»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'perl6: say 3;' or rakudo:, niecza:, std:, or /msg camelia perl6: ... | irclog: irc.perl6.org | UTF-8 is our friend!
Set by sorear on 25 June 2013.
Mouq Basically the parser pretends its a Perl 6 grammar, taking the .compile_time_value of *everything* 00:01
I mean...
No, I'm overthinking this 00:03
[Coke] colomon: any word from sorear lately 00:04
?
Mouq Anyway, lue, what exactly would M<> do and how would it fix your problem? 00:05
lue M<backlink:file:line> would do all the work of htmlifying for me, of course :) 00:06
lue suspects he would be the one to write this module, by the way.
Mouq Hm.
So.. According to spec, that would generate a Pod6::FormattingCode::Name::backlink.new(typename<backlink>, contents<file:line>) 00:08
lue ::Named:: , but yes, that's what it would be.
Mouq And then....? 00:09
lue github.com/perl6/Pod-To-HTML/blob/...ML.pm#L385 <--- I wonder if this would handle ::Named::backlink already
content processing would happen within the class, presumably? 00:10
Mouq That's what I don't get
Does the class have to custom handle each case of $?DOC ?
lue I think that function I linked to needs to explicitly support the M code, but that's it for ::HTML. 00:11
Mouq Is the class called by Pod::To::* ?
lue $?DOC ?
Mouq perlcabal.org/syn/S26.html#How+Pod+...+processed
so perl6 --doc=HTML would have a $?DOC eq 'HTML'
lue Mouq: M<> (and custom-named blocks) feel to me like a situation where you only use them in specific situations. 00:12
so yeah, the class itself would be checking $?DOC 00:13
00:14 perigrin_ left
lue r: =begin pod␤M<ok:hello>␤=end pod␤say $=pod; 00:15
camelia rakudo-parrot 1b5b32, rakudo-jvm 1b5b32, rakudo-moar 1b5b32: OUTPUT«Pod::Block::Named("name" => "pod").hash␤ Pod::Block::Para␤␤ Pod::FormattingCode("type" => "M").hash␤ ok:hello␤␤␤»
00:15 perigrin joined
lue looks like we need some support for M<> in rakudo 00:15
Mouq mmhm
00:16 pippo joined
Mouq What if we did this: 00:16
lue suspects DOC use might be easier to implement than DOC <phaser>, but is only guessing 00:18
00:18 pippo left
Mouq =config M<> { my ($file,$line) = .split(':'); qq[<a href='…] } 00:18
lue would that work in current rakudo? 00:19
Mouq Because there's this whole chain of APIs that we'd need to construct this feature
lue: Not right now :P
I'm guessing it would parse-fail
lue I should note that I'm just fine doing it the way the P5 generator does it, no need to make M<> and related work RIGHT NOW or anything :) 00:20
Mouq lue: What it would require is that instead of building the Perl 6 Pod::* objects in NQP, we build ASTs that build it for us 00:21
Which would simplify everything dramatically, I think
lue ASTs are likely the way to go when you go beyond simple document markup :) 00:22
Mouq But it requires some serious thought 00:24
lue My two remaining goals for syngen are to place backlinks on a more granular level (i.e. interpret fragments and use them), and dramatically speeding up the HTML generation process 00:25
Mouq lue: How well would my cheat work for you?
r: say $=pod.perl;␤␤=pod abcd 00:26
camelia rakudo-parrot 1b5b32, rakudo-jvm 1b5b32, rakudo-moar 1b5b32: OUTPUT«Array.new(Pod::Block::Named.new(name => "pod", config => ().hash, content => Array.new(Pod::Block::Para.new(config => ().hash, content => Array.new("abcd")))))␤»
Mouq ^^ should that work?
I mean, conceptually
That $=pod gets the Pod after it, because the documents Pod is constructed compile-time 00:27
lue I don't know; it's currently a <div class="smartlink"> surrounding a <p class="smartlink-file">
(the div should contain multiple p elements, though I haven't bothered getting that working yet, if ever) 00:28
a Pod6 way of generating the backlink list might make the multiple <p> per <div> easier though :D 00:29
=for backlinks\n=backlink L<>\n=backlink ...
Mouq Hm 00:30
"It'd be really helpful to have a module providing a generic ay to walk a Pod tree and invoke callbacks on each node"
lue (then again, detecting when that multiple backlinks per paragraph thing happens should be equally complex with Pod6 constructs or PUA unicode chars...) 00:31
00:32 dayangkun joined, dmol left
Mouq lue: With my cheat it's just =begin pod :class<smartlink>␤=pod :class<smartlink-file> L<linkylink>␤=pod :class<smartlink-file>␤=end pod 00:32
and then the css adds display: block; to .smartlink 00:33
lue yeah, that's essentially what I wrote, except with less syntax sugar :)
lue is more concerned with the Pod->HTML chokepoint at the moment :/
Mouq lue: Well, true, but mine's more generalizable :P and I think that the Pod conversion should be easier to extend 00:34
lue yeah, :class<> would be way more useful, at least in ::HTML land
Mouq So that you could do `perl6 --doc=HTML,CustomHTML file.pod` 00:35
And the results of Pod::To::HTML and Pod::To::CustomHTML could be merged. It would just require setting up some kind of API 00:36
like I said
:P
lue is trying to see how github.com/perl6/doc/blob/master/htmlify.pl does inline HTML conversion
00:39 pdcawley_ joined, pdcawley left
lue yech, htmlify does too much to the $pod it constructs for me to see clearly what's going on. 00:42
Mouq Huh. Pod::To::HTML can display images written '=Image i.imgur.com/YrZrpZj.jpg'
lue yeah, just noticed that :) . 00:43
That means if/when I get around to writing a sub parameters tutorial thing, I can use a pretty picture at one part instead of ASCII art. 00:44
00:46 stevan__ joined 00:48 lee_ joined 00:49 pdcawley joined
lue Come to think of it, when I last ran htmlify it wasn't much faster at making HTML docs that my way... so except for those using my thing with perl6-j, it appears to be more a rakudo/Pod::To::HTML issue. 00:49
00:49 pdcawley_ left, dayangkun left, stevan_ left, mavcunha left, ruoso left, Mouq left, [Sno] left, yihshih left, lee left, orevdiabl left, go|dfish left 00:50 dayangkun joined, ruoso joined, Mouq joined, [Sno] joined, yihshih joined, orevdiabl joined, go|dfish joined 00:52 mavcunha joined
Mouq lue: Would this work for you? gist.github.com/Mouq/9179799 00:54
colomon [Coke]: I haven't heard anything from sorear++ recently.
Mouq Aw, lue got netsplit
lue: Maybe? 00:55
lue Mouq: no, *you* got netsplit! :) 00:56
Mouq lue: Fair enough :)
lue looks like it'd work, though I like being able to specify I want a div tag or a p tag .oO(=for Div ?) 00:57
<{$node.config<tag>} class="{$node.config<class>}"> :P 00:58
Mouq lue: I don't really feel like that's in the spirit of Pod..
lue I feel like being able to at least choose div or span, or maybe using div instead of span (this ain't an inline block element in Pod6, after all) would be better that just span 01:00
s/that just/than just/
Mouq lue: Span is as un-opinionated as it gets, and you can always use CSS to make it act like a div 01:01
I'm trying to keep as much of this out of the actual Pod as possible
01:01 _thou left
lue Seems like this could require more than just us two discussing :) . For now I'm fine with doing it the way that I have it, though eventually getting rid of an entire stage of the process would be nice. 01:02
lue is, again, more concerned with the startling lack of speed in converting Pod to HTML, as that's now the longest stage in the process by far. 01:03
Mouq lue: Yeah… 01:05
lue File::Find::find is another timehog in the process.
Mouq How does one profile Perl6 code? 01:08
01:08 pmurias left
lue with --profile, 2>filename, and kcachegrind 01:09
Mouq raydiak++ doc links fixed :D 01:10
raydiak colomon: I just PR'd you a partial fix for a URI::Escape unicode issue
01:11 rurban joined
raydiak unbreak all the things! 01:11
colomon raydiak++
raydiak: should I trust you?
raydiak colomon: that's a very abstract question
Mouq "No profiling support"
raydiak people you can't trust always say "absolutely" to that question :) 01:12
colomon raydiak: feeling lazy enough tonight to do your pull request without testing. :)
lue Mouq: it works on perl6-p at least, don't know about -j and -m
raydiak colomon: oh, yeah I'm pretty confident in this one
lue (IIRC -m fails, don't know -j)
colomon raydiak: done
raydiak: it looked solid, though I have no real clue what's going on at that line. :) 01:13
raydiak colomon: it says "if this char is lower ascii", but assumes ordinals > 255 don't exist
colomon right, figured that was what was wrong with the old version 01:14
lue afk
raydiak specifically, it checks the highest bit of an 8-bit byte
well, 1-byte char...you know what I mean :) 01:15
colomon raydiak: thanks for pinging me, I see so many pull requests for p6 things I usually just ignore them.
raydiak Mouq, moritz, anyone else who syncs the html docs to feather: make sure to get a fresh URI::Encode and Pod::To::HTML before htmlifying and syncing again 01:16
colomon actually, as I read it, it was just testing the byte #7, right? So it would have gotten like half of the higher valued Unicode code points.
Mouq raydiak: will do
colomon bit #7, not byte, sigh.
01:17 rurban_ joined
Mouq raydiak: Also, do you have all the css etc. you have in doc pushed to github? The font seemed larger after you sync'd 01:17
raydiak colomon: yes, it still would have evaluated correctly for many unicode chars, we were just lucky enough to find a failing one :) 01:18
Mouq: everything looks to be committed and pushed and so forth 01:19
Mouq raydiak: ok
01:24 klapperl_ left
raydiak Mouq: possibly links I changed going from purple to blue have tricked your eyes? 01:25
01:26 klapperl joined, hoverboard left
raydiak I just went and looked and said "well, it does kinda look bigger in places", but then realized it was just the darker color of unvisited links adding visual weight 01:26
01:29 hoverboard joined
Mouq raydiak: Maybe 01:30
raydiak Mouq: if you notice anything definite let me know; I didn't (intentionally) do any design work recently 01:32
oh, and been meaning to say thanks Mouq++ for all the awesome doc work lately 01:33
Mouq np :) 01:34
Thank you for your work making it awesomer 01:35
raydiak np :) it's exciting to do things that people actually see 01:36
Mouq I have trouble believing that htmlify used to run quite this slow I just didn't notice 01:37
raydiak you aren't running it under moar, are you? 01:38
Mouq I am
raydiak idk how it compares w/parrot, but it's at least 1 or 2 orders of magnitude slower than doing it w/jvm 01:39
Mouq Huh... I was switching between them last week and thought it was comparable.
But I didn't pay as much attention as I should have/timed it 01:40
raydiak the last time I accidentally started a timed run under moar, I lost patience and ^C'd after 40 minutes 01:41
Mouq :|
raydiak on the same machine jvm does it in 8
Mouq Do you have a recent MoarVM?
raydiak yeah, was from nom within the past week 01:42
oh, and the vm itself was rebuilt at the same time 01:43
Mouq r: say 'infix:<(|)>' ~~ /<!after \\ > '|'/ 01:44
camelia rakudo-parrot 1b5b32, rakudo-jvm 1b5b32, rakudo-moar 1b5b32: OUTPUT«｢|｣␤␤»
Mouq r: say 'infix:<(\|)>' ~~ /<!after \\ > '|'/
camelia rakudo-parrot 1b5b32, rakudo-jvm 1b5b32, rakudo-moar 1b5b32: OUTPUT«Nil␤»
Mouq r: say 'infix:<(\|)>|routine/infix:<(|)>.html' ~~ /<!after \\ > '|'/ 01:45
camelia rakudo-parrot 1b5b32, rakudo-jvm 1b5b32, rakudo-moar 1b5b32: OUTPUT«｢|｣␤␤»
Mouq r: say ('infix:<(\|)>|routine/infix:<(|)>.html' ~~ /<!after \\ > '|'/).prematch
camelia rakudo-parrot 1b5b32, rakudo-jvm 1b5b32, rakudo-moar 1b5b32: OUTPUT«infix:<(\|)>␤»
TimToady <!after \\ > is two-pass parsing 01:49
what if there's a \\ before the \\ 01:50
?
