»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.
colomon m: say map *.base(36), ^50; 00:00
camelia rakudo-moar ce2631: OUTPUT«0 1 2 3 4 5 6 7 8 9 A B C D E F G H I J K L M N O P Q R S T U V W X Y Z 10 11 12 13 14 15 16 17 18 19 1A 1B 1C 1D␤»
grondilu m: say "9" before "A"
camelia rakudo-moar ce2631: OUTPUT«True␤»
grondilu m: say "9" before "12"
camelia rakudo-moar ce2631: OUTPUT«False␤»
grondilu :/ 00:01
colomon m: say "9" lt "12"
camelia rakudo-moar ce2631: OUTPUT«False␤»
colomon m: say "9" < "12"
camelia rakudo-moar ce2631: OUTPUT«True␤»
grondilu wants to create a infinite list of strings in lexicographic order 00:02
timotimo colomon: just idly clicking around in the call graph shows that report-box-diff takes three times as report-attributes-diff, which takes a bit longer than report-coordinate-diff
grondilu though 'map *.base(36), 1..*' would do
*thought
colomon m: say ('a', *.incr … *)['1000] 00:03
camelia rakudo-moar ce2631: OUTPUT«===SORRY!=== Error while compiling /tmp/pyucksOVDJ␤Two terms in a row␤at /tmp/pyucksOVDJ:1␤------> say ('a', *.incr ⏏… *)['1000]␤ expecting any of:␤ method arguments␤ postfix␤ i…»
colomon m: say ('a', *.incr … *)[1000]
camelia rakudo-moar ce2631: OUTPUT«===SORRY!=== Error while compiling /tmp/UurNBZoh9n␤Two terms in a row␤at /tmp/UurNBZoh9n:1␤------> say ('a', *.incr ⏏… *)[1000]␤ expecting any of:␤ method arguments␤ postfix␤ in…»
colomon m: my $a = 'a'l; say ('a', $a++ … *)[1000]
camelia rakudo-moar ce2631: OUTPUT«===SORRY!=== Error while compiling /tmp/SyZHX12yrF␤Two terms in a row␤at /tmp/SyZHX12yrF:1␤------> my $a = 'a'⏏l; say ('a', $a++ … *)[1000]␤ expecting any of:␤ postfix␤ infix stopper␤ …»
colomon m: my $a = 'a'; say ('a', $a++ … *)[1000] 00:04
camelia rakudo-moar ce2631: OUTPUT«===SORRY!=== Error while compiling /tmp/yQ9dEYFWGb␤Two terms in a row␤at /tmp/yQ9dEYFWGb:1␤------> my $a = 'a'; say ('a', $a++ ⏏… *)[1000]␤ expecting any of:␤ postfix␤ infix stopper␤ …»
grondilu m: say 'a', *.incr ... 'z'
camelia rakudo-moar ce2631: OUTPUT«===SORRY!===␤No such method 'incr' for invocant of type 'Str'␤»
colomon apparently cannot type
or remember method names
m: say ('a', *.succ … *)[1000]
grondilu m: say 'a', *.succ ... 'z'
camelia rakudo-moar ce2631: OUTPUT«===SORRY!=== Error while compiling /tmp/F505gfDY1K␤Two terms in a row␤at /tmp/F505gfDY1K:1␤------> say ('a', *.succ ⏏… *)[1000]␤ expecting any of:␤ method arguments␤ postfix␤ in…»
rakudo-moar ce2631: OUTPUT«a b c d e f g h i j k l m n o p q r s t u v w x y z␤»
colomon m: say 'a', *.succ ... 'zz'
camelia rakudo-moar ce2631: OUTPUT«a b c d e f g h i j k l m n o p q r s t u v w x y z aa ab ac ad ae af ag ah ai aj ak al am an ao ap aq ar as at au av aw ax ay az ba bb bc bd be bf bg bh bi bj bk bl bm bn bo bp bq br bs bt bu bv bw bx by bz ca cb cc cd ce cf cg ch ci cj ck cl cm cn co cp …»
grondilu colomon++
colomon errr, come to think of it though, that's not lexicographic order 00:05
00:05 erkan joined
colomon m: say ('a', * ~ 'a' … *)[100] 00:06
camelia rakudo-moar ce2631: OUTPUT«===SORRY!=== Error while compiling /tmp/wPQ0476j8i␤Two terms in a row␤at /tmp/wPQ0476j8i:1␤------> say ('a', * ~ 'a' ⏏… *)[100]␤ expecting any of:␤ postfix␤ infix stopper␤ infix…»
colomon m: say 'a', * ~ 'a' … *
camelia rakudo-moar ce2631: OUTPUT«===SORRY!=== Error while compiling /tmp/8K_pNKBdcv␤Two terms in a row␤at /tmp/8K_pNKBdcv:1␤------> say 'a', * ~ 'a' ⏏… *␤ expecting any of:␤ postfix␤ infix stopper␤ infix or met…»
grondilu well, what I do want is my strings to be as small as possible, and still be in some order
timotimo that's right, aa comes after a, not after z
colomon ah, well there's a definite order to the *.succ strings, they're just not in lexicographic order
grondilu how could I use more characters than just 'a'..'z'? 00:07
00:07 gfldex left
grondilu and I do need an infix comparison operator, though 00:07
colomon oh, that's easy
m: say 45.char 00:08
camelia rakudo-moar ce2631: OUTPUT«No such method 'char' for invocant of type 'Int'␤ in block <unit> at /tmp/F8MtE_QS0l:1␤␤»
colomon m: say 45.chr
camelia rakudo-moar ce2631: OUTPUT«-␤»
colomon m: say (32..100).map(*.chr)
camelia rakudo-moar ce2631: OUTPUT« ! " # $ % & ' ( ) * + , - . / 0 1 2 3 4 5 6 7 8 9 : ; < = > ? @ A B C D E F G H I J K L M N O P Q R S T U V W X Y Z [ \ ] ^ _ ` a b c d␤»
colomon m: say (32..1000).map(*.chr)
camelia rakudo-moar ce2631: OUTPUT« ! " # $ % & ' ( ) * + , - . / 0 1 2 3 4 5 6 7 8 9 : ; < = > ? @ A B C D E F G H I J K L M N O P Q R S T U V W X Y Z [ \ ] ^ _ ` a b c d e f g h i j k l m n o p q r s t u v w x y z { | } ~ ^H …»
colomon etc
timotimo colomon: which line does src/gen/m-CORE.setting:1096 correspond to on your build?
colomon guess you'll run out of unicode at some point
grondilu I want an infinite list
timotimo what about src/gen/m-CORE.setting:8118? 00:09
00:09 gfldex joined
timotimo you could write a class that does Positional and implement at_pos by doing .chr :P 00:09
colomon grondilu: I have a notion for you, but I'm answering timotimo first 00:10
timotimo: 1096: multi method Stringy(Mu:D $:) { self.Str }
timotimo huh, i wonder why that doesn't get spesh'd
grondilu :/ that would be a hassle. I don't get why '[before] *.base(36), 0..*' doesn't work :/
colomon timotimo: 8118: GatherIter, $state := {
timotimo and what does it need the $ for?
