IRCloggy #perl6 2014-03-10

Provider of IRC logs since 2005.

WARNING: As Freenode became unjoinable and lost all warnings in topics, we cannot log channels on Freenode anymore.

	2014-03-10
lue	<timotimo> unicode already specs a replacement character for encoding errors, iirc <-- isn't that just FFFD ? :)	00:00
timotimo	.u REPLACEMENT	00:00
yoleaux	U+FFFC OBJECT REPLACEMENT CHARACTER [So] ()	00:00
	U+FFFD REPLACEMENT CHARACTER [So] (�)	00:00
timotimo	apparently	00:00
	← btyler left	00:02
	← pippo left	00:04
	← fridim_ left	00:10
	→ fridim_ joined	00:11
	→ tgt joined	00:12
	← tgt left	00:16
	→ dayangkun joined	00:23
	→ LootingLemur joined	00:44
LootingLemur	Are we skipping 2014.02?	00:49
	← hoverboard left	00:50
timotimo	nobody volunteered to do a star release	00:53
	there was a compiler release, though. we don't advertise them as openly to not confuse people.	00:53
Mouq	I think there's a strong hope that we can do 2014.03 with JVM, and no one wanted to make an R* without it…	00:56
timotimo	hmm	00:56
	i would have accepted an R* without jvm	00:57
	Mouq too	00:58
Mouq	idk	00:58
timotimo	i didn't do it because i already did the compiler release that month :P	00:58
	and i'm lazy	00:58
colomon	lazy++	00:59
timotimo	more like ++lazy	00:59
	r: enum Foo <One Two>; say One eq "One";	01:02
	→ tgt joined	01:02
camelia	rakudo-parrot 1aeb7c, rakudo-jvm 1aeb7c, rakudo-moar 1aeb7c: OUTPUT«True␤»	01:02
timotimo	mhh, not bad.	01:02
	← pecastro left	01:05
	← tgt left	01:06
	→ hoverboard joined	01:10
	← LootingLemur left	01:16
	→ woosley joined	01:26
	← __thou left	01:29
	← hoverboard left	01:38
timotimo	r: say "".split("\|")	01:40
camelia	rakudo-parrot 1aeb7c, rakudo-jvm 1aeb7c, rakudo-moar 1aeb7c: OUTPUT«␤»	01:40
timotimo	r: say "".split("\|").perl	01:40
camelia	rakudo-parrot 1aeb7c, rakudo-jvm 1aeb7c, rakudo-moar 1aeb7c: OUTPUT«("",).list␤»	01:40
timotimo	hmm.	01:40
Mouq	r: say "timotimo".split("timo").perl	01:40
camelia	rakudo-parrot 1aeb7c, rakudo-jvm 1aeb7c, rakudo-moar 1aeb7c: OUTPUT«("", "", "").list␤»	01:40
timotimo	:)	01:40
	what are your next tuits going to be spent on, Mouq?	01:41
Mouq	timotimo: Right now I want to get https://github.com/Mouq/Grammar-Expression working	01:43
timotimo	ah, of course	01:43
	i remember looking at that and thinking to myself that that's going to be quite useful	01:43
Mouq	timotimo: But there's still =table, Pod::To::Markdown, and maybe some stuff to doc.perl6.org that I'd like to work on in the meantime, when I have time	01:44
	timotimo: Yourself?	01:44
	(Someone else can totally pick up Pod::To::Markdown, I only thought of how a nice it would be today, since Github doesn't support Pod6)	01:45
timotimo	i'm hoping to get the code rewritten that can turn embedded QAST::Blocks into QAST::Stmts if they don't define their own lexical variables	01:47
	currently, that already happens, but the effect of turning lexicals into locals doesn't cascade outwards, so it hardly gives a performance benefit yet	01:47
	→ btyler joined	01:48
timotimo	it ought to improve memory usage as well as give better run-time performance in nqp and hopefully also in rakudo	01:48
	though the rakudo optimizer can't do the same optimization yet, as blocks and variables have much more complicated semantics in perl 6 than they do in nqp	01:48
Mouq	Whoa, sweet :) ++timotimo	01:48
timotimo	hopefully! ;)	01:49
	and at some point i (or someone else) should introduce nqp::locallifetime to our moarvm MAST compiler	01:49
	and i'm kind of having an eye on the string instantiations in the moarvm heap during the core setting compilation	01:52
	https://gist.github.com/timo/e1af6d5c10a4e34d6cb0 - as you can see here, there's quite a bit of duplication	01:53
	i'm guessing that these instances come from some substring operation	01:53
Mouq	timotimo: I have no clue, but I can see why optimization would be useful here	02:01
	Mouq is at the insert-random-debug-statements-because-perl6-isn't-throwing-good-enough-errors stage of programming	02:01
timotimo	make sure to document these not-good-enough errors so that we can improve Perl 6	02:02
	→ tgt joined	02:02
Mouq	timotimo: Well, I have no idea what's even happening	02:03
	timotimo: So I don't know what error it should be giving	02:03
timotimo	ah, okay :)	02:03
Mouq	===SORRY!=== postcircumfix:<{ }> not defined for type Arithmetic	02:04
	apparently from the line last-term = True and last unless $infix.pos > $oldpos;	02:04
	← tgt left	02:07
timotimo	that's an interesting line	02:07
Mouq	*$last-term	02:08
	We don't have loop-labels yet, so it should really be last TERM unless $infix.pos > $oldpos;	02:08
timotimo	"True and" seems like you can always replace it with "so"	02:08
Mouq	and has lower prec than item =	02:09
	*"and" has lower prec than item assignment	02:09
	→ klapperl joined	02:09
timotimo	oh, ok	02:09
Mouq	OH. Duhh	02:10
	I'm calling last, of course my debug statement isn't getting run -_-	02:10
	lol	02:10
timotimo	yeah :)	02:10
	timotimo makes a good rubber duck, he has found	02:10
	→ hoverboard joined	02:12
	← klapperl_ left	02:13
	→ Sqirrel_ joined	02:14
	→ FROGGS_ joined	02:14
	← FROGGS left	02:18
	← Sqirrel left	02:18
lue	Could someone possibly enlighten me as to why this sometimes fails with "Cannot modify an immutable value" on JVM, or "No such method 'STORE' for invocant of type 'Any'" on MVM ? https://github.com/lue/synopsis-generator/blob/fragments/syngen.p6#L139	02:20
	Keep in mind, it doesn't fail the first time that line is encountered, and I've yet been unable to find out what the difference between a successful execution and the failed execution of that line is.	02:21
Mouq	r: (state%){0}.push(9) xx 2	02:23
camelia	rakudo-parrot 1aeb7c, rakudo-jvm 1aeb7c, rakudo-moar 1aeb7c: ( no output )	02:23
timotimo	something something autovivification?	02:23
Mouq	r: (state%){0}.push(Any) xx 2	02:23
camelia	rakudo-parrot 1aeb7c, rakudo-jvm 1aeb7c, rakudo-moar 1aeb7c: ( no output )	02:24
timotimo	r: my %test; %test{0}.push("hi"); %test{0}.push("bye"); say %test;	02:24
camelia	rakudo-parrot 1aeb7c, rakudo-jvm 1aeb7c, rakudo-moar 1aeb7c: OUTPUT«("0" => ["hi", "bye"]).hash␤»	02:24
timotimo	r: my %test{Int}; %test{0}.push("hi"); %test{0}.push("bye"); say %test;	02:24
camelia	rakudo-parrot 1aeb7c, rakudo-jvm 1aeb7c, rakudo-moar 1aeb7c: OUTPUT«Hash[Any,Int].new(0 => ["hi", "bye"])␤»	02:24
timotimo	hm, no clue	02:24
	does $BL get iterated over by push for some reason?	02:25
lue	The only difference I've seen so far is that the first successful run of that line in one synopsis didn't have a text fragment, but the failed one did have a text fragment. Testing this out on the commandline proved me wrong on that hunch :(	02:25
Mouq	lue: might --ll-exception help?	02:25
lue	timotimo: maybe it iterates over the @.textfrag attribute, but I don't know.	02:25
	Mouq: that was going to be my next try :) .	02:25
	colomon is busy doing $work using rakudo-moar	02:27
lue	Hm. I just had a thought to try changing the key to a true number, and afterwards I'll try ll-exception	02:27
Mouq	colomon: \o/	02:27
timotimo	colomon: cool. let us know if you stumble upon anything bad	02:28
colomon	timotimo: you'd better believe I will. ;)	02:29
JimmyZ	Mouq: you meant %placements{'0'}.push($BL); ?	02:31
timotimo	:)	02:31
	JimmyZ: no, it's a %hash{Int}	02:31
JimmyZ	Int is the value, not the key, methinks	02:32
lue	That's another thing to try actually, just use a plain ol' %hash	02:32
	JimmyZ: %hash{Int} is spec'd to set the type of the key to Int. (And besides, if it affected the key I'd've run into more obvious problems much sooner ☺)	02:32
timotimo	JimmyZ: see the code i treid before	02:32
	☺	02:33
JimmyZ	NYI? :P	02:33
timotimo	i'm pretty sure that's implemented	02:33
lue	JimmyZ: the issue is some random, obscure problem with the value part of the hash's pairs.	02:33
	I forgot to mention that things like %hash{0} //= Array.new() didn't work.	02:34
timotimo	is that something you just found out or something we knew all along?	02:34
JimmyZ	I'm curious, why not @array[0]	02:34
	:/	02:34
timotimo	maybe because it's sparse?	02:34
lue	timotimo: Something I knew all along, just thought I'd specify that Any.push isn't the crux of the issue. If I have that //= line, that's what fails instead.	02:35
