). She started teaching theml HTML, then CSS, then PHP ... soon she's going to flip them to RoR. I wonder how many will conclude that PHP is a good thing?
flussence no, p5 builds a parse tree first and then does stuff to that... (IIRC from the camel book)
doy well, it builds opcodes directly via a bison parser 17:49
flussence (although I've heard the p5 parser isn't fun to hack on either)
doy so the parse tree is still just implicit
flussence php's is more like trying to compile assembly code by using sed
doy i have no doubt 17:50
(:
[Coke] I am biased, of course, but I'd go for CF before php. 17:51
jnthn Time for a break from talk prep..
(yes, I didn't finish preparing for my NPW talk for tomorrow yet...)
flussence r: my \a = set <1 2 3>; say 1 ~~ a 18:04
p6eval rakudo 547fcb: OUTPUT«False␤»
flussence for some reason I get a segfault on the repl if I do "my \a = set <1 2 3>" and then "1 ~~ a" on another line, but not on the same line 18:06
GlitchMr Unlike Perl 5, ~~ is not "contains" operator. But still, it shouldn't crash 18:08
flussence also does it with "= any set" there (which is what I was going for) 18:09
PerlJam I don't think Perl 5 knows what operator ~~ is yet ;)
(unless p5p came to some concensus while I wasn't looking)
sisar r: my @a; say @a[1]; say @a[*-1]; 18:10
p6eval rakudo 547fcb: OUTPUT«Any()␤Cannot use negative index -1 on Array␤ in method gist at src/gen/CORE.setting:9934␤ in sub say at src/gen/CORE.setting:7250␤ in block at /tmp/xqsiYi23M9:1␤␤»
GlitchMr In Perl 5, you can say 1 ~~ [1, 2, 3] and it works
leont p5p and concensus on smartmatch? Who are you kidding!
GlitchMr I don't know why Perl 5 needs smartmatch 18:11
sisar what's with the "Cannot use negative index -1 on Array" error? What am i doing wrong ?
PerlJam GlitchMr: for the "switch" statement of course! :)
GlitchMr Smartmatching is too smart for Perl 5
sisar r: my Int @a; say @a[*-1]; 18:12
p6eval rakudo 547fcb: OUTPUT«Cannot use negative index -1 on Array+{TypedArray}␤ in method gist at src/gen/CORE.setting:9934␤ in sub say at src/gen/CORE.setting:7250␤ in block at /tmp/wuBlgCX9mc:1␤␤»
flussence sisar: @a[*] in that context is 0, because it's empty.
r: my @a; say @a[@a.elems - 1];
p6eval rakudo 547fcb: OUTPUT«Cannot use negative index -1 on Array␤ in method gist at src/gen/CORE.setting:9934␤ in sub say at src/gen/CORE.setting:7250␤ in block at /tmp/wzGMEkgVnn:1␤␤»
GlitchMr In most cases, smartmatching is just any() function, just like in List::MoreUtils 18:13
PerlJam sisar: did you think that "say @a[1]" would auto-vivify ?
TimToady In Perl 6 you can say [1, 2, 3] ~~ [1, 2, 3] and it works :)
GlitchMr $ [1, 2, 3] ~~ [1, 2, 3]
1
TimToady but it does something completely different 18:14
GlitchMr But well, ~~ in Perl 5 appears to be similar to Perl 6 operator. Except of those any() sort of operations aren't in Perl 6.
TimToady there are no implicit anys in p6
sisar PerlJam: yeah, say @[a1] should auto-vivify, shouldn't it ? I was thinking @a[*-1] would auto-vivify too, but now i see that is not possible.
TimToady and we actually have a type system
sisar *@a[1]
flussence the p6 ~~ is a lot nicer, since it just boils down to multi ACCEPTS methods underneath 18:15
GlitchMr Smartmatching in Perl 5 simply doesn't work well without typing
Of course it can get type of reference, but that's all
GlitchMr Also: $ 'red' ~~ ['blue', [['green', ['red']], 'purple']] 18:16
1
~~ is recursive?
That's too clever to be useful... 18:17
leont GlitchMr: In perl5? Sometimes􏿽x85
sisar flussence: Confirmed. I get the seg falut too on the REPL.