Mouq TimToady: Yeah... thank you 01:56
Mouq points out that perl6-m --target=ast doesn't work and hasn't for a while 02:02
or maybe ever
TimToady: Do you think my $level = 15;␤=begin item :numbered{$level}␤… should work? 02:08
*:numbered($level)
TimToady: Just at first glance with what you think Perl6 should be capable of 02:09
Well. Hm 02:13
Maybe BEGIN my $level = 15; maybe makes more sense 02:14
lue Mouq: perlcabal.org/syn/S26.html#Delimited+blocks 02:16
"All option keys and values must, of course, be constants since Pod is a specification language, not a programming language. Specifically, option values cannot be closures. See Synopsis 2 for details of the various Perl 6 pair notations."
Mouq Aww
Ok
That's probably for the best 02:17
lue makes things easier for rakudo at least :)
02:19 klapperl_ joined 02:22 klapperl left
Mouq r: my @*ABCD; 02:23
camelia ( no output )
Mouq nqp: my @*ABCD;
camelia ( no output )
02:26 fridim_ left
lue I'd like to find out what's causing the slowdown in Pod->HTML, but I can't get anything out of kcachegrind :( 02:29
Mouq lue: How fast is syngen with perl6-j? 02:31
lue Haven't checked recently, all I know is perl6-m is ≈1200 seconds. I can try to run it on perl6-j though.
Mouq complied perl6-p to be able to use its profiler; has no idea how to read it 02:32
lue you have to use something like kcachegrind 02:33
Mouq lue: My OSX is too new for Valgrind TT_TT 02:34
lue kcachegrind doesn't require valgrind, I don't think...
Mouq lue: I did push :class<> -> <span> to Pod::To::HTML btw 02:39
lue ok
Mouq dalek: hi 02:40
lue wishes he were better at interpreting what kcachegrind was telling him 02:43
For some reason I'm join() comes up as the offending function. I'm not sure if I should trust that finding though. 02:50
Mouq lue: Join is gonna be called all the time, that seems accurate 02:55
lue or perhaps it's gimme... I seriously don't know how to read all this :/ I'll leave it alone for now, perhaps someone more knowledgeable can help out later. 02:56
03:01 skids joined 03:13 pdcawley left, pdcawley joined
Mouq nqp: my @a = ['a','b','c']; for @a.kv -> $idx, $letter { say($idx, $letter) } 03:16
camelia nqp-jvm: OUTPUT«Assignment ("=") not supported in NQP, use ":=" instead at line 2, near " ['a','b',"␤ in panic (gen/jvm/stage2/NQPHLL.nqp:378)␤ in infix:sym<=> (gen/jvm/stage2/NQP.nqp)␤ in !protoregex (gen/jvm/stage2/QRegex.nqp:784)␤ in infix (gen/jvm/stage2/NQPHLL.nqp)…»
..nqp-moarvm: OUTPUT«Assignment ("=") not supported in NQP, use ":=" instead at line 2, near " ['a','b',"␤ at gen/moar/stage2/NQPHLL.nqp:369 (/home/p6eval/rakudo-inst-1/languages/nqp/lib/NQPHLL.moarvm:panic:120)␤ from <unknown>:1 (/home/p6eval/rakudo-inst-1/languages/nqp/lib/n…»
..nqp-parrot: OUTPUT«Assignment ("=") not supported in NQP, use ":=" instead at line 2, near " ['a','b',"␤current instr.: 'panic' pc 15934 (gen/parrot/stage2/NQPHLL.pir:5941) (gen/parrot/stage2/NQPHLL.nqp:425)␤»
Mouq nqp: my @a := ['a','b','c']; for @a.kv -> $idx, $letter { say($idx, $letter) }
camelia nqp-parrot: OUTPUT«Method 'kv' not found for invocant of class 'QRPA'␤current instr.: '' pc 81 ((file unknown):52) (/tmp/tmpfile:1)␤»
..nqp-moarvm: OUTPUT«Cannot find method 'kv'␤ at /tmp/tmpfile:1 (<ephemeral file>::40)␤ from gen/moar/stage2/NQPHLL.nqp:1090 (/home/p6eval/rakudo-inst-1/languages/nqp/lib/NQPHLL.moarvm::94)␤ from gen/moar/stage2/NQPHLL.nqp:1076 (/home/p6eval/rakudo-inst-1/languages/nqp/lib/…»
..nqp-jvm: OUTPUT«Method 'kv' not found for invocant of class 'NQPArray'␤ in (/tmp/tmpfile:1)␤ in (gen/jvm/stage2/NQPHLL.nqp:1099)␤ in eval (gen/jvm/stage2/NQPHLL.nqp:1085)␤ in evalfiles (gen/jvm/stage2/NQPHLL.nqp:1291)␤ in command_eval (gen/jvm/stage2/NQPHLL.nqp:1195…»
03:30 rurban left
TimToady turns out that @a.kv has to calculate keys first, which calls self.end, which forces finite semantics 03:30
if we invert all the .kv-ish methods to map the values rather than the keys, they can become properly lazy, and maybe run faster too 03:31
Mouq r: my @a = <a b c d e>; @a X, (my$)++ 03:35
camelia ( no output )
Mouq r: my @a = <a b c d e>; say @a X, (my$)++
camelia rakudo-parrot 1b5b32, rakudo-jvm 1b5b32, rakudo-moar 1b5b32: OUTPUT«a 0 b 0 c 0 d 0 e 0␤»
lue Mouq: what are you working on, ooc?
Mouq Oh, duh, X isn't blocky
lue: I finally convinced myself to work on getting tables to work right 03:36
lue Mouq: by the way, perl6-j on syngen takes 3601.609 seconds.
Mouq r: my @a = <a b c d e>; say gather for @a { take (state$)++; take $_} 03:37
camelia rakudo-parrot 1b5b32, rakudo-jvm 1b5b32, rakudo-moar 1b5b32: OUTPUT«0 a 1 b 2 c 3 d 4 e␤»
Mouq lue: Aww. And Moar++ :)
lue I think if I were using eval-server or whatever it might run faster... 03:38
Mouq lue: Are you still "$*EXECUTABLE …"? 03:39
lue yes, it runs whatever you ran the main script with.
Mouq I really don't know how to get around that... I imagine in that case the eval-server might help a lot 03:40
03:40 rurban joined
lue If Pod::To::HTML came with a way to parse pod files in-script instead of through shell(), then it would definitely be faster for JVM. 03:41
Mouq lue: Pod::To::HTML::pod2html? 03:42
lue don't you need to build up your own object (as I understand from doc/htmlify.pl)?
Mouq You have to convice perl6 to parse whatever you're looking at as Pod and put it in $=pod (which is a compile-time variable...) 03:44
lue I'd much prefer if there was a HTML::podfile or something
Mouq s/compile-time variable/variable that has to be known at compile-time/ 03:45
lue seems like the kind of thing that's not unreasonable for the module to do instead of me. :)
03:47 klapperl_ left
Mouq lue: I don't think it would be too difficult to take the Pod stuff from Rakudo and turn it into a Perl 6 grammar+actions, if that's what you mean 03:48
03:49 klapperl joined
Mouq That could actually be pretty useful 03:49
But none of the Pod things except <colonpair> reach back out into the rest of the Grammar AFAIK
lue :/ I'd love to avoid code duplication like that though. It'd preferable to do something like STD.parsefile($file, :rule<pod_doc>), though I think we're quite a ways away from that. 03:50
03:56 molaf_ left
Mouq r: DOC use ABCD 04:00
camelia ( no output )
04:03 kurahaupo_mobile joined 04:06 hoverboard left
dalek kudo/nom: e8cea1d | larry++ | src/core/List.pm:
List.end Considered Harmful

All the .pairs-like list methods calculated .keys first, then looked up everything using at_pos. Unfortunately, to determine the key list, the .keys method called self.end, which enforces finite semantics. This seems a bit suboptimal in a language that professes to support infinite lists. So now instead of mapping the keys, we map the values, and create the keys as needed. Performance seems roughly the same.		 04:08
04:09 hoverboard joined
TimToady actually, the main difference is the performance on lists that are big but not known to be infinite 04:10
lue that difference is for the better, right? :)
TimToady for ('a'..'zzzzzz').pairs -> $p {...} no longer hangs forever precalculating the list
that's the big win
04:11 kaleem joined
Mouq TimToady++ awsummm 04:11
TimToady should hunt down the rest of the .end calls to see if any of them make sense
r: .say for ('a'..'zzzzzz').pairs 04:13
camelia rakudo-parrot 1b5b32, rakudo-jvm 1b5b32, rakudo-moar 1b5b32: OUTPUT«(timeout)» 04:14
TimToady in an hour, that should return the initial values
04:15 kurahaupo_mobile left
Mouq r: say ('a'..'zzzzzz')[40] 04:17
camelia rakudo-parrot 1b5b32, rakudo-jvm 1b5b32, rakudo-moar 1b5b32: OUTPUT«ao␤»
Mouq r: say ('a'..'zzzzzz')[40]:delete 04:18
camelia rakudo-parrot 1b5b32: OUTPUT«Can not remove elements from a Range␤ in method gist at gen/parrot/CORE.setting:12260␤ in method gist at gen/parrot/CORE.setting:1042␤ in sub say at gen/parrot/CORE.setting:13157␤ in block at /tmp/tmpfile:1␤␤»
..rakudo-moar 1b5b32: OUTPUT«===SORRY!===␤Can not remove elements from a Range␤»
..rakudo-jvm 1b5b32: OUTPUT«Can not remove elements from a Range␤ in method gist at gen/jvm/CORE.setting:12239␤ in method gist at gen/jvm/CORE.setting:1039␤ in sub say at gen/jvm/CORE.setting:13118␤ in block at /tmp/tmpfile:1␤␤»
Mouq r: say Array.new('a'..'zzzzzz')[40]
camelia rakudo-parrot 1b5b32, rakudo-jvm 1b5b32, rakudo-moar 1b5b32: OUTPUT«ao␤»
Mouq r: say Array.new('a'..'zzzzzz')[40]:delete
camelia rakudo-parrot 1b5b32, rakudo-jvm 1b5b32, rakudo-moar 1b5b32: OUTPUT«(timeout)» 04:19
04:25 kurahaupo_mobile joined 04:35 rurban left 04:37 kaleem left
Mouq r: .say for ('a'..'zzzzzz').reduce({$^a,$^b}) 04:41
camelia rakudo-parrot 1b5b32, rakudo-jvm 1b5b32, rakudo-moar 1b5b32: OUTPUT«(timeout)»
04:48 Ben_Goldberg is now known as BenGoldberg
BenGoldberg rn: for (gather { .take for 1..10; say 'end' }) { .say; last }; say 'after' 04:49
camelia rakudo-jvm 1b5b32: OUTPUT«(timeout)» 04:50
..rakudo-parrot 1b5b32, rakudo-moar 1b5b32, niecza v24-109-g48a8de3: OUTPUT«1␤after␤»
BenGoldberg \o/
TimToady bows 04:51
BenGoldberg TimToady++
04:52 rurban joined
BenGoldberg j: say 'alive' 04:52
camelia rakudo-jvm 1b5b32: OUTPUT«alive␤»
BenGoldberg wonders why it timed out on jakudo...
