IRC log for #perl6, 2010-03-06

Perl 6 | Rakudo | Specs | Parrot | source cross referenced

All times shown according to UTC.

Time	Nick	Message
00:06		mjk joined #perl6
00:14		snarkyboojum joined #perl6
00:18	lichtkind	thanks
00:38		payload joined #perl6
00:39	cj	can someone recommend a book or document on compiler engineering?
00:42	m-locks	A. Aho, R. Sethi, J. Ullman: Compilers: Principles, Techniques, and Tools. Addison-Wesley, 1986.
00:42	cj	thanks, m-locks
00:42	m-locks	np
00:44	diakopter	cj: PLP3 (Programming Language Pragmatics 3rd edition)
00:44	cj	m-locks: you prefer the first edition?
00:44	diakopter	five stars
00:44	cj	thanks, diakopter :)
00:46	m-locks	they seem to have lotsa additional materials, PLP CD and stuff
00:56		colomon joined #perl6
01:11	colomon	rakudo: my Int $x = Mu;
01:11	p6eval	rakudo b348b3: OUTPUT«Type check failed for assignment␤current instr.: '&die' pc 16803 (src/builtins/Junction.pir:347)␤»
01:19	pugssvn	r29950 \| colomon++ \| [t/spec] Fudge S14-roles/basic.t (some very disturbing errors here).
01:19		jaldhar joined #perl6
01:22		wknight8111 joined #perl6
01:42	pugssvn	r29951 \| lwall++ \| [S32/IO.pod] put back :s, remove :z, :T, :B, :M, :A, :C
01:42		r29951 \| note that these all end up calling methods on IO, not strings
01:48	k23z__	can I get a rakudo commit bit ?
01:48		I'm wsdookadr on github
01:49	dalek	rakudo: 974d9a8 \| (Solomon Foster)++ \| t/spectest.data:
01:49		rakudo: Turn on S14-roles/basic.t.
01:49		rakudo: review: http://github.com/rakudo/rakud[…]8df1a96e0e4138836
01:50	diakopter	k23z__: I think they want an individual to build up a history of submitting good patches before one gets one of those
01:50	k23z__	ah ok sorry
01:50		can I write some tests for Perl6 ?
01:50		quietfanatic joined #perl6
01:50	diakopter	yes :)
01:51		you can have a commitbit to pugscode
01:51	k23z__	what tests could I write and what would I need ? (I have a compiled rakudo right here)
01:51	diakopter	if you don't have one already
01:51		the tests are stored in pugscode svn
01:52		re "what tests could I write", that's a difficult question
01:53		it requires a certain familiarity with the language, or a precise reading of the Synopses (the language specification)
01:53		really, both
01:53	colomon	Though you don't need to read the full spec, by any means -- just find a nice corner of it.
01:54	diakopter	and a knowledge of how the test suite is organized, so you can determine that the feature for which you're considering writing tests hasn't already been thoroughly covered.
01:59	k23z__	do you have guides for new devs ?
02:00		so the road is read the synopsis, the perl6 book on github -> write tests ?
02:00	diakopter	depends from where you're coming, and where you think you're going. :D (what languages you've used previously; what kinds of software you expect to write)
02:02		Perl 5 experience might help; it might not. The changes/additions (from 5->6) are numerous and varied in size (small,large) enough that existing Perl knowledge might hinder, since it might be more difficult to keep the languages distinct in your mind.
02:04		that's the case for me, anyway. It seems that those who have a thorough mastery of Perl 5 have a much easier time keeping them straight than someone like me, who isn't a Perl 5 expert.
02:05	TimToady	we all already know where Perl 5 ought to be different :)
02:07		course, there are still parts of Perl 6 that ought to be different (I'm looking at you IO.pod)
02:08	diakopter	k23z__: msg TimToady with your desired username & email address for pugscode svn (if you don't already have an account and you want to commit to the spec/testsuite)
02:09		I seem to have forgotten my commitbit password, otherwise I would add it
02:11	TimToady	do you still have a mail account at veradox?
02:14		_jaldhar joined #perl6
02:14	diakopter	ohhh that's the username
02:26	* diakopter	considers how to design an API for codegen routines that generate codegen routines
02:31		lue joined #perl6
02:32	lue	oi!
02:43		...chirp
02:43	colomon	hey all.
02:44	PerlJam	greetings colomon
02:44	lue	hello o/
02:44	colomon	submethods.t defines two classes in a... closure? in braces.
02:44		lives_ok {
02:44		class Foo { has $.foo_build; submethod BUILD() { $!foo_build++ } }
02:44		class Bar is Foo { has $.bar_build; submethod BUILD() { $!bar_build++ } }
02:44		}, "class definitions were parsed/run/compiled";
02:45		when the tests try to use those classes, they're not there.
02:45		any clue if they should be? Or are they local to those brackets?
02:45	lue	maybe because it's contained in the braces. That's my first guess :)
02:45	colomon	(If I take the braces away, the next twenty tests run fine...)
02:45	PerlJam	classes used to be package-scoped and I think they're lexical by default now.
02:46		(check the spec obviously :)
02:48		oh no, they're still package scoped (or should be)
02:48		S12:109
02:50	diakopter	std: package bar { class baz { class foo { }; say foo; }; say foo; }
02:50		chenfengyuan joined #perl6
02:50	p6eval	std 29951: OUTPUT«Undeclared routine:␤ 'foo' used at line 1␤ok 00:01 105m␤»
02:50		chenfengyuan left #perl6
02:50	diakopter	chenfengy didn't like what p6eval had to say
02:50	lue	how would you tell if a number is odd or even?
02:50		szabgabx joined #perl6
02:51	m-locks	even: number % 2 == 0
02:51	TimToady	% or !%
02:51	colomon	PerlJam: adding our doesn't even help.
02:51	diakopter	TimToady: see p6eval above
02:52		(I don't know whether that's correct)
02:52	lue	rakudo: say 15 % 2 == 0
02:52	p6eval	rakudo 974d9a: OUTPUT«0␤»
02:52	colomon	rakudo: { class Foo { method bar() { say "bar"; }; }; }; my $a = Foo.new; $a.bar
02:52	PerlJam	lue: you monkey patch .odd and .even methods :)
02:52	p6eval	rakudo 974d9a: OUTPUT«Could not find non-existent sub &Foo␤current instr.: '_block14' pc 29 (EVAL_1:0)␤»
02:52	TimToady	rakudo: say 15 !% 2
02:52	p6eval	rakudo 974d9a: OUTPUT«Confused at line 11, near "say 15 !% "␤current instr.: 'perl6;HLL;Grammar;panic' pc 500 (ext/nqp-rx/src/stage0/HLL-s0.pir:328)␤»
02:52	lue	those are supposed to exist!?
02:53	TimToady	alpha: say 15 !% 2
02:53	p6eval	alpha 30e0ed: OUTPUT«0␤»
02:54	diakopter	alpha: say 15 % 2 ?? "odd" !! "even"
02:54	TimToady	class foo is visible only inside baz
02:54	p6eval	alpha 30e0ed: OUTPUT«odd␤»
02:54	diakopter	oh
02:54	PerlJam	Hmm. .odd is actually mentioned in S04:332
02:54	lue	alpha: say 15.odd
02:54	TimToady	std: package bar { class baz { class foo { }; say foo; }; say baz::foo; }
02:54	p6eval	alpha 30e0ed: OUTPUT«Method 'odd' not found for invocant of class 'Int'␤in Main (file src/gen_setting.pm, line 324)␤»
02:54		std 29951: OUTPUT«ok 00:01 105m␤»
02:54	PerlJam	I guess I'm not the only one who channels his inner ruby programmer occasionally.
02:54	TimToady	std: package bar { class baz { class foo { }; say foo; }; say bar::baz::foo; }
02:55	diakopter	std: package bar { { class foo { }; say foo; }; say foo; }
02:55	p6eval	std 29951: OUTPUT«ok 00:01 105m␤»
02:55		std 29951: OUTPUT«ok 00:02 105m␤»
03:03		new_perl6_hacker joined #perl6
03:04	TimToady	it's quite ironic that when an Olympic skier falls on the course, it's called DNF
03:04	m-locks	haha
03:05	new_perl6_hacker	How does rakudo diffrentiate between a $string = "some string" and $code = "Code inside a string"
03:05	m-locks	well gotta sleep, gn y'all
03:06	colomon	o/
03:06	new_perl6_hacker	@m-locks its morning here
03:06	k23z__	!tell garu test
03:06	m-locks	it's 5am, about my bedtime ;P
03:06	new_perl6_hacker	Good night
03:06	TimToady	I have to get up at 5am to see if I'm in a panel discussion...
03:07	Tene	new_perl6_hacker: rakudo does not differentiate at all. Strings are strings, and there's no type for "a string that happens to incidentally look like Perl 6 code".
03:07	m-locks	yoink ->
03:07	colomon	new_perl6_hacker: but if you have { } inside a double-quote string, it is executed and the result inserted into the string.
03:07	PerlJam	new_perl6_hacker: what makes you think rakudo does differentiate those things?
03:08		colomon++ ah, good point
03:08	lue	I was going to program a little something featured in today's XKCD comic (http://xkcd.com/710/), but I think programming .odd and .even first'd be more fun
03:08		(and in P6 this time)
03:08	new_perl6_hacker	@PerlJam how will then find what methods I can invoke on strings and what methods I can invoke code objects
03:09	Tene	lue: make sure you generate graphviz output so you can graph your results. :)
03:09		new_perl6_hacker: the same way you discover what methods you can invoke on any type.
03:09	PerlJam	rakudo: "string".^methods.sy
03:09		er ...
03:09		rakudo: "string".^methods.say
03:10	p6eval	rakudo 974d9a: ( no output )
03:10		rakudo 974d9a: OUTPUT«maximum recursion depth exceeded␤current instr.: 'perl6;Seq;!fill' pc 13790 (src/builtins/Multi.pir:22)␤»
03:10	Tene	rakudo: {say "hi"}.^methods.perl.say
03:10	p6eval	rakudo 974d9a: OUTPUT«maximum recursion depth exceeded␤current instr.: 'perl6;Seq;!fill' pc 13790 (src/builtins/Multi.pir:22)␤»
03:10	Tene	ah, looks like that's broken ATM
03:10		alpha: {say "hi"}.^methods.perl.say
03:10	p6eval	alpha 30e0ed: OUTPUT«[{ ... }, { ... }]␤»
03:10	PerlJam	heh
03:10	Tene	anyway, it's in the class.
03:10	lue	graphviz...?
03:10	PerlJam	alpha: "String".^methods.say
03:10	p6eval	alpha 30e0ed:
03:10		..OUTPUT«sprintfScalarComplexStrpredsuccencodeperlACCEPTSWHICHatanminacosmaxbytesevalchopsinrindextanhIntvaluesasinhacosechabsendceilingp5chompkeyskvfloorasechordcapitalizepickroundevalfileintsplitmatchacoshwordsp5choplcjoinsinhcotanacotanatan2reducesrandtrimasecreversecotanhisapolar:d:eucf…
03:11	PerlJam	alpha: "String".^methods.join(" ").say
03:11	p6eval	alpha 30e0ed: OUTPUT«encode succ WHICH ACCEPTS perl sprintf Scalar Complex Str pred map :d cis :e :f log sech log10 min rand max :l truncate bytes sort sin sqrt asin Int cosh :s grep asinh acosech abs end uc ceiling p5chomp samecase floor substr asech ord round flip does int split match acosh
03:11		..acosec wo…
03:11	Tene	lue: http://graphviz.org/
03:11		lue: a simple text format for describing and rendering graphs.
03:12	new_perl6_hacker	And I will be able to execute that as &"string" or $string = "some code" ; &$string
03:13	TimToady	not like that
03:13	Tene	lue: http://upload.wikimedia.org/wi[…]h-all-30-no27.svg looks like it was generated with graphviz.
03:13		new_perl6_hacker: if you have code in a string that you want to execute, you'll need to use the 'eval' function, which invokes the compiler on the text.
03:13	TimToady	you have to work harder to do symbolic lookups in Perl 6
03:14		but then there's no strict refs anymore
03:15	new_perl6_hacker	Then all strings must ideally have 'eval' invokable on them
03:15	TimToady	eval is never ideal :)
03:15	lue	Tene: aah. I thought you were talking about P6 testing results :)
03:16	Tene	new_perl6_hacker: That's completely true, in exactly the same way as "all files can have the perl6 compiler run on them"
03:17	new_perl6_hacker	@tene , @TimToady thanks, got it now
03:17	* lue	.oO(bah. I wish I could upload whatever to wordpress.com . That'd make my life easier...)
03:18	Tene	new_perl6_hacker: No problem. Was there a specific task you were trying to perform or problem you were trying to solve, or just general curiosity?
03:18	new_perl6_hacker	I was looking at metaprogramming aspects
03:19		reflection basically
03:19	PerlJam	alpha: my @a = "a".."z"; @a."join"(" ")."say" # :)
03:19	p6eval	alpha 30e0ed: OUTPUT«a b c d e f g h i j k l m n o p q r s t u v w x y z␤»
03:20	new_perl6_hacker	But the example quoted by PerlJam earlier didn't show a 'eval' method, is that in the base class of string
03:21	Tene	You keep talking about classes. It's a function, not a method.
03:21	PerlJam	new_perl6_hacker: What other languages are you familiar with?
03:21	new_perl6_hacker	Ruby and Python
03:22	colomon	Tene: I think we added an eval method a while back, too.
03:22	PerlJam	I was getting a strong ruby vibe from you :)
03:22		mj41_ joined #perl6
03:22	PerlJam	(either that or smalltalk or something, but those are rared)
03:22		er, rarer
03:23	new_perl6_hacker	I think I'm getting confused, because I'm thinking everything from OOperspective .
03:23	colomon	rakudo: "say 'hello'".eval
03:23	p6eval	rakudo 974d9a: OUTPUT«hello␤»
03:24	PerlJam	alpha: "blah".^methods.map(*.name).sort.join(" ").say
03:24	p6eval	alpha 30e0ed: OUTPUT« :d :e :f :l :s ACCEPTS Array BUILD BUILDALL CREATE Capture Complex Hash Int Iterator PARROT REJECTS Scalar Scalar Str Str WALK WHENCE WHERE WHICH WHICH abs acos acosec acosech acosh acotan acotanh asec asech asin asinh atan atan2 atanh bless bytes can capitalize ceiling chars
03:24		..cho…
03:24	PerlJam	alpha: "blah".^methods.map(*.name).grep({ $_ ~~ /^e/ }).join(" ").say
03:24	p6eval	alpha 30e0ed: OUTPUT«encode eval elems evalfile exp end eigenstates␤»
03:25	PerlJam	there you go
03:26	Tene	alpha: say "foo".^can('eval')
03:26	p6eval	alpha 30e0ed: OUTPUT«eval␤»
03:26	Tene	alpha: say "foo".^can('eval').WHAT
03:26	p6eval	alpha 30e0ed: OUTPUT«␤»
03:28	PerlJam	hmm
03:28		why did WHICH appear in the list twice?
03:28		and Scalar and Str
03:30	lue	there's a WHENCE option? (001 !
03:30		rakudo: say 4.WHENCE
03:30	p6eval	rakudo 974d9a: OUTPUT«␤»
03:31	PerlJam	lue: WHENCE autovivifies an object
03:31		it's a closure from whence the object cometh
03:34	lue	'tis a closure whencefrom thy object cometh. (Old English translation free of charge :D )
03:36	new_perl6_hacker	:) just got a feeling that all programs(The mail program) go into a eval function inside the compiler.
03:37		The main program
03:37	Tene	That's right.
03:37		The compiler reads the program text into a string, invokes the compiler on it to get back a Code object, and then invokes that Code object.
03:38	new_perl6_hacker	This is great, so we basically have a compiler which we can use any where inside our program. This is a killer feature :)
03:40	lue	Our main goal is having the P6 interpreter run itself (be written in P6 code).
03:40		The key, IMO, is recursive definition :)
03:40	TimToady	well, it was a killer feature in Perl 3 or so
03:41	colomon	and in Lisp, Ruby, etc.
03:41		still, it's an awesome feature.
03:41	TimToady	but usually it's an escape valve indicating a missing feature elsewhere
03:43		we treat BEGIN the same say
03:43		Sarten-X joined #perl6
03:44	PerlJam	Why do we think all languages must be self-describable?
03:44	TimToady	and eval, in general, opens you up to all sorts of security problems, just like SQL injection
03:46		http://xkcd.com/327/
03:48	lue	TimToady: love that one :)
03:49	TimToady	PerlJam: we don't, just the good ones.
03:49		I blame Gödel...
03:49	PerlJam	and Escher and Bach ;)
03:50	lue	Don't forget Thought. That guy takes forever to figure things out. :)
03:52		.xkcd(fun with chess: http://xkcd.com/249/)
03:54		rakudo: say chr(0xFDD0)
03:54	p6eval	rakudo 974d9a: OUTPUT«﷐␤»
03:54	lue	.u �
03:54	phenny	U+FFFD REPLACEMENT CHARACTER (�)
03:54	PerlJam	you could do a good meta comic on photoshopping the chess board in that one
03:55	TimToady	don't photoshop the chess board, just photoshop me into xkcd :)
03:57	lue	WIN! (quick, somebody get Larry's head off of Yahoo Images!!!)