GlitchMr For me, ~~ in Perl 5 sounds like a mistake 18:18
diakopter r: my @a; my $c=now; @a[@a]=$_ for 1..5000; say now - $c;
GlitchMr And for me, given when in Perl 5 is mistake too
p6eval rakudo 547fcb: OUTPUT«6.4875409␤»
diakopter r: my @a; my $c=now; @a.push($_) for 1..5000; say now - $c;
leont I like when, actually
p6eval rakudo 547fcb: OUTPUT«(timeout)»
diakopter rakudo like @a[@a]= much better than push 18:19
leont And given an appropriate matcher library, smartmatch can be pretty cool
diakopter *likes
GlitchMr .push() should be better than
TimToady given/when is fine if you limit yourself to strings and numbers
leont It's the part that makes guesses that is awful
GlitchMr TimToady: isn't hash of subroutines for that?
diakopter GlitchMr: ?
PerlJam sisar: autovivification of array elements in rvalue context seems weird and slightly dangerous to me. 18:20
TimToady GlitchMr: sometimes a switch is more readable
diakopter GlitchMr: what do you mean by ".push() should be better than" 18:20
GlitchMr then* 18:21
diakopter what do you mean by "should"
@a[@a]= is far faster 18:22
r: my @a; my $c=now; @a[$_]=$_ for 0..4999; say now - $c; 18:24
p6eval rakudo 547fcb: OUTPUT«0.4320339␤»
diakopter but 5000 length queries are hella slow
(6 seconds) 18:25
where did all those crickets come fromt? 18:26
GlitchMr: why/how should push method be better? 18:27
GlitchMr It should be faster 18:32
diakopter do you mean "someone ought to make it faster?" or "I'm surprised it's not faster"
TimToady the two are not mutually exclusive :)
diakopter right, but it could be only one or the other 18:33
TimToady well, it *should* be only one or the other :)
diakopter I'm not surprised it's faster; method lookups/execution are slow 18:34
flussence they're both slow, but it's only 3x slower than the assignment method 18:35
diakopter r: my @a; my $c=now; unshift(@a,$_) for 1..3000; say now - $c; 18:38
p6eval rakudo 547fcb: OUTPUT«7.1017452␤»
diakopter r: my @a; my $c=now; splice(@a,0,0,$_) for 1..3000; say now - $c;
p6eval rakudo 547fcb: OUTPUT«7.945410␤»
GlitchMr So, multidispatch is slow? 18:42
flussence if you want to see *slow*, try `time perl6 -e "say +open('/usr/share/dict/words').lines"` 18:46
I ran the profiler and it seems like 30% of the time is just from .chomp
diakopter it's 15,000x slower than wc -l 18:49
r: my $a=sub {}; $a.$a 18:51
p6eval rakudo 547fcb: OUTPUT«Too many positional parameters passed; got 1 but expected 0␤ in sub at /tmp/BTUvRAiy46:1␤ in block at /tmp/BTUvRAiy46:1␤␤»
diakopter r: my $a=sub ($b) { say $b }; 15.$a 18:53
p6eval rakudo 547fcb: OUTPUT«15␤»
diakopter is that supposed to do that?
flussence r: my $c = method () { say self }; 15.$c; 18:54
p6eval rakudo 547fcb: OUTPUT«15␤»
flussence I think it's equivalent to that 18:54
hm, .lines seems a lot faster than .slurp.comb(/^^.*$$/) at any rate, and .slurp.split(/\n/) is even worse for memory usage 18:56
just poking around... is that commented out code in core/IO.pm:124 still relevant? 18:59
_edwin rosettacode.org/wiki/CRC-32#Perl_6 19:02
_edwin for some reason i seem to be drawn to write such absurdly slow code :} 19:03
_edwin maybe because in my dayjob I'm counting processor cycles all the time :/ 19:04
GlitchMr us2.php.net/manual/en/function.crc32.php 19:10
So... in PHP, crc32() can give negative numbers on 32-bit systems...