Mouq rn: for (gather { .take for 1..10; say 'end' }) { .say; last }; say 'after' 04:54
camelia rakudo-parrot 1b5b32, rakudo-jvm 1b5b32, rakudo-moar 1b5b32, niecza v24-109-g48a8de3: OUTPUT«1␤after␤»
04:57 hoverboard left, hoverboard joined
BenGoldberg rn: print do { ((state $ = 'a') ^= 'b') ?? 'foo' !! 'bar' } 04:57
camelia rakudo-parrot 1b5b32, rakudo-jvm 1b5b32, rakudo-moar 1b5b32: OUTPUT«bar»
..niecza v24-109-g48a8de3: OUTPUT«Stack overflow: IP: 0x50322a, fault addr: (nil)␤Stacktrace:␤ at Niecza.Kernel.UnboxAny<T> (Niecza.P6any) <0x0002b>␤ <...>␤ at Niecza.CtxJunctionBool.Get (Niecza.Variable) <0x00117>␤ at Niecza.CtxJunctionBool.Get (Niecza.Variable) <0x0025…»
Mouq still doesn't understand what you're trying to accomplish with ^= 04:58
04:59 kurahaupo_mobile left
lue it's not updated yet though, is it? 04:59
r: .say for ('a'..'zzzzzz').pairs
Mouq lue: nope
camelia rakudo-parrot 1b5b32, rakudo-jvm 1b5b32, rakudo-moar 1b5b32: OUTPUT«(timeout)» 05:00
BenGoldberg was just golfing down a bug to the shortest line that exhibits it.
05:00 araujo left
BenGoldberg n: print do { ((state $ = 'a') |= 'b') ?? 'foo' !! 'bar' } 05:00
camelia niecza v24-109-g48a8de3: OUTPUT«Stack overflow: IP: 0x56fd20, fault addr: (nil)␤Stacktrace:␤ at Niecza.Kernel.UnboxAny<T> (Niecza.P6any) <0x0002b>␤ <...>␤ at Niecza.CtxJunctionBool.Get (Niecza.Variable) <0x00117>␤ at Niecza.CtxJunctionBool.Get (Niecza.Variable) <0x002b…»
Mouq n: (state $ = False) | True 05:02
camelia ( no output )
Mouq n: say so (state $ = False) | True
camelia niecza v24-109-g48a8de3: OUTPUT«True␤»
Mouq n: (state $ = False) |= True
camelia ( no output )
Mouq n: say so ((state $ = False) |= True)
camelia niecza v24-109-g48a8de3: OUTPUT«Stack overflow: IP: 0x40797a86, fault addr: (nil)␤Stacktrace:␤ at Niecza.Kernel.UnboxAny<T> (Niecza.P6any) <0x0002b>␤ <...>␤ at Niecza.CtxJunctionBool.Get (Niecza.Variable) <0x00117>␤ at Niecza.CtxJunctionBool.Get (Niecza.Variable) <0x00…»
05:09 sftp_ joined 05:10 skids left 05:11 sftp left 05:18 hoverboard left 05:22 hoverboard joined
Mouq nqp: my@t:=[[1,2,3],[4,5,6],[7,8,9]]; say([].push(@t)[0][1]); 05:24
camelia nqp-moarvm, nqp-jvm: OUTPUT«2␤»
..nqp-parrot: OUTPUT«3␤»
Mouq nqp: my@t:=[[1,2,3],[4,5,6],[7,8,9]]; my @x; @x.push(@t); say(@x[0][1][1]) # much better 05:26
camelia nqp-moarvm, nqp-jvm, nqp-parrot: OUTPUT«5␤»
05:30 [Sno] left
TimToady r: .say for ('a'..'zzzzzz').pairs 05:42
camelia rakudo-parrot 1b5b32, rakudo-jvm 1b5b32, rakudo-moar 1b5b32: OUTPUT«(timeout)»
05:43 hoverboard left 05:50 kurahaupo_mobile joined 05:55 kurahaupo__ joined 05:56 kurahaupo_mobile left 06:00 kurahaupo__ left 06:05 kurahaupo_mobile joined
BenGoldberg r: ('a'..'zzzzzz').pairs.infinite.say 06:06
camelia rakudo-parrot e8cea1, rakudo-jvm e8cea1, rakudo-moar e8cea1: OUTPUT«Nil␤»
Mouq Updated :)
BenGoldberg r: for( ('a'..'zzzzzz').pairs ) { .say; last }
camelia rakudo-parrot e8cea1, rakudo-jvm e8cea1, rakudo-moar e8cea1: OUTPUT«===SORRY!===␤Word 'for' interpreted as 'for()' function call; please use whitespace instead of parens␤at /tmp/tmpfile:1␤------> for⏏( ('a'..'zzzzzz').pairs ) { .say; last }[…»
BenGoldberg r: for ( ('a'..'zzzzzz').pairs ) { .say; last }
camelia rakudo-parrot e8cea1, rakudo-jvm e8cea1, rakudo-moar e8cea1: OUTPUT«0 => "a"␤»
BenGoldberg r: for ( ('a'..'zzzzzz').pairs ) { .say; last if $_ ge 'aa' } 06:07
camelia rakudo-moar e8cea1: OUTPUT«(timeout)0 => "a"␤1 => "b"␤2 => "c"␤3 => "d"␤4 => "e"␤5 => "f"␤6 => "g"␤7 => "h"␤8 => "i"␤9 => "j"␤10 => "k"␤11 => "l"␤12 => "m"␤13 => "n"␤14 => "o"␤15 => "p"␤16 => "q"␤17 => "r"␤18 => "s"␤19 => "t"␤20 => "u"␤21 =>…»
..rakudo-jvm e8cea1: OUTPUT«(timeout)0 => "a"␤1 => "b"␤2 => "c"␤3 => "d"␤4 => "e"␤5 => "f"␤6 => "g"␤7 => "h"␤8 => "i"␤9 => "j"␤10 => "k"␤11 => "l"␤12 => "m"␤13 => "n"␤14 => "o"␤15 => "p"␤16 => "q"␤17 => "r"␤18 => "s"␤19 => "t"␤20 => "u"␤21 => …»
..rakudo-parrot e8cea1: OUTPUT«(timeout)0 => "a"␤1 => "b"␤2 => "c"␤3 => "d"␤4 => "e"␤5 => "f"␤6 => "g"␤7 => "h"␤8 => "i"␤9 => "j"␤10 => "k"␤11 => "l"␤12 => "m"␤13 => "n"␤14 => "o"␤15 => "p"␤16 => "q"␤17 => "r"␤18 => "s"␤19 => "t"␤20 => "u"␤21 …»
TimToady r: .say for ('a'..'zzzzzz').pairs 06:08
camelia rakudo-parrot e8cea1: OUTPUT«(timeout)0 => "a"␤1 => "b"␤2 => "c"␤3 => "d"␤4 => "e"␤5 => "f"␤6 => "g"␤7 => "h"␤8 => "i"␤9 => "j"␤10 => "k"␤11 => "l"␤12 => "m"␤13 => "n"␤14 => "o"␤15 => "p"␤16 => "q"␤17 => "r"␤18 => "s"␤19 => "t"␤20 => "u"␤21 …»
..rakudo-jvm e8cea1: OUTPUT«(timeout)0 => "a"␤1 => "b"␤2 => "c"␤3 => "d"␤4 => "e"␤5 => "f"␤6 => "g"␤7 => "h"␤8 => "i"␤9 => "j"␤10 => "k"␤11 => "l"␤12 => "m"␤13 => "n"␤14 => "o"␤15 => "p"␤16 => "q"␤17 => "r"␤18 => "s"␤19 => "t"␤20 => "u"␤21 => …»
..rakudo-moar e8cea1: OUTPUT«(timeout)0 => "a"␤1 => "b"␤2 => "c"␤3 => "d"␤4 => "e"␤5 => "f"␤6 => "g"␤7 => "h"␤8 => "i"␤9 => "j"␤10 => "k"␤11 => "l"␤12 => "m"␤13 => "n"␤14 => "o"␤15 => "p"␤16 => "q"␤17 => "r"␤18 => "s"␤19 => "t"␤20 => "u"␤21 =>…»
BenGoldberg r: for ( ('a'..'zzzzzz').pairs ) { .say; last if .value eq 'aa' } 06:11
TimToady r: for ('a'..'zzzzzz').pairs { .say; last if .value ge 'aa' }
BenGoldberg :)
camelia rakudo-parrot e8cea1, rakudo-jvm e8cea1, rakudo-moar e8cea1: OUTPUT«0 => "a"␤1 => "b"␤2 => "c"␤3 => "d"␤4 => "e"␤5 => "f"␤6 => "g"␤7 => "h"␤8 => "i"␤9 => "j"␤10 => "k"␤11 => "l"␤12 => "m"␤13 => "n"␤14 => "o"␤15 => "p"␤16 => "q"␤17 => "r"␤18 => "s"…»
rakudo-parrot e8cea1, rakudo-jvm e8cea1, rakudo-moar e8cea1: OUTPUT«0 => "a"␤1 => "b"␤»
TimToady 'b' ge 'aa' :)
and you don't need the supersitious parens :P 06:12
a method doesn't require (), unlike, say, lines() 06:13
06:14 SamuraiJack_ joined 06:21 darutoko joined 06:25 berekuk joined 06:29 BenGoldberg left 06:36 xenoterracide left 06:39 rurban left 06:41 mavcunha left 06:49 kurahaupo_mobile left 06:55 [Sno] joined
Mouq lue: "Blocks whose names contain at least one uppercase and one lowercase letter are assumed to be destined for specialized renderers or parser plug-ins." I hadn't seen this, reconsidering "=for Div" etc. 06:56
07:02 kurahaupo_mobile joined 07:03 dmol joined 07:10 rurban joined 07:15 rurban left, aborazmeh joined 07:21 FROGGS joined
FROGGS masak: do you think this should be specced to work? my regex foo { aaa }; say "aaa" ~~ foo # perhaps same for grammars 07:31
07:31 berekuk left 07:35 aborazmeh left
TimToady r: my regex foo { aaa }; say "aaa" ~~ &foo 07:36
camelia rakudo-jvm e8cea1: OUTPUT«(timeout)»
..rakudo-parrot e8cea1, rakudo-moar e8cea1: OUTPUT«｢aaa｣␤␤»
TimToady you need the noun form, not the verb form
otherwise it's going to try to call foo() first, then do the ~~ 07:38
[Coke] std: my $s = 'abcd'; while $s ~~ m:g/(.)/ { print $0 }; 07:41
camelia std 09dda5b: OUTPUT«ok 00:01 129m␤»
07:41 kurahaupo_mobile left
FROGGS TimToady++ 07:41
[Coke] ^^ RT #77144 is a request for that to warn.