03:58	colomon	Woah, I knew the Gathering was in process, but I didn't know someone had already taken Larry's head!
03:58	lue	http://xkcd.com/365/ yet another conference the poor guy can't go to :)
04:01		I think XKCD ought to do a comic about Perl for once (with guest appearance by TimToady of course :) )
04:01	colomon	what do you mean "for once"? There are at least two classic comics about Perl.
04:02		I have the t-shirt for one...
04:02	PerlJam	colomon: yeah, but one of them is called "lisp"
04:02	lue	I honestly cannot remember if there werl Perl comics on XKCD :)
04:02		s/werl/were/
04:03	colomon	http://xkcd.com/224/
04:03		http://xkcd.com/208/
04:04	new_perl6_hacker	If an exception happen inside an eval, should the main program see it?
04:04	TimToady	no, it's trapped in $!
04:06	lue	according to number 224, Larry has tight (if possibly indirect) connections with God (all hail!)
04:07		.xkcd(http://xkcd.com/18/ we need more of these jokes)
04:09	* lue	will paste a bunch of xkcd comic links to pastebin instead of flooding the channel
04:21	lisppaste3	lue pasted "XKCD comic links" at http://paste.lisp.org/display/95996
04:23		mj41_ joined #perl6
04:42	lue	heelooo-oooh
04:44		rv2733 joined #perl6
04:45		abra joined #perl6
04:56	lue	saluton? Hello? Bonjour? anyone?
04:59	* TimToady	is hiding behind a cricket
05:01	lue	ah, that's where you are.
05:11		quietfanatic_ joined #perl6
05:12	new_perl6_hacker	What is the maximum value of anything Inf can hold?
05:12	diakopter	-Inf
05:13	TimToady	how much memory do you have?
05:13	new_perl6_hacker	:)
05:13	lue	rakudo: say 0..3; say [0,3]; say [0,4)
05:13	* diakopter	sprixels along
05:13	p6eval	rakudo 974d9a: OUTPUT«Confused at line 11, near "say [0,4)"␤current instr.: 'perl6;HLL;Grammar;panic' pc 500 (ext/nqp-rx/src/stage0/HLL-s0.pir:328)␤»
05:14	lue	.oO(interval notation would be interesting for the "official P6 module" (as well as set-builder, possibly))
05:14	TimToady	no need
05:14		0..^4 is already there
05:15		and doesn't use up two (count 'em two) brackets
05:15	lue	ah well, that's why I didn't say "...interesting project for implementation in Rakudo" :)
05:16	TimToady	mathematicians do many unforgiveable things as far as computers are concerned, and that's one of 'em
05:16	lue	rakudo: say 0..Inf
05:16	p6eval	rakudo 974d9a: ( no output )
05:16	lue	rakudo: say 0..^Inf
05:17	p6eval	rakudo 974d9a: ( no output )
05:17	TimToady	rakudo: say (0..Inf).batch(10)
05:17	p6eval	rakudo 974d9a: OUTPUT«Method 'batch' not found for invocant of class 'Range'␤current instr.: '_block14' pc 29 (EVAL_1:0)␤»
05:17	TimToady	rakudo: say (0..Inf).iterator.batch(10)
05:17	p6eval	rakudo 974d9a: OUTPUT«0123456789␤»
05:17	TimToady	oughta do that automatically, methinks
05:18		rakudo: say (0...*).batch(10)
05:18	p6eval	rakudo 974d9a: OUTPUT«0123456789␤»
05:19	TimToady	that one works
05:19	lue	I can't help but think about the possibility of built-in SDL/OpenGL support (thru Parrot). The upside is that it'd be a MAJOR perk over other languages.
05:19		The downside is probable bloatedness
05:20	new_perl6_hacker	diakopter why does http://perlcabal.org/~fglock/perlito.html give an error message to say 'hello,word!';
05:21	japhb	lue: OpenGL support is just waiting on Tene to reenable the 'use OpenGL:from<parrot>;' notation (lost during the branch switch). kthakore++ is working on SDL
05:22	lue	\o/ Get ready for the flood of awesome (no doubt to be recorded by XKCD ;) )
05:22	Tene	japhb: I've currently got nothing scheduled tomorrow. I'll try to fit it in then.
05:22		japhb: although I should probably also look at the pcc refactor branch.
05:22	japhb	Tene, \o/
05:23		aww. I hate tuit conflicts. :-/
05:24	lue	I got a Norweigan Blue Parrot for the occasion (of OpenGL/SDL implementation) :)
05:24	japhb	:-)
05:24	Tene	japhb: with any luck, I'll feel conflicted and not work on either of them. :)
05:24	japhb	:-P
05:25	lue	Beautiful plumage, the Norweigan Blue.
05:25	japhb	Just to make you feel guilty, I'm writing WebGL code in JavaScript, because I can't have my OpenGL in Perl 6.
05:25		That reminds me, I need to throw some tuits at plumage
05:26	lue	Tene: if you don't work on anything parrot-related (due to conflict), then you're the pet shop owner :)
05:38		gfx joined #perl6
05:44	new_perl6_hacker	Is http://perlcabal.org/~fglock/perlito.html the sprixel project?
05:46	TimToady	no
05:49	lue	goodbye, see you during the hackathon tomorrow
06:08		athenot joined #perl6
06:12		perlster joined #perl6
06:33		quietfanatic joined #perl6
07:53		uniejo joined #perl6
07:54	uniejo	All Perl6 speaks today are moved into room 4A16
07:55		aesop joined #perl6
08:10		jonasbn joined #perl6
08:16		new_perl6_hacker joined #perl6
08:41		meppl joined #perl6
08:43		iblechbot joined #perl6
09:06		Su-Shee joined #perl6
09:20		uniejo joined #perl6
09:32	k23z__	err ..
09:32		room 4A16 ?
09:33		kfo joined #perl6
09:35		arlinius joined #perl6
09:35	arnsholt	k23z__: That's where we are, yes
09:36		The Copenhagen hackathon =)
09:37	Su-Shee	good morning
09:37	arnsholt	'morning Su-Shee
09:38	k23z__	hmm
09:51		mberends joined #perl6
09:52	jonasbn	k23z__: where are you at?
09:53		k23z__: did you make it here or do you require guidance?
09:53	mberends	heehee, the Copenhagen Perl 6 Hackathon has a good turnout, about 25 people :-)
09:56		jnthnandmasakbot++ is giving what would have been moritz_++'s grammar and regex talk
09:57	Su-Shee	25 people? nice. you can do a lot with 25 people in 48 hours.
09:57	mberends	Yet Another Example of JIT talk preparation
09:58	baest	:)
09:58	jonasbn	Su-Shee: we are not contribution right now, we are in the brainwash phase
09:58		contributing even
10:00		cotto joined #perl6
10:00		xabbu42 joined #perl6
10:04		payload joined #perl6
10:08	Su-Shee	well I hope you've got fun and it's a great weekend. :)
10:09	mberends	:)
10:18	k23z__	jonasbn: sorry for confusing you, I am not there , I am maybe 2000km away
10:21		rv2733 joined #perl6
10:26		orafu joined #perl6
10:42		eternaleye joined #perl6
10:47		fridim joined #perl6
10:48		xinming_ joined #perl6
10:51		krr joined #perl6
10:57	jonasbn	ah
10:57		hehehe
10:57	lisppaste3	baest pasted "Patch which ensures === returns bool not int" at http://paste.lisp.org/display/96001
10:58	baest	based on http://github.com/rakudo/rakud[…]1b5ca4855845517e5 and http://github.com/rakudo/rakud[…]efe612a7e10487525
10:59		fod joined #perl6
10:59		fod left #perl6
11:02	mberends	nom &
11:12		z3ndrag0n joined #perl6
11:30		kidd joined #perl6
11:43		uniejo joined #perl6
11:43		snarkyboojum joined #perl6
11:49		Su-Shee joined #perl6
11:51		clintongormley joined #perl6
11:56		snarkyboojum_ joined #perl6
12:03		masak joined #perl6
12:03	masak	oh hai from Copenhagen, #perl6!
12:03	arnsholt	Oh hai from the next row, masak!
12:04		quester joined #perl6
12:04		mberends joined #perl6
12:05	masak	mberends: lol I can see you both on- and offline! :D
12:05		jnthn++ is on now, by the way.
12:05		giving his famous "solved in Perl 6" talk.
12:06	mberends	masak: stop chattering and pay attention to the talk!
12:07	* masak	obliges :)
12:07		cognominal joined #perl6
12:07		payload joined #perl6
12:09	Su-Shee	oO(live from the benches in copenhagen.. ;)
12:09		payload joined #perl6
12:09	arnsholt	Su-Shee: Peanut gallery, more like ;)
12:10	pugssvn	r29952 \| mberends++ \| [docs/talks/p6.imp.2010.03.odp] slides of "Perl 6 Implementations in March 2010" from the Dutch Perl Workshop and the Copenhagen Perl 6 Hackathon 5th and 6th March 2010 (OpenOffice Presentation format)
12:13	ned	so jealous.
12:23	colomon	o/
12:30		wknight8111 joined #perl6
12:33	quester	rakudo: augment class Int { method even { ? ( self % 2 ) } }; 42.even.say;
12:33	p6eval	rakudo 974d9a: OUTPUT«Method 'even' not found for invocant of class 'Integer'␤current instr.: '_block14' pc 29 (EVAL_1:0)␤»
12:33	quester	rakudo: augment class Integer { method even { ? ( self % 2 ) } }; 42.even.say;
12:33	p6eval	rakudo 974d9a: OUTPUT«Can't augment something that doesn't exist␤current instr.: 'perl6;Perl6;Grammar;add_name' pc 34122 (src/gen/perl6-grammar.pir:242)␤»
12:34	* quester	wonders what class 42 really is...
12:35	masak	my talk is online now. it's linked at the bottom of http://feather.perl6.nl/~masak/
12:36		thanks to all the people on the channel who helped yesterday with the slides :)
12:36	mberends	rakudo: say 42.WHAT # for quester
12:36	p6eval	rakudo 974d9a: OUTPUT«Int()␤»
12:37	colomon	rakudo: say PARROT(42)
12:37	p6eval	rakudo 974d9a: OUTPUT«Integer␤»
12:37	colomon	rakudo: say PARROT(1+41)
12:37	p6eval	rakudo 974d9a: OUTPUT«Integer␤»
12:37	arnsholt	Ah, there are some issues with augmenting classes that come from Parrot, no?
12:39	colomon	Might be more accurate to simple say there are issues with classes that come from Parrot.
12:39		I don't know if the augmenting is bad too or just a window into the problems.
12:40	masak	the latter, methinks.
12:41	quester	mberends, colomon, masak: I think I've seen augments in some of the Rakudo internals. Hmm... can anyone suggest a good way of converting Integer to Int offhand?
12:42	masak	quester: what you're seeing is some kind of Parrot/Rakudo discrepancy that shouldn't really be visible.
12:42	colomon	rakudo: say pir::box__PI(42)
12:42	p6eval	rakudo 974d9a: OUTPUT«42␤»
12:43	colomon	rakudo: say PARROT( pir::box__PI(42))
12:43	p6eval	rakudo 974d9a: OUTPUT«Integer␤»
12:43	colomon	blast, that doesn't help.
12:43	quester	rakudo: pir::box__PI(42).perl.even.say;
12:43	p6eval	rakudo 974d9a: OUTPUT«Method 'even' not found for invocant of class 'String'␤current instr.: '_block14' pc 29 (EVAL_1:0)␤»
12:44	* quester	...String...?
12:44	colomon	.perl returns Str
12:44		well, String maybe.
12:44		you're finding all the cracks this morning.
12:44	quester	colomon: Thank you. Brain fail.
12:44	colomon	and my fuddled brain isn't remembering how to work around them.
12:45	masak	quester: 'String' is also a Parrot type.
12:46	quester	rakudo: "42".even.say;
12:46	p6eval	rakudo 974d9a: OUTPUT«Method 'even' not found for invocant of class 'Perl6Str'␤current instr.: '_block14' pc 29 (EVAL_1:0)␤»
12:46		iblechbot joined #perl6
12:47	quester	masak: Isnt Perl6Str supposed to be, um... hiddien away inside of Rakudo, too?
12:47	uniejo	rakudo: say (+pir::box__PI(42)).WHAT
12:47	quester	s/hiddien/hidden/
12:47	p6eval	rakudo 974d9a: OUTPUT«Num()␤»
12:47	masak	quester: yes, I think so.
12:47		quester: sometimes error messages leak through.
12:48	colomon	rakudo: say PARROT(42.Num.Int)
12:48	p6eval	rakudo 974d9a: OUTPUT«Int␤»
12:48	colomon	aha!
12:48		rakudo: augment class Int { method even { ? ( self % 2 ) } }; 42.Num.Int.even.say 41.Num.Int.even.saty
12:49	p6eval	rakudo 974d9a: OUTPUT«Confused at line 11, near "42.Num.Int"␤current instr.: 'perl6;HLL;Grammar;panic' pc 500 (ext/nqp-rx/src/stage0/HLL-s0.pir:328)␤»
12:49	masak	copypaste fail.
12:49	colomon	rakudo: augment class Int { method even { ? ( self % 2 ) } }; 42.Num.Int.even.say; 41.Num.Int.even.say;
12:49	p6eval	rakudo 974d9a: OUTPUT«0␤1␤»
12:49	TimToady	good morning
12:49	colomon	So augment works, but .even is backwards.
12:49		TimToady: o/
12:49	TimToady	who's winning?
12:50	pmichaud	hello, #perl6
12:50	colomon	\o
12:50	masak	pmichaud! \o/
12:50	TimToady	hah, I beat pmichaud
12:50	quester	colomon: Thank you. Err... yes, that would be another brain failure on my part, sorry...
12:50	masak	pmichaud, TimToady: all the peeps here say OH HAI
12:50		you'll soon be on-screen, I expect.
12:50	pmichaud	OH HAI PEEPS!
12:50	TimToady	OH HAI PEEPS!!!
12:50	pmichaud	(hah, I beat TimToady)
12:51	TimToady	hah, beat pmichaud++ again on !!!
12:51	masak	:)
12:51	jnthn	oh hai
12:51	masak	not on-screen yet. :)
12:51	TimToady	aww
12:51	masak	now!
12:51	colomon	oh, no, am I going to have to behave this morning?
12:51	pmichaud	OH HAI PEEPS!!!
12:51	masak	lol
12:51	colomon	o/
12:51	jnthn	:-)
12:52	masak	so in a sense, we're a lot more than three people who are able to answer panel questions...
12:52	TimToady	but correctly, that's another matter...
12:55	masak	lol http://twitter.com/Jippi/status/10071258756
12:55		jonasbn joined #perl6
12:55	jonasbn	http://twitter.com/Jippi/status/10071258756
12:56		argh masak beat me to it
12:56	TimToady	Perl 6 is trying to beat everyone to everything, more or less...
12:57	jonasbn	so actually the only scary thing is the high chance of success
12:58	pmichaud	I _wish_ that were the only scary thing :)
12:58	TimToady	jnthn++ is always scary
12:59	jnthn	..compliment? :-)
12:59		.oO( well, at least I got karma... )
12:59	masak	jnthn++ # scary!
13:00	mberends	jnthn ~~ scary # True
13:01	TimToady	rakudo: say "scary" but True
13:01	p6eval	rakudo 974d9a: OUTPUT«No applicable candidates found to dispatch to for 'infix:<does>'␤current instr.: 'infix:<but>' pc 226837 (src/gen/core.pir:2921)␤»
13:01	TimToady	alpha: say "scary" but True
13:01	p6eval	alpha 30e0ed: OUTPUT«The but operator can only be used with a role or enum value on the right hand side␤in Main (file <unknown>, line <unknown>)␤»
13:01	TimToady	well, at least it's a different error now...
13:02		or, maybe, it was a different error then...
13:06	colomon	Ah, Bool's not an Enum yet because Enums don't actually work yet, right?
13:06	masak	when I write 'our sub foo', the scope of the parameters in the signature of foo is a global scope rather than a lexical scope, yes?
13:06	TimToady	uh, no
13:06		only 'foo' is aliased
13:07	masak	so, how about this?