timotimo	hmm	02:35
	and --ll-exception doesn't help?	02:35
	← btyler left	02:35
lue	timotimo: haven't run --ll-exception yet, trying to see if pushing to %hash{1} changes anything	02:36
	(it takes a while to run through each time)	02:36
timotimo	mhm	02:36
	--ll-exception only causes better tracebacks; you should use it all the time when you're developing IMO :P	02:36
	lue realizes he forgot about the time-saving trick that makes it easier to debug faceplam	02:36
timotimo	i don't know what that is and i'd really like you to tell me :3	02:37
	Mouq finally finds the bug he was looking for: a ==Sorry== exception was getting put into an array and then thrown years later	02:38
timotimo	oops, that's kind of nasty	02:38
lue	timotimo: the trick? --backlinks=../roast/subdir-with-test-with-bl-that-crashes instead of --backlinks=../roast	02:38
timotimo	oh	02:39
	i got a feeling that that won't help me with all my debuggings in the future :)	02:39
lue	much easier when you're parsing a couple dozen test files than when parsing 918 :P	02:39
timotimo	yeah	02:39
lue	I almost wonder if my backlink generator shouldn't just parse the Pod6 file and then insert backlinks into the resulting tree, and save that to file. Only reason why I'm not doing that is to preserve as much of the file's formatting as possible.	02:40
timotimo	i think i'm going to bed now	02:42
JimmyZ	good night	02:42
Mouq	night timotimo!	02:42
	Mouq suddenly understands he was doing stuff completely wrong	02:48
	→ __thou joined	02:55
TimToady	at least with you it's sudden :)	03:01
	→ tgt joined	03:03
	← tgt left	03:07
lue	Well, that's not cool. Narrowing down the test files I process somehow fixes it :/	03:09
	→ Sibin joined	03:13
	lue afk	03:21
	[Coke] wonders when "package foo;" became illegal.	03:27
TimToady	2 Apr 2001	03:29
	← grep0r left	03:30
	← cooper left	03:30
[Coke]	you're good. :)	03:30
	I have code that worked about a year ago that used it.	03:30
	← hoverboard left	03:30
[Coke]	I wonder how long it took rakudo to catch up.	03:31
	(don't really need to know)	03:31
	← Sibin left	03:32
[Coke]	on the flip side, I can throw out my .combine function...	03:32
	→ cooper joined	03:33
	← cooper left	03:33
	→ cooper joined	03:33
	← cooper left	03:36
	→ Sibin joined	03:36
	← Sibin left	03:36
	→ cooper joined	03:37
	← cooper left	03:38
	→ cooper joined	03:38
Mouq	[Coke]: Sat May 7 18:20:10 2011 +0200	03:38
	git diff 1c0430601^..1c0430601	03:38
	:P	03:38
	→ grep0r joined	03:39
Mouq	Oh, nope	03:39
[Coke]	huh. also had to change a bunch of @'s for sets to $'s.	03:41
	yay, I have my kenken helper script working again. :)	03:42
	→ molaf joined	03:43
TimToady	\o/	03:43
lue	--ll-exception for r-j doesn't give me anything much different from before, trying r-m	03:55
	Mouq is having major Cursor issues	04:02
	→ tgt joined	04:04
TimToady	maybe it's about time to rewrite STD in Perl 6 :)	04:05
lue	TimToady: isn't it already though? :)	04:06
TimToady	well, it's a thin Perl 6 veneer over mostly Perl 5 semantics	04:06
	TimToady would be interested to see what an implementation of backtracking using real lazy lists would look like	04:07
	← xinming__ left	04:07
lue	Oh. I thought it was written for the ideal Perl 6 implementation, and so far only viv could handle it :)	04:07
TimToady	I'd probably use few hashes and more attributes if it were really ideal Perl 6	04:08
	*fewer	04:08
	← tgt left	04:08
TimToady	I guess we need to write vivi instead of viv	04:09
	← raydiak left	04:09
lue	.oO(./ⅥⅥ.p6 --to=r-m)	04:12
	← molaf left	04:13
	← __thou left	04:16
	Mouq is annoyed	04:18
Mouq	I don't think what I'm trying to do can be done with the current Rakudo :/	04:19
	Something to do with the NQP-Perl6 barrier	04:20
TimToady	is this an issue of Cursors not be quite as immutable as they're supposed to be in theory?	04:20
	*not being	04:21
	→ aje_ joined	04:21
aje_	moo	04:21
	Hello	04:22
lue	TimToady: looking at STD, I like how you called ｢｣ "perfect quotes" :)	04:22
TimToady	aje_: howdy mate	04:22
Mouq	It's an issue with creating new Cursors and losing their private attributes somehow	04:22
aje_	am new to the perl6 community and everything perl 6. just cloned rakudo* from github and am currently trying to build it. is rakudo the best place to start if I just want to hack around with perl 6 and try out the new language features?	04:23
TimToady	yup	04:23
aje_	OK	04:23
	STD/viv is not the place to start?	04:23
TimToady	nope	04:24
aje_	OK	04:24
TimToady	that's just a prototype we planned to throw away, on which the real parsers are based	04:24
aje_	aaah, OK	04:24
TimToady	and viv can only translate a very limited subset of P6 back to P5 in order to run	04:24
	and it cheats :)	04:25
aje_	I see... no worries then, just curious	04:25
	hehe	04:25
TimToady	are you familiar with P5?	04:25
lue	TimToady: ooc, if you did rewrite STD, would it be an actual rewrite, or would you just P6-ify the not-very-P6 bits of what's already there?	04:25
TimToady	it would mostly be rewriting CursorBase and LazyMap	04:26
	→ kaare_ joined	04:27
TimToady	no time for it currenlty	04:28
	*ntly	04:28
aje_	TimToady: Yes, from way back when I used to do Unix sysadmin stuff... but that must be at least 10 years ago	04:28
	← fridim_ left	04:31
	← Mouq left	04:35
	← grondilu left	04:48
	← telex left	04:49
	→ telex joined	04:50
	→ tgt joined	05:04
lue	I don't suppose there's a way to effectively replace "sub infix:<=>(Mu \a, Mu \b) is rw" without modifying rakudo itself, is there? :/ (This so I can see what's going on with the args it gets when it fails)	05:06
	← tgt left	05:09
	← araujo left	05:14
	← [Sno] left	05:23
	→ ggoebel11111 joined	05:57
	← ggoebel11110 left	05:58
	→ ggoebel joined	06:03
	→ tgt joined	06:05
	← kvakerman left	06:08
	← tgt left	06:10
	← BenGoldberg left	06:12
	→ woosley1 joined	06:20
	← woosley left	06:20
	← xragnar left	06:22
	→ xragnar joined	06:22
	← donk left	06:28
	→ water joined	06:30
	water → hoverboard	06:31
	→ [Sno] joined	06:39
	→ kaleem joined	06:52
	→ tgt joined	07:06
	← tgt left	07:10
	→ Monkey joined	07:15
	← Monkey left	07:15
	← xenoterracide left	07:18
	← dayangkun left	07:28
	← FROGGS_ left	07:30
	→ xenoterracide joined	07:35
moritz	good morning	07:38
	→ darutoko joined	07:41
	→ FROGGS joined	07:51
	→ dmol joined	07:55
	→ zakharyas joined	08:04
	→ tgt joined	08:07
aje_	yippie, rakudo star installed	08:07
	← hoverboard left	08:10
	← tgt left	08:11
FROGGS	timotimo: +1 to :fallback	08:13
	timotimo: I always have bogus data I have to turn into valid text (just for users to read), and sadly I can do that nicely in PHP but not in Perl 5, which is weird	08:14
	← vincent23 left	08:15
	→ vincent23 joined	08:21
	→ tgt joined	08:25
	← tgt left	08:29
	→ dayangkun joined	08:35
	← arnsholt left	08:38
	← dayangkun left	08:43
	→ pdcawley joined	08:43
sergot	morning o/	08:45
FROGGS	morning sergot	08:49
	→ fhelmberger joined	08:55
	→ dayangkun joined	08:55
	← pdcawley left	08:57
	→ pdcawley joined	09:11
	→ rindolf joined	09:26
	← kivutar left	09:33
	→ virtualsue joined	09:33
	→ dakkar joined	09:34
timotimo	o/	09:40
jnthn	o/	09:41
FROGGS	o/	09:41
JimmyZ	\o/	09:41
FROGGS	\o	09:41
jlaire	o\	09:42
	→ woosley joined	09:45
	← woosley1 left	09:45
	→ Mouq joined	09:46
	→ arnsholt joined	09:46
	← woosley left	09:46
arnsholt	o/	09:46
	→ woosley joined	09:46
arnsholt	jnthn: Is there any particular reason $VM on Rakudo/JVM doesn't load the stuff from nqp::backendconfig?	09:47
timotimo	hmm. i can't give .assuming a different invocant, can i?	09:48
FROGGS	arnsholt: because there is almost nothing in its backendconfig, no?	09:48
	j: say nqp::backendconfig	09:49
camelia	rakudo-jvm 1aeb7c: OUTPUT«("runtime.jars" => "/home/p6eval/rakudo-inst-2/languages/nqp/runtime/asm-4.1.jar:/home/p6eval/rakudo-inst-2/languages/nqp/runtime/asm-tree-4.1.jar:/home/p6eval/rakudo-inst-2/languages/nqp/runtime/jline-1.0.jar:/home/p6eval/rakudo-inst-2/languages/nqp/runtim…»	09:49
FROGGS	j: say $*VM	09:49
camelia	rakudo-jvm 1aeb7c: OUTPUT«("name" => "jvm", "properties" => {"java.vm.version" => "24.51-b03", "sun.jnu.encoding" => "UTF-8", "java.vendor.url" => "http://java.oracle.com/", "java.vm.info" => "mixed mode", "user.dir" => "/home/p6eval_eval", "sun.cpu.isalist" => "", "java.awt.graphic…»	09:50