colomon grondilu: because base 36 isn't in lexographical order any more than base 10 is 00:11
grondilu std: say [Before] map *.base(26), ^50;
camelia std 53b3ca6: OUTPUT«===SORRY!===␤Two terms in a row (preceding is not a valid reduce operator) at /tmp/UF1Rlr_fdf line 1:␤------> say [Before] ⏏map *.base(26), ^50;␤ expecting any of:␤ feed_separator␤ infix or meta-infix␤ infixed…» 00:12
grondilu m: say [before] map *.base(26), ^50;
camelia rakudo-moar ce2631: OUTPUT«False␤»
timotimo colomon: i don't have a line matching GatherIter,.*state
colomon timotimo: no, no, it's in GatherIter
timotimo oh
colomon the line is just
$state := {
timotimo derp :)
colomon $state := {
nqp::handle( $block(),
'TAKE', SEQ($takings := nqp::getpayload(nqp::exception()); yield(); nqp::resume(nqp::exception())));
$takings := $SENTINEL; yield();
};
grondilu needs to read the WP page about lexicographic order
colomon to be precise
timotimo yes, now i found it :)
and 1096 is inside class Mu, yes?
colomon timotimo: yes 00:13
m: sub lex($n) { 1001.chr x ($n div 1000) ~ ($n mod 1000).chr }; say (999..1111).map({ lex($_) }) 00:15
camelia rakudo-moar ce2631: OUTPUT«ϧ ϩ␀ ϩ ϩ ϩ ϩ ϩ ϩ ϩ ϩ ϩ ϩ␤ ϩ ϩ ϩ
00:15 gfldex left
colomon m: sub lex($n) { 127.chr x ($n div 126) ~ ($n mod 126).chr }; say (100..200).map({ lex($_) }) 00:15
camelia rakudo-moar ce2631: OUTPUT«d e f g h i j k l m n o p q r s t u v w x y z { | } ^H␀ ^H ^H ^H ^H ^H ^H ^H ^H ^H ^H␤ ^H ^H ^H
00:16 Jerry__ joined
colomon hurmph 00:16
00:17 gfldex joined 00:20 kurahaupo_ joined 00:21 Jerry__ left, Jerry90 joined 00:23 kurahaupo left
colomon grondilu: got it 00:25
grondilu: gist.github.com/colomon/196e74c18c504d1150c4 00:26
you can tweak @codes to control how quickly the string gets longer
00:29 BenGoldberg joined
grondilu that's quite complicated 00:29
00:30 jnap joined
grondilu it's hard for me to believe it's not easier to create an infinite ordered list of strings :/ 00:31
00:31 jnap left
grondilu I could of course use map *.Str, 0..* but I wanted to compress things a bit 00:32
BenGoldberg That wouldn't be in lexigraphical order.
grondilu I don't mind which kind of order as long as I have an order 00:33
BenGoldberg Ok, nmind.
grondilu m: say [<] map *.Str, 0..100; # just checking
camelia rakudo-moar ce2631: OUTPUT«True␤»
BenGoldberg m: say [lt] map *.Str, 0..100; 00:34
camelia rakudo-moar ce2631: OUTPUT«False␤»
grondilu m: say [<] map *.fmt("%x"), 0..100; # just checking
camelia rakudo-moar ce2631: OUTPUT«use of uninitialized value $/ of type Nil in string context in block <unit> at /tmp/7b6S9eUV_b:1␤␤Error while creating error string: No exception handler located for warn␤»
grondilu wth?
m: say 42.fmt("%x"); 00:35
camelia rakudo-moar ce2631: OUTPUT«2a␤»
BenGoldberg m: say *.fmt("%x");
grondilu m: say map *.fmt("%x"), ^10;
camelia rakudo-moar ce2631: OUTPUT«WhateverCode.new()␤»
rakudo-moar ce2631: OUTPUT«0 1 2 3 4 5 6 7 8 9␤»
grondilu m: say [+] map *.fmt("%x"), ^10;
camelia rakudo-moar ce2631: OUTPUT«45␤»
grondilu m: say [<] map *.fmt("%x"), ^10;
camelia rakudo-moar ce2631: OUTPUT«True␤»
grondilu m: say [<] map *.fmt("%x"), ^100; 00:36
camelia rakudo-moar ce2631: OUTPUT«use of uninitialized value $/ of type Nil in string context in block <unit> at /tmp/9mxEvrcN0K:1␤␤Error while creating error string: No exception handler located for warn␤»
BenGoldberg m: say [<] map *.fmt("%x"), ^99;
camelia rakudo-moar ce2631: OUTPUT«use of uninitialized value $/ of type Nil in string context in block <unit> at /tmp/gMUfE6yx4z:1␤␤Error while creating error string: No exception handler located for warn␤»
grondilu m: say [<] map *.fmt("%x"), ^50;
camelia rakudo-moar ce2631: OUTPUT«use of uninitialized value $/ of type Nil in string context in block <unit> at /tmp/FbFmoZlxcT:1␤␤Error while creating error string: No exception handler located for warn␤»
grondilu m: say [<] map *.fmt("%x"), ^20;
camelia rakudo-moar ce2631: OUTPUT«use of uninitialized value $/ of type Nil in string context in block <unit> at /tmp/6Mw_CpU73e:1␤␤Error while creating error string: No exception handler located for warn␤»
BenGoldberg m: say [<] map *.fmt("%x"), ^11;
camelia rakudo-moar ce2631: OUTPUT«use of uninitialized value $/ of type Nil in string context in block <unit> at /tmp/lc5aNFCnKB:1␤␤Error while creating error string: No exception handler located for warn␤»
grondilu m: say map *.fmt("%x"), ^20;
camelia rakudo-moar ce2631: OUTPUT«0 1 2 3 4 5 6 7 8 9 a b c d e f 10 11 12 13␤»
grondilu j: say [<] map *.fmt("%x"), ^20; 00:37
camelia rakudo-jvm ce2631: OUTPUT«(timeout)»
grondilu n: say [<] map *.fmt("%x"), ^20;
camelia niecza v24-109-g48a8de3: OUTPUT«Unhandled exception: Cannot parse number: a␤ at /home/p6eval/niecza/lib/CORE.setting line 1536 (die @ 5) ␤ at /home/p6eval/niecza/lib/CORE.setting line 3757 (ANON @ 10) ␤ at /home/p6eval/niecza/lib/CORE.setting line 3759 (NumSyntax.str2num @ …»
BenGoldberg m: say '9' < 'a';
camelia rakudo-moar ce2631: OUTPUT«use of uninitialized value $/ of type Nil in string context in block <unit> at /tmp/CnuFByDpmw:1␤␤Error while creating error string: No exception handler located for warn␤»
BenGoldberg Aha!