synopsebot Link: rt.perl.org/rt3//Public/Bug/Displa...l?id=77144
FROGGS masak: that is what you want perhaps :o) ----- ^
(not the RT link though) 07:42
07:42 berekuk joined
TimToady [Coke]: detecting list/scalar context confusion is going to be hard in a language that handles context lazily 07:51
[Coke] Then I'll reject the ticket, hokay? 07:52
Here's another ticket that requires a ruling, and perhaps a rethink in the 3.5 years since it was opened: RT #77340 07:53
synopsebot Link: rt.perl.org/rt3//Public/Bug/Displa...l?id=77340
07:53 Quom joined 07:54 robinsmidsrod joined 07:57 sftp joined 07:58 yihshih__ joined
TimToady given current understanding of how things itemize, I think Pm has it right, and we should reject the ticket 07:58
except that @%h<abc> can't work
07:58 ruoso_ joined
TimToady must be either @(%h<abc>) or %h<abc>.list or %h<abc>[] 07:58
[Coke] TimToady++ 07:59
07:59 pdcawley_ joined
[Coke] r: say say $*PERL 08:01
camelia rakudo-moar e8cea1: OUTPUT«("name" => "rakudo", "compiler" => {"name" => "rakudo", "ver" => "2014.02-18-ge8cea1d", "release-number" => "", "build-date" => "2014-02-24T06:25:33Z", "codename" => ""}).hash␤True␤»
..rakudo-parrot e8cea1: OUTPUT«("name" => "rakudo", "compiler" => {"name" => "rakudo", "ver" => "2014.02-18-ge8cea1d", "release-number" => "", "build-date" => "2014-02-24T06:32:25Z", "codename" => ""}).hash␤True␤»
..rakudo-jvm e8cea1: OUTPUT«("name" => "rakudo", "compiler" => {"codename" => "", "name" => "rakudo", "release-number" => "", "ver" => "2014.02-18-ge8cea1d", "build-date" => "2014-02-24T05:12:38Z"}).hash␤True␤»
[Coke] r: say $*PERL
camelia rakudo-parrot e8cea1: OUTPUT«("name" => "rakudo", "compiler" => {"name" => "rakudo", "ver" => "2014.02-18-ge8cea1d", "release-number" => "", "build-date" => "2014-02-24T06:32:25Z", "codename" => ""}).hash␤»
..rakudo-moar e8cea1: OUTPUT«("name" => "rakudo", "compiler" => {"name" => "rakudo", "ver" => "2014.02-18-ge8cea1d", "release-number" => "", "build-date" => "2014-02-24T06:25:33Z", "codename" => ""}).hash␤»
..rakudo-jvm e8cea1: OUTPUT«("name" => "rakudo", "compiler" => {"codename" => "", "name" => "rakudo", "release-number" => "", "ver" => "2014.02-18-ge8cea1d", "build-date" => "2014-02-24T05:12:38Z"}).hash␤»
[Coke] urk. I expected that to fail, so resent. 08:02
r: say $*PERL<version>
camelia rakudo-parrot e8cea1, rakudo-jvm e8cea1, rakudo-moar e8cea1: OUTPUT«(Any)␤»
08:03 FROGGS left, sftp_ left, pdcawley left, dayangkun left, ruoso left, Mouq left, yihshih left, orevdiabl left, go|dfish left 08:04 revdiablo joined
[Coke] rakudo: my Int (Str $x); say $x.WHAT # unspecced behavior? 08:08
camelia rakudo-parrot e8cea1, rakudo-jvm e8cea1, rakudo-moar e8cea1: OUTPUT«(Int)␤»
[Coke] rakudo: sub foo() { return 1,2,:c<3> }; say foo().perl 08:09
camelia rakudo-parrot e8cea1, rakudo-jvm e8cea1, rakudo-moar e8cea1: OUTPUT«(1, 2, "c" => Mu)␤»
[Coke] rakudo: say [1,2] X~ <a b>
camelia rakudo-parrot e8cea1, rakudo-jvm e8cea1, rakudo-moar e8cea1: OUTPUT«1a 1b 2a 2b␤»
[Coke] rakudo: my $*a = 42; foo; sub foo { say $*a; my $*a } 08:11
camelia rakudo-parrot e8cea1, rakudo-jvm e8cea1, rakudo-moar e8cea1: OUTPUT«(Any)␤»
08:11 go|dfish joined 08:12 FROGGS joined
[Coke] nom: class A { eval 'has $.x' }; say A.new(x => 3).x 08:15
camelia rakudo-parrot e8cea1, rakudo-jvm e8cea1: OUTPUT«===SORRY!===␤No such method 'name' for invocant of type 'NQPMu'␤Saw 1 call to deprecated code during execution.␤================================================================================␤Sub eval (from GLOBAL) called at:…»
..rakudo-moar e8cea1: OUTPUT«===SORRY!===␤Cannot find method 'name'␤»
08:15 kaleem joined
[Coke] nom: class A { EVAL 'has $.x' }; say A.new(x => 3).x 08:15
camelia rakudo-parrot e8cea1, rakudo-jvm e8cea1: OUTPUT«===SORRY!===␤No such method 'name' for invocant of type 'NQPMu'␤»
..rakudo-moar e8cea1: OUTPUT«===SORRY!===␤Cannot find method 'name'␤»
[Coke] nom: class A { BEGIN eval 'has $.x' }; say A.new(x => 3).x 08:16
nom: class A { BEGIN EVAL 'has $.x' }; say A.new(x => 3).x
moritz
.oO( we still have a 'nom' target? :-)
camelia rakudo-jvm e8cea1: OUTPUT«(timeout)»
..rakudo-parrot e8cea1: OUTPUT«===SORRY!===␤No such method 'name' for invocant of type 'NQPMu'␤Saw 1 call to deprecated code during execution.␤================================================================================␤Sub eval (from GLOBAL) called at:␤ /tmp/tmpfile, l…»
..rakudo-moar e8cea1: OUTPUT«===SORRY!===␤Cannot find method 'name'␤»
rakudo-jvm e8cea1: OUTPUT«(timeout)»
..rakudo-parrot e8cea1: OUTPUT«===SORRY!===␤No such method 'name' for invocant of type 'NQPMu'␤»
..rakudo-moar e8cea1: OUTPUT«===SORRY!===␤Cannot find method 'name'␤»
[Coke] nom: has $.x
camelia rakudo-parrot e8cea1, rakudo-jvm e8cea1, rakudo-moar e8cea1: OUTPUT«===SORRY!=== Error while compiling /tmp/tmpfile␤You cannot declare attribute '$.x' here; maybe you'd like a class or a role?␤at /tmp/tmpfile:1␤------> has $.x⏏<EOL>␤ …» 08:17
[Coke] whoops, meant for all that to go to a private window. 08:18
FROGGS :o)
[Coke] ok, only 83 more tickets that haven't been touched in 2+ years. 08:19
up to 34 testneeded tickets - those are often LHF for people looking to contribute. Feel free to grab one and submit a test. (usually the code is in the ticket.) 08:20
FROGGS I have some left in my inbox that I want to do 08:25
08:27 berekuk left
dalek kudo/nom: 299d708 | larry++ | src/core/List.pm:
combinations didn't itemize when k == 0

The null set was omitted from the powerset.		 08:28
moritz m: say (^5).list.keys; say (^5).list.keys;
camelia rakudo-moar e8cea1: OUTPUT«0 1 2 3 4␤0 1 2 3 4␤»
dalek ast: 2582eed | moritz++ | S32-exceptions/misc.t:
Tests for RT #67554
 08:31
synopsebot Link: rt.perl.org/rt3//Public/Bug/Displa...l?id=67554
[Coke] moritz++ 08:35
08:41 rindolf joined 08:44 rurban joined 08:49 rurban left, sunnavy_ is now known as sunnavy 08:53 salv0 joined 09:06 denis_boyun joined, rindolf left 09:09 zakharyas joined 09:16 denis_boyun left 09:18 Alula left 09:23 Alula joined 09:31 rindolf joined 09:33 grondilu joined 09:35 fhelmberger joined
grondilu on rosettacode.org/wiki/Almost_prime#Perl_6, TimToady wrote about a factors routine: « (to be included manually until we decide where in the distribution to put it) ». Couldn't it just be a Int method? 09:35
09:40 mavcunha joined 09:50 mavcunha_ joined 09:52 mavcunha left 09:53 dakkar joined 09:55 Quom left 10:18 pecastro left 10:19 zakharyas left 10:28 pecastro joined 10:32 Alula left, Alula joined
nwc10 is this known to happen: 10:35
rn: say 1.is-prime
camelia rakudo-parrot 299d70, rakudo-jvm 299d70, rakudo-moar 299d70, niecza v24-109-g48a8de3: OUTPUT«False␤»
nwc10 oh, wait. so something screwed locally. hmmm
live demo fail! :-)
jnthn Never happens to me :P 10:37
Comes out False on my local too, fwiw. 10:38
nwc10 I might have inconsistent versions of stuff
jnthn My local is on moar-conc, which is bleading in different ways than nom :)
10:43 zakharyas joined
moritz nwc10: maybe your 1 was twice as large as the usual 1 :-) 10:44
10:46 kaleem left 10:47 kaleem joined
jnthn r: multi infix:<==>(2, 1) { True }; say 2 == 1 10:48
camelia rakudo-parrot 299d70, rakudo-jvm 299d70, rakudo-moar 299d70: OUTPUT«True␤»
moritz jnthn: you are EVAL, erm, evil :-)
jnthn True evil is sneaking that into a module and adding "is export" :P 10:54
moritz aren't multis supposed to be exported by default anyway? :-)
jnthn S11 may well argue for that :)
It's a little more magic than I like tbh. 10:55
moritz same here
jnthn As in, it's not unusual to use multis inside a module as an implementation detail.
nwc10 i
jnthn nwc10: Most complex message so far today... 10:56
nwc10 yes, and it was meant for mutt
moritz [Coke]++ just quoted S11 for a ticket 10:57
Any C<proto> declaration that is not declared C<my> is exported by default.
Any C<multi> that depends on an exported C<proto> is also automatically
exported. Any autogenerated C<proto> is assumed to be exported by default.
are protos in the setting exported automatically? 10:58
if not, adding a multi a setting proto from a module wouldn't export it
jnthn Well, that's debatable. It sitll has to generate a proto in the nested scope. 10:59
*still
That is, it copies the outer proto with its candidate list and then appends to it.
11:03 _sri left 11:05 breinbaas joined, _sri joined 11:09 grondilu_ joined 11:10 pecastro_ joined
timotimo lue: are you running the pod-to-html stuff on moarvm? 11:17
11:17 grondilu left, pdcawley_ left, bjz left, silug__ left, silug__ joined 11:18 pdcawley joined, robinsmidsrod left 11:20 bjz joined, robinsmidsrod joined 11:21 pecastro left
timotimo doesn't have the patience for a full backlog right now :| 11:25
11:26 donaldh joined, donaldh left 11:28 donaldh joined 11:37 lizmat joined
timotimo [Coke]++ # amazing amounts of bug triage 11:41
11:46 dmol left
masak good afternoon, #perl6 11:48
masak is teaching Python today
masak .oO( hssssSSSS! )
nwc10 heresy! :-)
masak strangely enough, it feels very little like heresy. 11:49
jnthn will be teaching lots of C# in March. Doesn't feel like heresy either. :) 11:50
Though I do keep missing things from Perl 6. :) 11:51
masak think of us as scouts doing forays into other communities :)
jnthn The .Net folks are oddest generally, in so far as they tend to do .Net stuff and nothing else...
masak I hear COBOL folks are much the same :P 11:52
jnthn They've too much to write to have time to do anything else :P
masak bet it pays well, too. 11:53
jnthn Yeah, well. The 2038 bug will pay our generation's retirement well too. :)
masak "don't you dare die! then who will maintain our legacy bank software?"
jnthn: all we need to do is to create a massive 2038 scare :)
challenge accepted!
jnthn masak: Easy enough. Bad news sells. :) 11:54
masak .oO( "Ten things you wouldn't believe about our eventual destruction in the year 2038" )
jnthn 32 things, surely?
masak :P 11:55
-- signed, int
or is that 'signed, long' ?
jnthn Platform specific, naturally... :)
masak :) 11:59
timotimo moritz: would like support for "/today/" and "/yesterday/" as well as those"/summary/" URLs in the irclogs 12:04
donaldh today works, no? 12:05
jnthn yeah, today already does. 12:06
timotimo not in combination with /summary/
donaldh ah :) 12:07
12:08 zakharyas left
moritz timotimo: github.com/moritz/ilbot/issues/33 same problem with /text/ 12:10
timotimo: maybe add it to the ticket? (or even pull-request? :-) 12:11
timotimo thank you 12:12
moritz (all URLs are handled in ilbot.psgi)
timotimo i'd like the CSRF error for the rakudo.org/tickets links to go away 12:13
moritz CSRF error? 12:15
donaldh oh, panda bootstrap fails on JVM on OS X. Is it known to work / not work on JVM on other platforms?