timotimo	another thing is: $*VM on parrot is absurdly slow :(	09:50
	← aje_ left	09:51
	→ aje_ joined	09:52
	← Mouq left	09:52
	→ pecastro joined	09:55
arnsholt	There is that, yeah	09:55
	It's just weird that the JVM backend config has a separate op to get the information we need	09:56
	If nqp::backendconfig is sparse on information, we should add some more stuff to it, rather than add a JVM-specific op, no?	09:56
FROGGS	+1 from me :o)	09:57
	(dunno if that is worth much though)	09:57
jnthn	I think jvmproperties may be accessing something JVM-specific (that actually is called properties)...	09:57
arnsholt	It is, but is it really necessary to expose that fact to the user?	09:59
	The reason for this complaint is of course that it's a lot easier to add the "how to compile things" info to nqp::backendconfig on JVM, but I want that stuff in $*VM =)	10:01
timotimo	making something an op in nqp is not really "exposing a fact to the user" :)	10:03
FROGGS	timotimo: rakudo is the user :P	10:03
timotimo	:)	10:04
	← dayangkun left	10:10
arnsholt	timotimo: The user of NQP was what I meant, though =)	10:14
	← robinsmidsrod left	10:30
	→ robinsmidsrod joined	10:31
	→ woosley1 joined	10:36
	← woosley left	10:36
	→ pmurias joined	10:40
pmurias	if I need to #ifdef something in the setting for multiple backends what would be a good way to do that?	10:41
timotimo	i'm going to be procrastinating away writing today's p6weekly until a new cool item to write about pops up	10:43
	get busy, people! :)	10:43
jnthn	pmurias: There's already a #?if ... mechanism.	10:43
	pmurias: That's perhaps useful?	10:43
timotimo	jnthn: but that doesn't give logical operations, no?	10:43
	i think that's the core of the question	10:43
pmurias	jnthn: it only takes one backend	10:43
jnthn	oh	10:43
	But it can be negated	10:43
pmurias	but it can't be nested	10:44
jnthn	Nope	10:44
	→ woosley joined	10:44
	← woosley1 left	10:44
pmurias	and I want to disable a given section for both js and parrot	10:44
timotimo	time to pick up m4	10:45
	→ vincent21 joined	10:45
jnthn	Well, guess we could support #?if moar\|jvm	10:45
	← woosley left	10:45
	→ woosley joined	10:45
pmurias	or nesting	10:46
	← woosley left	10:46
jnthn	True.	10:46
pmurias	which one should I implement?	10:46
jnthn	Though the or is probably easier	10:47
	Anyway, it's just a simple Perl 5 script that does the thing.	10:47
	Don't feel strongly; I'd probably do the \|	10:49
	← vincent21 left	10:50
	← PZt left	10:51
	← virtualsue left	10:51
	← simcop2387 left	10:51
	→ simcop2387 joined	10:53
dalek	nqp-js: 06ccf9e \| (Pawel Murias)++ \| tools/build/Makefile-JS.in:	10:59
	nqp-js: Run test 2 during make test.	10:59
	nqp-js: review: https://github.com/pmurias/nqp-js/commit/06ccf9ec01	10:59
	nqp-js: 0772a15 \| (Pawel Murias)++ \| tools/build/gen-cat.pl:	10:59
	nqp-js: Implement #ifdef nesting in tools/build/gen-cat.pl.	10:59
	nqp-js: review: https://github.com/pmurias/nqp-js/commit/0772a157c0	10:59
	nqp-js: 3c6b592 \| (Pawel Murias)++ \| / (2 files):	10:59
	nqp-js: Use more of the setting, pass more tests.	10:59
	nqp-js: review: https://github.com/pmurias/nqp-js/commit/3c6b5929f1	10:59
pmurias	jnthn: nesting was easier to implement and avoid introducing more syntax	11:00
	* avoids	11:00
	← arnsholt left	11:04
	→ vincent21 joined	11:04
	→ rom1504 joined	11:12
daxim	mls, https://build.opensuse.org/project/monitor/home:ldieckow:branches:devel:languages:parrot	11:18
	I've started with 2014.01	11:19
	currently stuck with parrot 6.0.0 because I cannot figure out how to fdupes uninstalled %doc files	11:19
	I would like you to NOTICE ME SEMPAI because I've talked to you about the rakudo stack on obs before and it would be a shame to duplicate work on it	11:21
	→ Mouq joined	11:30
	← pmurias left	11:31
	← Mouq left	11:34
	→ sqirrel joined	11:39
	← Sqirrel_ left	11:39
	← dmol left	11:42
	→ pmurias joined	11:44
	← xenoterracide left	11:52
daxim	segfault when compiling nqp: http://paste.scsys.co.uk/323215 # who feels responsible?	11:53
FROGGS	I do	11:56
	daxim: do you have the MoarVM configure+build output somewhere?	11:56
	← skids left	11:57
daxim	certainly. build log <https://build.opensuse.org/package/live_build_log/home:ldieckow:branches:devel:languages:parrot/moarvm/openSUSE_13.1/x86_64> spec file <https://build.opensuse.org/package/view_file/home:ldieckow:branches:devel:languages:parrot/moarvm/moarvm.spec?expand=1> packages <http://software.opensuse.org/download.html?project=home%3Aldieckow%3Abranches%3Adevel%3Alanguages%3Aparrot&package=moarvm>	12:00
	→ brrt joined	12:01
	→ xenoterracide joined	12:02
	→ arnsholt joined	12:03
	→ scumdog1 joined	12:06
	← scumdog1 left	12:07
	→ btyler joined	12:10
	← benabik left	12:12
brrt	hi #perl6	12:14
colomon	o/	12:14
brrt	is your wheather as awesome as it is here?	12:14
	→ dayangkun joined	12:15
	→ fridim_ joined	12:16
daxim	FROGGS, do you want me to file a formal bug so the crash doesn't get dropped under the table?	12:17
jnthn	daxim: I'm too tied up with $dayjob task to look right now, but if there's a MoarVM segfault please file it.	12:18
colomon	brrt: high for the day is 43F, that's incredibly warm by the standards of the last two months...	12:22
brrt	43F in celcius….	12:23
	oh	12:23
	i've got 20 deg c at least	12:23
	68F	12:23
tadzik	12C here, very nice :)	12:23
colomon	brrt: Jan and Feb had only four or five days that got above 0C here.	12:23
	43F probably means half the town will be breaking out the light jackets and shorts.	12:24
brrt	:-)	12:24
	well, in europe its BBQ time at least	12:24
daxim	here the same info as above as github ticket: https://github.com/MoarVM/MoarVM/issues/81	12:27
	→ mavcunha joined	12:27
	← fridim_ left	12:31
masak	afty, #perl6	12:59
FROGGS	daxim++ # sorry, got distracted	13:01
	← kaare_ left	13:03
	→ kaare_ joined	13:03
FROGGS	daxim: I am moving the discussion to #moarvm	13:04
	→ guru joined	13:09
	guru → Guest58	13:10
	Guest58 → ajr_	13:10
	→ skids joined	13:23
tadzik	mwahaah	13:32
	uh, ww	13:32
masak	:)	13:33
	today's mini-challenge: make this the right window to say "mwhaah" in.	13:34
lizmat	waiting for someone to grab you by the mivocks?	13:35
moritz	which "word", consisting of six lower-case letters, produces the sha1 sum 12f4a69f0d00b48bd73901a75c693b67c877e69d? # for masak's mini-challenge	13:35
brrt	ugh	13:35
	26^6 options	13:35
lizmat	ah, that's so ascii thinking	13:36
moritz	brrt: just scroll up a bit	13:36
brrt	:-) moritz saved me so much time	13:36
masak	it would've been more fun if we had made you search. :)	13:37
moritz	mwahaah	13:37
FROGGS	hehe	13:37
	I got it too fwiw :o)	13:37
mls	daxim: I updated nqp/rakudo/moarvm in devel:languages:parrot last week	13:45
	I didn't upate parrot yet (because nqp no longer needs it for building)	13:46
tadzik	mls! \o/	13:46
mls	hi guys ;)	13:47
FROGGS	mls: nqp still needs parrot if you are going to provide nqp-p	13:47
mls	yes. I currently only build nqp-m	13:47
JimmyZ	wb, mls :)	13:48
moritz	not a bad choice (nqp-m only), we just need NativeCall and a bit more modules support	13:49
FROGGS	yes, I usually use nqp-m only, and when I am done hacking I test my stuff on the two others...	13:52
	JimmyZ too	13:52
jnthn	Same; nqp-m has fastest build/spectest time, so it's the shortest dev cycle. :)	13:52
mls	daxim: do you want write access to devel:languages:parrot?	13:53
FROGGS	btw, someone should rename that to devel:languages:perl6 some day	13:53
	excepts that parrot might stay in parrot	13:53
	-s	13:53
mls	yes. I should have named it "perl6" ;)	13:53
	but at that time I thought that there'll be many languages using parrot as a backend	13:54
JimmyZ	jnthn: just curious, what's your todo list after nativecall :P	13:54
	s/your/your next/	13:54
	← vincent23 left	13:56
	→ vincent23 joined	13:57
	← vincent23 left	13:57
FROGGS	my guess is getting m* in the shops	13:57
	→ vincent23 joined	13:57
	← vincent23 left	13:58
dalek	rakudo-star-daily: 66e0fce \| coke++ \| log/ (5 files):	13:58
	rakudo-star-daily: today (automated commit)	13:58
	rakudo-star-daily: review: https://github.com/coke/rakudo-star-daily/commit/66e0fce493	13:58
jnthn	Yeah...m*, async I/O, the Moar specializer, and native typed rw/array...	14:02
	office &	14:03