13:07		rakudo: my $a = 1; sub foo() { my $a = 2; our sub bar($x where $a) { say $x } }; bar(1); bar(2)
13:07	p6eval	rakudo 974d9a: OUTPUT«1␤Constraint type check failed for parameter '$x'␤current instr.: 'bar' pc 255 (EVAL_1:102)␤»
13:09	TimToady	looks like it's getting the wrong $a to me
13:09	* masak	submits rakudobug
13:09	TimToady	($a isn't a parameter though)
13:10	pmichaud	sub foo() { my $a = 2; our sub bar($x where $a) { say $x } }; bar(2);
13:10		rakudo: sub foo() { my $a = 2; our sub bar($x where $a) { say $x } }; bar(2);
13:10	p6eval	rakudo 974d9a: OUTPUT«Symbol '$a' not predeclared in bar␤current instr.: 'perl6;PCT;HLLCompiler;panic' pc 137 (compilers/pct/src/PCT/HLLCompiler.pir:101)␤»
13:10	pmichaud	looks like the where clause is attached to the wrong outer scope
13:10	TimToady	it does at that
13:11		Zn4rK joined #perl6
13:12	TimToady	std: sub foo() { my $a = 2; our sub bar($x where $a) { say $x } }; bar(2);
13:12	p6eval	std 29952: OUTPUT«ok 00:01 109m␤»
13:13	TimToady	STD might actually be looking up $a starting inside the inner sub's scope
13:13		in fact, it has to, in case it's an earlier parameter
13:13	masak	nod.
13:13	TimToady	std: sub foo() { my $a = 2; our sub bar($x where $x) { say $a } }; bar(2);
13:14	pmichaud	I think that rakudo somewhat gets that right already
13:14	p6eval	std 29952: OUTPUT«ok 00:01 107m␤»
13:14	masak	a parameter list actually introduces a lot of small nested scopes, doesn't it?
13:14	pmichaud	rakudo: sub foo($a, $x where $a) { say $x; }; foo(2,2);
13:14	p6eval	rakudo 974d9a: OUTPUT«2␤»
13:14	pmichaud	rakudo: sub foo($a, $x where $a) { say $x; }; foo(2,3);
13:14	p6eval	rakudo 974d9a: OUTPUT«Constraint type check failed for parameter '$x'␤current instr.: 'foo' pc 181 (EVAL_1:77)␤»
13:15	TimToady	so it sounds like it's splitting a hair that doesn't need splitting
13:15	pmichaud	it's just that the scope of the parameter block ends up with the wrong outer, I think
13:15	TimToady	I don't have a parameter block, I just use the one that will end up inside {}
13:15	Zn4rK	hej niklas
13:16	TimToady	scared me, thought it was christmas already
13:16	frettled	Now that's worrysome.
13:17	masak	jnthn++ mberends++ # you are so awesome
13:17	* pmichaud	agrees with masak++
13:19	TimToady	*++
13:19	masak	std: *++
13:19	p6eval	std 29952: OUTPUT«ok 00:01 104m␤»
13:19	TimToady	rakudo: say grep *++, 1,2,3
13:19	p6eval	rakudo 974d9a: OUTPUT«Cannot assign to readonly value␤current instr.: '&die' pc 16803 (src/builtins/Junction.pir:347)␤»
13:20	masak	that's correct, no?
13:20	TimToady	rakudo: my @foo = 1,2,3; say grep *++, @foo
13:20	p6eval	rakudo 974d9a: OUTPUT«Cannot assign to readonly value␤current instr.: '&die' pc 16803 (src/builtins/Junction.pir:347)␤»
13:20	colomon	wait, did that whatever just form a closure in Rakudo?
13:20	TimToady	maybe not
13:20	pmichaud	no, it's attempting to increment the Whatever constant :-)
13:20	TimToady	rakudo: say (*++).WHAT
13:20	p6eval	rakudo 974d9a: OUTPUT«Cannot assign to readonly value␤current instr.: '&die' pc 16803 (src/builtins/Junction.pir:347)␤»
13:20	pmichaud	rakudo: say (*).WHAT;
13:20	TimToady	I think the answer is no
13:20	p6eval	rakudo 974d9a: OUTPUT«Whatever()␤»
13:21	pmichaud	rakudo: say (*).succ
13:21	p6eval	rakudo 974d9a: OUTPUT«!whatever_dispatch_helper␤»
13:21	colomon	rakudo: say (*+1).WHAT
13:21	p6eval	rakudo 974d9a: OUTPUT«Method 'Num' not found for invocant of class 'Block'␤current instr.: 'perl6;Mu;' pc -1 ((unknown file):-1)␤»
13:21	pmichaud	(no, we aren't doing whatever closures just yet)
13:21	TimToady	alpha: say (*++).WHAT
13:21	p6eval	alpha 30e0ed: OUTPUT«Block()␤»
13:21	TimToady	alpha: my @foo = 1,2,3; say grep *++, @foo
13:21	p6eval	alpha 30e0ed: OUTPUT«123␤»
13:22	TimToady	alpha: my @foo = 1,2,3; say grep ++*, @foo
13:22	p6eval	alpha 30e0ed: OUTPUT«123␤»
13:22	TimToady	alpha: my @foo = 1,2,3; say map ++*, @foo
13:22	p6eval	alpha 30e0ed: OUTPUT«111␤»
13:22	TimToady	er...
13:22	masak	it's not forming a closure.
13:22		that's the whole explanation.
13:22	TimToady	alpha: my @foo = 1,2,3; say map +*, @foo
13:22	p6eval	alpha 30e0ed: OUTPUT«123␤»
13:24	TimToady	well, gee, the spec generalized it to all unaries and binaries, what, a whole week ago. why isn't it done yet? esp in alpha, that had a month head start...
13:24	masak	:)
13:24		std: * ?? * !! *
13:24	p6eval	std 29952: OUTPUT«ok 00:01 105m␤»
13:24	TimToady	but not to ternery :)
13:24	colomon	TimToady: should that be all binaries excepting the Range constructors?
13:25	TimToady	all ops except those that declare they will do it themselves
13:25		which Range will
13:25	colomon	How do you declare that?
13:26	TimToady	proto infix:<..> ($min, $max) is like(Whatever-and-stuff) {...}
13:26		where "stuff" is WhateverCode
13:27		:)
13:27	Zn4rK	niklas jag tror i m�ste g�
13:27	masak	Zn4rK: fel kanal?
13:27	TimToady	S02:936
13:29	* TimToady	wonders if 'fel' is related to olde english 'fell', as in 'a fell beast'
13:29	* masak	checks
13:30	masak	'fell' comes from Goth 'falljan', 'to cause to fall'.
13:30		oh wait.
13:30		that's the verb.
13:31	TimToady	to cause jan to fall :)
13:31	masak	the adjective comes from Old French, nominative of 'felon', wicked.
13:31	frettled	masak: so that fell was a FAIL :)
13:31	pmichaud	'felon' vs. 'fallen', I wonder?
13:31	TimToady	which might still be related to "fallen" if you go back
13:31		where's my AmHer...
13:32	masak	Swedish 'fel' ultimately comes from French 'faute', so seemingly not related.
13:32	arnsholt	Yeah, Norwegian "feil" is ultimately connected with Latin fallere
13:33	masak	we also have a verb, 'att fela', meaning 'to err'.
13:33	colomon	TimToady++
13:33	masak	so it seems to have more to do with failure than falling.
13:33	arnsholt	I think the f in falljan would have a different reflex in its Latin cognate
13:33	frettled	Rewind: maybe I should've said «that was a fell fel fail»
13:33	masak	frettled: :)
13:34		so, actually, Lolnglish is sort of reviving the Swedish 'fel' through its 'FAIL'. they even have the same etymological roots.
13:34	TimToady	felon was "yellow" or "bile" in I-E
13:35	masak	that's wicked, for sure.
13:35	jnthn	Grr. :-/ So Rakudo won't build on Win 7. :-(
13:35		(segfault when loadlib-ing dynops from a PBC...)
13:36	colomon	TimToady: back to the whatever closures... if I say something like $a = *+3; $b = $a+2 is that expected to leave $b with { $_ + 5 } effectively?
13:37	TimToady	good question
13:37	quester	Well, if you're done with the fell fel fail, I'm going to eat. I'm longing for some imitation Vietnamese noodle soup, off fo' faux pho &.
13:38	TimToady	I dunno
13:38	masak	quester: :)
13:38	* colomon	has been dreaming about math on closures for about ten months now.... ;)
13:39	TimToady	'course, 'round here people pronounce it Vietnamese, which doesn't rhyme
13:41		quester left #perl6
13:43	TimToady	if we did, we'd probably have to make Whatever a kind of junctional, outside-of-any type
13:44	pmichaud	well, that was the reason Rakudo originally came up with WhateverCode -- so that (*+3)+2 would still produce a code object of some sort.
13:44	TimToady	but instead of autothreading, autocurry
13:44		yes, but without writing out the actual subs
13:45	* TimToady	tries to grok * + (1\|2)
13:46	masak	o.O
13:47	pmichaud	alpha: my $a = * + (1\|2); say $a(3);
13:47	p6eval	alpha 30e0ed: OUTPUT«any(4)␤»
13:47	masak	hm.
13:47	pmichaud	hm.
13:47	TimToady	hm!
13:47		hm?
13:47	colomon	huh?
13:47	pmichaud	alpha: my $a = 3 + (1\|2); say $a;
13:47	p6eval	alpha 30e0ed: OUTPUT«any(4, 5)␤»
13:48	pmichaud	well, alpha sees * + (1\|2) as a junctional dispatch
13:48		so I'm not sure what ends up in $a
13:48		alpha: my $a = * + (1\|2); say $a.WHAT;
13:48	p6eval	alpha 30e0ed: OUTPUT«Junction()␤»
13:49	pmichaud	so, $a ends up as a junction of WhateverCode, I guess.
13:49	TimToady	.(3) oughta autothread over that, ya'd think
13:49	jnthn	Aye, it'd auto-thread.
13:50		kfo joined #perl6
13:52	* TimToady	wonders whether Whatever and WhateverCode are really strange parametric junctions...
13:53	TimToady	what's my brane doing all over the room in bits?
13:54	masak	we're simultaneously praising and hating {*} and #= right now in Copenhagen...
13:54	TimToady	#= is already dead, as far as I'm concerned
13:55	masak	it's still in S05.
13:55	pmichaud	yes, it doesn't get used much anymore
13:55	jonasbn	that would be undead then
13:55	TimToady	hmm
13:55	pmichaud	we could likely de-huffmanize #= now that it's less common
13:55	masak	yes, please.
13:55	pmichaud	i.e., put any parameters as part of the {*} token.
13:56		s/parameters/arguments/
13:56	pugssvn	r29953 \| lwall++ \| [S05] kill {*} and #=
13:56	masak	\o/
13:56	pmichaud	oh, is {*} gone?
13:57		that may end up being... unfortunate.
13:57		although I guess one can always do a direct closure instead.
13:57	jnthn	...grr, that's a pain.
13:57	TimToady	the presence of {*} almost always indicates the need to refactor
13:57	jnthn	You get away with it in STD because you only care about parsing, not actually Doing Stuff. :-)
13:57	TimToady	STD hasn't used {*} for months
13:58	masak	TimToady: I like it when you use that kind of phrase ('$X indicates the need to refactor'). there's really no good response to it :)
13:58	jnthn	We may be able to find another way...
13:58	pmichaud	jnthn: we already have another way
13:58	TimToady	well, we can still have an action call, but {*} #= won't be it, I suspect
13:58	jnthn	pmichaud: well, yes, but I'm lazy :-)
13:58	TimToady	see <O()>
13:59		oh, I had a problem with <O()> last night, thinking about default parse rules
13:59		suppose we make the body of a proto be the default
13:59		we get
13:59		proto token infix is binary is defequiv(%additive) { <sym> <O(\|%additive)> }
14:00	masak	'defequiv'?
14:00	TimToady	but then the O is wrong if they 'is equiv' it
14:00		that's redundant, if we put the <O>
14:00		proto token infix is binary { <sym> <O(\|%additive)> }
14:00		or
14:00		proto token infix is binary is defequiv(%additive) { <sym> <O(*)> }
14:00	colomon	jnthn: any luck with Hash attributes?
14:01	jnthn	colomon: I need to fix Rakudo building here before I can do much :-(
14:01	TimToady	but still not sure that the body of the proto is the right place for a default "is parsed"
14:01		it's a handy place tho
14:01	jnthn	(Get a segfault)
14:01	colomon	jnthn: oh noes!
14:02	masak	std: ?( $RIGHT === $HANDY )
14:02		new_perl6_hacker joined #perl6
14:02	p6eval	std 29952: OUTPUT«ok 00:01 108m␤»
14:02	masak	new_perl6_hacker: welcome back!
14:02		martinro joined #perl6
14:02	masak	martinro: hey :)
14:02	TimToady	s/new/newish/
14:02	new_perl6_hacker	thanks, I was there this morning ...
14:02	colomon	jnthn: I've been messing around a bit with the class and role tests a bit, and there are all sorts of weird little bugs...
14:03	jnthn	colomon: Yes.
14:04		There's a reason a bunch of S12-foo are still commented out. :-)
14:04	colomon	jnthn: but I won't bug you with them until you've got a Rakudo build again. :)
14:04	jnthn	Yeah...the issue is...nasty.
14:05		OH!!
14:06	colomon	The problem is you're using Windows? :)
14:06	masak	colomon: how intolerant! :)
14:06	jnthn	IIUC, the problem is that a realloc call is made...and the return value is ignored.
14:06		sigh
14:07	masak	jnthn: so freed memory space is subsequently overwritten?
14:07	colomon	that seems like a bad thing on any platform?
14:07	jnthn	well, the point ain't updated to the new location
14:07		colomon: Yes, it is.
14:07		C writing fail.
14:08		re-compiled Parrot now...maybe it's fixed it.
14:08	* jnthn	crosses fingers
14:08	masak	he actually crossed them, too.
14:08	colomon	masak: For the record, Windows machines were my primary computers from 1995-2008. But I hated Vista so much I decided to get my first Mac...
14:09	jnthn	omg, I think it helped...
14:09		colomon: I just stuck with XP. :-)
14:09		(Until Win7 arrived)
14:09	masak	colomon: Software sucks.
14:10	arnsholt	jnthn: I knocked on wood earlier today. It triggered my segfault again =)
14:10	TimToady	my problem on the proto body is that we also currently write the <infix> dispatch using the same name, so we'd essentially be ignoring the given body and substituting the dispatcher, and only using the proto body for the generic 'is parsed'
14:11	jnthn	OK, I have a patch that fixes my build. I can haz a productive again.
14:11	colomon	\o/
14:11	TimToady	jnthn++ has a fail fail
14:12		Psyche^ joined #perl6
14:12	masak	no, a fail fail isn't the same thing as fixing something.
14:12	TimToady	does that make it a fail fail fail?
14:12	masak	the fail had a fail, possibly.
14:12	jnthn	Well, it does now fail to fail.
14:13	masak	How much fail would a fail fail fail, if a fail fail could fail fail?
14:13	TimToady	cheezburgerz
14:14		speaking of which, the sky is starting to get light here
14:14	m-locks	we needs a LOL version of perl, like that lolpython
14:14	masak	didn't quietfanatic do an NCI in Perl 6 at one point?
14:15		TimToady: not much of a panel going on here, I'm afraid.
14:15	TimToady	yeah, I saw they took the slot for something else
14:15	masak	mostly a kind of newbie tutorial.
14:15	pmichaud	(newbie tutorial) +1
14:15	TimToady	ISFINE
14:15	pmichaud	that was what I had been planning for my slot anyway
14:16	masak	right now jnthn++ and mberends++ are doing the best to scare the newbies away by explaining their scary/awesome NCI code. :)
14:16	m-locks	heh
14:16	TimToady	was planning to spend the weekend several TZ to the east of here anyway, so as not to miss all the hackathon
14:16	pmichaud	same here
14:16		(still planning to do that)
14:17	TimToady	luv 2 scare dey little newbies way
14:17	m-locks	i've been translating those tutorials moritz__ and lichtkind have written, at least i can read them quite easily
14:17	colomon	I ended up with a very busy day scheduled for tomorrow, alas -- playing music for two church services in the morning, GMing in the afternoon.
14:18	TimToady	yes, well, I can read Russian till the time comes to understand it
14:18	jnthn	It wasn't scary, it was just beautifully terrifying.
14:18	arnsholt	That NCI stuff was very, very cool
14:19		new_perl6_hacker joined #perl6
14:19	TimToady	I'm only playing for one, but I play twice as loud, since I'm the drummer.
14:19	pmichaud	Much of Perl 6 is beautifully terrifying. See also "Camelia".
14:19	arnsholt	Hehe
14:19	new_perl6_hacker	http://perlcabal.org/syn/S17.html will there be support for event like activities on files and sockets etc
14:19	TimToady	I'd be terrified if I met a butterfly with a 3 meter wingspan...
14:20		or was it 5?
14:20	pmichaud	3, last I heard.
14:20	masak	TimToady: 5 feet, perhaps?
14:20	TimToady	mebbe she grew
14:20		she's young, after all
14:20	colomon	TimToady: I was asked to play a drum (bodhran) for this as well, and had to point out that I don't own one and don't know how to play it, either. :)
14:21	jnthn	Parrot r44705 fixes a nasty Rakudo-build segfault.
14:21		But doesn't seem to hit many people.
14:21		(You usually luck out.)
14:21	TimToady	colomon: just play it like a conga
14:21	colomon	jnthn: should we bump parrot in rakudo, then?