timotimo RT has detected a possible cross-site request forgery for this request, because the Referrer header supplied by your browser (rakudo.org:80) is not allowed by RT's configured hostname (rt.perl.org:443) or whitelisted hosts (www.bitcard.org:443, www.bestpractical.com:443). A malicious attacker may be trying to modify or access a search on your behalf. If you did not initiate this request, then you should alert
your security team.
moritz eeks
oh, and some of the search URLs produce other RT errors 12:16
timotimo p6weekly.wordpress.com/?p=68&sh...b399b58b78 - first draft of today's p6weekly post 12:23
nwc10 donaldh: not sure. hoelzro I think was working on it
moritz timotimo: not super exciting, but there are some new docs on doc.perl6.org 12:25
timotimo that's cool. which ones? 12:26
ah, i'll just look at the commit log
moritz IO::Handle, IO::FileTestable, Mix, MixHash, Mixy, QuantHash, Set, SetHash, Setty, regexes
Mouq++ did most of those (I only did regexes)
timotimo thank you!
moritz doc.perl6.org/language/regexes # ouch, something very wrong with table parsing or rendering
jnthn timotimo: "The most noticeable thing still missing is ThreadPool and its Scheduler." 12:28
timotimo: Actually ThreadPoolScheduler is working now, for non-time-based scheduling.
timotimo oh!
jnthn timotimo: So we're onto Promise itself.
timotimo excellent!
thanks for the correction
jnthn timotimo: Also, didn't you contribute something to this effort as well as me and JimmyZ? ;) 12:29
timotimo only tiny bits
timotimo doesn't like tooting his own horn too much :P
jnthn Still helpful bits :)
timotimo: In what's cooking, I expect I'll have promises working on Moar, and likely channels too, during the next week. 12:31
timotimo damnit!
jnthn timotimo: Probably supplies too. I think not the timer bits, though. :)
I know. Bloody progress. :P 12:32
timotimo i just lost most of the text >_<
changed from "Text" to "Visual" and hit ctrl-z
very much not recommended at all
nwc10 it's the old "this week" problem :-)
jnthn timotimo: I still have it in a tab here? 12:33
timotimo: Do you want the text or the HTML?
timotimo i still have that tab, too
jnthn ah, ok :)
timotimo i was able to copy the text + formatting into the wysiwyg editor
jnthn figures its lunch time... 12:34
timotimo bon appetit :)
p6weekly.wordpress.com/?p=68&sh...b3d9e6bc6d - updated with hopefully all corrections 12:40
p6weekly.wordpress.com/?p=68&sh...b3e4a495bf - more changes 12:42
jnthn "Mouq and lue have been working on Pod6 support" - may be worth saying "working on improving Pod6 support", so it's clear we aren't just getting started on it :) 12:43
moritz timotimo: regexes.pod is meant as reference more than a tutorial 12:44
timotimo ah, yes!
i had that change before i lost my stuff
moritz (I started of writing it as a tutorial, but it becomes more of a ref now)
grondilu_ std: my $x = 1; say $x*$x<=1 12:45
camelia std 09dda5b: OUTPUT«===SORRY!===␤Unable to parse quote-words subscript; couldn't find right angle quote at /tmp/Ek__fzleho line 1:␤------> my $x = 1; say $x*$x<⏏=1␤ expecting escape␤Parse failed␤FAILED 00:01 125m␤»
grondilu_ :/
12:46 grondilu_ is now known as grondilu, grondilu left, grondilu joined
grondilu std: my $x = 1; say $x*$x <= 1 12:46
camelia std 09dda5b: OUTPUT«ok 00:01 124m␤»
grondilu std: my $x = 1; say $x*$x <=1
camelia std 09dda5b: OUTPUT«ok 00:01 124m␤»
grondilu std: my $x = 1; say $x*$x<= 1
camelia std 09dda5b: OUTPUT«===SORRY!===␤Unable to parse quote-words subscript; couldn't find right angle quote at /tmp/91VugElieO line 1:␤------> my $x = 1; say $x*$x<⏏= 1␤ expecting escape␤Parse failed␤FAILED 00:01 125m␤»
timotimo p6weekly.wordpress.com/?p=68&sh...b3f3db24d3 - most up to date by far! 12:47
timotimo recommends to click the "save draft" button regularly in wordpress' editor 12:49
masak timotimo: changing to "Visual" is very much not recommended at all. :)
timotimo aye. 12:50
masak timotimo: I tend to hit "Preview".
WYSIWYG is a big lie, *especially* when there's an underlying textual representation.
grondilu std: $_*$_<= 1
camelia std 09dda5b: OUTPUT«===SORRY!===␤Unable to parse quote-words subscript; couldn't find right angle quote at /tmp/7Quo3EsW6h line 1:␤------> $_*$_<⏏= 1␤ expecting escape␤Parse failed␤FAILED 00:01 124m␤»
masak grondilu: whitespace.
moritz grondilu: "use whitespace around infix operators"
masak grondilu: if you skip the whitespace before an infix, it *becomes* a postfix or postcircumfix. 12:51
grondilu I thought this was only a best practice, not mandatory
masak grondilu: it's only best practice
but for a good reason :)
moritz grondilu: it's a best practise you should follow, because if not, you always need to be aware of potential ambiguties 12:52
timotimo p6weekly.wordpress.com/?p=68&sh...b4074773d7 <- mentioned the lack of rakudo star release manager
moritz github.com/moritz/perl6-wtop/blob/...ctices.pod
grondilu shouldn't LTM fall back to the only interpretation that actually parses or something? I mean, there is no > so the < should not be interpreted as the beginning of a postfix, should it? 12:53
timotimo why have i never heard of perl6-wtop before?
masak grondilu: no, LTM doesn't work like that.
jnthn grondilu: LTM isn't a magic unicorn :P
timotimo notices he doesn't have any bread in the home 12:54
masak grondilu: think of it as being greedy on individual tokens.
moritz grondilu: that would be backtracking
jnthn grondilu: It works at *token* level.
moritz and backtracking usually leads to very unintuitive behavior
grondilu ok
12:55 tgt joined
timotimo moritz: can you use definition lists for the section "backslashed, predefined character classes" perhaps? 12:55
moritz timotimo: I don't know how to write those in pod6 12:56
timotimo OK
timotimo is AFK for a bit and would be happy about more feedback to the post
13:00 tgt left
timotimo added numeric formatting test suite to the LHF 13:01
13:05 daniel-s_ joined 13:07 denis_boyun joined 13:12 zakharyas joined 13:20 btyler joined 13:28 colomon joined
nwc10 timotimo: I think your wording for "You can safely disregard the huge Cross Site Request Forgery warning that may pop up and just “click here to resume your request”" is unfortunate 13:29
I think something more exact like "we are aware that you may get a warning about a Cross Site Request Forgery from rakudo.org to rt.perl.org. We're working on how to avoid this happening, but for now please use “click here to resume your request” for this specific warning" 13:31
because it reads like you're telling people to ignore that warning always. Which you aren't, but "people are stupid" 13:32
or at least, play stupid on TV.
13:32 kst`` joined 13:33 DrEeevil joined, Shozan joined 13:34 Kelder joined, araujo joined, araujo left, araujo joined 13:35 yath_ joined, kshannon_ joined, JimmyZ_ joined, dagurval_ joined, cxreg2 joined, segomos_ joined, cosimo_ joined, BooK_ joined, Psyche^_ joined, pmichaud_ joined 13:38 exodist_ joined 13:39 eternaleye_ joined 13:40 btyler_ joined, genehack_ joined
timotimo p6weekly.wordpress.com/?p=68&sh...b4bbec4c9e 13:40
thanks
13:40 Khisanth left, yves left, BooK left, genehack left, geekosaur left, Bucciarati left, pmichaud left, yath left, mathw left, Ayiko left, atrodo left, mathw joined 13:41 colomon left 13:42 eternaleye_ is now known as eternaleye 13:44 Alina-malina left 13:46 woolfy1 joined, yogan_ joined, geekosaur joined, Bucciarati joined, atrodo joined 13:47 avuserow3 joined, yves_ joined, denis_boyun left, btyler left, pdcawley left, _sri left 13:48 denis_boyun_ joined, _sri joined, lue joined, Alina-malina joined
hoelzro morning, #perl6 13:49
13:49 atta joined, Khisanth joined, ajr joined 13:50 ajr is now known as Guest17304, Shozan is now known as SHODAN, Guest17304 is now known as ar_, ar_ is now known as ajr_ 13:51 pdcawley joined, kaare_ left
masak hoelzro! \o/ 13:53
13:54 xenoterracide joined
nwc10 hoelzro: offhand, do you know the answer to: oh, panda bootstrap fails on JVM on OS X. Is it known to work / not work on JVM on other platforms? 13:56
which was irclog.perlgeek.de/perl6/2014-02-24#i_8335189
hoelzro nwc10: I haven't worked on panda yet
I did work to make ufo work on the different backends, though
donaldh Ah, panda bootstrap is recursively copying .work when running local-fetch for panda itself.
nwc10 hoelzro: sorry, I didn't realise that. 13:57
hoelzro np =)
nwc10 I'm very pleased that people are working on this stuff
(as I've got too much other stuff on the go)
donaldh Failed to create directory '.work/1393250114_7//.work/1393250114_7/.work/1393250114_7/.work/ ....... : File name too long
I'll work on a fix. 13:58
hoelzro donaldh: I think someone reported that dir() returns relative (rather than absolute) paths on various backends 13:59
iirc, it's supposed to return absolute
that sounds like it could result in that recursion issue 14:00
moritz iirc it's supposed to return Path objects 14:01
hoelzro timotimo++ # post
moritz: it does
14:01 yoleaux joined, ChanServ sets mode: +v yoleaux
hoelzro someone mentioned something about abs vs rel, though 14:02
14:08 kbaker_ joined 14:09 virtualsue joined, colomon joined
FROGGS donaldh: have a look at the patch in my comment: github.com/tadzik/panda/pull/68 14:12
14:13 smls joined
FROGGS donaldh: and also have a look at this for the .work problem: 14:13
gist.github.com/Mouq/cd9363ce6f20c699af77 14:14
14:14 jnap joined
donaldh FROGGS: Thanks. Yeah, I suspected lazy. 14:14
FROGGS donaldh: if you have time to fix it please go ahead
timotimo p6weekly.wordpress.com/2014/02/24/...e-2014-02/ - published 14:17
JimmyZ_ timotimo++ 14:22
14:22 JimmyZ_ is now known as JImmyZ, JImmyZ is now known as JimmyZ
timotimo pokes perlbug-admin again about the CSRF issue 14:24
14:24 skids joined
smls Hi 14:28
yoleaux 16 Feb 2014 01:00Z <japhb> smls: That mockup looks really good. Aside from my general feeling that the green has been a bit too chartreuse for a while, I think it's excellent work. +1 to make it so.
16 Feb 2014 18:40Z <raydiak> smls: I think you did an awesome job, though I also would consider a different book icon. also, the colors are still up in the air, so feel free to play more with them. many people don't like the current combinations, and I'm not very sold on them myself. make sure you get moritz++'s input too. smls++!
16 Feb 2014 18:41Z <raydiak> smls: as for the structure, layout, etc, my first impression is that it is quite well-considered :)
17 Feb 2014 03:01Z <raydiak> smls: I changed the colors today, in case that affects anything you are/will be working on
smls oh, feedback 14:29
timotimo smls! :D
smls Has raydiak done any more doc.perl6.org style related work in the meantime? 14:30
I think I'll have a go at implementing those mockups in HTML/CSS this week, and while I'm at it try to make it look a bit more professional too (cf. the book icon and menu). 14:33
timotimo much appreciated!