	→ kbaker joined	14:03
FROGGS	loop labels would alos be nice... but perhaps somebody else can do that, since it is kinda done for nqp-p, and needs doing for nqp-m and nqp-j	14:04
rurban__	parrot -O2 unfortunately doesn't work yet on nqp. found a label bug	14:11
	→ __thou joined	14:11
	[Coke] wonders if jvm or moar are ready to add the star pile for testing modules.	14:12
	→ virtualsue joined	14:13
	→ jnap joined	14:15
FROGGS	[Coke]: I think so	14:17
	[Coke]: even when NativeCall and their deps will fail, it still makes sense to fix possible bugs in other modules	14:17
	rurban__: the default for parrot is to do -O1 when configures with --optimize?	14:18
	← sqirrel left	14:19
	← xenoterracide left	14:24
rurban__	yes	14:33
	-O1 is equal to -O	14:33
	oh, sorry. --optimize does not call parrot with -O. so not	14:33
	I just tested optimized nqp pbc's via parrot -L. -X. -O2	14:34
	how much nqp packfiles are already optimized or not	14:34
	→ PZt joined	14:37
FROGGS	ahh, I see, thanks	14:37
	← kaare_ left	14:49
	← telex left	14:49
	← ribasushi left	14:49
rurban__	how much constant propagation and constant folding will improve nqp and thus rakudo performance	14:49
	I estimate 2-3% only	14:49
	→ kaare__ joined	14:51
	→ telex_ joined	14:51
	→ ribasushi joined	14:51
	telex_ → telex	14:52
	← kaleem left	14:55
moritz	rakudo already has a constant folder, fwiw	15:00
	so it'll likely only affect the compiler itself	15:01
	→ treehug88 joined	15:09
	← rindolf left	15:13
	→ araujo joined	15:13
	← mavcunha left	15:14
	← jnap left	15:15
rurban__	it affects the generated nqp code. lots of constants which can be propagated	15:15
	fixed the bug, testing now again	15:15
	→ SamuraiJack_ joined	15:17
rurban__	yes, all nqp tests pass now with parrot -O2	15:17
	→ jnap joined	15:18
moritz	was it an nqp or a parrot bug?	15:19
	perlpilot → PerlJam	15:19
timotimo	cool!	15:21
rurban__	moritz: an old premature imcc optimization	15:25
	running more tests, then rakudo tests and updated benchmarks (but I guess it will be only 3-5% max)	15:26
timotimo	is anything perl 6 related going to participate in the gnome outreach programme?	15:27
moritz	if somebody has a great idea for that, I'd be willing to mentor (if it fits my areas of expertise)	15:28
timotimo	well, it's not limited to programming, so documentation could be a viable thing to do	15:28
	(and that's also an area of expertise of yours :) )	15:29
	anyway, i'll start work on today's p6weekly post. i'd love to hear some suggestions :)	15:30
moritz	timotimo: did you note my suggestion for the "how to contribute" section from (end of) last week?	15:30
timotimo	i'm not sure	15:30
	feel free to restate it	15:31
moritz	r: 1	15:31
camelia	rakudo-parrot 1aeb7c, rakudo-jvm 1aeb7c, rakudo-moar 1aeb7c: ( no output )	15:31
moritz	simplify that to 'rakudo-{parrot,jvm,moar} 1aeb7c:'	15:31
timotimo	ah yes	15:31
	did you see my implementation attempt?	15:31
moritz	no	15:32
timotimo	https://gist.github.com/timo/c2119931370ee31360e7	15:32
	alas, it's not complete	15:33
	it doesn't handle situations yet when there's niecza in the mix or something similar	15:33
moritz	and, uhm, I hate to spoil this, but camelia is written in p5 :(	15:33
timotimo	not important. this script can pre-generate all the combinations and write them to a text file	15:34
	you would then just have to run that script whenever you rebuild the rakudos and nieczas and stuff	15:34
	→ treehug8_ joined	15:35
moritz	well, I'd prefer it camelia's workflow didn't get more complicated	15:36
timotimo	fair enough	15:36
	i'll put it up on the blog	15:36
	← treehug88 left	15:36
moritz	thanks	15:36
	← arnsholt left	15:38
timotimo	can perl6-m now completely build parrot with all dependencies and stuff? now that the precomp bug is fixed?	15:39
moritz	build parrot?	15:40
timotimo	er	15:40
	moritz confused	15:40
timotimo	panda	15:40
	all those animals starting with pa :\|	15:40
pmurias	is there a lot of stuff missing for a perl6 camelia?	15:41
timotimo	pmurias: how would you like your work for this week described?	15:42
	you're adding more and more files to the core setting and making more and more spectests pass?	15:42
PerlJam	.oO( pmurias: the most awesome Perl 6 work ever accomplished! )	15:42
timotimo	for last week, actually	15:43
	well, last week and today :)	15:43
moritz	pmurias: probably something to make the IRC module reconnect on timeouts and other stuff	15:43
	→ denis_boyun joined	15:44
tadzik	hmm	15:45
timotimo	tadzik: you got something cool to report?	15:45
tadzik	I wonder how is Net::IRC::Bot doing	15:45
	timotimo: no :(	15:45
timotimo	that's fine	15:45
tadzik	I did absolutely nothing, aside from the usual maintenance	15:45
moritz	tadzik: last time I used it, it, like, mostly worked quite well	15:45
timotimo	if you keep that up, though, be prepared to receive a strongly worded letter :P	15:45
tadzik	:>	15:46
	I know, I know	15:46
	want to give me a project idea?	15:46
	I'd like to do something in Perl 6 userspace	15:46
timotimo	huh? but there are no perl 6 users!	15:46
masak	tadzik: mishu!	15:46
tadzik	oh, I could rework MuEvent and make it exploit all the concurrency awesomeness	15:46
moritz	timotimo: port camelia to Perl 6	15:46
masak	ooh	15:47
tadzik	moritz: me? :)	15:47
	I wanted to start with synopsebot	15:47
	see if it can run long-term	15:47
timotimo	http://perlcabal.org/syn/S15.html - is this supposed to 404?	15:47
tadzik	masak: mishu is a great idea, but :)	15:47
	I was already sick for the entire last week, and if I was to take a week of and go to sweden now, $work may get... disappointed	15:48
masak	tadzik: oh, you wanted to meet up for it? awesome, I didn't realize.	15:49
tadzik	I wouldn't want it any other way :)	15:49
masak	\o/	15:49
	tadzik: even hearing that, I'm very excited. looking forward to meeting you, whenever :D	15:49
timotimo	tadzik: do you know what the exact status of panda on moar is? now that the precomp bug has been fixed?	15:50
tadzik	I found that I can trick myself into being productive by traveling somewhere to be productive there	15:50
	it worked for all the last conferences with hackathons :)	15:50
	timotimo: I think it works :P	15:50
	let me check again	15:50
timotimo	to be honest, i wasn't even really sure what exactly broke :(	15:50
	i know you couldn't install URI, or at least i thought that was it	15:50
tadzik	ah, URI is tricky	15:50
	and I think some thinking is required for how to spec META.infos and all this stuff to make it not-crazy	15:51
	what I mean is: modules with per-backend code are tricky	15:51
moritz	URI has always been the big test for precompilation on rakudo-p	15:51
tadzik	oh, maybe it's mime-base64 that has those problems	15:51
	either way, it's something that wants to be solved	15:52
moritz	that's also tricky, yes :-)	15:52
hoelzro	I got MIME::Base64 precompiling not an hour ago	15:52
timotimo	\o/	15:52
moritz	URI is tricky because it has lots of interdependencies, and partially used with require	15:52
timotimo	hoelzro: can you summarize last week's stuff for me?	15:52
hoelzro	I had to trick it, though	15:52
moritz	and MIME::Base64 is tricky because it has a parrot-specific backend	15:52
hoelzro	timotimo: I'm afraid I've been in and out of the Perl 6 world over the last few weeks =/	15:52
timotimo	that's fine	15:53
vincent21	https://gist.github.com/timo/c2119931370ee31360e7	15:53
hoelzro	latest version of the Moar report: https://gist.github.com/hoelzro/9467700	15:53
pmurias	timotimo: it's not awesome at all ;) I'm porting over the nqp-js to use the nqp build system so it can be merged into the repo	15:53
timotimo	pmurias: ah, so just nqp so far?	15:53
	that's fine, too! :)	15:53
pmurias	yes, and those tests are passing using a different build system	15:54
hoelzro	crap, forgot summary.txt =(	15:55
tadzik	ah	15:56
	panda cannot connect to feather for some reason	15:56
	and it only happens on my box; I guess I'll debug that	15:57
	← jnap left	15:57
tadzik	what was that strace incantation for child threads?	16:00
	ach, -3	16:00
	-f that is	16:01
	and mystery solved: moarvm cannot ipv6	16:01
hoelzro	ok, summary.txt is in the latest moar report	16:02
	hoelzro .oO( I should set up the moar report to run automatically... )	16:03