14:22	jnthn	colomon: Can do, or can wait for the next reason to bump.
14:22		Doesn't see to hit anyone else.
14:22	colomon	TimToady: I can't really properly play conga, either. Though I did pretend to play quad toms in high school.
14:22	jnthn	Also I didn't check that no other Parrot changes cause breakage.
14:22	masak	jnthn: s/Thought/Thou/. otherwise, nice message :)
14:22		hm. s/Though/Thou/, I mean.
14:22	jnthn	oh crap
14:22		I'm tired! :-)
14:22	colomon	I can try bumping parrot here, mostly just watching the boy on a Saturday morning.
14:23	new_perl6_hacker	http://perlcabal.org/syn/S17.html will there be support for event like activities on files and sockets etc????
14:23	TimToady	my boy don't take much watchin' anymore
14:23		he'll proably sleep the night through, and then some...
14:23		b
14:24	baest	rakudo: use Test; my @a = 2,3; is(@a, (3, 2))
14:24	p6eval	rakudo 974d9a: OUTPUT«not ok 1 - ␤# got: [2, 3]␤# expected: (3, 2)␤»
14:24	baest	rakudo: use Test; my @a = 2,3; is(@a, (2, 3))
14:24	p6eval	rakudo 974d9a: OUTPUT«not ok 1 - ␤# got: [2, 3]␤# expected: (2, 3)␤»
14:24	colomon	uh, that's weird
14:24		Parrot revision r44705 required (currently r44371)
14:24		borup joined #perl6
14:24	colomon	To automatically checkout (svn) and build a copy of parrot r44705,
14:24		try re-running Configure.pl with the '--gen-parrot' option.
14:24		Or, use the '--parrot-config' option to explicitly specify
14:24		the location of parrot_config to be used to build Rakudo Perl.
14:24	baest	yeah, I thought so :)
14:25	TimToady	rakudo: use Test; my @a = 2,3; is(@a, [2,3])
14:25	p6eval	rakudo 974d9a: OUTPUT«ok 1 - ␤»
14:25		jferrero joined #perl6
14:25	colomon	that's what I get when I run perl Configure.pl --gen-parrot
14:26	masak	baest: it's the difference between a Seq and an Array, I think.
14:26	baest	TimToady: yes, I guessed that, but I've seen several test in spectest which does the is(@a, (2,3) style
14:26	TimToady	probably from back when lists promoted to arrays
14:27	baest	ok, so that should be change in the tests, right?
14:27	TimToady	yeah, unless we change our minds again :)
14:27	masak	lists don't promote to arrays?
14:28	pmichaud	all of the list/array/iterator/seq stuff is "up in the air" at the moment, I think.
14:28		well, not all, but a lot.
14:28	masak	happy juggling.
14:28	pmichaud	at the moment, parcels promote to Seq
14:28		(in item conetxt)
14:28	jnthn	colomon: Got example to hand of the has %.foo issue?
14:28	TimToady	S03:3105
14:28		grr
14:28	pmichaud	my $a = (1,2,3); say $a.WHAT;
14:28	TimToady	S02:3105
14:28	pmichaud	rakudo: my $a = (1,2,3); say $a.WHAT;
14:28	TimToady	sb Seq
14:28	p6eval	rakudo 974d9a: OUTPUT«Seq()␤»
14:29	colomon	rakudo: class A { has %.foo; }; my $a = A.new;
14:29	p6eval	rakudo 974d9a: ( no output )
14:29	pmichaud	rakudo: my $a = (1,2..5,3); say $a.WHAT;
14:29	p6eval	rakudo 974d9a: OUTPUT«Seq()␤»
14:29	colomon	rakudo: class A { has %.foo; }; my $a = A.new; say $a.foo.WHAT
14:29	p6eval	rakudo 974d9a: OUTPUT«Method 'WHAT' not found for invocant of class 'Hash'␤current instr.: '_block14' pc 29 (EVAL_1:0)␤»
14:29	colomon	jnthn: there you go
14:29	pmichaud	I think likely a BUILD issue
14:29	TimToady	WHAT!
14:29	jnthn	> class A { has %.foo; }; my $a = A.new; say $a.foo.WHAT
14:30		Hash()
14:30		Fixed...spectesting.
14:30	masak	jnthn++
14:30		jnthn is back in the saddle!
14:30	TimToady	no, that one is ++jnthn
14:30	masak	:)
14:30	colomon	ye-haw!
14:30	* masak	hasn't gotten into the habit of giving prefix karma yet
14:30	TimToady	now fix the rest of 'em before we ask :P
14:31	masak	when's --ircnick applicable?
14:31	jnthn	When I made the silly bug in the first place. :-)
14:31	masak	jnthn: no, it must be before something.
14:32	TimToady	I think it has to be a spec error
14:32	masak	making a silly bug in the first place merits a jnthn--
14:32	jnthn	:-P
14:32		I'll try and be more perfect.
14:32		I blame my lapse on insufficient pivo.
14:33	masak	there is no 'try'... and no 'more perfect' either :P
14:33	TimToady	--gen-parrot
14:33	masak	:)
14:33	jnthn	rakudo: try { more perfect }; say "see?"
14:33	p6eval	rakudo 974d9a: OUTPUT«see?␤»
14:33	masak	++jnth :)
14:33		*jnthn
14:33	TimToady	um
14:34		std: try { more perfect }; say "see?"
14:34	p6eval	std 29953: OUTPUT«Undeclared routines:␤ 'more' used at line 1␤ 'perfect' used at line 1␤ok 00:01 107m␤»
14:34	jnthn	Yeah, we don't warn yet on undefined routines.
14:34	masak	I won't report that.
14:34	jnthn	...becuase you already did last week? :-)
14:35	masak	quite possibly.
14:35	jnthn	:-P
14:35	masak	I don't know anymore! the RT is too big!
14:35	TimToady	well, now, whose fault is that?
14:35	masak	when I said I'd deluge you with tickets, I expected you all to frigth back!
14:35	pmichaud	14:31 <jnthn> When I made the silly bug in the first place. :-)
14:36		masak: I did close a few tickets last week :)
14:36	masak	++pmichaud
14:36	* TimToady	is trying to figure out how to frigth
14:36	masak	hm, I guess that's for future ticket closing.
14:36		TimToady: it's related to how babby is formed.
14:36	TimToady	oh, okay
14:36	pmichaud	anyway, I'm hoping we can close some tickets over the next couple of hackathon days :)
14:37	jnthn	I'm truly sorry for your lots
14:37	masak	jnthn: 'truely' :)
14:37	TimToady	chroolie
14:37	jnthn	...ok, now I don't know which is the lol spelling and which is the English.
14:37	masak	who does? :)
14:37	jnthn	Ceiling cat.
14:37	baest	rakudo: my @a; say @a.list
14:37	p6eval	rakudo 974d9a: OUTPUT«Null PMC access in elements()␤current instr.: 'perl6;SeqIter;get' pc 13979 (src/builtins/Routine.pir:113)␤»
14:38	* masak	submits rakudobug
14:38	jnthn	oh no, not that bug.
14:38	* TimToady	gives jnthn++ another no so he can have noes
14:38	masak	lolland
14:39	jnthn	colomon: yay, patch works
14:39		I gotta go finish moving house now...will push, maybe you can enable some spectests.
14:39	masak	rakudo: my @a = 1, 2, 3; say @a.list
14:39	p6eval	rakudo 974d9a: OUTPUT«123␤»
14:39	colomon	I'm spectesting the new parrot now... but I had to blow away my parrot directory to get --gen-parrot to work with it!
14:39	masak	rakudo: my @a = 1; shift @a; say @a.list
14:40	p6eval	rakudo 974d9a: OUTPUT«␤»
14:40	colomon	jnthn++
14:40	mberends	travel Copenhagen -> Lund -> Copenhagen &
14:40	pmichaud	.list is still suspect-ish
14:40	colomon	Does it work? Should it exist? What should it do? ;)
14:40	pmichaud	exist, yes.
14:41		it basically returns the "list" form of the invocant
14:41		its exact specification and implementation need some rethinking
14:41		I'm likely to work on that today.
14:41	colomon	pmichaud++
14:41	pmichaud	(unless other events here overtake me again, as they did yesterday)
14:41	jnthn	pushed
14:41		gone
14:41		o/
14:42	masak	o/
14:42	colomon	\o
14:42	TimToady	o\
14:43	dalek	rakudo: 8505d39 \| Worthington@.(none)++ \| src/builtins/Mu.pir:
14:43		rakudo: Fix has %.noms.
14:43		rakudo: review: http://github.com/rakudo/rakud[…]1ca9a2b4134823ffb
14:44		robertml joined #perl6
14:46	colomon	Nice author name there...
14:47	pugssvn	r29954 \| baest++ \| Put list inside [] to do array comparison
14:49		r29955 \| baest++ \| Put list inside [] to do array comparison, fix error in test which was chop readonly string and fudge an empty array test
14:55	baest	colomon: do you remember yesterday, your change regarding returning Bool in ===, I've noticed that you didn't update Num. Also there was a minor thing with the syntax, so here's a patch if you want it: http://paste.lisp.org/display/96001
14:57	colomon	baest: fixing it for the other types as well?
14:57	baest	colomon: yes, I only found it for str, int and num
14:57	colomon	rakudo
14:58		rakudo: say 4 === 5
14:58	p6eval	rakudo 974d9a: OUTPUT«0␤»
14:58	colomon	rakudo: say (4 === 5).WHAT
14:58	p6eval	rakudo 974d9a: OUTPUT«Bool()␤»
14:58		IllvilJa joined #perl6
14:59	baest	rakudo: say (4.0 === 5.0).WHAT
14:59	p6eval	rakudo 974d9a: OUTPUT«Bool()␤»
14:59	baest	weird
14:59	colomon	give me a minute here, and I'll see what I can do.
14:59		PacoLinux joined #perl6
15:00	dalek	rakudo: 43f5a52 \| (Solomon Foster)++ \| build/PARROT_REVISION:
15:00		rakudo: Bump parrot.
15:00		rakudo: review: http://github.com/rakudo/rakud[…]6b579295329edefd3
15:00	baest	colomon: of course, the other part is '&prefix:?' instead of 'prefix:?' which I could see in Str that was the preferred version
15:01	colomon	I don't know the status of the leading &, but it definitely needs to be prefix:<?> now, I think...
15:02	pmichaud	any instances of 'prefix:?' are wrong and should be removed asap.
15:02	baest	colomon: I'm not sure, but I just figured the other way around because of: http://github.com/rakudo/rakud[…]efe612a7e10487525
15:02	colomon	Indeed.
15:03		hmmmm....
15:03	pmichaud	if you want to call the prefix:<?> function from PIR, it's "&prefix:<?>"
15:03		anything else is likely bogus.
15:03		(same for other operators -- ampersands and angles are required)
15:05	* colomon	is trying to get the === test file up and running again...
15:06	colomon	rakudo: say 1 !=== 4
15:06	p6eval	rakudo 974d9a: OUTPUT«Confused at line 11, near "say 1 !==="␤current instr.: 'perl6;HLL;Grammar;panic' pc 500 (ext/nqp-rx/src/stage0/HLL-s0.pir:328)␤»
15:06	baest	general question. If i have a minor patch should a post to a pastebot, push it to a clone on github or some other way
15:06	colomon	pmichaud: any reason !=== can't just be not === ?
15:06	pmichaud	that works for now; normally !op is meta
15:06	TimToady	that's what the metaop is supposed to do
15:07	colomon	!!!!
15:07		I didn't realize ! was a meta op
15:07	pmichaud	it is.
15:07		we haven't implemented that in rakudo yet
15:07		so in the meantime we've just created them explicitly as we need them
15:08	colomon	permission to do that in this case?
15:08	pmichaud	sure, that's fine.
15:08	TimToady	S03:3927
15:08	* colomon	thinks ! sounds like meta op LHF....
15:09	pmichaud	it shouldn't be too bad; we already have the op= metaops.
15:09		martinro joined #perl6
15:11	TimToady	rakudo: say 1,2 X 3,4
15:11	p6eval	rakudo 974d9a: OUTPUT«Confused at line 11, near "say 1,2 X "␤current instr.: 'perl6;HLL;Grammar;panic' pc 500 (ext/nqp-rx/src/stage0/HLL-s0.pir:328)␤»
15:12	TimToady	rakudo: say 1,2 Z 3,4
15:12	p6eval	rakudo 974d9a: OUTPUT«1324␤»
15:12	TimToady	odd that X isn't there but Z is
15:13		rakudo: say 1,2 X+ 3,4
15:13	p6eval	rakudo 974d9a: OUTPUT«Confused at line 11, near "say 1,2 X+"␤current instr.: 'perl6;HLL;Grammar;panic' pc 500 (ext/nqp-rx/src/stage0/HLL-s0.pir:328)␤»
15:13	colomon	TimToady: my fault, I guess.
15:13		I wanted Z, so I made it happen.
15:13	TimToady	what, for implementing Z
15:13		:
15:13		:)
15:13	colomon	:)
15:13		krr joined #perl6
15:15	colomon	okay, I've got !=== working here, which makes S03-operators/value_equivalence.t run, but it still has a 3 fails.
15:16		=== doesn't work on array references right now. (err, test says array, actually is probably a Parcel?
15:16		)
15:16		my @a = (1,2,3);
15:16		no, I guess that's an Array.
15:17	arnsholt	What's the procedure for bringing back test files?
15:17		I've found a file that can be brought back, as long as a couple of tests are marked TODO
15:17	colomon	mark 'em todo and bring 'em back!
15:17	pmichaud	...array reference?
15:18	arnsholt	colomon: Righties
15:18	colomon	my @a = (1,2,3);
15:18		my @b = (1,2,3);
15:18		ok (\@a === \@a), "=== on array references (1)";
15:18		ok (\@b === \@b), "=== on array references (2)";
15:18		both those fail
15:18	pmichaud	the correct term is now "capture"
15:18		(I think)
15:19	colomon	arnsholt: todo tests which are running are much more valuable than tests that aren't running.
15:19		m--locks joined #perl6
15:20	arnsholt	Indeed. The correct procedure is putting "#?rakudo todo" above the ok/is call, no?
15:20	colomon	with a reason, if possible
15:20		#?rakudo todo 'scalar reference equivalence'
15:20	arnsholt	Yeah, pondering that now
15:20	colomon	for instance.
15:21	pugssvn	r29956 \| colomon++ \| [t/spec] Fudge value_equivalence.t.
15:22	baest	colomon: ok, and after a patch is created what is the best way to post it? I have 2 files ready for bringing back along with a minor change
15:22	arnsholt	Do I have to call make spectests to do the fuding of the spectests?
15:22	colomon	patch to the tests or patch to rakudo itself?
15:23	baest	colomon: rakudo, I committed the test before
15:23		tests
15:23	colomon	pmichaud: what is the preferred way of receiving Rakudo patches?
15:23	pmichaud	rt
15:23	colomon	you can send them to rakudobugs....
15:23		that is, rt.
15:23	baest	pmichaud, colomon: thanks
15:24		:)
15:24		will do
15:24	pmichaud	afaik, http://wiki.github.com/rakudo/[…]to-create-a-patch is still valid.
15:24		(if not, it needs to be updated)
15:25		steps 4-6, anyway.
15:25		steps 1-3 aren't required
15:25		actually, just 4 and 6
15:26	baest	pmichaud: sorry, I forgot about that page
15:26	pmichaud	the main item is that we'd like commits to be squashed (easier to review), and that it go to RT so it can be tracked.
15:30	avar	Why are you recommending that commits be squashed instead of people working on topic branches? Then you can just review the branch as is without destroying commit information
15:31		People can also just send in "please review x..y"
15:31	pmichaud	reviewing a branch is hard
15:31	avar	why? you just diff / log from where the branch diverged to where its head is at now
15:31	pmichaud	I have to pull the branch
15:32		then do the diff
15:32		also, I'm not sure that the individual commits are that important
15:32		at any rate, this is the process we're using for now and I'm not in a rush to try to learn a new one until after April
15:34		also, with "please review x..y", there's no record of the patch should it be not applied (and the github account and/or branch disappears)
15:34	dalek	rakudo: ab3f44b \| (Solomon Foster)++ \| src/ (2 files):
15:34		rakudo: Quick implementation of !===. (Should be a meta op, but this will work for now.)
15:34		rakudo: review: http://github.com/rakudo/rakud[…]b03433d8500718bb6
15:34		rakudo: 7882db4 \| (Solomon Foster)++ \| t/spectest.data:
15:34		rakudo: Turn on S03-operators/value_equivalence.t.
15:34		rakudo: review: http://github.com/rakudo/rakud[…]69253cae690d44b74
15:35	pugssvn	r29957 \| colomon++ \| [t/spec] Defudge passing tests.
15:35		Psyche^ joined #perl6
15:37		martinro joined #perl6
15:39	baest	hmm git format-patch doesn't do anything....
15:39	colomon	should have made a file
15:39		you might need to tell it what commits go into the patch, though.
15:39	arnsholt	Hmm. Does Test.pm in the new master have todo?