That doesn't makes sense. 19:11
supernovus gah, apparently DateTime.now still believes my local timezone to be +0700 instead of -0700. :( 19:11
GlitchMr But it's PHP
sorear when you get right down to it the return value of crc32 should be an element of GF(2^32), not int or unsigned it 19:12
returning negative numbers doesn't make less sense than returing positive numbers greater than 1
or ℤ/2ℤ[x] 19:14
diakopter r: (method { self.(self) }).(method { self.(self) }) # hee hee 19:15
p6eval rakudo 547fcb: OUTPUT«maximum recursion depth exceeded␤current instr.: 'print_exception' pc 88369 (src/gen/CORE.setting.pir:42327) (src/gen/CORE.setting:9039)␤called from Sub '<anon>' pc 174 ((file unknown):118) (/tmp/9w7_9BuMOm:1)␤called from Sub '<anon>' pc 196 ((file unknown):134) (/…
diakopter sorear: niecza p6eval is dieded 19:16
sorear diakopter: I know
diakopter r: $_=method {self.(self)};$_.$_ 19:17
p6eval rakudo 547fcb: OUTPUT«maximum recursion depth exceeded␤current instr.: 'print_exception' pc 88369 (src/gen/CORE.setting.pir:42327) (src/gen/CORE.setting:9039)␤called from Sub '<anon>' pc 116 ((file unknown):83) (/tmp/e3WOOmTLLj:1)␤called from Sub '<anon>' pc 138 ((file unknown):99) (/tm…
[Coke] r: my $ℤ= 3; say $;
p6eval rakudo 547fcb: OUTPUT«===SORRY!===␤Non-declarative sigil is missing its name␤at /tmp/vaPJsqD5uA:1␤»
[Coke] r: my $ℤ= 3; say $ℤ;
p6eval rakudo 547fcb: OUTPUT«3␤»
[Coke] huh. spif.
.u ℤ
phenny U+2124 DOUBLE-STRUCK CAPITAL Z (ℤ)
brrt would cqrs be a good fit for a shop inventory system? 19:20
TimToady r: my \ℤ = 3; say ℤ; 19:23
p6eval rakudo 547fcb: OUTPUT«3␤»
jnthn .u ℤ
phenny U+2124 DOUBLE-STRUCK CAPITAL Z (ℤ)
brrt what does the \ sigil do?
jnthn (gives me a box here...) 19:23
brrt: Introduces a name you can from then on refer to without the sigil 19:24
Also as a result of having no sigil it imposes no context
brrt can barely wrap his had arround that 19:26
oh
i see
[Coke] r: my \say = 3; say say; 19:29
p6eval rakudo 547fcb: OUTPUT«===SORRY!===␤Confused␤at /tmp/wR6Au7GVKw:1␤»
[Coke] r: my \if = 3; say if;
p6eval rakudo 547fcb: OUTPUT«␤»
[Coke] std: my \if = 3; say if;
p6eval std 77327a4: OUTPUT«===SORRY!===␤Unsupported use of bare 'say'; in Perl 6 please use .say if you meant $_, or use an explicit invocant or argument at /tmp/SzPydU71Ox line 1:␤------> my \if = 3; say⏏ if;␤Confused at /tmp/SzPydU71Ox line 1:␤------> [3…
[Coke] I suspect that my \<keyword> should be disallowed. 19:30
"this ain't tcl"
TimToady this is why you should use Unicode symbols
brrt r: my \whatever = 42; say whatever;
p6eval rakudo 547fcb: OUTPUT«42␤» 19:31
brrt r: my \whatever = 1,2,3; say whatever
p6eval rakudo 547fcb: OUTPUT«1 2 3␤»
brrt nice
jnthn TimToady: "i" and "f" are both in Unicode :P
brrt r: my \whatever = Foo => 'bar'; say whatever
p6eval rakudo 547fcb: OUTPUT«"Foo" => "bar"␤»
brrt yay, i like perl6 again 19:32
TimToady so fickle... 19:33
japhb_ TimToady, I was thinking the exact same thing. ;-) 19:34
sorear, I know you know about the niecza being unavailable on p6eval. If you haven't had a chance to look at it personally, I can add that it won't compile locally for me right now either; seems to be related to the gamma function commit yesterday: 19:37
lib/Builtins.cs(11,7): error CS0246: The type or namespace name `SpecialMathFunctions' could not be found. Are you missing a using directive or an assembly reference?
japhb_ That's the error after 'git pull; make clean; make' 19:37
TimToady [Coke]: note that most keywords in p6 will win under LTM because they tend to include a trailing whitespace. 19:38
r: my \do = 42; say do do;
p6eval rakudo 547fcb: OUTPUT«42␤»
japhb_ my \do = 42; say do do ; 19:39
r: my \do = 42; say do do ;
p6eval rakudo 547fcb: OUTPUT«===SORRY!===␤Error while compiling block : Error while compiling op call: Error while compiling block : Error while compiling block : Can only use get_how on a SixModelObject␤»
supernovus r: my \foo = 1,2,3,4; say foo.WHAT;
p6eval rakudo 547fcb: OUTPUT«Parcel()␤»
jnthn whoa...