tadzik	argh, gdbing perl6-m is pain, or am I doing it wrong?	16:07
timotimo	http://p6weekly.wordpress.com/?p=86&shareadraft=531de39172131 - first draft. i've surely missed some interesting and/or important changes during the week?	16:09
	← fhelmberger left	16:10
tadzik	huh	16:10
	moarvm is doing something very weird with ipv6	16:11
	https://gist.githubusercontent.com/tadzik/9468083/raw/349b9d4b2bcbe1ebb2a94bb568a44a58599e0f66/gistfile1.txt	16:11
	it connects to correct host in the first place, but then it closes it, and connects to...incomplete address	16:12
TimToady	timotimo: no one will know what moar-conc means	16:13
timotimo	good point.	16:13
	→ benabik joined	16:13
timotimo	actually ... shouldn't it be possible to (ab)use the backtracking of regexes to make the prefix & suffix summarization much shorter code-wise?	16:16
	hm, not quite. it would probably not find summarizable things if it had to re-order them first	16:17
tadzik	argh argh	16:20
	timotimo: halp	16:20
timotimo	but i don't know how to sockets :(	16:20
tadzik	but you know how to gdb	16:20
	and I can't run moarvm with this thing	16:20
	it needs some wizardry with 'set args' I figured	16:21
timotimo	oh	16:21
	gdb and args is easy	16:21
	just figure out how exactly moarvm is invoked	16:21
	and then gdb --args /foo/bar/moarvm --libpath=something something.moarvm foo.pm6	16:22
	(a friend recently showed me this and blew my mind completely and utterly)	16:23
[Coke]	you can do that from the command line!?	16:23
tadzik	oh	16:23
benabik	Huh. I had always just used `gdb foo` and then `run bar baz` to debug `foo bar baz`	16:23
tadzik	me too	16:23
	oh?	16:23
timotimo	YES! :D	16:23
tadzik	me not	16:23
[Coke]	timotimo++ timotimo++ timotimo++ # holy crap	16:24
tadzik	I have no idea what's surprising me anymore	16:24
	but something is	16:24
	but with the help of puts() I found where moarvm fails anyway :D	16:24
jnthn	tadzik: perl6-m is just a shell script; it's better to gdb moar itself and r it with the stuff in the shell script plus your args, I think. At least, that's what I do with msvc...	16:25
	← [Sno] left	16:26
benabik	GNU utilities always have the least useful man pages. I know they'd prefer we use the info pages, but those don't always seem to be installed. At least a summary of all the options would be nice.	16:26
	← brrt left	16:26
timotimo	jnthn: no, even better is to gdb --args moar foo bar baz	16:26
	:)	16:26
benabik	timotimo: I doubt that works with msvc	16:26
jnthn	timotimo: Or what :) Either way, without the shell script :P	16:26
tadzik	yeah, I got to that part :)	16:27
jnthn	benabik: I tend to set it up inside VS :)	16:27
tadzik	verdict: resolve_host_name can't ipv6	16:27
jnthn	tadzik: I never actually tried out the ipv6 bit of Moar sockets...	16:27
	→ dmol joined	16:27
jnthn	I thought the code I did would handle to it...	16:27
	s/to//	16:27
	timotimo: "He has also said before that Parrot’s concurrency model is vastly superior to MoarVM’s, so I’m very excited to see an implementation of Perl 6 concurrency on top of parrot." - might want to mention that's a highly disputed claim. :)	16:29
	timotimo: The font size seems to shrink in the last paragraph...	16:30
tadzik	memcpy(dest, result->ai_addr, sizeof(struct sockaddr));	16:30
	that looks wrongish	16:30
timotimo	how is "While not everybody agrees with him, I'm very excitedly looking forward to an implementation of Perl 6 concurrency on top of parrot that kicks MoarVM's butt."?	16:30
tadzik	oh, not	16:30
	I should talk less :P	16:30
jnthn	timotimo: Well, "so we can see if the claim holds up" is more the reason :)	16:31
timotimo	fair enough	16:31
jnthn	It'll need a volunteer to write that implementation, 'cus I sure ain't working on it.	16:31
timotimo	sure enough	16:31
tadzik	ah, yes, gotcha	16:32
jnthn	tadzik: Found the bug? :)	16:33
	→ jnap joined	16:33
tadzik	I think	16:33
	I think sockaddr only works for v4, and you need sockaddr_in6 for v6	16:33
	jnthn has some teaching to do over the coming weeks, BUT has got finished with writing course material, which is good for the tuit supply :)	16:33
jnthn	tadzik: ah, ok :)	16:34
	← zakharyas left	16:34
tadzik	or maybe not :	16:34
timotimo	Should i encourage my readers to volunteer to do the Rakudo * 2014.02 release even though it's already half way into 03?	16:34
tadzik	I never bothered to learn ipv6 sockets. "Baah, we'll never use ipv6 anyway"	16:34
jnthn	timotimo: Probably not...	16:34
tadzik	and now I'm the one who noticed the bug... :)	16:34
timotimo	hm, ok.	16:34
jnthn	timotimo: Better use of time is doing stuff to make sure the 2014.03 one supports more than just Parrot :)	16:35
timotimo	aye. but that's not necessarily something a semi-outsider could just pick up	16:35
	jnthn wonders if it's github or the train wifi that is being flaky...	16:37
	← jnap left	16:38
	→ jnap joined	16:38
	← SamuraiJack_ left	16:39
tadzik	github is	16:41
jnthn	ah, k	16:41
	uh-oh...here's the underground station where connection gets interesting...	16:41
tadzik	:P	16:41
	→ vincent22 joined	16:43
nwc10	timotimo: true, for "not necessarily". Some people probably can. (and possibly more easily than they can get the confidence to roll a release)	16:44
	I think that it's the right thing to suggest trying - help work on ecosystem modules	16:44
	even if it's just golfing problems down to bug reports	16:44
timotimo	anything in particular you'd like me to mention?	16:44
nwc10	I don't know. I hope that jnthn, [Coke] or holzero would have a better idea	16:45
	and I have a train to catch	16:45
	to see the damian	16:45
tadzik	aww yiss	16:45
	I seem to have fixed it	16:45
	should socket spectests work on rakudo?	16:45
	rakudo-moar, that is	16:46
jnthn	tadzik: yes	16:46
	tadzik: Passed for me last I checked (day or two ago)	16:46
tadzik	okayo	16:46
[Coke]	what are we looking at ? things to do for people who are just joining the #perl6 community?	16:46
timotimo	[Coke]: i'll mention mojo6 if that's okay with you	16:46
	[Coke]: yes, i'd like to encourage people who have so far just been looking/lurking to have a closer look at the community and implementations	16:47
	→ SamuraiJack joined	16:48
	← ajr_ left	16:49
	→ isBEKaml joined	16:49
tadzik	oh, hahaha	16:51
	now that I fixed ipv6, it turns out that the moduleDB doesn't support it :P	16:51
timotimo	hah, excellent :)	16:51
	still, decent ipv6 support is definitely a necessity for Perl 6	16:51
jnthn	tadzik: The patch is still welcome :)	16:52
timotimo	not only because the numbers are equal :)	16:52
tadzik	sure :)	16:52
timotimo	http://p6weekly.wordpress.com/?p=86&shareadraft=531dedc52b185	16:52
tadzik	ok, I can fetch stuff from ipv6.google.com, so I say it works	16:53
	→ rindolf joined	16:55
	→ btyler_ joined	16:56
	← btyler left	16:57
	← vincent21 left	16:57
	← jnap left	16:57
	← virtualsue left	16:58
	→ jnap joined	17:00
	← mcglk left	17:03
	← FROGGS left	17:03
	→ mcglk joined	17:04
	→ mcglk_ joined	17:05
	→ mcglk__ joined	17:06
timotimo	post published	17:08
	← mcglk left	17:09
	→ mcglk joined	17:10
	← mcglk_ left	17:10
tadzik	\o/	17:10
	timotimo++	17:10
jnthn	yay, think I've a patch that fixes the Supplies issue without regressing roles tests :)	17:11
	← mcglk__ left	17:11
	→ raiph joined	17:11
tadzik	aww yiss	17:12
	now both panda-moar and moduledb like ipv6	17:12
	I hope they still like ipv4 too	17:12
	==> Successfully installed Term::ANSIColor	17:12
jnthn	\o/	17:12
tadzik	I like how fast that is. May actually convince people to trust panda a bit, instead of installing things manually :P	17:13
isBEKaml	panda is actually a bumbling animal. Why would people trust it with important stuff?	17:13
tadzik	duude, chiillll	17:14
jnthn	'cus it's unbearably cute	17:14
	← mcglk left	17:14
isBEKaml	It only reinforces its image with those Kung fu series. :)	17:14
	→ mcglk joined	17:14
tadzik	I think pandas are always high on eucaliptus	17:14
	oh wait, that's koalas	17:15
isBEKaml	mate, you got your geography mixed up.	17:15
jnthn	Pandas only do bamboo...which I'm not sure you can get high on, though admittedly I never tried...	17:16
[Coke]	(mojo6) sure, point them to me on this channel.	17:16
	← jnap left	17:17
	← mcglk left	17:18
	← SamuraiJack left	17:19
dalek	rakudo/moar-conc: c9d478e \| jonathan++ \| src/vm/moar/ops/container.c:	17:21