15:40	pmichaud	I'm pretty sure it does/must.
15:40	arnsholt	Then I'm doing something wrong
15:41	baest	colomon: yes, git format-patch -M 054b1b6d602c823b318f81e473f6fde5f12becf6 does nothing even though I can see the commit with git log
15:41	arnsholt	nm. Think I found it
15:41	colomon	baest: last time I did it, I used git format-patch -1
15:42	baest	colomon: wheee, success, thanks!
15:42	pmichaud	afk for a bit
15:43		m6locks joined #perl6
15:44	arnsholt	I had no reason in the todo fudge, which resulted ins a no-arg call to todo
15:44		Which isn't allowed, apparently
15:46		payload left #perl6
15:46	pugssvn	r29958 \| arnsholt++ \| [t/spec] Fudged two tests in S06-traits/is-rw.t
15:47	arnsholt	There. Now someone with the privs can bring S06-traits/is-rw.t back in spectest.data
15:47	colomon	arnsholt: give me a min...
15:49	pugssvn	r29959 \| colomon++ \| [t/spec] Add a bunch of tests to make sure === and !=== return Bool.
15:52	dalek	rakudo: 28f5203 \| (Solomon Foster)++ \| t/spectest.data:
15:52		rakudo: Turn on S06-traits/is-rw.t. arnsholt++.
15:52		rakudo: review: http://github.com/rakudo/rakud[…]cc4964e3ba1c54d90
15:57	cognominal	Performing 'gmake realclean' ...
15:57		Makefile:3660: compilers/imcc/Rules.mak: No such file or directory
15:57		gmake: *** No rule to make target `compilers/imcc/Rules.mak'. Stop.
15:57		...on Snow Leopard
15:57		when running perl Configure.pl --gen-parrot on a fresh rakduo
15:58		jferrero joined #perl6
15:58	pugssvn	r29960 \| colomon++ \| [t/spec] === tests on Rat and Num.
15:58	colomon	cognominal: is that latest?
15:59		and fresh fresh? I got that something like that when I tried updating, but when I blew away the parrot directory --gen-parrot worked fine for me.
15:59	cognominal	ok, will do a rm --rf rakudo/parrot
16:01	arnsholt	Is there any particular reason why there are two files named pi.t?
16:02		They do almost the same thing, too
16:02	colomon	arnsholt: no idea.
16:05	* colomon	is startled to see that there already was a !=== implementation, which couldn't possibly have been running thanks to the lack of !=== in the grammar....
16:12	cognominal	thx colomon
16:12		it works
16:13	colomon	\o/
16:14		payload joined #perl6
16:17		payload1 joined #perl6
16:24		snarkyboojum_ joined #perl6
16:25	pugssvn	r29961 \| colomon++ \| [t/spec] Unfudge test that now works.
16:26	dalek	rakudo: ba6cd4e \| (Solomon Foster)++ \| src/builtins/ (2 files):
16:26		rakudo: Change === on Num to return Bool, and clean up === on Int. baest++
16:26		rakudo: review: http://github.com/rakudo/rakud[…]e34808f48b2f2a929
16:26		simcop2387 joined #perl6
16:29		jmarti joined #perl6
16:41	m6locks	oh, i didn't know linus torvalds created git
16:41		nice
16:42	TimToady	we all have our faults
16:45	m6locks	...but it's not compiling on that shell box :(
16:47		abra joined #perl6
16:48		simcop2387 joined #perl6
16:50		justatheory joined #perl6
16:54	arnsholt	S02-literals/quoting-unicode.t can be brought back without changes
16:54	quietfanatic	phenny: tell masak It wasn't so much of an NCI as a wrapper around Parrot's NCI, combined with a C parser that figured out the hard stuff for you.
16:54	phenny	quietfanatic: I'll pass that on when masak is around.
16:56		Trashlord joined #perl6
16:56	quietfanatic	phenny: tell masak Except that the parser wasn't very good and there weren't any better ones I could use.
16:56	phenny	quietfanatic: I'll pass that on when masak is around.
16:57	arnsholt	rakudo: my $a = 3; say "$a".WHAT;
16:57	p6eval	rakudo ba6cd4: OUTPUT«Int()␤»
16:57	arnsholt	Known issue or new?
17:04	lue	oh hello there
17:05	Trashlord	hi
17:05	quietfanatic	alpha: Q:PIR{$P0 = loadlib '/usr/lib/libcrypt.so'␤$P1 = dlfunc $P0, 'crypt', 'ttt'␤set_hll_global ['Link';'C';'NCI'], '$crypt', $P1}; &C::crypt = sub ($p1, $p2) {$Link::C::NCI::crypt($p1, $p2)}; say C::crypt('Password', '$1$Salt');
17:05	p6eval	alpha 30e0ed: OUTPUT«$1$Salt$3By6sIxwjQONPKbAQWYfQ1␤»
17:06	lue	.hack(And our total this hour is.... $3000000 worth of bug fixes!)
17:09	pugssvn	r29962 \| arnsholt++ \| [t/spec] Fudged a test in S02-literals/string-interpolation.t
17:10	arnsholt	S02-literals/string-interpolation can be brought back as well now
17:10	colomon	arnsholt++
17:11	arnsholt	I'm digging to see if I can find out why the failing test fails, but it'll probably take a while to figure out
17:11	TimToady	rakudo: my $a = 3; say (~$a).WHAT
17:11	p6eval	rakudo ba6cd4: OUTPUT«Str()␤»
17:12	arnsholt	Funky
17:12	TimToady	interestingly, had the same bug in Perl 5 once upon a time
17:12		you implement interpolation with infix:<~>, but forget the degenerate case
17:12	colomon	arnsholt: running spectest with those files turned on, will push as soon as it passes. :)
17:12	arnsholt	Spiffy
17:13	TimToady	er, not Perl 5, Perl 1 or so... :)
17:13	lue	I am going to be in a whirlwind of coding, I can tell.
17:13		simcop2387 joined #perl6
17:13	lue	And I'm ready, with multiple-tabbed Konsole, Kate, Konversation... and Firefox.
17:14	TimToady	um, are any of those coding tools?
17:14	lue	not specifically designed for it, but it's all I need for the hackathon :)
17:15		I have The Guide in one tab of Firefox, and I know where my towel is. I'm ready!
17:15	TimToady	depending on your location, the main thing you need for this hackathon might be an alarm clock :)
17:17	uniejo	Just make it a location aware alarm clock. F.ex. phone with gps.
17:18	lue	I don't need an alarm clock (and a location aware one'd be useless). I'm currently in the publishing corporations of Ursa Minor Beta :)
17:21		(ANFSCD) I'm assuming .even and .odd are actual methods for numbers (it makes an appearance in the spec after all)
17:22	baest	rakudo: [4, 3, 1].grep(* > 2);
17:22	p6eval	rakudo ba6cd4: OUTPUT«Method 'Num' not found for invocant of class 'Block'␤current instr.: 'perl6;Mu;' pc -1 ((unknown file):-1)␤»
17:22	baest	is that the correct syntax for it, though?
17:23	colomon	baest: I think yes, but * > 2 is NYI.
17:23	dalek	rakudo: c054780 \| (Solomon Foster)++ \| t/spectest.data:
17:23		rakudo: Turn on S02-literals/string-interpolation.t and S02-literals/quoting-unicode.t. arnsholt++
17:23		rakudo: review: http://github.com/rakudo/rakud[…]e5c19297eea404c66
17:23	baest	colomon: ok, so no rakudobug for this?
17:24	colomon	baest: it's a well-known issue.
17:24	TimToady	lue: there are no .even/.odd methods
17:24	colomon	rakudo: [4, 3, 1].grep({ $_ > 2 }).perl.say
17:24	p6eval	rakudo ba6cd4: OUTPUT«GatherIterator.new()␤»
17:24	colomon	rakudo: [4, 3, 1].grep({ $_ > 2 }).batch(10).perl.say
17:24	p6eval	rakudo ba6cd4: OUTPUT«(4, 3)␤»
17:24	TimToady	there are already lots of ways to spell even and odd in Perl
17:24	lue	aw. I was hoping almost (then why is it in a code example in the spec?)
17:25	TimToady	$x % 2, $x +& 1
17:25		rakudo: say 3 % 2, 3 +& 1
17:25	p6eval	rakudo ba6cd4: OUTPUT«11␤»
17:26	TimToady	there's two different odd functions
17:26	colomon	wasn't there also ?% or something?
17:26	TimToady	!% 2 is even
17:26		rakudo: 3 !% 2
17:26	p6eval	rakudo ba6cd4: OUTPUT«Confused at line 11, near "3 !% 2"␤current instr.: 'perl6;HLL;Grammar;panic' pc 500 (ext/nqp-rx/src/stage0/HLL-s0.pir:328)␤»
17:27	TimToady	but nyri
17:27		alpha 3 !% 2
17:27		alpha: 3 !% 2
17:27	p6eval	alpha 30e0ed: ( no output )
17:27	TimToady	alpha: say 3 !% 2
17:27	p6eval	alpha 30e0ed: OUTPUT«0␤»
17:28	lue	alpha: say 4 !% 2
17:28	p6eval	alpha 30e0ed: OUTPUT«1␤»
17:29		jagd joined #perl6
17:29	lue	ah, so 1 if even, 0 if odd?
17:29	TimToady	!% means "is not indivisibly by"
17:29		s/y/e/
17:29	lue	alpha: say 5 !% 3; say 6 !% 3;
17:29	TimToady	or "does not have a remainder"
17:29	p6eval	alpha 30e0ed: OUTPUT«0␤1␤»
17:30		dual joined #perl6
17:30	betterworld	I'd say !% means "is divisible by" ;)
17:30	TimToady	it means "is divisible by!"
17:31	lue	that seems easy to program in (away!)
17:31	betterworld	oh, I did not see the "in" ;)
17:33	colomon	lue: the tricky bit is that all the !op operators are supposed to be generated automatically.
17:33	TimToady	not exactly
17:33		it's a higher-order function
17:33	colomon	"generated automatically when needed" ? ;)
17:33	TimToady	so they don't need to be generated, just composed with 'not'
17:34		likewise (we hope) for all the other metaops
17:34		we can't generate XRS[!%]= in advance
17:34		(and we might add a Z for zipwith)
17:36	colomon	TimToady: don't show XRS[!%]= to the doubting Thomases, or we will be universally reviled...
17:36	TimToady	so anyway, the code emitter just has to turn the base op into &infix:<op> or whatever, and then pass it to the right higher-order function (reduce, zip, not, whatever)
17:36	colomon	TimToady: though if it were that easy, I'd already have done it for a bunch of those.... :)
17:38	* lue	wants to work on something :/
17:38	TimToady	put STD's metaop parsing into rakudo
17:38		(just kidding...kinda)
17:39	lue	metaops do not exist at all then :)
17:40	TimToady	alpha: say 3 R/ 30
17:40	p6eval	alpha 30e0ed: OUTPUT«10␤»
17:40	TimToady	R/ is pronounced "goes into" :)
17:41	lue	any value to metaops, and do they exist in other languages? I have never had to use them, obviously :)
17:41	colomon	pmichaud implied I ought to be about to handle !op...
17:43		lue: metaops are a huge slice of awesome.
17:43	lue	I have never heard of them in any other language, which is why I don't know. :)
17:43	colomon	Well, the most widely spread one we have is op=
17:44		nihiliad joined #perl6
17:44	colomon	That's in most C-family languages.
17:44	lue	what's the most `omniscient' kind of operator?
17:46	colomon	I'll leave that to the theologians.
17:47	lue	.hack(how many people at Copenhagen now?)
17:48	colomon	TimToady: any hints on how to find the op= rule in STD.pm?
17:48		TimToady: never mind!
17:48		errr... maybe?
17:49	TimToady	it's handled specially in STD, actually, at the end of infixish
17:49		line 3172 or so
17:49	colomon	danke!
17:50	TimToady	but it's the only one that is handled after the fact
17:50	colomon	do you call it infix_postfix_meta_operator like Rakudo does? (sorry, refreshing pugs directory tree takes a long time...)
17:50	TimToady	for the others, see things like token infix_prefix_meta_operator:sym<!> {
17:51	lue	.oO(o hell, firefox froze on perlcabal again :( )
17:51	colomon	lue: use chrome for perlcabal. seriously
17:51	TimToady	i've never had trouble with ff on perlcabal, fyi...
17:51		memory issue?
17:52		uniejo joined #perl6
17:52	lue	colomon:Well, I want to use chrome, but unsure of it being available in the Fedora/PPC repos
17:52	colomon	TimToady: I don't know, but I was seeing it as well when I used Firefox for getting the synopses. Loading pages could take half a minute, and freeze all of Firefox in the process. uuuuugly.
17:52	lue	TimToady: PPC G3 Pismo (less that 1GHz if I remember), 384 MiB RAM (twice that swap space)
17:53		I know it must be an issue with all the iframes and my memory though :)
17:53	* TimToady	reminds himself that the Cray-1 ran at only .1 GHz
17:54		masakmoritzmbere joined #perl6
17:54	lue	point one Giga-Hertz! wow.
17:54	masakmoritzmbere	this is the masak-moritz-mberends mechbot, stopping by.
17:54	TimToady	hibot!
17:55	lue	masakmoritzmberends --masak="awaken"
17:55	masakmoritzmbere	lue: I am awake, all three of us.
17:55	arnsholt	masakmoritzmbere: Oh hai
17:55		Were you guys heading out for dinner at some point?
17:55	masakmoritzmbere	arnsholt: mberendsbot and masakbot just found moritz, and we immediately joined together into a grander structural unit.
17:56		arnsholt: yes. spicy food.
17:56	moritz__	re
17:56	TimToady	couldn't you just be m*bot or some such
17:56	moritz__	shouldn't it be m.*bot? :-)
17:56	masakmberends	oh, moritz__ fell off.
17:56	colomon	(spicy food)++
17:57	masakmberends	arnsholt: anyway, you're welcome to be assimilated, and we'll create a spicyfood-eating mechbot of five.
17:58	lue	colomon: unless you can point me to Chrome for Linux/PPC that I don't have to compile, I'm staying with firefox :)
17:58	arnsholt	masakmberends: Spicy food sounds excellent. When and where?
17:58	masakmberends	arnsholt: dunno yet.
17:58		somewhere spicy. possibly Indian.
17:58	pugssvn	r29963 \| moritz++ \| [t/spec] unfudges for rakudo
17:58		r29964 \| moritz++ \| [t/spec] re-fudge quoting.t for rakudo
17:58		r29965 \| moritz++ \| [t/deprecated-syntax.pod] undef is gone
17:58	masakmberends	would be a definite plus if it's within the Copenhagen area.
17:59	pugssvn	r29966 \| moritz++ \| [t/spec] enhance precedence.t, and fudge for rakudo
17:59	TimToady	oh, I was going to suggest one here
17:59	colomon	moritz++ # yow!
17:59	moritz__	colomon: 10 hours worth of train travel :-)
17:59	colomon	and git svn, eh?
18:00	borup	masakmberends: there arent that many indian places in copehagen
18:00	* uniejo	commits to food ;-)
18:00	lue	colomon: don't recommend what I'll never be able to use :D
18:00	masakmberends	borup: any suggestions for other spicy/good food?
18:00		wolf2k_ubuntu joined #perl6
18:00	borup	masakmberends: would japanese work
18:01	dalek	book: 41c5229 \| moritz++ \| src/regexes.pod:
18:01		book: [regex] minor grammar fix
18:01		book: review: http://github.com/perl6/book/c[…]18ffe62e5810179ef
18:01		book: bd06566 \| moritz++ \| src/operators.pod:
18:01		book: New chapter on operators (work in progress)
18:01		book: review: http://github.com/perl6/book/c[…]12538507b3be5061b
18:01		book: bf30c2b \| moritz++ \| docs/announce/2010-03:
18:01		book: [docs] stub announce for 2010-03 release
18:01		book: review: http://github.com/perl6/book/c[…]100cdd987be187c4a
18:01		book: 3e7d952 \| moritz++ \| src/regexes.pod:
18:01		book: [regex] impmrove description of rindex(); also comment on the editorial note by chromatic++
18:01		book: review: http://github.com/perl6/book/c[…]c3515fee47d73fd79
18:01	colomon	Clearly moritz__ needs to take more 10 hour train trips...
18:01	masakmberends	borup: for the masak part of the mechbot, it would, yes.
18:01	lue	o noes the flood again!!
18:02		Grimnir_ joined #perl6
18:02	Grimnir_	OH HAI
18:02	TimToady	OH HAI OH
18:02	lue	OHIO
18:02	masakmberends	E I E I O
18:02	borup	masakmberends: uniejo suggests wagamama
18:02	TimToady	OH HAI OH GOES EYE MOOSE
18:02	borup	masakmberends: very close to the main station
18:02	masakmberends	borup: the whole bot agrees: Japanese is a fine choice.
18:03		borup: excellent.