grondilu oh yeah
n: say [<] map *.fmt("0x%x"), ^20; 00:38
camelia niecza v24-109-g48a8de3: OUTPUT«True␤»
BenGoldberg n: say '9' < 'a';
camelia niecza v24-109-g48a8de3: OUTPUT«Unhandled exception: Cannot parse number: a␤ at /home/p6eval/niecza/lib/CORE.setting line 1536 (die @ 5) ␤ at /home/p6eval/niecza/lib/CORE.setting line 3757 (ANON @ 10) ␤ at /home/p6eval/niecza/lib/CORE.setting line 3759 (NumSyntax.str2num @ …»
BenGoldberg n: say '9' < '0xa';
camelia niecza v24-109-g48a8de3: OUTPUT«True␤»
BenGoldberg p6: say '9' < 'a';
grondilu yeah I must not forget 0
camelia rakudo-parrot ce2631: OUTPUT«use of uninitialized value $/ of type Nil in string context in any at gen/parrot/BOOTSTRAP.nqp:1221␤␤use of uninitialized value $/ of type Nil in string context␤Cannot call 'Real'; none of these signatures match:␤:(Mu:U \v: *%_)␤ in any at …»
..rakudo-moar ce2631: OUTPUT«use of uninitialized value $/ of type Nil in string context in block <unit> at /tmp/tmpfile:1␤␤Error while creating error string: No exception handler located for warn␤»
..niecza v24-109-g48a8de3: OUTPUT«Unhandled exception: Cannot parse number: a␤ at /home/p6eval/niecza/lib/CORE.setting line 1536 (die @ 5) ␤ at /home/p6eval/niecza/lib/CORE.setting line 3757 (ANON @ 10) ␤ at /home/p6eval/niecza/lib/CORE.setting line 3759 (NumSyntax.str2num @ …»
grondilu yeah I must not forget 0x
camelia ..rakudo-jvm ce2631: OUTPUT«use of uninitialized value $/ of type Nil in string context in any at gen/jvm/BOOTSTRAP.nqp:1214␤␤Unhandled exception: use of uninitialized value $/ of type Nil in string context␤ in warn (gen/jvm/CORE.setting:732)␤ in warn (gen/jvm/CORE.setting…»
BenGoldberg niecza's error msg is sensible. rakudo's is less so. 00:39
grondilu isn't there an extension of the printf format for base 36?
00:40 Jerry90 left
BenGoldberg m: say 11.base(16); 00:40
camelia rakudo-moar ce2631: OUTPUT«B␤»
00:40 Jry joined
grondilu since Z is the last letter, it could be 0Z. so *.fmt("0Z%Z") 00:40
BenGoldberg m: say '0' lt '00'; 00:44
camelia rakudo-moar ce2631: OUTPUT«True␤»
grondilu also, if I define infix:«<=>» on a custom type, aren't infix:«<» and infix:«>» automatically defined as well?
timotimo colomon: did you have time to try the jit with your code yet?
grondilu P5's overload do something like that IIRC
*does
colomon grondilu: quick answer is no
BenGoldberg m: say [<] map 'a' x *, ^10;
camelia rakudo-moar ce2631: OUTPUT«use of uninitialized value $/ of type Nil in string context in block <unit> at /tmp/UaKTibGN0A:1␤␤Error while creating error string: No exception handler located for warn␤»
colomon timotimo: I have not had time, sorry
BenGoldberg m: say [lt] map 'a' x *, ^10; # d'oh
camelia rakudo-moar ce2631: OUTPUT«True␤»
timotimo that's fine
BenGoldberg m: say [lt] map 'a' x *, ^1000;
timotimo i'm distracted at the moment anyway
camelia rakudo-moar ce2631: OUTPUT«True␤»
BenGoldberg Ta-da! Lexigraphical order achieved. 00:45
colomon BenGoldberg: I already did that one, and grondilu said he wanted reasonable length.
00:49 ClarusCogitatio joined, xenoterracide joined 00:50 MilkmanDan left
grondilu
.oO( use MONKEY_TYPING; augment class Int { multi method base($base, :$lexicographic!) {...} } )
 00:50
00:50 Akagi201_ joined
colomon grondilu: you can't make base lexicographic. 00:50
grondilu damn it 00:51
I just want a base method so that $a.base($base) < $b.base($base) if and only if $a < $b 00:52
in other words a base method that preserves integer order
colomon do you mean $a.base($base) lt $b.base($base) ? 00:53
00:53 MilkmanDan joined
grondilu yes 00:53
colomon that can't work with any normal base, that was the point of my above comment.
I don't understand why you think it might?
grondilu oh
BenGoldberg You mean a numeric encoding, not necessarily a base. Use the right terms, and people will be less confused ;)
00:54 dayangkun joined
grondilu ok 00:54
fair point
BenGoldberg m: sub encode($n is copy) { return '0' unless $n; $n = $n.base(2); 1 ~ encode($n.chars-1) ~ $n.substr(1) }; say [lt] map { encode($_) }, ^200;
camelia rakudo-moar ce2631: OUTPUT«True␤»
colomon certianly a numeric encoding can be done, both BenGoldberg and I have suggested them
BenGoldberg m: sub encode($n is copy) { return '0' unless $n; $n = $n.base(2); 1 ~ encode($n.chars-1) ~ $n.substr(1) }; say map { encode($_) }, ^200;
camelia rakudo-moar ce2631: OUTPUT«0 10 1100 1101 1110000 1110001 1110010 1110011 11101000 11101001 11101010 11101011 11101100 11101101 11101110 11101111 111100000000 111100000001 111100000010 111100000011 111100000100 111100000101 111100000110 111100000111 111100001000 111100001001 1111000…»
colomon BenGoldberg++ # that's a lovely one 00:55
BenGoldberg There's a name for it, but I forget what ;)
That one is asymtotically optimal, if you don't mind that it produces bits :) 00:56
grondilu well you beat me. This one is too smart for me to get it quickly. I'll digest it and maybe I'll find a version that uses a larger alphabet. 00:57
BenGoldberg m: sub encode($n is copy) { return '0' unless $n; $n = $n.base(2); 1 ~ encode($n.chars-1) ~ $n.substr(1) }; say [lt] map { ('0b'~encode($_)).Int.base(37) }, ^200;
camelia rakudo-moar ce2631: OUTPUT«base must be between 2 and 36, got 37␤ in method Str at src/gen/m-CORE.setting:13249␤ in method Stringy at src/gen/m-CORE.setting:1096␤ in sub infix:<lt> at src/gen/m-CORE.setting:1517␤ in sub at src/gen/m-CORE.setting:17995␤ in block <unit…» 00:58
BenGoldberg m: sub encode($n is copy) { return '0' unless $n; $n = $n.base(2); 1 ~ encode($n.chars-1) ~ $n.substr(1) }; say [lt] map { ('0b'~encode($_)).Int.base(36) }, ^200;
camelia rakudo-moar ce2631: OUTPUT«False␤»
BenGoldberg I'm sure there's a way to do it, I'm just not thinking of it right now. 00:59
grondilu you wanted to use :2[encode($_)]
(though it does not help much)
BenGoldberg I meant, to make it base 36 for shorter strings, and still be a lexigraphical encoding. 01:01
grondilu so my challenge is: find a sub f(Int $ --> Str) {...} such that f($a) lt f($b) if and only if $a < $b and with f($a) as short as possible.