	rakudo/moar-conc: Fix role type checking bug in assignment.	17:21
	rakudo/moar-conc:	17:21
	rakudo/moar-conc: Mostly unbusts Supply on Moar.	17:21
	rakudo/moar-conc: review: https://github.com/rakudo/rakudo/commit/c9d478ef62	17:21
	→ mcglk joined	17:22
	→ Rotwang joined	17:22
	→ mcglk_ joined	17:23
vendethiel	r: 0 but "" ==> .perl	17:24
camelia	rakudo-parrot 1aeb7c, rakudo-jvm 1aeb7c, rakudo-moar 1aeb7c: OUTPUT«[31m===[0mSORRY![31m===[0m Error while compiling /tmp/tmpfile␤Sorry, do not know how to handle this case of a feed operator yet.␤at /tmp/tmpfile:1␤------> [32m0 but "" ==> .perl[33m⏏[31m<EOL>[0m␤ …»	17:24
vendethiel	r: say (0 but "").perl	17:24
camelia	rakudo-parrot 1aeb7c, rakudo-jvm 1aeb7c, rakudo-moar 1aeb7c: OUTPUT«␤»	17:24
	→ mcglk__ joined	17:24
	→ mcglk___ joined	17:26
vendethiel	uuh :(	17:26
	← mcglk left	17:26
	mcglk___ → mcglk	17:26
TimToady	r: say ("" but 0).perl	17:27
camelia	rakudo-parrot 1aeb7c, rakudo-jvm 1aeb7c, rakudo-moar 1aeb7c: OUTPUT«""␤»	17:27
	← mcglk_ left	17:27
	→ mcglk___ joined	17:28
	← mcglk__ left	17:28
	→ mcglk_ joined	17:29
	→ mcglk__ joined	17:30
	→ SamuraiJack joined	17:31
	← mcglk left	17:31
	→ mcglk joined	17:32
[Coke]	no buts about .perl	17:32
	← mcglk___ left	17:32
tadzik	oh-oh, interesting fact	17:32
	rakudo-moar is apparently the first rakudo to support IPv6	17:32
TimToady	r: say ("foo" but "bar").perl	17:32
camelia	rakudo-parrot 1aeb7c, rakudo-jvm 1aeb7c, rakudo-moar 1aeb7c: OUTPUT«"foo"␤»	17:32
TimToady	r: say ("foo" but "bar").gist	17:33
camelia	rakudo-parrot 1aeb7c, rakudo-jvm 1aeb7c, rakudo-moar 1aeb7c: OUTPUT«foo␤»	17:33
TimToady	r: say ("foo" but "bar").Str	17:33
camelia	rakudo-parrot 1aeb7c, rakudo-jvm 1aeb7c, rakudo-moar 1aeb7c: OUTPUT«bar␤»	17:33
	← mcglk_ left	17:33
tadzik	or maybe other platforms have clever sockets that fallback to v4	17:33
	→ mcglk_ joined	17:34
	→ FROGGS[mobile] joined	17:34
	← mcglk__ left	17:34
	→ mcglk__ joined	17:35
	← mcglk left	17:36
	← kbaker left	17:38
	← isBEKaml left	17:38
	← mcglk_ left	17:38
geekosaur	many but not all linux distributions make use of a kernel network stack mechanism where opening an ipv4 socket automatically opens a compatible associated ipv6 socket	17:39
	→ mcglk joined	17:39
	← mcglk__ left	17:40
tadzik	oh, neat	17:41
	→ mcglk_ joined	17:41
dalek	perl6-roast-data: e4a6668 \| coke++ \| perl6_pass_rates.csv:	17:42
	perl6-roast-data: today (automated commit)	17:42
	perl6-roast-data: review: https://github.com/coke/perl6-roast-data/commit/e4a66689a6	17:42
	← dakkar left	17:43
	← mcglk left	17:44
	← mcglk_ left	17:46
	→ [Sno] joined	17:46
	← pdcawley left	17:52
	→ mcglk joined	17:53
	→ mcglk_ joined	17:54
	← pecastro left	17:55
	→ mcglk__ joined	17:55
	← mcglk left	17:57
	← mcglk_ left	17:59
	→ mcglk joined	17:59
	← mcglk__ left	18:00
	← lizmat left	18:00
	→ lizmat joined	18:00
	→ mcglk_ joined	18:02
	← mcglk left	18:04
	← mcglk_ left	18:06
	→ vincent_ joined	18:09
	← vincent_ left	18:09
	→ spider-mario joined	18:13
pmichaud	good afternoon, #perl6	18:14
jnthn	o/ pmichaud	18:15
	→ kbaker joined	18:15
pmichaud	did anyone else see http://lists.alioth.debian.org/pipermail/pkg-rakudo-devel/2014-March/000367.html ?	18:15
	→ guru joined	18:17
	guru → ajr_	18:17
jnthn	pmichaud: Hadn't seen that, no	18:18
pmichaud	I'm reaching out to Alessandro to see if he can tell us the best way to pass the packaging role to one of us.	18:18
jnthn	pmichaud++	18:19
pmichaud	if there's anyone that would be interested in taking over that role/responsibility, please let us know asap	18:21
	right now I'm out for a walk :)	18:21
	→ jnap joined	18:27
	→ dwarring joined	18:30
	→ mcglk joined	18:32
jnthn	detrain, dinner &	18:33
	→ mcglk_ joined	18:34
	← mcglk left	18:37
	→ arnsholt joined	18:37
	← mcglk_ left	18:39
[Coke]	jnthn: shouldn't that be "derail" ?	18:39
	← jnap left	18:40
	→ FROGGS joined	18:42
	→ denis_boyun_ joined	18:44
	← denis_boyun left	18:45
	← darutoko left	18:47
	→ cognominal__ joined	18:55
	← cognominal left	18:56
FROGGS	tadzik / [Coke] / timotimo: I usually do: which perl6-m, cat /path/to/perl6-m, and then copy most of it, then: gdb --args <most of it> -e 'something that segfaults'	19:00
	→ LLamaRider joined	19:04
tadzik	I wouldn't mind a gdb-perl6 script :)	19:04
LLamaRider	[Coke]++ best of luck with the Mojolicious port! I'm eagerly looking forward to it :)	19:05
	does someone know if NativeCall reached a state where it can expose variables to P6? I remember that had blocked me some time back when trying to port libxml.	19:07
	→ hoverboard joined	19:07
FROGGS	timotimo++ # blög	19:12
timotimo	LLamaRider: i think if you want to do something like that you'll have to write a tiny piece of Inline::C	19:18
	→ jnap joined	19:19
FROGGS	yes, there is no way to access defines for example without writing C code	19:20
	okay, there could be some magic that writes code for you in theory	19:20
	← btyler_ left	19:26
[Coke]	LLamaRider: glad to hear it. Just need... tuits. :)	19:28
	→ Sqirrel joined	19:33
vendethiel	FROGGS: (mind giving me the link again ? blog)	19:35
FROGGS	vendethiel: http://p6weekly.wordpress.com/	19:37
vendethiel	FROGGS++, timotimo++	19:37
FROGGS	:o)	19:37
moritz	timotimo++	19:40
lue	apparently, binding an Array.new (instead of assigning it) seems to work for the issue I described last night.	19:44
FROGGS	jnthn: I am working currently at "panda bundle"	19:48
lue	Which means that it's the assignment operator itself that was breaking there, though I have no clue why the hash wasn't receptive at that specific time.	19:48
FROGGS	jnthn: it can already detect dependencies, be registering my own module loader, which records every module loading, and delegates to the normal one	19:48
	jnthn: how could I get my hands on the things a module provides? like class definitions and such?	19:49
timotimo	its WHO, no?	19:49
FROGGS	on what do I call it?	19:50
timotimo	no, not on WHAT	19:50
	:P	19:50
	dunno :\|	19:50
FROGGS	perhaps I should just compare GLOBAL of before and after loading a module	19:50
LLamaRider	timotimo: then in P6 the equivalent of XS is a mixture of NativeCall and Inline::C (that's a P6 port?)?	19:51
timotimo	not quite, XS is much more than that	19:51
	as in: XS lets you directly fiddle with the interpreter's internal state	19:51
	in NativeCall, you can't do that	19:51
	since NativeCall is cross-backend	19:52
LLamaRider	oh, so a direct port might even be impossible. Great news xD	19:52
timotimo	you can, however, build extension libraries, like what rakudo does to gets the nqp::p6* ops in place	19:52
	but i'm not sure how easy it is to make that "packagable"	19:52
	on moar and parrot i think you have to supply the extop on the commandline, but perhaps only at build-time?	19:54
	and those are specific per VM, of course	19:54
	→ dlinnyj joined	19:54
LLamaRider	I'm not sure what is controversial about passing a C variable into P6 through NativeCall. Shouldn't be that different than function names or?	19:54
timotimo	i think the lib we're using doesn't supply that functionality, but i could easily be wrong.	19:55
	← hoverboard left	19:55
FROGGS	are we talking about something define'd in land?	19:57
timotimo	no, define is never possible without invoking a compiler (or at least the preprocessor)	19:58
	→ hoverboard joined	19:58
FROGGS	okay, so far we agree	19:58
	and how (on earth) should somebody access a variable?	19:59
timotimo	well, you get a scalar that is repr'd "CVariable" and that directly points at the memory of the variable	20:00
FROGGS	ahh	20:00
	I am doing this crap in SDL, getting a memory address by a function, and then unpack that memory...	20:01
	shudder	20:01
LLamaRider	yes, it ought to be something shudder-invoking :D	20:01
timotimo	the only hard part i can really think of right now is how to find the memory address reliably	20:01
	perhaps you can even use the CStruct repr for single variables at that point	20:02
FROGGS	m: BEGIN my $g = GLOBAL::.keys; use Test; say GLOBAL::.keys (-) $g	20:03
camelia	rakudo-moar 1aeb7c: OUTPUT«set(Test)␤»	20:03
timotimo	heh	20:03
	that's lovely and a tiny bit scary	20:03
	← SamuraiJack left	20:07
	→ jnap1 joined	20:13
	← jnap left	20:14
	← vincent22 left	20:19
	jnthn back	20:20