18:03		wknight8111 joined #perl6
18:03	lue	nihongo? excellent choice
18:03		(to lazy to fiddle with char map/input method)
18:03	uniejo	http://www.wagamama.dk/locations/map
18:03	dalek	rakudo: 14a83eb \| moritz++ \| t/spectest.data:
18:03		rakudo: more passing test files
18:03		rakudo: review: http://github.com/rakudo/rakud[…]5e4786e7d8035b5ea
18:03	TimToady	cut and paste this: 日本語
18:04	Grimnir_	you gotta love utf8-able terminals
18:04	TimToady	well, really, にほんの食べ物
18:04	lue	(must message Fedora to support UTF8 in their tty's... note to self)
18:04	jnthn	masakmberends: oh hai...where is located the bot? :-)
18:04	* jnthn	did find his room :-)
18:04	lue	jnthn! o/
18:04	Grimnir_	I was at a few perl 6 speaks today. it looks like a very interesting language
18:04	masakmberends	jnthn: room 420.
18:05	jnthn	aww...Japanese, not Indian? :-/
18:05		OK, coming :-)
18:05	Grimnir_	what are the possibilites for parallel computing and the like? what can perl6 do?
18:05	k23z__	wow, do you guys all day go to workshops and stuff like that ?
18:05		I'm at home, I even took a nap
18:05	lue	And flood the IRC with short messages
18:05	k23z__	lue: me ?
18:05	lue	and masak is back in full glory!
18:05	masak	\o/
18:05	phenny	masak: 16:54Z <quietfanatic> tell masak It wasn't so much of an NCI as a wrapper around Parrot's NCI, combined with a C parser that figured out the hard stuff for you.
18:05		masak: 16:56Z <quietfanatic> tell masak Except that the parser wasn't very good and there weren't any better ones I could use.
18:05	TimToady	Grimnir_: mostly just made sure that it's easy to express various sorts of parallelism naturally, and then let the computer figure it out
18:06	masak	jnthn: oh, so your room existed in the territory but not on the map?
18:06	moritz_	Grimnir_: I think not much is implemented, but there are lots of possibilities; threads, events, object pipes, hyper operators and so on
18:06	TimToady	various constructs make various promises about not-caring about interactions
18:06	Grimnir_	TimToady: what does that mean? the code has to let the computer figure out
18:06	lue	k23z: _we_ flood the irc with messages (I'm not there physically, but in spirit)
18:06	moritz_	k23z__: I spent 10.5hrs on the train today to get to the workshop
18:06	TimToady	hyperops are basically vector ops
18:06		feed operators are pipes
18:06	Grimnir_	I would suggest to do a CSP-implementation or very similar in perl6. that would be awesome
18:06	TimToady	junctional logic can be evaluated in any order
18:06		masak joined #perl6
18:07	lue	.hack(And the total this hour is.... £3 worth in bug fixes ($32839423492 worth in the US))
18:07	masak	you guys should see me and moritz_ right now. :) 15 minutes after we meet, we're sitting side by side, IRC-ing. :P
18:07	lue	:)
18:07	Grimnir_	it would be cool if it would support it with a pretty nice syntax
18:07		akl joined #perl6
18:07	lue	(I got japanese input working once, what happened to it...)
18:08	TimToady	rakudo: say (1,2,3) »+« (4,5,6)
18:08	p6eval	rakudo c05478: OUTPUT«Confused at line 11, near "say (1,2,3"␤current instr.: 'perl6;HLL;Grammar;panic' pc 500 (ext/nqp-rx/src/stage0/HLL-s0.pir:328)␤»
18:08	TimToady	alpha: say (1,2,3) »+« (4,5,6)
18:08	p6eval	alpha 30e0ed: OUTPUT«579␤»
18:08	arnsholt	masak: Has the collective reached a decision re: fud?
18:08	masak	do I have everyone's permission to stir up a potential hornet's nest?
18:08		(don't ask to ask...)
18:09		arnsholt: are you coming by?
18:09	lue	hornet nest taken. masak cries. hornet nest given back. all is well. (that's a YES :) )
18:09	arnsholt	masak: I wouldn't mind coming by
18:10		There are signals that usually mean I should eat soon =)
18:10	Grimnir_	perl6: say ("hello world")
18:10	p6eval	elf 29966, pugs, rakudo c05478: OUTPUT«hello world␤»
18:10	arnsholt	And I thought hornets' nests were your job? ;)
18:10	colomon	rakudo: say "hello world"
18:10	p6eval	rakudo c05478: OUTPUT«hello world␤»
18:10	masak	alpha: class A { has @.b; }; say A.new(b => [1,2,3]).b.perl
18:10	p6eval	alpha 30e0ed: OUTPUT«[[1, 2, 3]]␤»
18:11	TimToady	Grimnir_: your parens were not function call parens, unlike in perl 5
18:11	masak	ok, so the hornet's nest is that I really don't like the double [] here. :(
18:11		I've been talking to jnthn about it most of the afternoon.
18:11		he's quite good at explaining that I'm wrong.
18:11		I'm mostly being stubborn.
18:11	lue	neither TimToady. I remember him confusing them with the purpose of the «» brackets
18:11		(the outer [] anyway)
18:12	masak	arnsholt: come over, then!
18:12	moritz_	arnsholt: so, was that a "yes" to coming?
18:12	Grimnir_	at the perl6 speak today I didn't see any "\n", does that mean "say" automatically writes a newline?
18:12	TimToady	yes
18:12		it's strange, but it's one of the most used functions in p6
18:12	arnsholt	masak/moritz_: On my way then. Be there soonish
18:12	masak	Grimnir_: as does Perl 5.10, incidentally.
18:12		++arnsholt
18:12	arnsholt	I'll SMS masak when I'm in the neighbourhood
18:12	masak	oki
18:13	Grimnir_	is there a normal print function in perl6 too?
18:13	TimToady	yes
18:13		rakudo: print "foo"
18:13	p6eval	rakudo c05478: OUTPUT«foo»
18:13	Grimnir_	ok, that's why
18:13	uniejo	rakudo: class Food {sub eat {}}; Food.new().WHERE.say
18:13	p6eval	rakudo c05478: OUTPUT«47750022470056␤»
18:13	Grimnir_	I haven't used any of the special perl 5.10 features, it seems
18:13	TimToady	rakudo: say "foo"
18:13	p6eval	rakudo c05478: OUTPUT«foo␤»
18:13	TimToady	note the ␤ character there
18:14		should be a little NL or some such
18:14		martinro joined #perl6
18:14	TimToady	if your terminal does utf-8 right
18:14	lue	say creates a \n automagically, and print doesn't. That's it.
18:14	TimToady	that's what?
18:14	colomon	TimToady: any notion what the O() should be for infix_prefix_meta_operator?
18:14	Grimnir_	perl6: @test@.say
18:15	masak	TimToady: am I unreasonable in thinking that .new(:b[1,2,3]) should initialize to just one layer of array?
18:15	Grimnir_	ach
18:15	p6eval	rakudo c05478: OUTPUT«Confused at line 11, near "@test@.say"␤current instr.: 'perl6;HLL;Grammar;panic' pc 500 (ext/nqp-rx/src/stage0/HLL-s0.pir:328)␤»
18:15		..elf 29966: OUTPUT«Parse error in: /tmp/fAERKl76hs␤panic at line 1 column 0 (pos 0): Can't understand next input--giving up␤WHERE: @test@.say␤WHERE:/\<-- HERE␤ STD_red/prelude.rb:99:in `panic'␤ STD_red/std.rb:76:in `scan_unitstopper'␤ STD_red/std.rb:224:in `comp_unit'␤ STD_red/std.rb:210:in
18:15		..`_UNI…
18:15		..pugs: OUTPUT«*** ␤ Unexpected "@."␤ expecting "::"␤ Variable "@test" requires predeclaration or explicit package name␤ at /tmp/NHJqMu8t1L line 1, column 6␤»
18:15	Grimnir_	perl6: "test".say
18:15	p6eval	elf 29966, pugs, rakudo c05478: OUTPUT«test␤»
18:15	Grimnir_	nice
18:15	lue	sorry TimToady. that's $_ [ :) ]
18:15	TimToady	lol
18:15		masak: it seems to me that it's a binding, not an assignment
18:16	masak	TimToady: yes.
18:16	TimToady	so no, it's not unreasonable :)
18:16	masak	TimToady: but jnthn++ says that binding isn't the right thing for obj init, in general.
18:16	TimToady	otoh, jnthn may be right about the OO semantics :)
18:16	masak	because it creates a rift in timespace, basically.
18:17	k23z__	moritz_: wooo..
18:17		moritz_: that's commitment man
18:17	TimToady	otgh, maybe it copies in values but [...] is an object, so gets the same one
18:17	masak	YES!
18:18		whatever that means, I'm for it!
18:18		k23z___ joined #perl6
18:18	Grimnir_	I loved the quicksort example in Perl6. is perl6 in general that cool?
18:18	TimToady	no, usually it's cooler :P
18:18	masak	Grimnir_: much cooler.
18:18	Grimnir_	cooler? that's very cool
18:18	TimToady	picnic cooler, even
18:19	colomon	which quicksort? was it jnthn's new version?
18:19	Grimnir_	it was the one, jnthn showed at open source days today, yes
18:19	colomon	ah, that's indeed a very sweet little piece of code.
18:19	Grimnir_	even though the pivot should be taken at random :)
18:20		you could do it with one parameter
18:20	TimToady	without looking, I assume it's taking hte first elem as the pivot
18:20	moritz_	rakudo: say (1, 2, 3).max
18:20	colomon	Grimnir_: please don't encourage TimToady to figure out how to make that a part of the language....
18:20	p6eval	rakudo c05478: OUTPUT«3␤»
18:20	Grimnir_	colomon: it's so simple it's not important :)
18:20		rakudo: say (1,2,3).pick
18:20	TimToady	we could .pick a random pivot though, which would work better for a presorted list on average
18:20	p6eval	rakudo c05478: OUTPUT«2␤»
18:21	Grimnir_	rakudo: say (1,2,3).pick
18:21	p6eval	rakudo c05478: OUTPUT«2␤»
18:21	TimToady	or .pick(3).mid
18:21	Grimnir_	rakudo: say (1,2,3).pick
18:21	p6eval	rakudo c05478: OUTPUT«1␤»
18:21	Grimnir_	rakudo: say (1,2,3).pick.mid
18:21	p6eval	rakudo c05478: OUTPUT«Method 'mid' not found for invocant of class 'Integer'␤current instr.: '_block14' pc 29 (EVAL_1:0)␤»
18:21	moritz_	rakudo: say max(1..5)
18:22	p6eval	rakudo c05478: OUTPUT«-Inf␤»
18:22	Grimnir_	rakudo: say (1,2,3).mid
18:22	p6eval	rakudo c05478: OUTPUT«Method 'mid' not found for invocant of class 'Parcel'␤current instr.: '_block14' pc 29 (EVAL_1:0)␤»
18:22	colomon	I was just going to say, what's .mid? (I mean, I think I understand the meaning, but....
18:22	TimToady	rakudo: say [max](1..5)
18:22	p6eval	rakudo c05478: OUTPUT«Not enough positional parameters passed; got 0 but expected at least 1␤current instr.: 'max' pc 239375 (src/gen/core.pir:6883)␤»
18:22	colomon	rakudo: say max(1..5)
18:22	p6eval	rakudo c05478: OUTPUT«-Inf␤»
18:22	TimToady	wants a closure first
18:22	Grimnir_	alpha: say (1,2,3).mid
18:22	colomon	WTF?
18:22	p6eval	alpha 30e0ed: OUTPUT«Method 'mid' not found for invocant of class 'List'␤in Main (file src/gen_setting.pm, line 324)␤»
18:22	* masak	submits rakudobug for max(1..5)
18:22	TimToady	there is no mid
18:22	moritz_	mid soooo reminds me of QBasic
18:22		iblechbot joined #perl6
18:22	TimToady	but if there were, it would be between min and max :)
18:22	moritz_	(where it took the function of substr())
18:22	colomon	rakudo: (1..5).max.say
18:23	p6eval	rakudo c05478: OUTPUT«5␤»
18:23	TimToady	rakudo: (1..5).minmax
18:23	p6eval	rakudo c05478: ( no output )
18:23	TimToady	rakudo: (1..5).minmax.say
18:23	p6eval	rakudo c05478: OUTPUT«15␤»
18:23	TimToady	rakudo: (1..5).minmax.average :)
18:24	p6eval	rakudo c05478: OUTPUT«Confused at line 11, near "(1..5).min"␤current instr.: 'perl6;HLL;Grammar;panic' pc 500 (ext/nqp-rx/src/stage0/HLL-s0.pir:328)␤»
18:24	TimToady	rakudo: (1..5).minmax.sum.say
18:24	p6eval	rakudo c05478: OUTPUT«Method 'sum' not found for invocant of class 'Parcel'␤current instr.: '_block14' pc 29 (EVAL_1:0)␤»
18:24	lue	TimToady: you put a smiley in the code with average :)
18:24	masak	alpha: (1..5).minmax.sum.say
18:24	p6eval	alpha 30e0ed: OUTPUT«Method 'sum' not found for invocant of class 'List'␤in Main (file src/gen_setting.pm, line 324)␤»
18:25	masak	TimToady: should there be a .sum method?
18:25	colomon	no.
18:25	moritz_	please not
18:25		just like .max is superfluous
18:25	masak	is not! :)
18:25	* moritz_	hits masak
18:25	lue	rakudo: say (1..5).minmax.average
18:25	p6eval	rakudo c05478: OUTPUT«Method 'average' not found for invocant of class 'Parcel'␤current instr.: '_block14' pc 29 (EVAL_1:0)␤»
18:25	* masak	slaps moritz_ with a seasoned trout
18:26	* moritz_	ducks deep, so that masak hits mberends
18:26	masak	mberends: sorry! :/
18:26	lue	masak: use this herring! (I need it back afterwards to cut down a tree)
18:26	masak	lue: this herring is damaged.
18:26		I can't use it.
18:27	TimToady	if it's bleeding, it might be a red herring
18:27	* masak	.oO( use Herring; )
18:27	colomon	my wife dreams of making an album of all herring songs...
18:27	Grimnir_	lol
18:28	masak	colomon: but would that really be worth herring?
18:28	lue	This halibut is good enough for Jehovah! Take it masak!
18:28		athenot joined #perl6
18:28	moritz_	http://stackoverflow.com/quest[…]with-rakudo-perl6
18:28	TimToady	the scales would be in tuna, I suppose
18:29	* colomon	is experiencing pun-lock
18:29	colomon	masak: it would be if I were playing bass
18:29	masak	:)
18:30	TimToady	they make me sing bass, but I'm really a barracuda
18:30		simcop2387 joined #perl6
18:30	lue	colomon: if your album is good enough, you might become a starfish!
18:30		.oO(fish puns are about as fun as cheese puns)
18:30	TimToady	er
18:32	colomon	hey, my crazy hacked rakudo compiled!
18:33	TimToady	put it on ice, and ship it
18:33	colomon	> say 1 !&& 3
18:33		Could not find non-existent sub &infix:<>
18:34	TimToady	well of course you couldn't find it--it's non-existent, duh!
18:34		you should look for an existent sub
18:34	colomon	but also note how it is &infix:<> -- the nothing operator!
18:35	TimToady	std: 1 !&& 3
18:35	p6eval	std 29966: OUTPUT«ok 00:01 107m␤»
18:35	TimToady	hmm, thought that one was too fiddly
18:35	lue	alpha: multi sub infix:<>{say "oh hello";}; say 54 #this is 5 <nothing> 4
18:35	p6eval	alpha 30e0ed: OUTPUT«54␤»
18:35	lue	alpha: multi sub infix:< >{say "oh hello";}; say 5 4 #this is 5 <nothing> 4
18:35	TimToady	or not iffy enough
18:35	p6eval	alpha 30e0ed: OUTPUT«Confused at line 10, near "4 #this is"␤in Main (file <unknown>, line <unknown>)␤»
18:36	lue	alpha: multi sub infix:< >($a,$b){say "oh hello $a and $b";}; say 5 4
18:36	p6eval	alpha 30e0ed: OUTPUT«Confused at line 10, near "4"␤in Main (file <unknown>, line <unknown>)␤»
18:36	lue	alpha: multi sub infix:< >($a,$b){say "oh hello $a and $b";}; say 5 4;
18:36	p6eval	alpha 30e0ed: OUTPUT«Confused at line 10, near "4;"␤in Main (file <unknown>, line <unknown>)␤»
18:36	masak	lue: two terms in a row. 5 4.
18:36	TimToady	std: multi sub infix:< >($a,$b){say "oh hello $a and $b";}; say 5 4;
18:36	p6eval	std 29966: OUTPUT«===[0mSORRY!===[0m␤Two terms in a row at /tmp/n8HbYo9IDp line 1:␤------> a,$b){say "oh hello $a and $b";}; say 5 ⏏4;␤ expecting any of:␤ bracketed infix␤ infix or meta-infix␤ statement modifier loop␤FAILED 00:01 109m␤»
18:37	masak	even STD.pm agrees with me.
18:37	lue	what the heck does two terms in a row mean?