14:33 salv0 left
timotimo i'm glad to see all areas of our documentation (at least those that exist already, excluding perl6-book) are getting some love right now 14:33
lue: moarvm is known to be horribly, horribly, horribly slow at concatenating strings 14:34
smls about the colors, I actually didn't mind the previous green/yellow scheme
but we can always have some more color bikeshedding later, once it's in HTML/CSS ;P
on a completely unrelated note: 14:35
perl6: my $a = 5; my $b = \$a; $$b = 10; say $a
camelia niecza v24-109-g48a8de3: OUTPUT«10␤»
..rakudo-parrot 299d70, rakudo-jvm 299d70, rakudo-moar 299d70: OUTPUT«5␤»
smls ^^ do Capture's no longer provide P5 reference-like syntrax for simple cases? 14:36
how would I write that instead?
FROGGS perl6: my $a = 5; my $b := $a; $b = 10; say $a
camelia rakudo-parrot 299d70, rakudo-jvm 299d70, rakudo-moar 299d70, niecza v24-109-g48a8de3: OUTPUT«10␤»
FROGGS even one char shorter :o) 14:37
timotimo www.youtube.com/watch?v=85LJh4sFi_M - oooh beautiful and stunning!
smls ok
what are Capture's used for now? 14:38
timotimo they represent the result of applying a list of things to a signature, iiuc 14:39
smls this doc suggest they're basically references on steroids: search.cpan.org/~lichtkind/Perl6-Do...apture.pod
but it seems outdated
timotimo This ReleasePerl6-Doc-0.36 [Download] [Browse] 19 Mar 2008
Latest ReleasePerl6-Doc-0.47 [Download] [Browse] 05 Dec 2010
the current release doesn't contain the Capture FAQ document any more 14:40
14:40 salv0 joined
smls ok 14:40
timotimo but even that one is pretty outdated
perl6 has been moving a lot since 2010 14:41
i wonder at what point a LTS release of rakudo would be interesting 14:42
14:44 virtualsue left
nwc10 how L is the Long in LTS? 14:46
It's been talked about privately for Perl 5. I don't think that it works. All the distros use different dates for their own LTSes
so either you'll never get them actually syncing to the release you'd like them to be LTS 14:47
14:47 virtualsue joined
nwc10 or you just add more lag on distribution update, because they aren't using this year's, but the upstream LTS from 18 months ago 14:47
[Coke] Parrot tried having LTS releases, and I think there was a disconnect between bundlers and developers. And rakudo is *still* changes often enough that I wouldn't recommend it until we hit 6.0.0" 14:48
er, I was trying to edit that to "Perl 6 1.0" 14:49
14:49 zakharyas left 14:50 shinobicl joined
timotimo mhm 14:50
nwc10 I think no sooner than 6.0.0 14:51
and quite possibly give it a year or so to shake down after that
timotimo the idea would be to only merge performance improvements and obvious bugfixes, but no spec-change-related things or stuff that likely breaks compatibility with user's programs
masak what [Coke]++ said.
14:52 SamuraiJack_ left
moritz atm it would also imply LTS versions of panda and essential modules 14:52
timotimo aye
moritz and you really don't want to do that right now :-)
nwc10 I thought that the *design* of Perl 6 was that the language was lexically modifiable
timotimo yes, i agree that "right now" is much too soon
in half a year, though ... ;)
nwc10 meaning that the parser can easily support more than one version of the language concurrently 14:53
timotimo oh. like ... "use backwardscompatible '2014.06'"?
nwc10 I thought it was more use v6.1.0
or at least, v6.0.0 was meaning "keep me on the v6.0.0 spec"
timotimo i completely forgot about that feature for a moment 14:54
14:57 rurban joined
[Coke] I wonder if we need spec version/implementation version. 15:01
spec version cannot be sufficient to say what code you want to run, I imagine.
vendethiel timotimo: are you boldly guessing a perl6.0.0 release date ? :d 15:02
timotimo :P
15:07 vendethiel left, vendethiel joined 15:10 ajr_ left, ajr joined, ajr is now known as Guest96668 15:11 Guest96668 is now known as ajr_ 15:15 _thou joined 15:19 zakharyas joined 15:29 [Sno] left 15:38 virtualsue left, rurban left 15:41 pmurias joined 15:46 rurban joined 15:50 xinming joined, ajr_ left, rurban left 15:51 ajr_ joined 16:05 yihshih joined, yihshih__ left 16:06 zakharyas left 16:08 rurban joined 16:09 zakharyas joined 16:11 rurban left 16:13 zakharyas left 16:20 treehug88 joined 16:23 nebuchadnezzar joined, mtk left 16:25 mtk joined 16:27 takadonet joined
takadonet morning 16:27
16:28 telex left, dmol joined 16:29 rurban joined 16:30 telex joined 16:32 denis_boyun_ left 16:38 clsn joined 16:42 smls left
[Coke] timotimo: (bug triage) and yet it never ends. 16:42
16:42 smls joined
[Coke] ponders a mini doc on helping out with the bug queue. 16:43
(in the mean time, feel free to ping me if you want to help)
16:43 zakharyas joined
[Coke] you==not timotimo specifically. 16:43
16:44 BenGoldberg joined
[Coke] the XSS error happens when you are NOT logged into RT, I think. 16:44
nwc10: do you know what fivers do with their spam tickets? Is it something other than clicking the big S? 16:45
[Coke] is happy to hit their recent spam tickets also when he's in there.
nwc10 pretty sure it's "click the big S" 16:46
and yes, clicking S is appreciated
kid51 is usually the most active on this
[Coke] nwc10: done. 16:47
16:58 yihshih__ joined, kshannon joined, segomos joined 16:59 kaleem left 17:00 smls left, clsn left, dmol left, mtk left, nebuchadnezzar left, yihshih left, ajr_ left, shinobicl left, xenoterracide left, segomos_ left, kshannon_ left, SHODAN left, dakkar left, ggoebel1117 left, raydiak left, Yappo__________7 left, ingy1 left, xfix left, BenGoldberg left, telex left, _thou left, salv0 left, colomon left, yoleaux left, avuserow3 left, eternaleye left, exodist_ left, BooK_ left, JimmyZ left, yath_ left, bjz left, go|dfish left, rurban_ left, lee_ left, lizmat left, breinbaas left, mavcunha_ left, FROGGS left, revdiablo left, ruoso_ left, sftp left, darutoko left, klapperl left, pnu left, zamolxes_ left, tokuhirom left, aindilis left, itz left, ribasushi left, nwc10 left, sunnavy left, pochi_ left, erkan left, dylanwh left 17:01 fhelmberger left 17:02 zakharyas left, ggherdov left 17:03 xenoterracide joined
TimToady p6: say [1,2] X~ <a b> 17:05
camelia niecza v24-109-g48a8de3: OUTPUT«1 2a 1 2b␤»
..rakudo-parrot 299d70, rakudo-jvm 299d70, rakudo-moar 299d70: OUTPUT«1a 1b 2a 2b␤»
17:06 lizmat joined, breinbaas joined, mavcunha_ joined, FROGGS joined, revdiablo joined, ruoso_ joined, sftp joined, darutoko joined, klapperl joined, pnu joined, zamolxes_ joined, tokuhirom joined, aindilis joined, itz joined, ribasushi joined, nwc10 joined, sunnavy joined, pochi_ joined, erkan joined, dylanwh joined, Shozan joined, clsn joined, BenGoldberg joined, telex joined, _thou joined, salv0 joined, colomon joined, yoleaux joined, avuserow3 joined, eternaleye joined, exodist_ joined, BooK_ joined, JimmyZ joined, yath_ joined, bjz joined, go|dfish joined, rurban_ joined, lee_ joined, kornbluth.freenode.net sets mode: +v yoleaux
TimToady I assume that one is still considered a rakudobug 17:06
17:06 kshannon_ joined, xinming_ joined, telex left
[Coke] I just saw that bug yesterday, yes. 17:07
rt.perl.org/rt3/Ticket/Display.html?id=74072
TimToady I assumed you were testing for still-bughood, but then you said something about test-neededs
so I wasn't sure
17:07 xfix joined
[Coke] ETOOMANYTESTS 17:07
that one already has a test, even. t/spec/S03-metaops/cross.t 17:08
TimToady muchos gracias
[Coke] <nod>
TimToady hmm, was just looking at that code yesterday, maybe I can figure something out... 17:09
17:09 ggherdov joined
[Coke] m: say $] 17:09
camelia rakudo-moar 299d70: OUTPUT«===SORRY!=== Error while compiling /tmp/_cHikCl3Xj␤Unsupported use of $] variable; in Perl 6 please use $*PERL_VERSION␤at /tmp/_cHikCl3Xj:1␤------> say ⏏$]␤ expecting any of:␤ argument list␤ …»
17:11 kshannon left, xinming left, Alina-malina left
timotimo $PERL_VERSION? 17:11
17:11 pmurias left
timotimo not $*PERL<version>? 17:11
[Coke] rt.perl.org/Ticket/Display.html?id=120996 is a dupe, merged it. 17:12
17:12 colomon left 17:16 pmurias joined 17:19 telex joined 17:20 salv0 left 17:29 japhb_ joined 17:37 bluescreen100 joined
vendethiel r: my $*a = 1; sub b { $*a = 2; c(); }; sub c { $*a = 3; say "c: {$*a.perl}"; }; say $*a.perl; b(); say $*a.perl; 17:37
camelia rakudo-parrot 299d70, rakudo-jvm 299d70, rakudo-moar 299d70: OUTPUT«1␤c: 3␤3␤»
vendethiel uw0t
17:38 Alina-malina joined, FROGGS left
timotimo why does this confuse you? 17:39
17:40 takadonet left
TimToady if you want the subs to establish their own dynamic scope, you have to say 'my $*a = 3' 17:40
or 'temp $*a;' 17:41
vendethiel I'd expect the subs not to run before I call them
as opposed to, say, a simple block 17:42
geekosaur ... what 17:43
17:44 hoverboard joined, nebuchadnezzar joined, SAAJWDK joined, kurahaupo_mobile joined, dmol joined, mtk joined, dakkar joined, ggoebel1117 joined, raydiak joined, Yappo__________7 joined, ingy1 joined
segomos vendethiel: it ran in the order you called for 17:45
17:45 spider-mario joined
vendethiel geekosaur: is this a question of some sorts? 17:45
I'm just blind, apparently
definitely
geekosaur I cannot figure out why you think it ran before you called them 17:46
segomos it output 1, you called b which called c and output 'c: 3' and then output 3
vendethiel geekosaur, segomos : [18:45] <14vendethiel> I'm just blind, apparently # my bad
segomos me too 17:47
vendethiel r: my $*a = 1; sub b { my $*a = 2; c(); }; sub c { $*a = 3; say "c: {$*a.perl}"; }; say $*a.perl; b(); say $*a.perl;
camelia rakudo-parrot 299d70, rakudo-jvm 299d70, rakudo-moar 299d70: OUTPUT«1␤c: 3␤1␤»
vendethiel i liek it
segomos interesting
vendethiel expected 17:48
17:48 colomon joined
segomos that is similar to the 'this' confusion people have in javascript 17:49
17:49 denis_boyun joined
masak not really, no. 17:51
well, depends on your threshold for similarity, I guess :P 17:52
jnthn Mars looks close tonight, so I'll svn ci 17:53
uh, wrong window
masak segomos: the main confusion people have about 'this' in JavaScript is that each individual routine invocation gets its own 'this', and so you cannot simply use 'this' from an inner function in a method, for example.