	← dlinnyj left	20:21
jnthn	FROGGS: If you're at module loader level, you can look into the pre-merged GLOBALish which is accessible through the module context ($module_ctx or so I think at some point in ModuleLoader)	20:23
FROGGS	jnap1: yes, that what I am going to do, compare GLOBALish from before and after	20:24
	because I don't want to implement more of load_module's logic	20:25
	jnthn: this is already there: https://github.com/tadzik/panda/commit/17b66240d4c7af4da0d779f272d2bb4bf29716f9#diff-4	20:25
jnthn	FROGGS: No, GLOBALish in $module_ctx, before global merging happens, is that diff I think.	20:28
	Also there's EXPORT.	20:28
FROGGS	ahh, hmm	20:28
jnthn	FROGGS: To be clear, I'm suggesting you might be able to intercept it before merge_globals happens...	20:28
	→ pecastro joined	20:31
	← benabik left	20:37
	→ BenGoldberg joined	20:39
FROGGS	jnthn: but how?	20:40
jnthn	FROGGS: Dunno, it'll need some changes to ModuleLoader I guess	20:42
FROGGS	yeah...	20:42
jnthn	FROGGS: I think that probably what is today ModuleLoader just wants to end up being "guts" support for writing a CompUnitRepo.	20:42
	if I understand The Plans correctly anyway :)	20:43
FROGGS	that sounds about right	20:43
	it is The One And Only when it comes to compiler only stuff, but after that it just does the dirty work of loading an already located file (or perhaps piece om mem some day)	20:44
	← rindolf left	20:49
FROGGS	jnthn: btw, nextsame is not specced to return, right?	20:49
timotimo	that's right	20:49
FROGGS	ahh, have an idea...	20:49
timotimo	callsame does that	20:49
FROGGS	ahh	20:49
	of course	20:50
	"next"-same	20:50
	→ xenoterracide joined	20:50
FROGGS	sometimes it is too obvious :o)	20:50
	timotimo++	20:50
jnthn	Yeah, nextsame is "pass control on to the candidate and don't come back to me again" :)	20:50
timotimo	:)	20:50
	.o( don't you dare come crawling back to me )	20:50
FROGGS	hehe	20:50
	NEXT	20:50
jnthn	tssk. Finally I have a bit of free time but...so tired...	20:51
	← kaare__ left	20:51
timotimo	don't push yourself too hard :)	20:51
jnthn	yeah. Gotta teach tomorrow	20:51
	Thing is, the next issue in my queue is a parallel GC bug, which I probably should attempt when more awake ;)	20:52
FROGGS	ohh noes!	20:52
timotimo	oh :(	20:52
jnthn	Well, this is the first time we're exercising the parallel GC with anything resembling a real world workload.	20:53
timotimo	right	20:53
jnthn	We've already taken the non-parallel case from "argh SEGV" to "haven't seen a GC bug in a good while". I'm quite sure we can do the same with the parallel GC too. :)	20:53
timotimo	yays	20:53
jnthn	These things just take work/patience is all.	20:54
	timotimo unfortunately has no clue how to do it himself	20:54
	← LLamaRider left	20:56
jnthn	Well, guess I'm concious enough to stub some nativecall bits maybe... :)	20:57
timotimo	that sounds nice :)	20:58
FROGGS	very nice :o)	20:58
	I do like NativeCall... and it could be quite fun to hack on SDL again	20:59
japhb	Someone asked earlier about the state of Net::IRC::Bot. The concurrency branch seems to be working fine. Before I merge, I want to 1) run some load testing to make sure it's as concurrent as I want it to be (I have some ideas for this, but feel free to volunteer tests!), and 2) I want to test on more than one Rakudo backend. It sounds like this will be coming soon, thanks to jnthn++. :-)	20:59
jnthn	japhb: Yeah, still some work to go, but it's coming along.	21:00
FROGGS	daxim: I can't reproduce your bug in aa openSUSE-13.1_x86_64 virtualbox :/	21:01
	an*	21:01
	← dayangkun left	21:01
	← skids left	21:02
	← Sqirrel left	21:04
FROGGS	Nominal type check failed for parameter '<anon>'; expected CompUnitRepo but got DepTracker	21:05
	so callsame is not very happy when called in a method, when "self" changes	21:05
	← BenGoldberg left	21:07
FROGGS	m: class Foo { method bar { } }; class Bar { method bar { callsame } }; Bar.bar	21:07
camelia	rakudo-moar 1aeb7c: ( no output )	21:07
FROGGS	hmpf	21:07
	m: class Foo { method bar { } }; class Bar if Foo { method bar { callsame } }; Bar.bar	21:07
camelia	rakudo-moar 1aeb7c: OUTPUT«[31m===[0mSORRY![31m===[0m Error while compiling /tmp/LyLGKxSHtp␤Unable to parse class definition␤at /tmp/LyLGKxSHtp:1␤------> [32mclass Foo { method bar { } }; class Bar [33m⏏[31mif Foo { method bar { callsame } }; Bar.[0m␤ expecting …»	21:07
FROGGS	m: class Foo { method bar { } }; class Bar is Foo { method bar { callsame } }; Bar.bar	21:07
camelia	rakudo-moar 1aeb7c: ( no output )	21:08
FROGGS	m: class Foo { method bar { 42 } }; class Bar is Foo { method bar { callsame } }; say Bar.bar	21:08
camelia	rakudo-moar 1aeb7c: OUTPUT«42␤»	21:08
FROGGS	damn	21:08
	← raiph left	21:09
	← FROGGS[mobile] left	21:11
	→ dayangkun joined	21:18
	→ tgt joined	21:18
	← dwarring left	21:28
	← kbaker left	21:29
	→ Mouq joined	21:32
	← telex left	21:33
	→ BenGoldberg joined	21:34
	→ telex joined	21:34
timotimo	should we try to do something cool with BuildID?	21:42
	like, whenever you build a MoarVM (and the extensions in nqp and rakudo) stuff will get uploaded to some file host so that reproducibility is "easy"?	21:43
	→ treehug88 joined	21:44
	← Rotwang left	21:44
	← treehug8_ left	21:46
	→ rurban joined	21:51
	→ jnap joined	21:52
	← jnap1 left	21:53
rurban	nwc10: Is this parrot master? parrot -O2 should be faster now	21:54
Mouq	tadzik: https://github.com/tadzik/Task-Star/pull/1	22:05
dalek	nqp: 8a414b7 \| jonathan++ \| / (2 files):	22:06
	nqp: Map nativecall ops.	22:06
	nqp:	22:06
	nqp: Nothing works yet; the next step is to build stuff up in MoarVM to get	22:06
	nqp: towards passing t/nativecall/01-basic.t.	22:06
	nqp: review: https://github.com/perl6/nqp/commit/8a414b763a	22:06
lue	Hm. I'm afraid I'll have to go with an alternate route to place these backlinks by fragment; it's taking way too much longer to get through that step :/ .	22:08
	← hoverboard left	22:11
	← ajr_ left	22:13
	← spider-mario left	22:26
	← rurban left	22:27
	← tgt left	22:28
	← treehug88 left	22:31
	→ skids joined	22:34
Mouq	r: my &deb = {note $^c.MATCH; $^c}; say "ab:cd" ~~ /[<alpha>+] +% [\: &deb]/	22:37
camelia	rakudo-jvm 1aeb7c: OUTPUT«｢ab:｣␤ alpha => ｢a｣␤ alpha => ｢b｣␤␤java.lang.RuntimeException: No such attribute '$!pos' for this object␤ in method ACCEPTS at gen/jvm/CORE.setting:12747␤ in method ACCEPTS at gen/jvm/CORE.setting:798␤ in block at /tmp/tmpfile:…»	22:37
	..rakudo-parrot 1aeb7c: OUTPUT«｢ab:｣␤ alpha => ｢a｣␤ alpha => ｢b｣␤␤maximum recursion depth exceeded␤current instr.: 'print_exception' pc 140541 (src/gen/p-CORE.setting.pir:59872) (gen/parrot/CORE.setting:11098)␤called from Sub '' pc 41459 (src/gen/p-CORE.setting…»	22:37
	..rakudo-moar 1aeb7c: OUTPUT«｢ab:｣␤ alpha => ｢a｣␤ alpha => ｢b｣␤␤P6opaque: no such attribute '$!pos'␤ in method ACCEPTS at src/gen/m-CORE.setting:12802␤ in method ACCEPTS at src/gen/m-CORE.setting:798␤ in block at /tmp/tmpfile:1␤␤»	22:37
Mouq	Firstly: ^^ that's awesome.	22:38
	Secondly: That's the same bug that prevents Grammar::Expression from being possible	22:38
	Well, maybe not *possible, but it definitely makes it a whole lot harder	22:39
	r: use NQPCursorRole:from<NQP>; my &deb = {note $^c.MATCH; $^c but NQPCursorRole}; say "ab:cd" ~~ /[<alpha>+] +% [\: &deb]/	22:40
camelia	rakudo-moar 1aeb7c: OUTPUT«===SORRY!===␤Unexpected named parameter 'line' passed␤»	22:40
	..rakudo-parrot 1aeb7c: OUTPUT«===SORRY!===␤"load_bytecode" couldn't find file 'NQPCursorRole.pbc'␤»	22:40
	..rakudo-jvm 1aeb7c: OUTPUT«===SORRY!===␤java.nio.file.NoSuchFileException: NQPCursorRole␤»	22:40
Mouq	The bug has to do with returning Cursors from code back to the Cursor that called the code	22:46
	→ hoverboard joined	22:46
	hoverboard → Guest27230	22:47
Ulti	even golfed Perl6 looks nice http://codegolf.stackexchange.com/questions/13152/shortest-code-to-produce-infinite-output/23706#23706	22:47
synopsebot	Link: https://rt.perl.org/rt3//Public/Bug/Display.html?id=23706	22:47
Ulti	its also amongst the shortest other than horrible things like golfscript	22:48
Mouq	r: .say xx *	22:50
camelia	rakudo-parrot 1aeb7c, rakudo-jvm 1aeb7c, rakudo-moar 1aeb7c: ( no output )	22:50
Mouq	r: 0.say xx *	22:50
camelia	rakudo-parrot 1aeb7c, rakudo-jvm 1aeb7c, rakudo-moar 1aeb7c: ( no output )	22:50