18:37	TimToady	it means there's no such thing as a null infix operator in Perl 6
18:38	lue	That's why I tried a space :)
18:38	TimToady	it won't let you
18:38	diakopter	std: multi sub infix:<\ >($a,$b){say "oh hello $a and $b";}; say 5 \ 4;
18:38	p6eval	std 29966: OUTPUT«Bareword found where operator expected at (eval 246) line 24, near "q<\>,␤ }, 'RE_method'),␤ }, 'RE_ast'),␤};␤␤our $CATEGORY = 'infix';␤␤sub infix__S_500__PEEK { $_[0]->_AUTOLEXpeek"␤ (Might be a runaway multi-line <> string starting on line 17)␤Can't create infix:sym<\
18:38		..…
18:38	masak	lue: space is usually not significant.
18:38	TimToady	it if it could, I wouldn't tell you how
18:38	diakopter	wheh
18:39	lue	alpha: multi sub circumfix:<jnthn jnthn>($a){say "do something with $a";}; jnthn 42 jnthn
18:39	p6eval	alpha 30e0ed: OUTPUT«do something with 42␤»
18:39	lue	(still thinking of the omnisub :D )
18:39	TimToady	because it's very, very, very, very important to Perl 6 to know whether it's expecting a term or an infix
18:39	diakopter	TimToady: Bareword found where operator expected
18:39	TimToady	so it can catch most errors
18:40	colomon	> say 3 !&& 3
18:40		Could not find non-existent sub &infix:<&&>
18:40		that's a bit better....
18:40	TimToady	except for the double negative :)
18:41		maybe we could add in a "noway, nohow"
18:41	colomon	TimToady: I think we can take for granted that my work on Rakudo's grammar will not fix Parrot's weird error messages. :)
18:41	diakopter	std: multi sub infix:<\ >(){}
18:41	p6eval	std 29966: OUTPUT«Bareword found where operator expected at (eval 246) line 24, near "q<\>,␤ }, 'RE_method'),␤ }, 'RE_ast'),␤};␤␤our $CATEGORY = 'infix';␤␤sub infix__S_500__PEEK { $_[0]->_AUTOLEXpeek"␤ (Might be a runaway multi-line <> string starting on line 17)␤Can't create infix:sym<\
18:41		..…
18:41	TimToady	I'm just gonna keep carping till someone does...
18:41	* colomon	doesn't think he has a commit-bit for Parrot...
18:42	masak	TimToady: I'm with you.
18:42	* masak	never says no to a good carp
18:42	moritz_	colomon: you can still submit patches :-)
18:44	TimToady	diakopter: I'm ignoring you. :)
18:44		szabgabx joined #perl6
18:44	k23z__	<kbsmasher> I have tried: __asm { nop } but it didn't do anything.
18:45	lue	I'm going to log out for a moment, to enable japanese input.
18:45	TimToady	japanese input, yum
18:45		quick, talk about lue
18:45	diakopter	bleh
18:46	TimToady	or make a cheese pun
18:46	masak	TimToady: even if you /ignore diakopter, you will see the STD.pm output he produces. :)
18:46	colomon	> say 4 !&& 4
18:46		Could not find non-existent sub &infix:<sol&&>
18:46		Aha!
18:46	TimToady	yes, well, add_macro is all chewing gum and bailing wire, and I don't feel obligated to make it a real thing just yet
18:47	masak	why would anyone be chewing bailing wire?
18:48	TimToady	and I'm not /ignoring him, just ignoring him. :)
18:48	* diakopter	thought it was baling wire
18:48	masak	/oh
18:48		it's whaling wire, actually.
18:48	TimToady	okay, it's kleenex and scotch tape then.
18:49	Grimnir_	what's scotch tape?
18:49	diakopter	a trademarked brand of 3M
18:49	jnthn	The stuff that doesn't hold the universe together.
18:49	masak	Grimnir_: http://en.wikipedia.org/wiki/Scotch_Tape
18:49	colomon	Grimnir_: sellotape
18:50		lue joined #perl6
18:50		simcop2387 joined #perl6
18:50	Grimnir_	oh, that's just called tape in danish
18:51	masak	wagamama, here we come!
18:51	colomon	> say 4 !&& 5
18:51		Could not find non-existent sub &infix:<sol!&&>
18:51		yay!
18:51	lue	風の楽土 (I hope I wrote winds of rakudo correctly!)
18:51		colomon: the sun is nonexistant? Noes!
18:52	TimToady	that's more like rakudo of wind
18:52		wind's rakudo
18:52	lue	bah, I'm not a student of japanese
18:52	colomon	lue: don't panic! just use your sub-etha to thumb a ride someplace else.
18:53	TimToady	is an etha like a metha or a propa?
18:53	lue	I did pick up a few phrases thru translating things in (and obsessing over) The Legend of Zelda games :)
18:53		colomon: I would, but my pandimensional wireless card is broken :/
18:54		(stupid Verizon... :D )
18:54	quietfanatic	ゼルダの伝説
18:54	colomon	lue: oh. sucks to be you.
18:55		oh wise ones
18:55		I seem to have hacked !op into the grammar.
18:55	lue	時のオカリナ
18:56	colomon	any hints on how to generate the metaop on demand, then?
18:56		there's addmethod (is that the right name), is there also an addsub?
18:58	lue	oi! I have a security update to sudo (I'm scared of why they needed to fix it)
18:58	TimToady	maybe it prevents insecure people from using it
19:00	quietfanatic	I'd be more worried if I didn't occasionally see a security update for sudo.
19:01	TimToady	lessee, if I'm really about 3 or 4 TZ east of here, I should take a geezer nap now.
19:01		zzz &
19:02	lue	o no! the copenhagen air has slept TimToady
19:04	colomon	pmichaud, jnthn: why is $opname stored as a .lex in gen_assign_metaop, rather than just passed as a parameter?
19:05	k23z__	anyone in here down for some hardcore assembly hacking let me know
19:05		should I add hardcore oldschool assembly hacking
19:06		jagd joined #perl6
19:07		jagd joined #perl6
19:08	k23z__	hey related to assembly I saw some code used back in the day by reverse engineers written in Perl
19:08		even a basic disassembler :)
19:09		will Perl6 be compiled to bytecode and if so , will it be hard to write a disassembler for it ?
19:09		or for Parrot binary code in general
19:09		because I remember there was something (also back in the day ...2002 maybe) that decompiled Java like crazy
19:09		it received no resistance whatsoever ...
19:10	avar	where here are the spectests rakudo uses: http://svn.openfoundry.org/pugs/t/ ?
19:11	colomon	http://svn.pugscode.org/pugs/t/spec
19:11		not that I quite understood the question, but that's where the Perl 6 spectests live.
19:11	avar	ah, the pugs svn moved
19:12		isn't this pretty much the only thing the pugs svn is used for now?
19:12	* lue	&
19:13	colomon	avar: not at all, there are a zillion projects in pugs svn.
19:13		The official perl 6 specs live there, as does Larry's STD.pm official grammar.
19:13		and a bunch of other projects as well.
19:14		I did a svn up there earlier today, looked like over 200 files have been changed in the last month or two.
19:14		> say 4 !&& 5
19:14		Null PMC access in invoke()
19:15		martinro joined #perl6
19:15		jagd joined #perl6
19:16	colomon	and wow, I don't have any idea how to debug that. :(
19:19	lisppaste3	colomon pasted "Slightly broken implementation of !op -- help?" at http://paste.lisp.org/display/96015
19:20		explorer__ joined #perl6
19:25	colomon	oh, wait, should be prefix:<!>... stupid spectest, finish already so I can try that!
19:25		jagd joined #perl6
19:26		lichtkind joined #perl6
19:26		avar joined #perl6
19:29	quietfanatic	colomon: So you're actually generating an &infix:<!op> at compile time, as opposed to putting &infix_prefix_meta_operator:<!>(&infix:<op>, $a, $b) in the code?
19:30	colomon	yes.
19:30		I don't know which approach is better, I'm just imitating the way op= is currently handled.
19:30	quietfanatic	Ah. Just seems more complicated that way.
19:31	colomon	though I'm suddenly suspicious your way is better....
19:31	quietfanatic	Mine is just the naive way to do it (I think)...
19:32	colomon	I dunno, your way sounds elegant to me.
19:32	quietfanatic	Because I'm not familiar with grammars and compilers and all that.
19:32	lisppaste3	colomon annotated #96015 "slightly improved but still Null PMC access version" at http://paste.lisp.org/display/96015#1
19:32	quietfanatic	My way might be a little slower at runtime.
19:33	colomon	your way might work. seems like it's worth a try, anyway.
19:36	* colomon	is going mad with the power of git!
19:36	lichtkind	colomon: please dont do so :)
19:37		with 30 entries now: http://www.perlfoundation.org/[…]ndex.cgi?timeline
19:49	colomon	quietfanatic: the only problem being, I'm not quite sure how to implement your suggestion...
19:49	quietfanatic	hmm...
19:49	colomon	> say 4 !&& 5
19:49		Could not find non-existent sub &infix_prefix_meta_operator:<!>
19:49		== progress
19:50	quietfanatic	Yeah, well then you need to write &infix_prefix_meta_operator:<!> :)
19:50	colomon	sure, it's how you call it and how you implement it that's the issue here. :)
19:50	quietfanatic	Ah.
19:50		ShaneC joined #perl6
19:51	quietfanatic	Well, it needs to have as parameters infix:<&&>, 4, and 5, right?
19:51	colomon	right
19:51	quietfanatic	s/infix/&infix/
19:51		in some order or other
19:52		what should the parameters be when a user writes a new meta op?
19:52	colomon	hmmm?
19:52		btw, I'm compiling an attempt to guess how to do it as I type.
19:52	quietfanatic	I mean...hold on, consulting the spec
19:54		The spec doesn't seem to have an example for overloading metaops, just infixes and such.
19:55	colomon	holy smokes, that almost sort of worked.
19:55		multi sub infix_prefix_meta_operator:<!>($a, $b, $c) {
19:55		say $a.WHAT;
19:55		say $b.WHAT;
19:55		say $c.WHAT;
19:55		}
19:55		> say 4 !&& 6
19:55		Str()
19:55		Int()
19:55		Int()
19:55	quietfanatic	Great! Except why is $a Str()?
19:56	colomon	probably that's how I'm passing it.
19:56		but... hmm...
19:56		rakudo: "say"(34);
19:56	p6eval	rakudo 14a83e: OUTPUT«invoke() not implemented in class 'Perl6Str'␤current instr.: '_block14' pc 29 (EVAL_1:0)␤»
19:56	quietfanatic	rakudo: &::<say>(34)
19:56	p6eval	rakudo 14a83e: OUTPUT«Confused at line 11, near "&::<say>(3"␤current instr.: 'perl6;HLL;Grammar;panic' pc 500 (ext/nqp-rx/src/stage0/HLL-s0.pir:328)␤»
19:57	quietfanatic	alpha: &::<say>(34)
19:57	p6eval	alpha 30e0ed: OUTPUT«Syntax error at line 10, near "&::<say>(3"␤in Main (file <unknown>, line <unknown>)␤»
19:57	quietfanatic	ah, it seems symbolic lookup is not yet implemented.
19:57	colomon	you can call a method specified by a string, surely there's a way to do that with a sub?
19:58	quietfanatic	Well, it's the syntax that's NYI, not the implementation (if I know what I'm talking about)
19:58		alpha: (Q:PIR{%r = get_hll_global [], '&say'}).(34)
19:58	p6eval	alpha 30e0ed: OUTPUT«error:imcc:syntax error, unexpected ']' (']')␤ in file 'EVAL_22' line 71␤Null PMC access in invoke()␤in Main (file <unknown>, line <unknown>)␤»
19:58	quietfanatic	alpha: (Q:PIR{%r = get_hll_global '&say'}).(34)
19:58	p6eval	alpha 30e0ed: OUTPUT«Null PMC access in invoke()␤in Main (file <unknown>, line <unknown>)␤»
19:59	diakopter	alpha: (Q:PIR{%r = get_hll_global 'say'})(34)
19:59	p6eval	alpha 30e0ed: OUTPUT«34␤»
20:00	lisppaste3	colomon annotated #96015 "very different attempt which also slightly fails" at http://paste.lisp.org/display/96015#2
20:01	colomon	oh hai, I think if you guys can figure out how to do that with an infix op we're good to go here.
20:01		but I've got to run shower now...
20:02	quietfanatic	How do you pass it the op itself, instead of the name of the op?
20:07		payload joined #perl6
20:07		athenot joined #perl6
20:13		athenot_ joined #perl6
20:14		athenot_ joined #perl6
20:14	quietfanatic	colomon: Just a guess; should that line 'my $opsub := "&infix:<$sym>";' be 'my $opsub = $<infix>.ast;' or something like that?
20:15		maybe not.
20:17		s/=/:=/
20:28		jagd left #perl6
20:35		synth joined #perl6
20:48		martinro joined #perl6
20:50	mikehh	looks like I got a seg fault after t/spec/S14-roles/basic.rakudo - Non-zero wait status: 11 in spectest_smolder #32531 - Ubuntu 9.10 amd64 (gcc with --optimize) - parrot r44708
20:50		id does not report it in the smolder report
20:50		justatheory joined #perl6
20:51	mikehh	s/id/it/
20:53	baest	is qw// dead and replaced by <>?
20:53	quietfanatic	baest: yes
20:54	baest	quietfanatic: thanks
20:59		prism joined #perl6
20:59	baest	I'm looking at split.t. split returns a GatherIterator, but that doesn't respons well to being compared to Parcels with is_deeply. Any ideas?
21:02	colomon	baest: I think that's all supposed to change in pmichaud's next round of fixes.
21:03		So I'd not worry about it at the moment.
21:03	baest	colomon: oh ok
21:05	* diakopter	answers my question from yesterday (how to design an api for codegen-generators): macros. duh.
21:05	pugssvn	r29967 \| baest++ \| Use <> instead of qw// and skip places with rx:Perl5 which isn't implemented
21:09		rv2733 joined #perl6
21:13		krr joined #perl6
21:20		k23z__ joined #perl6
21:20		k23z__ joined #perl6
21:32	TimToady	baest: actually, qw// should also still work; <> is just the shortcut, like '' and q// still both work
21:33		rakudo: say qw/foo bar/
21:33	p6eval	rakudo 14a83e: OUTPUT«Confused at line 11, near "say qw/foo"␤current instr.: 'perl6;HLL;Grammar;panic' pc 500 (ext/nqp-rx/src/stage0/HLL-s0.pir:328)␤»
21:33	TimToady	std: say qw/foo bar/
21:33	p6eval	std 29967: OUTPUT«ok 00:01 107m␤»
21:33	diakopter	pugs: say qw/foo bar/
21:33	p6eval	pugs: OUTPUT«foobar␤»
21:34		newbie joined #perl6
21:37		jferrero joined #perl6
21:45	baest	TimToady: oh, ok. Hmm, I just a test, but if it's the same it shouldn't matter and now at least, it is parseable by rakudo
21:45		+changed
22:03		uniejo joined #perl6
22:06		wolf2k_ubuntu joined #perl6
22:07		mberends joined #perl6
22:08		am0c joined #perl6
22:10	k23z__	hey what if perl6 offered permutations as a core language feature ?
22:10		is it unreasonable ?
22:11	mberends	k23z__: that feels like something that belongs a level above the core language, probably in a module
22:11		masak joined #perl6
22:12	k23z__	can Perl6 modules be written in Parrot code ?
22:12		what is the equivalent of XS in Perl6 ?
22:12	mberends	yes, in PIR
22:12	masak	the equivalent of XS is Parrot NCI.
22:12	k23z__	will it be blazing fast if written in PIR ?
22:12	TimToady	define blazing
22:13	mberends	XS is another matter, we're currently working on NCI
22:13	lichtkind	mberends: hej
22:13	mberends	hi lichtkind
22:13	masak	k23z__: it might possibly be a bit faster than compiled Perl 6 in some cases. no order-of-magnitude wins, though.
22:13	k23z__	TimToady: well I'm writing assembly code for generating permutations that'll be called from XS which will be called from Perl5 , will I be able to do this with Perl6 ? http://github.com/wsdookadr/Ru[…]aster/SJT/SJT.asm
22:14	lichtkind	and masaks now there too :)
22:14		mberends: seen http://www.perlfoundation.org/[…]ndex.cgi?timeline ?
22:15	mberends	lichtkind: I'm now looking at the current one, had seen it a few days ago
22:15	lichtkind	mberends: thats not the same :)
22:15	TimToady	with a decent jit it might go fast, but parrot's jit has suffered repeated bitrot
22:15	masak	jitrot.
22:15	jnthn	Parrot's jit got ripped out, iirc.
22:15	k23z__	TimToady: assembly code is my definition of blazing fast
22:15	masak	jitrip.
22:15	mberends	lichtkind: oh yes, exact dates!