BenGoldberg m: sub encode($n is copy) { return '0' unless $n; $n = $n.base(36); $n.substr(0,1) ~ encode($n.chars-1) ~ $n.substr(1) }; say map { encode($_) }, ^200;
camelia rakudo-moar ce2631: OUTPUT«0 10 20 30 40 50 60 70 80 90 A0 B0 C0 D0 E0 F0 G0 H0 I0 J0 K0 L0 M0 N0 O0 P0 Q0 R0 S0 T0 U0 V0 W0 X0 Y0 Z0 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 110A 110B 110C 110D 110E 110F 110G 110H 110I 110J 110K 110L 110M 110N 110O 110P 110Q 110R 110S 110T…»
BenGoldberg m: sub encode($n is copy) { return '0' unless $n; $n = $n.base(36); encode($n.chars-1) ~ $n }; say map { encode($_) }, ^200; 01:02
camelia rakudo-moar ce2631: OUTPUT«0 01 02 03 04 05 06 07 08 09 0A 0B 0C 0D 0E 0F 0G 0H 0I 0J 0K 0L 0M 0N 0O 0P 0Q 0R 0S 0T 0U 0V 0W 0X 0Y 0Z 0110 0111 0112 0113 0114 0115 0116 0117 0118 0119 011A 011B 011C 011D 011E 011F 011G 011H 011I 011J 011K 011L 011M 011N 011O 011P 011Q 011R 011S 011T…»
grondilu m: sub encode($n is copy) { return '0' unless $n; $n = $n.base(36); $n.substr(0,1) ~ encode($n.chars-1) ~ $n.substr(1) }; say [lt] map { encode($_) }, ^200;
camelia rakudo-moar ce2631: OUTPUT«False␤»
grondilu maybe I should ask on stackexchange or something :) 01:03
BenGoldberg 's brain is tired.
01:05 colomon_ joined 01:06 colomon left, colomon_ is now known as colomon
grondilu or more precisely: given an alphabet @abc and a subset Word of Str where /^^<@abc>+$$/; find a sub f(Int $n --> Word) {...} such that f is bijective and f($a) lt f($b) iff $a < $b 01:07
(if such a bijection exists, which I'm not sure of) 01:08
colomon you left out the "short as possible" bit there... 01:09
good news: I just got a working smoke run
grondilu no need since I imposed an alphabet and a bijection
colomon bad news is I had to manually kill two run-away test file runs to get the smoke test to finish 01:10
grondilu: 'a' x $n is bijective and in your alphabet 01:11
but it's the least short vaguely sane approach
grondilu yeah but you pick your own alphabet
and it's injective, not surjective
I mean what's f^-1("b")?
01:12 FROGGS_ joined
grondilu (what I meant is that 'a' x $n works only if @abc = 'a') 01:12
colomon ah, I see
yeah, I don't see any obvious reason to believe that is possible. 01:14
01:15 FROGGS left
grondilu I think it's possible and it's pretty easy to prove 01:15
both sets are totally ordered with a minimal element each.
so you can remove the minimal elements and associate them. Then reiterate. 01:16
colomon :\ 01:18
grondilu you don't agree?
colomon nope
consider your alphabet
do you agree that there are an infinite number of strings that start with "a"? 01:19
grondilu there is
there are indeed I mean
colomon then how can you ever map a string that starts with 'b' to a number
?
by your method, I mean
grondilu damn it 01:20
dalek ast: 7fc8efd | TimToady++ | S32-exceptions/misc.t:
update to new P5var warnings
 01:21
grondilu you're right my method can not be surjective, it would give [\~] 'a' xx * again 01:22
I guess there's soething about lexicographic orders I simply don't fundamentally understand. 01:24
colomon grondilu: I think you might even be able to reverse your proof to prove why it's impossible to do.
dalek kudo/nom: 54cd8fb | TimToady++ | src/ (3 files):
update to new P5var warnings

Also a start on implementing sigil-only variables, which parse but produce an erroneous QAST at the moment. They do not, however, interfere with existing code, so in they go without a branch. :)
grondilu you mean by absurdo or something? Possibly indeed. 01:25
BenGoldberg m: sub encode($n is copy) { return '0' unless $n; $n = $n.base(2); 1 ~ encode($n.chars-1) ~ $n.substr(1) }; say map { :2(reverse(encode($_))).base(36).reverse }, ^200;
camelia rakudo-moar ce2631: OUTPUT«Type check failed in binding $str; expected 'Str' but got 'Array'␤ in sub unbase at src/gen/m-CORE.setting:7102␤ in block <unit> at /tmp/PFOwnT4B8D:1␤␤»
BenGoldberg is confused. 01:26
colomon grondilu: suppose there *is* an bijective mapping for the first N numbers.
find the longest string in those N strings. 01:27
BenGoldberg m: say 'abc'.reverse; # what's the right way to do this?
camelia rakudo-moar ce2631: OUTPUT«abc␤»
colomon then make a string of a's one character longer than that long string.
BenGoldberg: 'abc'.flip
grondilu BenGoldberg: flip
BenGoldberg Ok.
colomon grondilu: that string of all a's will be lexographically before at least N/4 of the strings you already have 01:28
grondilu colomon: well done
so I must give up on requiring a bijection. Nevermind. The problem remains. Just need to rephrase it. 01:29
BenGoldberg You need a "universal coding", I think. 01:33
The encoding() function I gave was a variation of the "Elias Omega Coding", modified to be able to represent 0 01:34
grondilu Interstingly enough, the set of strings with the lexicographic order is totally order, and yet when we take all successive smallest elements, we don't exhaust the set. I thoutht that was not possible.
grondilu checks out Elias Omega Coding 01:35
BenGoldberg Also, besides reading the wikipedia page, you might pester the folks on ##math. 01:36
colomon grondilu: the problem is that a longer string still fits in between shorter strings.
I reckon the set of strings in an alphabet is uncountable. 01:37
grondilu BenGoldberg++ neat ideas for RC tasks 01:38
colomon: if the alphabet is finite, the set of strings is very much countable. 01:39
actually, even with an infinite alphabet it's still countable
colomon hmmm, yeah, that seems right 01:40
grondilu as long as the strings are finite
BenGoldberg m: sub encode($n is copy) { return '0' unless $n; $n = $n.base(10); 1 ~ encode($n.chars-1) ~ $n }; say [<] map { encode($_) }, ^200; # success!
camelia rakudo-moar ce2631: OUTPUT«True␤»
BenGoldberg m: sub encode($n is copy) { return '0' unless $n; $n = $n.base(36); 1 ~ encode($n.chars-1) ~ $n }; say [<] map { encode($_) }, ^200; # success! 01:41
camelia rakudo-moar ce2631: OUTPUT«use of uninitialized value $/ of type Nil in string context in block <unit> at /tmp/XnsVMzS3q2:1␤␤Error while creating error string: No exception handler located for warn␤»
BenGoldberg m: sub encode($n is copy) { return '0' unless $n; $n = $n.base(36); 1 ~ encode($n.chars-1) ~ $n }; say [lt] map { encode($_) }, ^200; # success!
camelia rakudo-moar ce2631: OUTPUT«True␤»
BenGoldberg m: sub encode($n is copy) { return '0' unless $n; $n = $n.base(36); 1 ~ encode($n.chars-1) ~ $n }; say map { encode($_) }, ^200; # success!