japhb_: That one'd want reporting.
supernovus hmm, better than what I get locally, which is "Segmentation fault (core dumped)"
jnthn r: do
japhb_ jnthn, OK, will do
p6eval rakudo 547fcb: OUTPUT«===SORRY!===␤CHECK FAILED:␤Undefined routine '&do' called (line 1)␤»
sorear japhb_: poke colomon, looks like colomon forgot to commit makefile changes 19:40
colomon doh!
there you go, sorry about that. 19:41
dalek ecza: 7ae412a | (Solomon Foster)++ | Makefile:
Add SpecialMathFunctions.cs to the build.
TimToady std: my \do = 42; say do do ; 19:42
p6eval std 77327a4: OUTPUT«ok 00:00 42m␤»
TimToady std: say do do ; 19:43
p6eval std 77327a4: OUTPUT«ok 00:00 42m␤»
TimToady r: say do do ;
p6eval rakudo 547fcb: OUTPUT«===SORRY!===␤Error while compiling block : Error while compiling op call: Error while compiling block : Error while compiling block : Can only use get_how on a SixModelObject␤»
moritz r: say BEGIN BEGIN 3
p6eval rakudo 547fcb: OUTPUT«3␤»
diakopter r: my \if = (my \say = 1 ); say if if if if {} if 19:44
p6eval rakudo 547fcb: ( no output )
diakopter no clue there
TimToady std: my \if = (my \say = 1 ); say if if if if {} if 19:45
p6eval std 77327a4: OUTPUT«===SORRY!===␤Missing semicolon at /tmp/vdW6AgbtuV line 1:␤------> my \if = (my \say = 1 ); say if if⏏ if if {} if␤Parse failed␤FAILED 00:00 44m␤»
diakopter I cannot fathom how rakudo's parsing that 19:47
I've tried all kinds of replacing if with 1
japhb_ colomon, Working like a charm now, thank you! 19:48
colomon thank you for bringing it to my attention
japhb_ jnthn, bug sent
np
Does someone need to kick the rebuild for p6eval, or will it try automatically? 19:49
moritz it's done by cron 19:50
japhb_ moritz, even when a previous build is stuck? 19:50
moritz define "stuck"
if it fails, no problem 19:51
moritz if it hangs... no idea, never had that problem 19:51
sorear I don't think it's hanging
japhb_ fair enough
Before colomon's fix, the build died quickly here, I was just wondering about the message that p6eval was giving (that sounded like it was trying to compile the setting and just never completing) 19:52
sorear japhb_: since the build failed, each time you ran p6eval it tried to compile the setting 19:53
and then failed a few seconds in, killed with signal SIGXCPU
japhb_ nodnod
diakopter r: if 1 { say 4 } if 1 { say 5 } 19:55
p6eval rakudo 547fcb: OUTPUT«4␤5␤»
diakopter jnthn: why does the above work? 19:55
sorear std: if 1 { say 4 } if 1 { say 5 }
p6eval std 77327a4: OUTPUT«===SORRY!===␤Strange text after block (missing comma, semicolon, comment marker?) at /tmp/t4yEuVbNZ5 line 1:␤------> if 1 { say 4 }⏏ if 1 { say 5 }␤ expecting any of:␤ horizontal whitespace␤ statement_control␤Parse failed␤FAILE…
hoelzro hey Perl6 folk 19:58
I finally managed to build Rakudo * 2012.09 on my laptop!
\o/
[Coke] yay! 19:59
hoelzro now I'm back with questions ;)
$*EXECUTABLE_NAME is supposed to be the name of the script I'm running, right? 20:00
like $0 from Perl 5?
sorear I think that it's actually $^X 20:01
$0 is $*PROGRAM_NAME or something
hoelzro ah, that's what I'm looking for =)
because it was returning $^X and I was thinking "wtf"
sorear: thanks!
flussence all the systemish vars are listed at the top of S28 20:02
moritz doc.perl6.org/language/variables#Sp...+Variables 20:03
jnthn diakopter: I think that one's RT'd already...some oddity with end of statement.
diakopter oh
jnthn Though I thought I sync'd some of that stuff with STD when doing the LTM work... 20:04
Must be some difference somewhere.