Mouq	Oh, it's actually lazy now :P :D	22:50
	r: eager 0.say xx *	22:50
camelia	rakudo-parrot 1aeb7c, rakudo-jvm 1aeb7c, rakudo-moar 1aeb7c: ( no output )	22:50
Mouq	:?	22:50
	r: (0.say xx *)[15]	22:50
camelia	rakudo-parrot 1aeb7c, rakudo-jvm 1aeb7c, rakudo-moar 1aeb7c: OUTPUT«0␤0␤0␤0␤0␤0␤0␤0␤0␤0␤0␤0␤0␤0␤0␤0␤»	22:50
Ulti	my plan is to do pretty Perl6 golfs though rather than the golf style we had with Perl5	22:51
	though I guess xx * is kind of neat	22:51
	but maybe less obvious what it does	22:51
Mouq	r: sub term:<∞> {(state$)++.say; ∞}()	22:54
camelia	rakudo-jvm 1aeb7c: OUTPUT«0␤0␤1␤2␤3␤4␤5␤6␤7␤8␤9␤10␤11␤12␤13␤14␤15␤16␤17␤18␤19␤20␤21␤22␤23␤24␤25␤26␤27␤28␤29␤30␤31␤32␤33␤34␤35␤36␤37␤38␤39␤40␤41␤42␤43␤44␤45␤46␤47␤48␤49␤50␤51␤52…»	22:54
	..rakudo-moar 1aeb7c: OUTPUT«(timeout)0␤0␤1␤2␤3␤4␤5␤6␤7␤8␤9␤10␤11␤12␤13␤14␤15␤16␤17␤18␤19␤20␤21␤22␤23␤24␤25␤26␤27␤28␤29␤30␤31␤32␤33␤34␤35␤36␤37␤38␤39␤40␤41␤42␤43␤44␤45␤46␤47␤48␤49␤50…»	22:54
	..rakudo-parrot 1aeb7c: OUTPUT«(timeout)0␤0␤1␤2␤3␤4␤5␤6␤7␤8␤9␤10␤11␤12␤13␤14␤15␤16␤17␤18␤19␤20␤21␤22␤23␤24␤25␤26␤27␤28␤29␤30␤31␤32␤33␤34␤35␤36␤37␤38␤39␤40␤41␤42␤43␤44␤45␤46␤47␤48␤49␤5…»	22:54
Ulti	o___O	22:56
TimToady	that doesn't seem right	22:57
lue	TimToady: what, the two 0's? I agree :)	22:57
Mouq	n: sub term:<∞> {(state$)++.say; ∞}()	22:57
camelia	niecza v24-109-g48a8de3: OUTPUT«(timeout)0␤1␤2␤3␤4␤5␤6␤7␤8␤9␤10␤11␤12␤13␤14␤15␤16␤17␤18␤19␤20␤21␤22␤23␤24␤25␤26␤27␤28␤29␤30␤31␤32␤33␤34␤35␤36␤37␤38␤39␤40␤41␤42␤43␤44␤45␤46␤47␤48␤49␤50…»	22:58
Mouq	m: sub term:<∞> {(state$=0)++.say; ∞}()	22:58
camelia	rakudo-moar 1aeb7c: OUTPUT«(timeout)0␤0␤1␤2␤3␤4␤5␤6␤7␤8␤9␤10␤11␤12␤13␤14␤15␤16␤17␤18␤19␤20␤21␤22␤23␤24␤25␤26␤27␤28␤29␤30␤31␤32␤33␤34␤35␤36␤37␤38␤39␤40␤41␤42␤43␤44␤45␤46␤47␤48␤49␤50…»	22:58
Mouq	Huh	22:58
TimToady	oh, wait, postdec	22:59
Mouq	m: sub term:<∞> {(++state$).say; ∞}()	22:59
camelia	rakudo-moar 1aeb7c: OUTPUT«(timeout)1␤1␤2␤3␤4␤5␤6␤7␤8␤9␤10␤11␤12␤13␤14␤15␤16␤17␤18␤19␤20␤21␤22␤23␤24␤25␤26␤27␤28␤29␤30␤31␤32␤33␤34␤35␤36␤37␤38␤39␤40␤41␤42␤43␤44␤45␤46␤47␤48␤49␤50␤51…»	22:59
TimToady	hmm, still strange	22:59
	skids notes "note()" is only documented as an aside under warn()	23:00
TimToady	is it slurping the ++ inside the once somehow?	23:00
	but then it wouldn't work the other times	23:00
	it's more like the top sub and the 2nd-level subs don't share the variabel	23:01
	*ble	23:01
	skids: is that a warning?	23:02
skids	Just that it's ard to look up in the synopsis. Because almost every section has "note that" somewhere :-)	23:03
TimToady	m: sub term:<∞> {(state$=0)++.say}()	23:03
camelia	rakudo-moar 1aeb7c: OUTPUT«0␤»	23:03
skids	m: sub term:<∞> {(++state$).say; ∞; ∞ }()	23:04
camelia	rakudo-moar 1aeb7c: OUTPUT«(timeout)1␤1␤2␤3␤4␤5␤6␤7␤8␤9␤10␤11␤12␤13␤14␤15␤16␤17␤18␤19␤20␤21␤22␤23␤24␤25␤26␤27␤28␤29␤30␤31␤32␤33␤34␤35␤36␤37␤38␤39␤40␤41␤42␤43␤44␤45␤46␤47␤48␤49␤50␤51…»	23:04
TimToady	m: sub term:<∞> {(state$=0)++.say}; ∞; ∞	23:04
camelia	rakudo-moar 1aeb7c: OUTPUT«0␤1␤»	23:04
tadzik	oh, look Czech PW	23:05
Mouq	m: sub term:<∞> {(++state$).say; ∞ xx ∞ }()	23:06
camelia	rakudo-moar 1aeb7c: OUTPUT«(timeout)1␤1␤2␤3␤4␤5␤6␤7␤8␤9␤10␤11␤12␤13␤14␤15␤16␤17␤18␤19␤20␤21␤22␤23␤24␤25␤26␤27␤28␤29␤30␤31␤32␤33␤34␤35␤36␤37␤38␤39␤40␤41␤42␤43␤44␤45␤46␤47␤48␤49␤50␤51…»	23:06
	→ Psyche^_ joined	23:06
Mouq	m: sub term:<∞>($w) {$w.say; (++state$).say; ∞('a') xx ∞('b') }()	23:07
camelia	rakudo-moar 1aeb7c: OUTPUT«===SORRY!===␤Cannot find method 'orig'␤»	23:07
TimToady	m: sub term:<∞> {(++state$x); say $x; ∞ }()	23:07
Mouq	m: sub term:<∞>($w) {$w.say; (++state$).say; ∞ 'a' xx ∞ 'b' }() #?	23:07
camelia	rakudo-moar 1aeb7c: OUTPUT«(timeout)1␤1␤2␤3␤4␤5␤6␤7␤8␤9␤10␤11␤12␤13␤14␤15␤16␤17␤18␤19␤20␤21␤22␤23␤24␤25␤26␤27␤28␤29␤30␤31␤32␤33␤34␤35␤36␤37␤38␤39␤40␤41␤42␤43␤44␤45␤46␤47␤48␤49␤50␤51…»	23:07
	rakudo-moar 1aeb7c: OUTPUT«[31m===[0mSORRY![31m===[0m Error while compiling /tmp/bSd4XVkJik␤Two terms in a row␤at /tmp/bSd4XVkJik:1␤------> [32mterm:<∞>($w) {$w.say; (++state$).say; ∞ [33m⏏[31m'a' xx ∞ 'b' }() #?[0m␤ expecting any of:␤ postfix␤ …»	23:07
skids	m: sub term:<∞> {(++state$).say; ∞ Z ∞ }()	23:08
camelia	rakudo-moar 1aeb7c: OUTPUT«(timeout)1␤1␤2␤3␤4␤5␤6␤7␤8␤9␤10␤11␤12␤13␤14␤15␤16␤17␤18␤19␤20␤21␤22␤23␤24␤25␤26␤27␤28␤29␤30␤31␤32␤33␤34␤35␤36␤37␤38␤39␤40␤41␤42␤43␤44␤45␤46␤47␤48␤49␤50␤51…»	23:08
aje_	wtf	23:08
Mouq	Oh, I like ∞ Z ∞ :)	23:08
TimToady	well, it's just two items	23:09
Mouq	aje_: We're paying homage to Cantor :)	23:09
aje_	Cantor?	23:09
Mouq	https://en.wikipedia.org/wiki/Georg_Cantor	23:10
aje_	aaah :)	23:10
	→ rurban joined	23:10
	← Psyche^ left	23:10
TimToady	I guess with respect to א₀ the extra first value doesn't hurt	23:11
Mouq	TimToady: ((∞ Z ∞) Z (∞ Z ∞)) Z ((∞ Z ∞) Z (∞ Z ∞)) …	23:11
TimToady	we need a D metaop for diagonalization	23:12
	Mouq &	23:13
lue	TimToady: isn't there a non-RTL version of aleph for math? or am I severely misremembering?	23:15
	ℵ₀ :) [you could also probably use an LTR override for normal aleph]	23:17
	Mouq returns to note that there's also ∞ X ∞ if you want to get all Xponential about it	23:18
Mouq	:P	23:18
TimToady	Cantor's multidimensional dust	23:19
	unfortunately I think that still leaves you in ℵ₀	23:20
lue	r: my $a = 5; $a = +& 6; say $a;	23:20
camelia	rakudo-parrot 1aeb7c, rakudo-jvm 1aeb7c, rakudo-moar 1aeb7c: OUTPUT«[31m===[0mSORRY![31m===[0m Error while compiling /tmp/tmpfile␤Unsupported use of & variable␤at /tmp/tmpfile:1␤------> [32mmy $a = 5; $a = +&[33m⏏[31m 6; say $a;[0m␤ expecting any of:␤ …»	23:20
lue	r: my $a = 5; $a = +\| 6; say $a;	23:20
camelia	rakudo-parrot 1aeb7c, rakudo-jvm 1aeb7c, rakudo-moar 1aeb7c: OUTPUT«[31m===[0mSORRY![31m===[0m Error while compiling /tmp/tmpfile␤Variable '&prefix:<\|>' is not declared␤at /tmp/tmpfile:1␤------> [32mmy $a = 5; $a = +[33m⏏[31m\| 6; say $a;[0m␤»	23:20
lue	r: my $a = 5; $a = +^ 6; say $a;	23:20
camelia	rakudo-parrot 1aeb7c, rakudo-jvm 1aeb7c, rakudo-moar 1aeb7c: OUTPUT«-7␤»	23:20
lue	This is what happens when you forget to type out the other $a; weird bugs :)	23:21
BenGoldberg	r: say 5 +& 6, 5 +\| 6, 5 +^ 6;	23:22
camelia	rakudo-parrot 1aeb7c, rakudo-jvm 1aeb7c, rakudo-moar 1aeb7c: OUTPUT«473␤»	23:22
	TimToady is thinking, 5 is prime, 6 is 2 3, how come they don't & to 0?*	23:24
TimToady	I guess you can have a base prime where each digit is 0..*	23:25
	and you multiply all the digits instead of adding them :)	23:25
	except then 0 and 1 don't exist, go figure	23:26
	← rurban left	23:31
	← denis_boyun_ left	23:33
	← pmurias left	23:45
	← nebuchadnezzar left	23:45
	→ nebuchadnezzar joined	23:46
lee__	dir() is supposed to take 2 args right?	23:49
	https://gist.github.com/leedo/216f4be4ee626d2c84cb	23:49
timotimo	a starting directory and a test method, right?	23:52
	you have to give it as a named i believe	23:52
	r: say &dir.signature	23:52
camelia	rakudo-parrot 1aeb7c, rakudo-jvm 1aeb7c, rakudo-moar 1aeb7c: OUTPUT«:(Cool $path = { ... }, :test($test) = { ... })␤»	23:52
timotimo	there	23:52
lee__	ah-ha! thanks	23:53
	guess http://doc.perl6.org/routine/dir could use some updating	23:54
timotimo	er	23:54
	are you sure?	23:54
	it seems correct to me	23:54
lee__	oh, you're right. my mind is stuck in p5 land	23:55
timotimo	i wonder if using a regex instead of the "none" junction would be faster	23:55
	← pecastro left	23:55
timotimo	i think it may try to create the junction anew for each dir call	23:55
	but i didn't verify that	23:55
TimToady	surely none could do constant folding if given constants	23:57
timotimo	aye.	23:57
	though it'll be thunked anyway and probably cloned for each invocation	23:57
TimToady	surely we avoid cloning things that are constant clones :)	23:58
timotimo	who knows :)	23:58
tadzik	timotimo: do you know the story why directories starting with '.' are "hidden"? :)	23:58
timotimo	i think it was because file names were compared to some ascii value or something?	23:59
	and . fell outside the range?	23:59
tadzik	nah	23:59
TimToady	is this related to why unicorns didn't get onto the ark?	23:59
tadzik	in all the unix tools, they were checking if the file is not '.' nor '..'	23:59
timotimo	ah	23:59
tadzik	so to save time, they just skipped whatever started with .	23:59
timotimo	but not properly :D	23:59

Logs Search ←Prev date Next date→

Channels Documentation