camelia rakudo-moar ce2631: OUTPUT«0 101 102 103 104 105 106 107 108 109 10A 10B 10C 10D 10E 10F 10G 10H 10I 10J 10K 10L 10M 10N 10O 10P 10Q 10R 10S 10T 10U 10V 10W 10X 10Y 10Z 110110 110111 110112 110113 110114 110115 110116 110117 110118 110119 11011A 11011B 11011C 11011D 11011E 11011F 11…»
grondilu m: sub encode($n is copy) { return '0' unless $n; $n = $n.base(36); 1 ~ 01:42
camelia rakudo-moar ce2631: OUTPUT«===SORRY!=== Error while compiling /tmp/ZiBlULt6dB␤Bogus statement␤at /tmp/ZiBlULt6dB:1␤------> urn '0' unless $n; $n = $n.base(36); 1 ~⏏<EOL>␤ expecting any of:␤ postfix␤ infix stopper␤…»
grondilu oops
BenGoldberg m: sub encode($n is copy) { return 'a' unless $n; $n = $n.base(36); 'b' ~ encode($n.chars-1) ~ $n }; say map { encode($_) }, ^200; 01:43
camelia rakudo-moar ce2631: OUTPUT«a ba1 ba2 ba3 ba4 ba5 ba6 ba7 ba8 ba9 baA baB baC baD baE baF baG baH baI baJ baK baL baM baN baO baP baQ baR baS baT baU baV baW baX baY baZ bba110 bba111 bba112 bba113 bba114 bba115 bba116 bba117 bba118 bba119 bba11A bba11B bba11C bba11D bba11E bba11F bb…»
grondilu m: sub encode($n is copy) { return '0' unless $n; $n = $n.base(36); 1 ~ encode($n.chars-1) ~ $n }; say [lt] map { encode($_) }, ^200;
camelia rakudo-moar ce2631: OUTPUT«True␤»
grondilu BenGoldberg++
m: sub encode($n is copy) { return '0' unless $n; $n = $n.base(36); 1 ~ encode($n.chars-1) ~ $n }; say map { encode($_).chars - .chars }, ^200; 01:44
camelia rakudo-moar ce2631: OUTPUT«0 2 2 2 2 2 2 2 2 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 …»
grondilu lol they are actually longer than the decimal encoding, though :)
BenGoldberg m: sub encode($n is copy) { return 'a' unless $n; $n = $n.base(36); 'b' ~ encode($n.chars-1) ~ $n }; say map { encode(10 ** $_) }, ^20;
camelia rakudo-moar ce2631: OUTPUT«ba1 baA bba12S bba1RS bba27PS bba3255S bba3LFLS bba45YC1S bba51NJCHS bba5GJDGXS bba64LDQPDS bba719XTF1TS bba7CRE66I9S bba83JLXPT2PS bba8ZG3D62R5S bba99UGXNORJLS bbaA2QGPCKVNG1S bbaARCN1HSSIGHS bbaB7LIEEXZX4KXS bbaC23Z405FZ79TDS␤»
BenGoldberg m: sub encode($n is copy) { return 'a' unless $n; $n = $n.base(36); 'b' ~ encode($n.chars-1) ~ $n }; say map { encode(10 ** $_).chars - $_ }, ^20; 01:45
camelia rakudo-moar ce2631: OUTPUT«3 2 4 3 3 3 2 2 2 1 1 1 0 0 -1 -1 -1 -2 -2 -2␤»
BenGoldberg m: sub encode($n is copy) { return 'a' unless $n; $n = $n.base(36); 'b' ~ encode($n.chars-1) ~ $n }; say map { encode(10 ** $_).chars / $_ }, ^20;
camelia rakudo-moar ce2631: OUTPUT«Type check failed for return value; expected 'Int' but got 'Failure'␤ in any return_error at src/vm/moar/Perl6/Ops.nqp:646␤ in method floor at /home/p6eval/rakudo-inst-1/languages/perl6/runtime/CORE.setting.moarvm:1␤ in method Str at src/gen/m-COR…»
BenGoldberg m: sub encode($n is copy) { return 'a' unless $n; $n = $n.base(36); 'b' ~ encode($n.chars-1) ~ $n }; say map { encode(10 ** $_).chars / ($_+1) }, ^20;
camelia rakudo-moar ce2631: OUTPUT«3 1.5 2 1.5 1.4 1.333333 1.142857 1.125 1.111111 1 1 1 0.923077 0.928571 0.866667 0.875 0.882353 0.833333 0.842105 0.85␤»
grondilu well, I must add as a requirement that it must do better than the decimal encoding. That is f($a).chars <= $a.chars 01:46
at least for most $a
BenGoldberg When the numbers get big enough, it will be smaller than the base 10 version.
grondilu ok, I can believe that.
(though I was planning on using that for relatively small numbers :) ) 01:47
BenGoldberg m: sub encode($n is copy) { return 'a' unless $n; $n = $n.base(36); 'b' ~ encode($n.chars-1) ~ $n }; say first { encode(10 ** $_).chars < $_ }, ^20; 01:48
camelia rakudo-moar ce2631: OUTPUT«14␤»
BenGoldberg m: sub encode($n is copy) { return 'a' unless $n; $n = $n.base(36); 'b' ~ encode($n.chars-1) ~ $n }; say first { encode(10 ** $_).chars <= $_ }, ^20;
camelia rakudo-moar ce2631: OUTPUT«12␤»
BenGoldberg So, encoding(10**12) has a length of 12. 01:49
grondilu that's pretty big
BenGoldberg m: sub encode($n is copy) { return 'a' unless $n; $n = $n.base(10); 'b' ~ encode($n.chars-1) ~ $n }; say first { encode(10 ** $_).chars <= $_ }, ^40;
camelia rakudo-moar ce2631: OUTPUT«Nil␤»
BenGoldberg m: sub encode($n is copy) { return 'a' unless $n; $n = $n.base(10); 'b' ~ encode($n.chars-1) ~ $n }; say first { encode(10 ** $_).chars <= $_ }, ^100;
camelia rakudo-moar ce2631: OUTPUT«Nil␤»
grondilu I guess I'll go with *.fmt("0x%x") then 01:50
01:50 nbrown__ left
BenGoldberg m: sub encode($n is copy) { return 'a' unless $n; $n = $n.base(36); 'b' ~ encode($n.chars-1) ~ $n }; say map { encode(10 ** $_).chars <= $_ }, 12..13; 01:51