sirrobert hallo 20:26
sorear hi
jnthn o/ sirrobert 20:26
sirrobert how do I write an lvalue method? 20:27
there's something I saw once, but I can't find it... rw-return or return-rw maybe
jnthn return-rw,
sirrobert trying to write a customer setter/getter
hmm
r: class A { method v (Str $v) { return-rw $v } }; A.new.v = 4; 20:29
p6eval rakudo 547fcb: OUTPUT«Not enough positional parameters passed; got 1 but expected 2␤ in method v at /tmp/I0JuUdkuTr:1␤ in block at /tmp/I0JuUdkuTr:1␤␤»
sirrobert r: class A { method v (Str $v?) { return-rw $v } }; A.new.v = 4;
p6eval rakudo 547fcb: OUTPUT«Cannot assign to a readonly variable or a value␤ in block at /tmp/q2QRroqHSj:1␤␤»
sirrobert keep getting that 20:30
(ignore that the type was wrong in that example.. the error is that it can't assign.
r: class A { method v (Str $v?) { return-rw $v } }; A.new.v = 'foo;;
p6eval rakudo 547fcb: OUTPUT«===SORRY!===␤Confused␤at /tmp/YXhUR1B0qz:1␤»
colomon sirrobert: what do you think you're assigning there? 20:31
n: class A { method v (Str $v?) { return-rw $v } }; A.new.v = 4;
p6eval niecza v22-9-g7ae412a: OUTPUT«Unhandled exception: Writing to readonly scalar␤ at /tmp/AV7BC97rWo line 1 (mainline @ 6) ␤ at /home/p6eval/niecza/lib/CORE.setting line 4205 (ANON @ 3) ␤ at /home/p6eval/niecza/lib/CORE.setting line 4206 (module-CORE @ 576) ␤ at /home/p6eval/niecza/lib…
colomon n: class A { method v (Str $v?) { return-rw $v } }; my $a; A.new.v($a) = 4; say $a;
p6eval niecza v22-9-g7ae412a: OUTPUT«Unhandled exception: Nominal type check failed in binding '$v' in 'A.v'; got Any, needed Str␤ at /tmp/DWz1Sv6kkO line 0 (A.v @ 1) ␤ at /tmp/DWz1Sv6kkO line 1 (mainline @ 7) ␤ at /home/p6eval/niecza/lib/CORE.setting line 4205 (ANON @ 3) ␤ at /home/p6eval…
sirrobert in that example, just some var. but this doesn't work either:
jnthn Yeah, doing it to a (readonly) parameter won't cut it
sirrobert: What're you actually trying to do, by the way? :) 20:32
colomon n: class A { method v (Str $v is rw) { return-rw $v } }; my $a; A.new.v($a) = 4; say $a;
p6eval niecza v22-9-g7ae412a: OUTPUT«===SORRY!===␤␤Strange text after block (missing comma, semicolon, comment marker?) at /tmp/zTBkTLy1QT line 1:␤------> thod v (Str $v is rw) { return-rw $v } }⏏; my $a; A.new.v($a) = 4; say $a;␤␤Parse failed␤␤»…
sirrobert r: class A { has $.v; method v ($v) { return-rw $.v } }; A.new.v = 'foo;
p6eval rakudo 547fcb: OUTPUT«===SORRY!===␤Confused␤at /tmp/6wxkeBIzbG:1␤»
sirrobert that's closer to what I want to do.