jnthn: *lol*
segomos it's scoping $*a to 'sub b', its caller. in js people have confusion over 'this' being scoped to whatever called the function (which could be anything)
jnthn ...boy must that one look weird without context :)
masak jnthn: don't you dare try to explain!
jnthn: it's funnier this way :P
tadzik jnthn: ...what is a right window for that? :D
segomos lol
jnthn I think this is the first time lag has turned out to be a positive :P 17:54
masak segomos: no, 'this' is still scoped to the callee in JS.
segomos: are you sure you have the right definition of 'caller' in you mind?
segomos: or maybe you're completely misunderstanding how dynamic lookup works -- I cannot tell. 17:55
17:59 FROGGS joined
dalek ecs: 5ef5f5c | larry++ | S32-setting-library/Containers.pod:
Add Range.sample for TheDamian++
 18:00
[Coke] docs.perl.org - it would be nice if you got an indication that a search returned no results. 18:01
moritz raydiak: ^^
[Coke] er, doc.perl6.org/
also, I'd like to be able to hit "ord<enter>" in the search box and have it go somewhere. 18:02
18:02 dakkar left, awwaiid left
moritz TimToady: it seems your last spec commit added much more than Range.sample 18:03
TimToady: like, duplicated the Buf section
segomos masak: no i get what you're saying, but 'this' in js isn't always equal to .caller
TimToady oops
18:03 telex left 18:04 pmurias left
masak segomos: there is some weapons-grade confusion going on here. why would 'this' be equal to .caller? 18:04
segomos i'm generalizing the way the scope of a variable works
18:05 SamuraiJack_ joined
masak yes, that is probably where we need to begin. 18:05
segomos masak: i think because i'm half assedly paying attention -
masak segomos: 'lexical lookup' vs 'dynamic lookup' -- do these terms sound familiar?
segomos yea
masak segomos: well, the thing with 'this' is that it does neither. it's simply invocation-specific. that is, the *invocation* sets 'this'.
sometimes to a sensible object (like with methods and constructors), sometimes to the global object (like with functions), and sometimes to something you specify (like with call/apply). 18:06
18:06 berekuk joined
segomos yea i'm familiar with how it works 18:06
masak ok.
then I was hoping you would also say that it's nothing like dynamic lookup :) 18:07
food &
dalek ecs: 73c60fc | larry++ | S32-setting-library/Containers.pod:
remove accidental duplication
18:07 pmurias joined 18:08 molaf joined 18:09 telex joined
segomos masak: ahh :) sorry, I'm getting frustrated with this SAP soap thing and only half reading (apparently) 18:11
18:14 synopsebot left, synopsebot joined
FROGGS segomos: I'll fell the joy (pain, really) of SAP soap in near future too :o) 18:14
masak segomos: s'ok. I wasn't sure which one of us was confused, to be honest. 18:19
moritz was confused
:-)
18:20 SamuraiJack_ left 18:22 ajr joined 18:23 ajr is now known as Guest6564, Guest6564 is now known as ajr_ 18:33 xragnar joined 18:39 tgt joined 18:42 Khisanth left, khisanth_ joined 18:44 ajr__ joined, pmurias left, khisanth_ is now known as Khisanth, ajr_ left 18:46 pmurias joined
segomos FROGGS: ahh, i think i just got BOBJ QAAWS working which means i can do web type development instead of using these hokey tools 18:53
FROGGS segomos: my project will probably start in april or may, before that I'wont know what I am supposed to do 18:57
18:59 SamuraiJack_ joined
nwc10 www.google-melange.com/gsoc/org2/g...oc2014/tpf 19:02
19:02 treehug88 left, treehug8_ joined 19:03 hoverboard is now known as EyeOfTheTiger
timotimo we wanted to get the peephole optimizer task removed from the ideas page 19:03
can someone do that?
19:03 kurahaupo_mobile left
nwc10 surely it's a wiki, so anyone could? 19:04
timotimo ... i suppose at least registration is required?
i thought there'd probably be a requirement to hve some privileges
19:05 darutoko left 19:06 kurahaupo_mobile joined 19:07 EyeOfTheTiger is now known as hoverboard 19:09 ingy1 is now known as ingy 19:16 SamuraiJack_ left 19:19 Rotwang joined
lue timotimo: interesting. I was looking at join last night thinking to myself "kcachegrind says it's only called 4 times, it can't be what's *really* slowing everything down". 19:25
timotimo t.h8.lv/p6bench/2014-02-20-rakudo.html ← have a closer look 19:26
at some point we ought to have concurrency benchmarks, too :) 19:27
19:31 denis_boyun__ joined
lue You should probably leave parrot out of those, unless you want a depressing flatline on all the charts :) 19:31
19:31 denis_boyun left
masak .oO( this... is a flatlined Parrot! ) 19:32
timotimo parrot still beats out moarvm when it comes to string operations
[Coke] On the contrary, don't we want stats on parrot vs. moar?
grondilu moritz: in github.com/moritz/perl6-wtop/blob/...tices.pod, you could possibly mention infix:<**> as an other exception to the rule about putting spaces around an infix operator.
timotimo and otherwise it's usually only like 2x faster
nwc10 I thought that too
lue [Coke]: not for concurrency I don't think :)
nwc10 and what's curious is why parrot wins on while_concat
timotimo it's not curious
nwc10 OK. Why isn't it curious? 19:33
timotimo moar's way of concatenating strings is absolutely abysmally pessimized
our rope implementation is thoroughly troubled
masak grondilu: you mean infix:<**> in the Regex slang?
grondilu no, I mean the exponentiation operator. 19:34
lue jnthn mentioned just recently that the ropes are forced into being simple strings (because the ropey parts of the impl. are broken), which is apparently absolutely terrible for performance.
grondilu it's an infix, isn't it?
timotimo it's doing the pessimized "copy all the things" thing
19:34 rurban left
timotimo and also, we're forcing all strings to use 32bit codepoint storage even if it would fit into 8bit 19:35
i think parrot is better at that very point than moar
jnthn I don't think the 32-bit/8-bit is the big deal. 19:36
It's the building up an intricate data structure, only to throw it away again.
timotimo how about the way we're accessing the codepoints for copying?
jnthn timotimo: Yes, that is more the problem. 19:37
timotimo so, a fast path for two 32bit non-rope strings ought to be a bit of a help?
one that uses memcpy?
jnthn timotimo: Yes. 19:38
timotimo seems like an easy thing to whip up in a few minutes right now :)
19:38 kaare_ joined
[Coke] jnthn: do you have any need of curated tickets from RT to help you find things to work on, or are you (as I suspect) all set in the find stuff to do area? 19:38
jnthn [Coke]: I've got a fairly sizable bunch of things to do in the near future. 19:40
19:40 kurahaupo_mobile left, kurahaupo_mobile joined
jnthn [Coke]: At least one RT ticket being one of the most urgent. 19:40
[Coke] hokay. 19:41
19:41 molaf left
[Coke] Where did the node.js backend stuff end up? 19:43
19:43 dmol left, panchiniak joined
FROGGS jnthn: what is the sanest way to access a BOOThash on jvm? 19:45
nqp::backendconfig()<foo> does not work
"No such method 'at_key' for invocant of type 'BOOTHash'" 19:46
jnthn FROGGS: From perl 6 or nqp?
Oh, then Perl6
Well, nqp::atkey(.., 'foo')
FROGGS in rakudo's core_epilog.pm to be exact
hmmm, I think I tried that....
jnthn Try it righter
:P
FROGGS yeah :P
dalek kudo-star-daily: 5029d43 | coke++ | log/ (5 files):
today (automated commit)
 19:48
FROGGS [Coke]: btw, I tried to fix that EVAL for pugs... but I gave up after a while
19:49 tgt left 19:51 rindolf left
FROGGS jnthn: seems to work(-ish), there are more accesses that need to be nqp::atkey, likely that I've just not seen that the first change succeeded 19:51
[Coke] FROGGS: me too, which is why I opened a ticket. :) 19:54
FROGGS yeah
[Coke] eval looks special. (other stuff, it was a simple mod to Prim.hs)
FROGGS :/
[Coke] casts "Summon au"
FROGGS expecto au?
20:00 tgt joined 20:01 rindolf joined 20:05 pdcawley left
masak you can only cast au if the thought you're holding is happy enough. 20:05
timotimo %) 20:06
20:08 pdcawley joined 20:09 kivutar joined
FROGGS *g* 20:09
so, flying on a broomstick doesn't ount, ehh?
count*
I need to reparo my keyboard -.-
masak .oO( we all knew that au was brilliant, but we hadn't realized until now she's a patronus ) 20:10
pmurias: another thing that's usually written in all-caps: heredoc terminating strings. 20:17
by the way, it's an interesting language evolution that heredoc is spelled "here-doc" in perldoc. 20:18
20:18 tgt left, tgt joined 20:26 Sqirrel joined
timotimo lue: i have a moarvm patch that you can try with your pod6 htmlify stuff - shouldn't need to re-build rakudo or nqp or the module. 20:29
lue I'd be willing to try it. Do I need latest MoarVM?
timotimo it's based on MoarVM master, but it should apply cleanly to older MoarVM as well 20:30
20:30 kurahaupo_mobile left
timotimo the branch you want to use is flatten_fastpath in moarvm 20:30
lue oh, so I can just switch to it then, no manual patch application like I was thinking :) . 20:31
timotimo that's right
i figured that'd be much nicer for everyone
can't tell if the spectests are clean yet. 20:35
lue O.O before: about 1246.0347964 seconds after: 244.9768101 seconds. (Most of that is hung up in File::Find::find; all steps after the last use of find() blinked past before I knew what was happening :D) 20:40
timotimo++ ! \o/
Huh, all of a sudden I get links pointing to github.com/perl6/roast/blob/master...tax.t#L33\ (notice the backslash) can't tell if that's me or the MVM patch.
timotimo sounds good 20:41
hm. interesting.,
FROGGS interesting\
:P
timotimo could you be so kind as to golf it?
lue I'll revert to pre-patch moar, and see if it comes up again. 20:42
(that'll take about 15 minutes to run through though) 20:43
timotimo hah
can you reduce the workload?
lue Yes, actually. I'll just have to copy over a single synopsis to use. 20:44
timotimo spectests are clean, annoyingly
so if it's my patch that's to blame, we could really use a spectest 20:45
well, a regression test really
lue *headdesk* "Obtained 0 synopses." well, no wonder it went by so fast :P This also means the backslashes are my fault, almost surely. 20:47
timotimo huh?
so the performance measurement was improper, too?
lue I forgot to switch back to the pod6ify branch of specs before testing you patch. Sorry :( 20:48
timotimo hehehe
that's fine. it should still give improved performance
i'll run the proper benchmarks soon
lue The backslash is because, as I thought, it's a leftover from moving from double quotes to a heredoc :) 20:49
(link still points to the correct line and file, interestingly) 20:50
masak lue: did you know heredocs can be double-quoted, too? :) 20:51
lue what do you mean? 20:52
20:52 araujo left
masak qq:to/FOO/ 20:52
lue ah, I just meant a literal "" :)
masak yeah, I figured.