camelia rakudo-moar ce2631: OUTPUT«True True␤»
BenGoldberg m: sub encode($n is copy) { return 'a' unless $n; $n = $n.base(36); 'b' ~ encode($n.chars-1) ~ $n }; say map { encode(10 ** $_) }, 12..13;
camelia rakudo-moar ce2631: OUTPUT«bba7CRE66I9S bba83JLXPT2PS␤»
BenGoldberg Hmm... all of the 'overhead' is right at the front. Maybe we can encode it! 01:52
colomon timotimo: oooh! moar_jit is rakudobrew config 01:53
recent addition. :) 01:54
02:24 kyun joined
kyun NMAKE : fatal error U1077: “C:\ProgramData\Oracle\Java\javapath\java.EXE”: return code“0x1” 02:30
Oh, I build jvm backend faild
colomon timotimo: gist.github.com/colomon/8e89203b8d...nt-1290523 02:34
02:35 noganex joined 02:38 noganex_ left 02:45 MilkmanDan left 02:49 MilkmanDan joined 03:01 kyun_ joined 03:03 kyun left, kyun_ is now known as kyun, mberends left 03:12 mberends joined 03:14 pochi_ joined 03:20 tphilipp joined, tphilipp left, rindolf joined 03:26 salv0 left 03:27 salv0 joined 03:50 xenoterracide left 03:59 raiph left 04:00 mr-foobar joined, rindolf left 04:01 rindolf joined 04:04 kaare_ joined 04:06 Jry left 04:09 xenoterracide joined 04:10 [Sno] left 04:21 tphilipp joined 04:24 tphilipp left 04:26 xenoterracide left 04:27 anaeem1 joined 04:32 xenoterracide joined 04:37 tphilipp joined, tphilipp left 04:39 tphilipp joined, tphilipp left 04:40 tphilipp joined, tphilipp left, tphilipp joined 04:41 tphilipp left, rindolf left, tphilipp joined, rindolf joined, tphilipp left 04:48 Jry joined 04:54 jack_rabbit joined 04:55 hagiri joined
hagiri hi all 04:55
;) 04:56
04:56 kaare_ left 04:57 BenGoldberg left
jack_rabbit ,r my $v=0;loop ($i=0;$i+12 < @nums.elems;$i++){my $w=[*] @nums[$i..($i+12)];if ($w > $v) {$v = $w;}};say $v; 04:57
r: my $v=0;loop ($i=0;$i+12 < @nums.elems;$i++){my $w=[*] @nums[$i..($i+12)];if ($w > $v) {$v = $w;}};say $v; 04:58
camelia rakudo-jvm 54cd8f: OUTPUT«(timeout)»
..rakudo-{parrot,moar} 54cd8f: OUTPUT«===SORRY!=== Error while compiling /tmp/tmpfile␤Variable '$i' is not declared␤at /tmp/tmpfile:1␤------> my $v=0;loop ($i⏏=0;$i+12 < @nums.elems;$i++){my $w=[*] @␤ expecting any of:␤ postfi…»
jack_rabbit r: my $v=0;loop (my $i=0;$i+12 < @nums.elems;$i++){my $w=[*] @nums[$i..($i+12)];if ($w > $v) {$v = $w;}};say $v;
camelia rakudo-{parrot,jvm,moar} 54cd8f: OUTPUT«===SORRY!=== Error while compiling /tmp/tmpfile␤Variable '@nums' is not declared␤at /tmp/tmpfile:1␤------> y $v=0;loop (my $i=0;$i+12 < @nums.elems⏏;$i++){my $w=[*] @nums[$i..($i+12)];if (␤ exp…»
jack_rabbit r: my @nums = [1..1000]; my $v=0;loop (my $i=0;$i+12 < @nums.elems;$i++){my $w=[*] @nums[$i..($i+12)];if ($w > $v) {$v = $w;}};say $v; 04:59
camelia rakudo-{parrot,jvm,moar} 54cd8f: OUTPUT«0␤»
jack_rabbit r: my @nums = 1..1000; my $v=0;loop (my $i=0;$i+12 < @nums.elems;$i++){my $w=[*] @nums[$i..($i+12)];if ($w > $v) {$v = $w;}};say $v;
camelia rakudo-{parrot,jvm,moar} 54cd8f: OUTPUT«(timeout)»
05:00 xenoterracide left
jack_rabbit Is there a reason the comparison operators need spaces on either side of them? 05:02
well.. I guess only sometimes? 05:04
Is this a bug?
r: my @nums = 1..50; my $v=0;loop (my $i=0;$i+12 < @nums.elems;$i++){my $w=[*] @nums[$i..($i+12)];if ($w>$v) {$v=$w;}};say $v;
camelia rakudo-{parrot,jvm,moar} 54cd8f: OUTPUT«2209723830420986880000␤»
jack_rabbit is fine, but 05:05
r: my @nums = 1..50; my $v=0;loop (my $i=0;$i+12<@nums.elems;$i++){my $w=[*] @nums[$i..($i+12)];if ($w>$v) {$v=$w;}};say $v;
camelia rakudo-{parrot,jvm,moar} 54cd8f: OUTPUT«===SORRY!=== Error while compiling /tmp/tmpfile␤Missing block␤at /tmp/tmpfile:1␤------> ++){my $w=[*] @nums[$i..($i+12)];if ($w>⏏$v) {$v=$w;}};say $v;␤ expecting any of:␤ postfix␤ …»
jack_rabbit is an error.
05:07 hagiri left
TimToady %foo<bar> is the postfix notation for a literal key, so less-than requires a space 05:09
05:09 rindolf left, rindolf joined
TimToady Perl 6 is not optimized for golf 05:10
though when you think about it, real golf is not optimized for golf either... 05:13
05:13 Zikomos joined
TimToady you can't use your putter to get from a term to an infix, you need a longer club :) 05:14
moritz and much space on either side of the club :-)
jack_rabbit Makes sense. I was just curious. 05:15
05:15 ecocode_ joined 05:18 rindolf left, rindolf joined 05:25 kaare_ joined 05:29 telex left
jack_rabbit so, there's no way I can tell to assign an infinite list to an array variable. Someone suggested a sub that returns the list. 05:30
But with this: sub rnd {(rand * 2).Int} ... *)};
rnd[0] varies.