r: class A { has $.v; method v ($v) { return-rw $.v } }; A.new.v = 'foo';
p6eval rakudo 547fcb: OUTPUT«Not enough positional parameters passed; got 1 but expected 2␤ in method v at /tmp/Bk6BQzD21a:1␤ in block at /tmp/Bk6BQzD21a:1␤␤»
sirrobert dang it 20:33
heh
r: class A { has $.v; method v ($v?) { return-rw $.v } }; A.new.v = 'foo';
p6eval rakudo 547fcb: OUTPUT«maximum recursion depth exceeded␤current instr.: 'print_exception' pc 88369 (src/gen/CORE.setting.pir:42327) (src/gen/CORE.setting:9039)␤called from Sub 'v' pc 151 ((file unknown):178492324) (/tmp/RoXftqPBKG:1)␤called from Sub 'v' pc 174 ((file unknown):178492341) …
sirrobert look, so here's the problem I want to solve =)
jnthn sirrobert: Yes, but why do you want to write the custom accessor in the first place? That kinda changes the answer :)
sirrobert I am handling semantic versioning as a set of properties: major, minor, patch (1.25.2)
sorear sirrobert: $.v means self.v, the infinite recusion is thus obvious 20:34
sirrobert I want to be able to set with $foo.version = '1.2.3'
sorear sirrobert: You meant $!v
colomon r: class A { has $.a; method v() { return-rw $!a; }; }; A.new.v = 'foo';
p6eval rakudo 547fcb: ( no output )
colomon r: class A { has $.a; method v() { return-rw $!a; }; }; my $a = A.new; $a.v = 'foo'; say $a.a 20:34
p6eval rakudo 547fcb: OUTPUT«foo␤»
colomon sirrobert: there you go
sirrobert sorear: ohhhhh... I keep forgetting that $.v is a virtual method call
thanks all
skids should there be an "is rw" on the method now that there is return-rw? 20:35
sirrobert How do I do some processing on the value that's getting assigned? 20:35
sorear use a Proxy 20:36
skids sirrobert: use a proxy.
colomon sirrobert: I think you might... yeah proxy
sirrobert docs url?
(haven't heard of a proxy in perl6)
or you just mean a private attr with a different name? 20:37
sorear n: Proxy.new(FETCH => -> $ { }, STORE => -> $, $new { say $new }) = 5
p6eval niecza v22-9-g7ae412a: OUTPUT«5␤»
sorear n: Proxy.new(FETCH => -> $ { 7 }, STORE => -> $, $new { say $new })++
p6eval niecza v22-9-g7ae412a: OUTPUT«8␤»
sirrobert hmm
skids docs are a bit sparse. Working example (you don't need the Zavolaj stuff in there): gist.github.com/2771728
sirrobert skids: thanks, reading 20:38
sorear r: Proxy.new(FETCH => -> $ { 7 }, STORE => -> $, $new { say $new })++
p6eval rakudo 547fcb: OUTPUT«8␤»
colomon sirrobert: perlcabal.org/syn/S06.html#Lvalue_subroutines
sirrobert colomon: thanks
(didn't realize you could do method !foo () {...} for private methods...?)
ok, makes more sense now. Thanks all 20:39
dalek osystem: 1e70a74 | (Timothy Totten)++ | META.list:
Added Netstring library.
 20:41
kudo-debugger: c1d805d | jonathan++ | lib/Debugger/UI/CommandLine.pm:
fail shouldn't count as a throw.
 20:42
sirrobert proxy is kinda like p5 tie 20:43
skids Pretty much.
jnthn sirrobert: If you haven't heard the news already: latest debugger in git repo has step out and step over. It also allows looking at self and attributes. :) 20:44
sirrobert: Thanks for the feature requests :)
sirrobert jnthn++: woot! no more work today, gonna play with it =) 20:45
thanks
sirrobert jnthn: that's going to be GREAT for skipping those loooong regex parsings heh 20:46
grondilu is thi debugger panda-compliant? I don't see it on modules.perl.org.
grondilu meant modules.perl6.org
jnthn grondilu: It's build is, afaik, just a little too demanding for Panda at the moment. 20:47
grondilu ok, I'll wait 20:48
jnthn grondilu: It's bundled in with the Rakudo Star monthlies. 20:49
I think tadzik++ mumbled something about custom build steps at some point, so at some point I hope it'll be Pandable
sjn \o 20:50
jnthn hi sjn 20:50
sjn: In Stockholm? 20:51
sjn is at claes++'s apartment, laxing to music, preparing some trouble for tomorrow's NPW talk :)
jnthn: yep, you?
jnthn laxing to music? sounds a bit fishy... 20:52
sjn: Yes, I'm here
In a hotel that should be 10ish minutes walk to the venue in the morning :)
sjn is done laxing, and proceeds with his re-laxing.
jnthn
.oO( hmm, that pun probably only works in Swedish... )
sjn jnthn: yes, it does. :) 20:53
diakopter haha 20:53
sjn you could say that was a red herring of a pun :)
jnthn :D 20:53
sorear jnthn: lax=salmon in swedish?
phenny: sv en "lax"? 20:54
phenny sorear: "salmon" (sv to en, translate.google.com)
jnthn sorear: yeah
skids wonders how many words for salmon there are in finnish
sorear I saw Lachs on a menu in .de, so I was kinda prepared for that
colomon hmmm, origin of lox? 20:55
sorear two years or so ago I was in a USian restraunt and I saw a menu item, "FRIED LOX"
sorear to which I say to the group "Fried LOX? I've heard of fried Kool-Aid, but this is just ridiculous" 20:56
skids Sorry I didn't understand that. I have a herring problem. 20:57
sorear LOX also means liquid oxygen...
jnthn Oh... 20:58
jnthn didn't know that.