22:16	TimToady	yes, well, you're talking NCI there, not parrot, currently
22:16	masak	rakudo: class A { has $.b = my $foo = 42; say $foo }
22:16	p6eval	rakudo 14a83e: OUTPUT«42␤»
22:16	k23z__	is NCI at an equivalent speed with assembly code ?
22:16	masak	alpha: class A { has $.b = my $foo = 42; say $foo }
22:16	quietfanatic	Parrot's NCI is, I think, easier to use than Perl5's XS.
22:16	p6eval	alpha 30e0ed: OUTPUT«Lexical '$foo' not found␤in Main (file src/gen_setting.pm, line 324)␤»
22:16	k23z__	quietfanatic: maybes some docs please ? with lots of examples ?
22:16	TimToady	NCI is an API, not an engine
22:17	masak	rakudo: class A { has $.b = my $foo = 42; method x { say $foo } }; A.new.x
22:17	p6eval	rakudo 14a83e: OUTPUT«Null PMC access in getprop()␤current instr.: '&infix:<=>' pc 16638 (src/builtins/Junction.pir:228)␤»
22:17	* masak	submits rakudobug
22:17	diakopter	masak: :)
22:17	TimToady	it's like asking what's faster, an elevator, or an elevator door?
22:17	mberends	lichtkind: the whole timeline looks so very good :)
22:17	quietfanatic	k23z__: It's somewhere in Parrot's documentation.
22:17	lichtkind	mberends: thanks
22:18	k23z__	TimToady: to me it's all the same
22:18	diakopter	my elevator is a door
22:18	lichtkind	mberends: so we finally get to the translation of the tut :)
22:18	masak	my door is an elevator.
22:18	k23z__	TimToady: I don't understand your comparison because I don't know as much about p6 as you do
22:18	diakopter	... no one does...
22:18	masak	rakudo: class A { my $foo = 42; method x { say $foo } }; A.new.x
22:18	TimToady	has nothing to do with p6
22:18	p6eval	rakudo 14a83e: OUTPUT«42␤»
22:18	lichtkind	masak: lives on mars und uses the super secret jumprooms :)
22:18	TimToady	it's a category error
22:18	* k23z__	is not much of a category theorist :D
22:19	quietfanatic	k23z__: He means NCI isn't code that runs, it's a connection from one kind of code to another.
22:19	masak	What's faster, an elevator, or a category error?
22:19	k23z__	masak: :)
22:19	masak	I mean, it's like comparing apples and category errors.
22:20	TimToady	APIs and oranges
22:20	diakopter	my door is an orange
22:20	k23z__	can you guys make it so that I'll be able to write assembly inline in Perl6 pretty please with sugar on top ?
22:20	masak	TimToady: :)
22:20	k23z__	like __asm { ... }
22:20	masak	diakopter: :)
22:20	quietfanatic	inline? Well, I don't think we have that yet.
22:21	TimToady	in any case, that's a severe architecture dependency
22:21	quietfanatic	The best you could do would be to compile your assembly to a library, and then link to that library from Parrot
22:21	diakopter	k23z__: just write some tests...
22:21	quietfanatic	...which is what NCI does.
22:22		nihiliad joined #perl6
22:23	TimToady	APIs and O()-ranges
22:24	pugssvn	r29968 \| moritz++ \| [t/spec] test that lexicals escape thunks
22:24	jnthn	.oO( what the thunk? )
22:25	diakopter	rakudo: say 1..my$a # times out I think
22:25	moritz_	speakiing of O() - STD.pm uses both <EXPR(item %item_assignment)> and <O(\|%term)> - what's the difference?
22:25	p6eval	rakudo 14a83e: ( no output )
22:25	TimToady	<O> sets the actual precedence on the token we're returning
22:26		the arg to EXPR is a cap on how tight we require the precedence to be
22:26		in this subexpressoin
22:26	arnsholt	One of the Parrot devs (I think) commented about Rakudo copying a lot in a comment on my segfault bug: http://trac.parrot.org/parrot/[…]et/1499#comment:8
22:26	TimToady	ion
22:26	arnsholt	(In case that's interesting to someone)
22:26	TimToady	that's how we stop on the comma when you say my $a = 1, 2
22:27	Tene	k23z__: That's certainly possible to do (__asm {}), but it's going to be a decent while before that's a priority for any current Perl 6 devs that I know about.
22:27	lue	hello!
22:27	masak	arnsholt: why would Rakudo copy a lot in a comment on your segfault bug? :P
22:27		that sounds like a metabug to me.
22:27	arnsholt	masak: A bit of a tangent, from him digging around in the Rakudo code, I imagine
22:28	Tene	k23z__: for performance, working on Parrot's GC and then getting a good JIT for Parrot are much higher on the priority list for most people who might work on that, afaik.
22:28	TimToady	k23z__: but feel free to work on it yourself :D
22:29	masak	snarkyboojum: ping
22:29	Tene	k23z__: Rakudo has the ability right now to use inline PIR, which is Parrot's bytecode.
22:29		rakudo: Q:PIR { print "lolol" }
22:29	TimToady	for permuting integers, that'll likely be faster than anything you could write in p6
22:29	p6eval	rakudo 14a83e: OUTPUT«lolol»
22:30	lue	I want to do something worthwhile during the hackathon. :/ Lurking is fun, but helping is as well :)
22:30	TimToady	at least, until rakudo gets native ints
22:30	lue	(If I can't work, I'll just learn Japanese >:) )
22:31	k23z__	is Perl6 going to be like Java ?
22:31		if it has a vm ..
22:31		and stuff like that
22:31	masak	k23z__: yes.
22:32	k23z__	I wonder how fast is it , if it was registered in the Language shootout then I could compare it with other stuff
22:32		but maybe speed is not the most important concern at this stage of development ?
22:32	moritz_	slow. At the moment at least
22:32	Tene	k23z__: A module that would support inline ASM would take the text provided, compile it into executable code, use something like dlsym to get a handle to the code and then execute it. There's going to be a big calling conventions mismatch between Perl 6 calling conventions and ASM, and I'm sure several other issues.
22:32	diakopter	about 5000-10000 slower
22:32	moritz_	k23z__: that's correct
22:32	diakopter	5000-10000x slower, I mean
22:33	k23z__	Tene: calling conventions in terms of ? order of params on stack ?
22:33		Tene: what else ?
22:33	Tene	k23z__: Perl 6 isn't stack-based. Parrot isn't stack-based. You'll need to write code to marshall the args into something reasonably accessible from ASM.
22:33		IllvilJa joined #perl6
22:34	k23z__	I don't know any other model than stack-based
22:34		how does Parrot work when calling some fucntion ?
22:34	moritz_	it's continuation passing style
22:34	diakopter	Perl 6 isn't stack-based?
22:34	moritz_	http://en.wikipedia.org/wiki/C[…]ion-passing_style
22:34		(at least parrot)
22:34	Tene	Rakudo doesn't have any native data types right now, as well. You'd need to take Perl 6 Integer objects and put them into registers, etc. I'm not all that familiar with ASM, so I don't really know what it would look like.
22:35		diakopter: a Perl 6 sub isn't required to set up a C-like stack frame for every invocation.
22:35	k23z__	why is the Perl6 using a vm ?
22:35		are there some particular advantages that you are fond of ?
22:35	moritz_	k23z__: yes; it's needed
22:35	arnsholt	k23z__: Using a VM you get a lot more platform independence
22:35	diakopter	Tene: nor is any language whose subroutines can be inlined when deemed safe....?
22:36	arnsholt	There's also GC, and so on
22:36	Tene	diakopter: which means that you need to care about that when talking about inline ASM
22:36	moritz_	k23z__: in C you don't even have objects, so you need some kind of runtime system
22:37	Tene	k23z__: Perl 6 is an extremely dynamic language. It's likely that any implementation is going to implement something like a VM, at least internally, so we're formally doing that as a separate project, for a variety of reasons.
22:37		Parrot's goal is to be a VM good for any dynamic language.
22:37	k23z__	hm ok I get the idea
22:37	Tene	k23z__: There are a few other alternative implementations in progress, and any implementation that can pass the spec test suite "Is Perl 6"
22:38	lue	Rakudo is only the most likely candidate. A few years ago it was Pugs.
22:38		most likely candidate _to be the first_
22:39	k23z__	I am also interested in speed, XS was very good for Perl5 because you get to mess with the internals, with the core, you can make it do whatever you want
22:40		I hope a similar mechanism be devised for p6
22:40	TimToady	I hope p6's mechanism is far, far better than XSW
22:40		XS
22:40	Tene	k23z__: right now, it looks like any API like that is going to be implementation-dependant. It's possible to access C code from Rakudo right now, but it's not very well-documented.
22:41	k23z__	Tene: ah , neat :)
22:41	masak	k23z__: the hope for Perl 6 is that the internals will be something that you won't often feel the need to mess with, and that messing with the operators, the metamodel, the slangs and macros will be enough.
22:41	moritz_	rakudo: my %p; %p<a> = [%p<a>, 4]; say %p.perl
22:41	p6eval	rakudo 14a83e: OUTPUT«{"a" => [Proxy.new(), 4]}␤»
22:41	lue	Speaking of which, is SDL/OpenGL useable yet? :)
22:41	mberends	lue: almost :)
22:41	moritz_	masak: that's RT #64080
22:42		and "fixed" now :-)
22:42	diakopter	k23z__: Perl 6 specifies native types, which (assuming a good optimizer) should provide quite efficient data structures/algorithms.
22:42	Tene	k23z__: the process is: write a C library, load that .so or .dll with the loadlib op, and then fetch symbols in it with the dlfunc op.
22:42	TimToady	and it should be possible to specify APIs from the Perl side, so you don't have to write grotty glue code
22:42		(see XS)
22:43	k23z__	Tene: yes that seems to be very much like Perl5 does it..
22:43	TimToady	the hard work is getting the arguments in and out
22:43	Tene	k23z__: here is a wrapper library in PIR for a GUI library I was playing with a while back: http://github.com/tene/parrot-[…]er/Elementary.pir
22:44		k23z__: you notice that you need to explicitly specify the signature
22:44		k23z__: there have been a few projects to parse C header files and generate library wrappers like this automatically.
22:44		None are currently good enough to rely on, that I've heard about.
22:45	arnsholt	Are there no plans for a general Perl 6 C API?
22:45	masak	moritz_: I won't hit you, I'm out of fish.
22:46	Tene	arnsholt: I have never heard of one. Every time that question has been asked that I've seen, the consensus has been that any such functionality is likely to be implementation-dependant.
22:46	TimToady	for any two well-defined languages, we should be able to come up with a well-defined implementation-independent interface
22:47	diakopter	hmm. where're we gonna find a well-defined language?
22:47	TimToady	and the fact is, when I hooked Perl 5 up to Java, I found Java supplied sufficient type info to do that, even though Perl didn't.
22:47	mberends	arnsholt: it's not a high priority to make own-foot-shooting-guns, but eventually it will come ;;)
22:47	TimToady	and that's one of the reasons Perl now has better type info
22:47	k23z__	Tene: implementation-dependant doesn't matter very much since there is only one implementation(Rakudo)
22:48	diakopter	k23z__: that is extremely incorrect.
22:48	TimToady	perl6: say "hi, k23z__!"
22:48	moritz_	k23z__: http://perl6.org/compilers/
22:48	p6eval	elf 29968, pugs, rakudo 14a83e: OUTPUT«hi, k23z__!␤»
22:48	TimToady	there's at last three
22:49		kfo joined #perl6
22:49	Tene	k23z__: That's not quite the question he asked.
22:49	masak	k23z__: I'm not sure how you can have reached this point in #perl6-hood thinking there is only one Perl 6 implementation.
22:49	Tene	k23z__: the fact is, though, even if there were only one implementation right now, Perl 6 has a very strong community focus on establishing a spec separate from any implementation.
22:49	diakopter	loose reading of lue> Rakudo is only the most likely candidate.
22:50		snarkyboojum joined #perl6
22:50	masak	snarkyboojum: \o
22:50	Tene	k23z__: We very much don't want to repeat the mistakes of Perl 5, where the only spec is the one existing implementation.
22:50	k23z__	masak: I thought pugs development has stopped ?
22:50	masak	k23z__: your point being?
22:50	TimToady	but not mildew or sprixel
22:50	moritz_	it still exists
22:50	Tene	k23z__: That's right, pugs is not under current active development. That doesn't mean it doesn't exist, though.
22:51	TimToady	doesn't mean someone won't pick it up again and run with it either
22:51	moritz_	rakudo: &foo # what should that do?
22:51	p6eval	rakudo 14a83e: OUTPUT«Could not find non-existent sub &foo␤current instr.: '_block14' pc 29 (EVAL_1:0)␤»
22:51	moritz_	std: &foo
22:51	p6eval	std 29967: OUTPUT«ok 00:01 108m␤»
22:51	masak	moritz_: that's a question pmichaud asked as well.
22:51	* arnsholt	has considered looking at Pugs
22:51	snarkyboojum	masak: g'day :)
22:51	masak	moritz_: I'm of the opinion that it should not warn.
22:51	arnsholt	It's ludicrously hubristic, but might be fun
22:51	k23z__	masak: 2-1 = 1
22:51	masak	snarkyboojum: I'm hacking on Tardis.
22:51	moritz_	masak: I know; he gave me RT #73074
22:52		and I have no idea what to do with it
22:52	masak	k23z__: I'm not sure you've convinced me you're not a troll.
22:52	snarkyboojum	masak: nice! how'd the talk go?
22:52	masak	k23z__: forgive me if I don't put much energy into replying to you from now on.
22:52	k23z__	???
22:52	masak	snarkyboojum: it felt nice to give it. it's the first talk that I wouldn't mind giving again sometime.
22:53	snarkyboojum	masak: excellent :) slides up anywhere?
22:53	masak	snarkyboojum: I just pushed what I've committed so far to Tardis.
22:53	k23z__	ok I am not going to be using this channel any more , bye #perl6
22:53	snarkyboojum	masak: sweet - will take a look
22:53	masak	snarkyboojum: bottom of http://feather.perl6.nl/~masak/
22:53		k23z__ left #perl6
22:53	diakopter	masak: I think k23z__ was not being a troll.
22:53	TimToady	now he'll probably brag he got kicked out of #perl6 :)
22:54		I think k23z__ isn't sure himself
22:54	diakopter	masak: he was here yesterday too
22:54	masak	diakopter: I didn't say he was. I said he hadn't managed to convince me he wasn't.
22:55		diakopter: there's a subtle difference.
22:55	diakopter	I didn't say you said/thought he was.
22:55	masak	:)
22:55	TimToady	masak didn't say he thought you said that. :P
22:55	masak	I'm not saying I think he isn't, by the way.
22:56	TimToady	ain't communication wunnerful
22:56	* Khisanth	bows to masak's skills
22:56	diakopter	I was saying (in different words) that he had managed to convince me he wasn't being a troll.
22:56	TimToady	now you tell us
22:56	masak	diakopter: is the proof contained in P? :)
22:57	diakopter	but that's what I said
22:57	masak	that's what she said!
22:57	diakopter	_I think_ that _he wasn't being a troll_
22:57	* masak	is not convinced masak is not a troll
22:57	TimToady	no, you just said you thought he wasn't, you didn't say he'd convinced you of it :)
22:57		k23z__ joined #perl6
22:58	masak	TimToady: that's what diakopter said, but maybe not what he thought he said :)
22:59	diakopter	what else could have convinced me of that?
22:59	pugssvn	r29969 \| moritz++ \| [t/spec] test that operators can be accessed by their infix:<foo> and &infix:<foo> names
22:59	TimToady	hi, k23z__
22:59	k23z__	masak: I actually rejoined because I think you have mistaken me for a "troll"
22:59	masak	k23z__: yes. I might very well have been mistaken.
22:59	moritz_	now we're at leat one step meta. Congratulations everyone.
22:59	masak	k23z__: I just wish I could rule it out more readily.
23:00	k23z__	if you label me a troll again I will never enter this channel
23:00		that's all I had to say
23:00	* k23z__	goes back to code
23:00	masak	k23z__: you're welcome to help. should we resume our fruitful discussion about implementations, and Rakudo not being the only one, in which mostly everyone on the channel disagreed with you?
23:00	diakopter	k23z__: I'm glad you have questions about Perl 6
23:00	TimToady	nuclear blackmail? then stay away, by all means
23:01	masak	k23z__: I label you a troll.
23:01	TimToady	take your marbles and run while you can
23:01	dalek	rakudo: 9a94503 \| moritz++ \| t/spectest.data:
23:01		rakudo: turn on new test in S03-operators/names.t
23:01		rakudo: review: http://github.com/rakudo/rakud[…]ae48a2a1997b67654
23:01	lue	日本語　ゼルダの伝説　風の卓と (felt this conversation needed a sharp turn to SCD)
23:01	masak	lue: you finally got your kana. :)
23:01	TimToady	actually, I think moritz++ has the right idea
23:01	quietfanatic	lue: err, I think that's the wrong spelling of タクト
23:02	diakopter	I do not believe it was a threat of nuclear blackmail
23:02		I believe it was someone merely pointing out that they felt offended by being labeled something they weren't intending to be.
23:02