05:30 telex joined
jack_rabbit Is there some way to "memoize" this? 05:30
05:31 Zikomos left, rindolf left, rindolf joined
jack_rabbit this, rather: sub rnd {{(rand * 2).Int} ... *}; 05:34
moritz m: my @rand := (rand * 2) xx *; say @rand[^5] for ^2 05:36
camelia rakudo-moar 54cd8f: OUTPUT«1.00679510450107 1.94302298007086 1.43152089303529 0.375055139916747 0.156311243017144␤1.00679510450107 1.94302298007086 1.43152089303529 0.375055139916747 0.156311243017144␤»
05:36 [Sno] joined
moritz jack_rabbit: like this? 05:36
an array variable is exactly right for memoizing this 05:37
and, if you can't assign, use binding (:=)
jack_rabbit hmm. the interpreter just spins like it's trying to realize the entire list. 05:38
only if there are no statements following the assignment, though. 05:39
your example works fine. 05:40
Is there a way to do the assignment without a follow-up statement in the interpreter? 05:44
05:53 rindolf left, rindolf joined 05:56 _thou left 06:01 rindolf left, rindolf joined 06:02 SamuraiJack joined 06:12 rindolf left, flussence joined, rindolf joined 06:13 isacloud____ joined 06:18 mberends left 06:19 kyun_ joined 06:21 kyun_ is now known as kyunsiu, itz left, kyun left, kyunsiu is now known as kyun 06:22 gfldex left 06:23 itz joined 06:27 mj41 joined 06:28 itz_ joined 06:30 itz left 06:32 Vlavv left 06:36 virtualsue joined 06:45 itz joined
sergot o/ 06:46
06:49 itz_ left 06:52 kyun_ joined, kyun left
sergot m: say 1[1] 06:54
camelia rakudo-moar 54cd8f: OUTPUT«use of uninitialized value of type Nil in string context in block <unit> at /tmp/L2iPiNIiD9:1␤␤use of uninitialized value of type Nil in string context in block <unit> at /tmp/L2iPiNIiD9:1␤␤Error while creating error string: No exception handler …»
06:54 _thou joined, virtualsue left
sergot m: say 1; 06:55
camelia rakudo-moar 54cd8f: OUTPUT«1␤»
06:56 kyun_ is now known as kyun 06:58 _thou left 06:59 kurahaupo_ left
sergot m: <a a a b b c>.Set.say 07:01
camelia rakudo-moar 54cd8f: OUTPUT«set(a, b, c)␤»
sergot m: <a a a b b c>.Set.fmt('%s').say 07:03
camelia rakudo-moar 54cd8f: OUTPUT«a␤b␤c␤»
sergot m: <a b c>.fmt('%s').say
camelia rakudo-moar 54cd8f: OUTPUT«a b c␤»
07:12 FROGGS_ is now known as FROGGS
sergot TimToady: could you give us the answer here: rt.perl.org/Public/Bug/Display.html?id=121947 ? will be great :) 07:17
07:18 kurahaupo joined 07:30 rindolf left, rindolf joined 07:33 virtualsue joined 07:45 donaldh joined 07:51 domidumont joined 07:54 darutoko joined 08:04 rindolf left, rindolf joined 08:09 zakharyas joined 08:13 Ven joined 08:23 fhelmberger joined 08:29 skarn joined 08:31 virtualsue left 08:35 dakkar joined 08:36 tomaw- joined 08:40 SamuraiJack left 08:42 _thou joined 08:47 _thou left 08:51 JimmyZ joined
JimmyZ planeteria.org/perl6/ is for sale ... 08:51
moritz oh noez 08:52
08:59 Jry left 09:03 virtualsue joined 09:07 JimmyZ left 09:10 virtualsue left 09:16 Ven left, donaldh left, salv0 left, araujo left, xinming left, zoosha left, pmichaud left 09:17 Ven joined, donaldh joined, salv0 joined, araujo joined, xinming joined, zoosha joined, pmichaud joined, Pleiades` joined, thilp_ joined, tadzik joined, pochi_ left, Possum left, _sri left, moritz left, sjohnson left 09:19 rindolf left, rindolf joined 09:22 pochi joined 09:23 moritz joined, Possum joined 09:27 Exodist joined, stux|RC-only joined 09:30 _sri joined, yeltzooo joined, silug joined, \00 joined, ingy joined, sjohnson joined 09:34 pecastro joined 09:38 nebuchad` joined 09:46 lizmat_ joined 09:47 fling joined, Gruber joined, bonsaikitten joined, Juerd_ joined 09:48 oetiker_ joined, eviltwin_b joined 09:49 nhayashi joined 09:50 geekosaur left, Grrrr left, eviltwin_b is now known as geekosaur, Juerd_ is now known as Juerd 09:51 jnthn left, jnthn joined, Ven left 10:11 ilogger2 joined, ChanServ sets mode: +v ilogger2 10:12 ivan`` joined 10:15 aborazmeh joined 10:20 virtualsue joined 10:23 colomon_ joined 10:24 anaeem___ joined 10:25 psch joined 10:26 jnthn_ joined, japhb_ joined 10:27 rindolf joined 10:29 nbrown__ joined 10:30 _thou joined 10:33 ClarusCogitatio_ joined 10:34 colomon_ is now known as colomon 10:35 donaldh joined, _thou left 10:38 Ven joined
donaldh .tell jnthn NQPMatch objects are getting through add_inlining_info_if_possible in QAST::Want objects 10:38
yoleaux donaldh: I'll pass your message to jnthn.
lizmat_ r: $ 10:39
donaldh .tell jnthn this is an odd one - gist.github.com/donaldh/f5502df973ec2b41a9c2 fixes it, but why?
yoleaux donaldh: I'll pass your message to jnthn.
camelia ( no output )
moritz donaldh: is this about what keeps the parse tree alive / in memory? 10:40
lizmat_ std: $
camelia std 53b3ca6: OUTPUT«ok 00:00 121m␤»
10:40 virtualsue left
donaldh moritz: causes matches and the CORE.setting text to get serialized into the compiled setting. 10:41
lizmat_ there are spectests for bare sigils that are supposed to fail with "anonymous $sigil variable outside of declaration"
judging from std, this test seems faulty rather than the code
moritz lizmat_: then it seems those are outdated
lizmat_ well, it's just that they started failing after my latest pull 10:42
Ah, I see TimToady is busy with 54cd8fb6d9f43e147c06665f357ccccfb6f749cb
lizmat_ fixed the tests
moritz TimToady++ # getting rid if "Non-declarative sigil is missing its name" 10:43
it always took me a minute to understand what that meant 10:44
dalek ast: 60951ad | (Elizabeth Mattijsen)++ | S03-operators/context.t:
Fix tests now that bare sigils are ok
 10:45
moritz hoelzro: re "Include the offending variable name in uninitialized warning", commit 5f583edc15d779d2b200ca2850abca889107cc32: -- I intentially removed that, because it reported wrong variable names too often 10:47
hoelzro: like vars from the setting, which are of no interest to the user 10:48
10:50 mattp___ joined, immortal joined, rindolf left 10:51 Tene_ joined, smash_ joined, brother| joined, btyler_ joined, cxreg joined, rindolf joined, anocelot joined, revdiablo joined, hoelzro_ joined, sergot_ joined, jnthn joined
lizmat std: $@ 10:51
camelia std 53b3ca6: OUTPUT«ok 00:01 121m␤»
moritz m: $@ 10:52
camelia rakudo-moar 54cd8f: OUTPUT«===SORRY!===␤Unknown QAST node type NQPMu␤»
moritz that's LTA
masak submits rakudobug 10:53
lizmat m: sub f { f(|$) }
camelia rakudo-moar 54cd8f: OUTPUT«===SORRY!===␤Cannot find method 'returns'␤»
lizmat std: sub f { f(|$) }
camelia std 53b3ca6: OUTPUT«ok 00:01 125m␤»
moritz TimToady: ^^ return self if $name eq '$' || $name eq '@' || $name eq '%' || $name eq '&'; this probably leaves $/.ast undefined/NQPMu
hoelzro: perl6/doc htmlify.p6 spits out many "use of uninitialized value of type Nil in string context in any enter at src/gen/m-Metamodel.nqp:3080" warnings; do you know where those come from? 10:55
10:55 betterwo1ld left, Tene left, anocelot_ left, jnthn_ left, sergot left
masak moritz: it's a recent change in rakudo. 10:56
m: say ~Nil
camelia rakudo-moar 54cd8f: OUTPUT«use of uninitialized value of type Nil in string context in block <unit> at /tmp/A16vG05IMp:1␤␤␤»
masak things like this are probably the culprit:
m: say ~(42 if 0)
camelia rakudo-moar 54cd8f: OUTPUT«use of uninitialized value of type Nil in string context in block <unit> at /tmp/JKr1gM4VQY:1␤␤␤»
10:56 Celelibi joined
lizmat std: "$" 10:56
camelia std 53b3ca6: OUTPUT«ok 00:01 122m␤»
lizmat m: "$"
camelia ( no output ) 10:57
masak moritz: TimToady ruminated over them in the backlog, that `if` is a tricky one here -- whether it should give back () or Nil
10:57 betterworld joined, rindolf left 10:58 rindolf joined
dalek ast: 40c5a41 | (Elizabeth Mattijsen)++ | S32-exceptions/misc.t:
More test fixups after TimToady++'s work
 10:58
11:00 nbrown__ left 11:02 mj41 joined, dakkar joined 11:03 JTABCU5B joined, Celelibi left, dg joined, timotimo joined 11:04 Celelibi joined
lizmat r: my @a=({:a<1>}, {:b<1>}, {:a<1>}); my $with = &[evq]; @a .= uniq(:$with)); say @a 11:07
camelia rakudo-moar 54cd8f: OUTPUT«===SORRY!===␤Cannot find method 'returns'␤»
..rakudo-{parrot,jvm} 54cd8f: OUTPUT«===SORRY!===␤No such method 'returns' for invocant of type 'NQPMu'␤»
lizmat this seems to segfault in some situations during spectesting
lizmat is investigating 11:08
FROGGS evq? 11:09
nwc10 lizmat: it might well be irclog.perlgeek.de/moarvm/2014-08-29#i_9262222
FROGGS r: my @a=({:a<1>}, {:b<1>}, {:a<1>}); my $with = &[eqv]; @a .= uniq(:$with)); say @a
lizmat src/Perl6/Actions.nqp:6940
camelia rakudo-{parrot,jvm,moar} 54cd8f: OUTPUT«===SORRY!=== Error while compiling /tmp/tmpfile␤Unexpected closing bracket␤at /tmp/tmpfile:1␤------> ); my $with = &[eqv]; @a .= uniq(:$with)⏏); say @a␤»
FROGGS r: my @a=({:a<1>}, {:b<1>}, {:a<1>}); my $with = &[eqv]; @a .= uniq(:$with); say @a
camelia rakudo-{parrot,jvm,moar} 54cd8f: OUTPUT«{"a" => "1"} {"b" => "1"}␤» 11:10
11:10 vike left
lizmat FROGGS: what is the diff ? 11:11
FROGGS s/evq/eqv/ and s/))/)/
masak m: say (1, 2, 3) ~~ (1, 2, Int)
camelia rakudo-moar 54cd8f: OUTPUT«False␤»
masak why doesn't the above match?