.oO( In California, would FRIED LAX mean the airport burnt down? )
dalek kudo-debugger: 2a155b7 | jonathan++ | bin/perl6-debug.nqp:
Don't single-step sigspace.

It feels like it adds negative value to the debugging experience.
_edwin rosettacode.org/wiki/CRC-32#Perl_6 20:59
NativeCall crc32 \o/
_edwin somewhat makes up for my superslow CRC-32 (sorry for double line) 21:00
flussence wonders if it's possible to access the linux kernel hash/crypto functions using NativeCall... 21:00
skids _edwin: You should admire how slow the native CRC implementations are in the Sum package :-) 21:01
I will be getting around to NCI interfaces soon I hope.
skids flussence: that I doubt, but openssl has quite a few hash functions and it is everywhere unix flavored these days. 21:03
_edwin skids, hm...just wanted to install Sum and got this: gist.github.com/3881510 21:04
skids Hrm... I'll have to try a panda install. star09 is working for me. What rakudo version? 21:05
flussence skids: right, I guess openssl knows how to pass through to the kernel stuff for hardware accel too... 21:07
_edwin This is perl6 version 2012.09.1-7-gfb11f13 built on parrot 4.4.0 revision RELEASE_4_4_0
skids flussence: if any library does, it would be openssl, I guess. I'll probably and up writing ties to rhash and mhash as well. Windows will require a different developer as I'm allergic to it. 21:08
Well, I'll try a panda install when I get home and see if I can figure out what's broken.
skids Hrm it is probably tripping up on the POD hooks for validating the manpage synopsis -- maybe it is stripping pod first. 21:13
skids decommutes.
bbkr_ how can I bump module version in Star ? 21:16
_edwin I really wonder how the J solution for CRC-32 works 21:21
sorear _edwin: www.jsoftware.com/help/dictionary/dx128.htm#3 21:23
128!:3 is the standard library routine for CRCs
_edwin 128!:3 is a standard library call??? these guys are crazy :) 21:26
but thanks for explaining. i thought it was a combination of primitives and wondered how it could be so short 21:27
hoelzro hey guys, I wrote a blog post about building Rakudo * on Arch Linux: hoelz.ro/blog/building-rakudo-perl-...arch-linux 21:30
please let me know if I wrote anything incorrect!
sorear _edwin: a programming system with good native support for polynomials over GF(2) can do a CRC with some padding, a cast, a remainder, and a cast 21:31
I thought J had native polynomails but looks like I'm wrong
something like SAGE could probably do it
_edwin true, but it still would need a bit of program text to turn a string into a polynomial 21:32
sorear , #: a. i. 'Hello World' 21:35
__sri i suppose everyone here has already read this awesome book, but just in case, the architecture of PyPy www.aosabook.org/en/pypy.html 21:38
dalek kudo-debugger: 1462774 | jonathan++ | bin/perl6-debug.nqp:
Make m:P5/.../ regexes debuggable too.
 21:40
PerlJam sri: I haven't read it (yet :-). Thanks. 21:41
_edwin the paragraph about RPython makes it sound somewhat similar to NQP 21:42
__sri the chapters about GHC and LLVM are also pretty cool 21:43
oh and the CLR chapter... actually... it's all great :) 21:44
jnthn sleep & 22:08
skids Hrm. Are POD contents supposed to be in precompiled modules or not -- it seems there is the remnant of the pos structure, but contents are empty. 22:49
s/pos/pod/
flussence skids: Pod blocks add a lot to parsing time, keeping them around would probably defeat the purpose of precompiling 23:10
sorear ...huh? 23:11
if you precompile, there is no parsing time.
flussence I see a giant base64'd binary thing in pir output and some chopped up pod strings in an array, maybe the binary holds all the info needed to reassemble the text on demand, I dunno. 23:22

Powered by Cro and the Raku® Programming Language.

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

Thank you!