20:53 zaydon joined 20:54 mcglk joined 21:02 panchiniak is now known as sovaco-de-cobra 21:03 sovaco-de-cobra is now known as panchiniak 21:04 araujo joined
lue timotimo: 1246.0347964 seconds vs. 1222.1842851 seconds (can't tell if that's significant; a previous pre-patch run was at 1237.8338873) 21:13
timotimo ho-hum. 21:14
21:16 araujo left
timotimo in that case, something else should be a bit slower 21:16
could be that you use join and that doesn't hit the fast path
21:16 araujo joined
timotimo and string_repeat doesn't usually hit the fast path either 21:17
lue I think I tossed out my kcachegrind stuff at some point, couldn't make much of it anyway. [at least, *I* couldn't :)]
timotimo i have the same problem 21:19
21:19 d4l3k_ joined 21:20 Pleiades` joined
lue I think at some point it's just going to take time processing the synopses, though hopefully less than the 10 or so minutes it does currently take. 21:20
21:21 dalek left, d4l3k_ is now known as dalek
lue Oh, dur. I've been using perl6-p for the profiling stuff. If perl6-m actually had profiling support, I'd get actually useful data from there :) . 21:22
timotimo you can still cachegrind that :P
21:23 rindolf left
lue I tried using callgrind once, IIRC it was useless (could try again though) 21:23
timotimo yeah :(
diakopter has been getting opinions/input on what people need for a profiler for moarvm, so at some point he'll have something for us to work with 21:24
lue I just want something that'll show off bottlenecks in code. Not sure how useful that suggestion is though :) 21:29
timotimo lue: can you tell if you're using join more often than concatenate?
lue If my attempt to run callgrind on perl6-m --doc=HTML fails, I'll profile perl6-p and tell you.
timotimo thank you 21:31
how often MVM_string_repeat happens would also be interesting 21:32
those would all require individual fastpaths unfortunately
lue maybe to the point where a better string impl is needed, perhaps? :)
timotimo yes. 21:33
please make a better string impl if you can :P
lue I just read about ropes not too long ago (to better know the more technical bits), so I don't know how good I'd be :P
21:34 virtualsue joined
lue This perl6-m callgrind might take a while :) 21:35
timotimo yeah, probably :(
lue Perhaps I shouldn't've picked S02, the largest of the synopses... 21:36
timotimo heh :)
lue It takes ≈1m30s on perl6-m, so add into that the 20..50x slowdown I read valgrind taking, and I expect at least 10 more minutes to go into this. 21:37
masak r: "foo" ~~ /(bar)? foo/; say ~$0 21:39
camelia rakudo-jvm 299d70, rakudo-moar 299d70: OUTPUT«use of uninitialized value of type Any in string context␤␤»
..rakudo-parrot 299d70: OUTPUT«use of uninitialized value of type Any in string context in block at /tmp/tmpfile:1␤␤␤»
masak r: "foo" ~~ /(bar)? foo/; say $0.^name
camelia rakudo-parrot 299d70, rakudo-jvm 299d70, rakudo-moar 299d70: OUTPUT«Any␤»
masak r: "foo" ~~ /(bar)? foo/; say ~($0 // "")
camelia rakudo-parrot 299d70, rakudo-jvm 299d70, rakudo-moar 299d70: OUTPUT«␤»
timotimo hm, should that be Nil?
moritz r: say so "foo" ~~ /(bar)? foo/; 21:42
camelia rakudo-parrot 299d70, rakudo-jvm 299d70, rakudo-moar 299d70: OUTPUT«True␤»
21:44 d4l3k_ joined
lue timotimo: trying my hand at ropes could possibly be interesting, but not only do I need to be familiar with the MVM codebase, I also would much prefer experimenting in a separate, useless project first :) [Also, if I must use C, that's another thing...] 21:45
timotimo you'd have to be using C, aye
21:45 dalek left
timotimo and the code for ropes is kinda complicated 21:45
21:45 d4l3k_ is now known as dalek
timotimo it uses the visitor pattern to work on ropes, which is pretty nice 21:46
but it's not obvious how exactly things work
and especially where the gremlins are
lue Not saying I couldn't use C, just that it'd take some time to beat the C++11 outta me :P 21:47
timotimo hah 21:48
also you can't use C99
21:48 awwaiid joined
lue Yeah, I caught that in #moarvm. That means none of that mixing declarations and code I got bitten by in some project I had to repair a while ago, right? 21:49
timotimo yeah, declaration has to be at the beginning of blocks :( 21:50
i've made that mistake quite often
lue Good thing I'm not in charge of moar, else I'd say "use a better compiler" :P 21:51
timotimo hah
msvc++ could in theory be "better"
allegedly, MS's response to "MSVC is lagging far behind the standards!" is "use msvc++ instead, dummies" 21:52
moritz can anybody tell me what's wrong with this table? 21:54
github.com/perl6/doc/blob/master/l...s.pod#L253
21:54 treehug8_ left
moritz it's not rendered as a table at all, neither by Pod::To::Text nor by Pod::To::HTML 21:54
timotimo does it get confused by initial empty lines? 21:55
moritz removing it doesn't unconfuse it 21:56
timotimo also, i think the table above that for the short and long unicode thingies ought to have many more empty lines
moritz let's start with the simpler one :-)
timotimo since i believe seeing at least one empty line in between puts it into "one row per paragraph" mode rather than "one row per line" mode
masak 'night, #perl6 21:57
timotimo 'night masak
lue ♞ masak o/
moritz: did you remove the final empty line too?
moritz lue: yes
fwiw the table in lib/Whatever.pod works fine
and it also has newlines before and after the table body 21:58
timotimo hmm 21:59
moritz aaaargh
lue Oh! Try wrapping | and + in V<>s ! (may need :allow<V> on the table)
moritz if I add an 'a' before each row "fixes" it 22:00
timotimo does :allow actually work in tables now?!
lue maybe -
timotimo: it does in code blocks at least, not sure if it works everywhere though
22:00 zakharyas joined
lue (theoretically every cell is a nested =para, so you shouldn't need :allow) 22:00
*maybe - too
22:00 colomon left
moritz what is V<>? 22:01
timotimo "verbatim"
lue: not in the implementation i've seen :)
moritz how does it differ from C<>?
22:01 colomon joined
timotimo doesn't change formatting 22:01
moritz when I add V<...> around the first column, it's rendered as a table, including the V<>
lue the point is to not have whitespace on at least one side of the symbol, because they're looked at as cell dividers here. 22:02
timotimo yes, as i said, formatting codes inside tables are sadly NYI
really sad about that :(
so maybe use "+" instead?
that'll be rendered for "humans", but it at least won't look weird
moritz so, currently there's no way to render a table that has | as a column?
timotimo: use + where? 22:03
lue not one surrounded by whitespace at least
timotimo V<...> would counter that problem, but is NYI
moritz: use \" around the operators
lue thinks Pod6 tables need some refactoring in general; there're too many styles to say "Pod6 will figure it out, don't worry" 22:04
tadzik that's how I wrote it, heh 22:05
timotimo has a much saner implementation
timotimo somewhat
tadzik it's interpreting S26 differently, but correcter I think
timotimo tadzik: did you see i showed it to Mouq++ and he liked it?
moritz timotimo: then I have \" around the operators in the output :(
tadzik timotimo: didn't notice, no 22:06
22:06 skids left
timotimo moritz: yes, you will. but it won't be as confusing to the user as having a visible V<..> around it 22:06
hm. how do &foo; get treated there?
would that perhaps work as a blatant hack?
moritz it's not *that* kind of an operator
but rather regex unicode class combination operators 22:07
as in <:alpha+:digit>
timotimo i know.
moritz found a hack!
using non-breaking spaces around the operator :-)
timotimo hah, clever
moritz a sad kind of clever
timotimo agreed
dalek c: fbc8306 | moritz++ | lib/regexes.pod:
[regexes] fix tables / work around rakudo limitations
 22:08
lue
.oO( =begin table :tablespec<+-+|+-+|+++++> / :tablespec<┏━┓┃┛━┗┃┳┫┻┣╋> ...)
 22:09
timotimo %)
22:10 Sqirrel left 22:14 zaydon left 22:18 bluescreen100 left 22:25 Rotwang left 22:26 kbaker_ left 22:32 Mouq joined
Mouq moritz: "so, currently there's no way to render a table that has | as a column?" :/ sorry 22:34
moritz: It's also a problem I've noticed e.g. there's no way to do L<infix:<(\|)> | infix ∪ > or whatever 22:35
The parsing just needs to be better
lue timotimo: I just got callgrind output for perl6-m --doc=HTML S02-bits.pod6 (took just over 1h), and I got a lot of stuff on MVM, but it looks like no perl6-specific stuff 22:36
Mouq I won't have time tonight, but I'm working on getting timotimo's table implementation into Rakudo 22:37
Actually, I may as well commit and push to a branch
It at least builds
lue Mouq: in that case, I don't think you space around the L<> bar
*you need space 22:38
Mouq lue: The problem is that the Pod::To::* modules have to parse the inside of L<>, and that's stupid 22:39
lue oh, is that what the \| is for?
22:39 zakharyas left
Mouq L<> should produce a Pod::FormattingCode(type<L>, link<whatev>, content([Pod::*])) 22:40
22:40 dmol joined 22:41 rurban joined
Mouq This isn't something that S26 says, but I can't imagine an argument to the contrary 22:41
lue Mouq: would L<infix:<(\|)>|infix ∪> break things? 22:42
Mouq lue: With Pod::To::HTML it produces content => 'infix:<(\', link => ')>|infix ∪' 22:43
lue couldn't you use V<|> or E<0x7C> where needed? 22:44
Mouq Unfortunately no. 22:45
Not with the way things are now, anyway.
I mean 22:46
22:46 pippo joined
Mouq yes, that would be the way to do it, rather than \| 22:46
but it still wouldn't work
dalek kudo/pod-table: 25761c4 | Mouq++ | src/Perl6/ (2 files):
WIP, converting timotimo++'s pod-table parser to the necessary NQP
 22:50
lue not even E<> ? That's unfortunate.
Mouq r: say "\c[NL]" 22:57
camelia rakudo-parrot 299d70, rakudo-jvm 299d70, rakudo-moar 299d70: OUTPUT«===SORRY!=== Error while compiling /tmp/tmpfile␤Unrecognized character name NL␤at /tmp/tmpfile:1␤------> say "\c[NL⏏]"␤»
Mouq r: say "\c[NEL]"
camelia rakudo-parrot 299d70, rakudo-jvm 299d70, rakudo-moar 299d70: OUTPUT«===SORRY!=== Error while compiling /tmp/tmpfile␤Unrecognized character name NEL␤at /tmp/tmpfile:1␤------> say "\c[NEL⏏]"␤»
22:57 virtualsue left
Mouq r: say "\c[INTERROBANG]" 22:58
camelia rakudo-parrot 299d70, rakudo-jvm 299d70, rakudo-moar 299d70: OUTPUT«‽␤»
23:01 colomon left
Mouq r: my $code = 'INTERROBANG'; say nqp::codepointfromname($code) 23:03
camelia rakudo-parrot 299d70, rakudo-jvm 299d70, rakudo-moar 299d70: OUTPUT«8253␤»
Mouq r: my $code = 'INTERROBANG'; say chr(nqp::codepointfromname($code)) 23:04
camelia rakudo-parrot 299d70, rakudo-jvm 299d70, rakudo-moar 299d70: OUTPUT«‽␤»
Mouq Any other way to do this?
23:04 pdcawley left 23:05 btyler_ left 23:06 pdcawley joined
lue r: say nqp::namefromcodepoint(8253); 23:06
camelia rakudo-jvm 299d70: OUTPUT«===SORRY!===␤No such method 'jast' for invocant of type 'NQPMu'␤»
..rakudo-parrot 299d70, rakudo-moar 299d70: OUTPUT«===SORRY!===␤Error while compiling op namefromcodepoint: No registered operation handler for 'namefromcodepoint'␤»
lue Aw :)
23:06 treehug88 joined 23:07 berekuk left 23:13 ajr__ left, vendethiel left 23:15 colomon joined 23:17 rurban left 23:20 vendethiel joined 23:21 tgt left 23:28 hoverboard left 23:44 spider-mario left 23:45 skids joined 23:47 hoverboard joined 23:49 dmol left 23:53 kivutar left

Powered by Cro and the Raku® Programming Language.

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

Thank you!