lizmat aha!
11:11 Ven left
masak or rather, what happens in that match? 11:11
FROGGS m: &[baz]
camelia rakudo-moar 54cd8f: OUTPUT«===SORRY!===␤Cannot find method 'returns'␤»
FROGGS LTA 11:12
masak submits rakudobug
FROGGS masak++
masak I reiterate my question, though: why doesn't `(1, 2, 3) ~~ (1, 2, Int)` match? 11:13
:)
FROGGS m: say (1, 2, 3) ~~ (1, 2, 3) # doesnt it only call Accepts on the list?
camelia rakudo-moar 54cd8f: OUTPUT«True␤»
lizmat m: say &[evq].WHAT
camelia rakudo-moar 54cd8f: OUTPUT«===SORRY!===␤Cannot find method 'returns'␤»
lizmat that's the root of the problem, I would think
FROGGS m: say (1, 2, 3) »~~« (1, 2, Int) # masak: you meant that? 11:15
camelia rakudo-moar 54cd8f: OUTPUT«True True True␤»
masak FROGGS: well, you can do that and it works.
11:15 vike joined
masak FROGGS: but I'm more wondering why ~~ on the whole parcel doesn't DWIM and smartmatch 3 against Int. 11:16
lizmat hmmm.... it;s just another LTA error message, the failing test code has eqv, not evq :-(
FROGGS that feels weird...
11:18 telex joined, brother| is now known as brother, rindolf left 11:19 cognome_ joined, rindolf joined
masak or rather, since `(1, 2, 3) ~~ (1, 2, Int)` is False but `3 ~~ Int` is True, what kind of matching does happen, element-wise? 11:20
m: say (1, 2, (3, 4)) ~~ (1, 2, (3, 4))
camelia rakudo-moar 54cd8f: OUTPUT«False␤»
masak o.O
FROGGS hmmm
masak clearly my intuition about this is way off. 11:21
FROGGS rakudo/src/core/Parcel.pm:11: multi method ACCEPTS(Parcel:D: $topic) { self.list.ACCEPTS($topic) }
11:21 Alina-malina joined, synopsebot joined, avuserow_ joined
FROGGS m: say (1, 2, 3) ~~ (1, 2, *) 11:22
camelia rakudo-moar 54cd8f: OUTPUT«True␤»
11:22 apejens joined
FROGGS masak: it checks using === on the elems 11:22
11:22 Woodi joined, Ulti joined
masak m: say (3, 4) === (3, 4) 11:24
camelia rakudo-moar 54cd8f: OUTPUT«False␤»
11:24 lestrrat joined
masak FROGGS: that's just sad. 11:24
I vote for === being changed to ~~ for the elements. 11:25
dalek ast: d755436 | (Elizabeth Mattijsen)++ | S32-exceptions/misc.t:
Add RT number
masak furthermore, I vote for the abolition of *, which is seldom used and mostly muddling up the semantics of list/parcel matching.
it has no "model integrity".
FROGGS masak: I am already spectesting that
masak and it's more-than-subsumed by the awesome matching powers of signatures.
11:26 oetiker joined
jnthn masak++ # reporting the issue that bit me today :) 11:27
yoleaux 28 Aug 2014 16:00Z <japhb> jnthn: When you're awake again, can you look at gist.github.com/japhb/a452a0a8dd7bd5682e1c ? I'm concerned by *both* the weird NPE when threading, but also the fact that having threads <= items causes apparent lockup. I would think the tasks would just share threads in the pool in the normal fashion, making for slow results, but no lock.
28 Aug 2014 22:46Z <flussence> jnthn: Is this a bug, or am I grossly misunderstanding how S17 works? irclog.perlgeek.de/perl6/2014-08-28#i_9260839
10:38Z <donaldh> jnthn: NQPMatch objects are getting through add_inlining_info_if_possible in QAST::Want objects
10:39Z <donaldh> jnthn: this is an odd one - gist.github.com/donaldh/f5502df973ec2b41a9c2 fixes it, but why?
dalek ast: 23abe1b | (Elizabeth Mattijsen)++ | S32-list/uniq.t:
Mark test as skip because of segfault
jnthn omfg, that's a lot of messages
11:27 bowtie joined, mhasch joined, bowtie is now known as Guest36225
nwc10 jnthn: there's also a few nopastes from me in #moarvm 11:27
jnthn Yeah, noticed... 11:28
That channel is rather easier to backlog :)
11:28 Ulti_ joined, [Coke]_ joined 11:29 carlin joined, arnsholt_ joined, diakopte1 joined 11:30 apejens_ joined, perlpilot joined, Ulti left, apejens left
timotimo o/ 11:31
colomon timotimo: I got some JIT timings for you last night. 11:33
gist.github.com/colomon/8e89203b8d...nt-1290523
also, channel in general: I figured out that what's been blocking smoke testing for the last few days are hangs in Test-ClientServer and Chess 11:35
I got a successful run last night
11:35 dayangkun joined
FROGGS ohh! 11:35
colomon by killing those tests manually
FROGGS do tell
ahh
colomon 02-simple-socket.t hangs in the former
pgn.t in the latter 11:36
dalek ast: 2b3a12a | (Elizabeth Mattijsen)++ | S32-exceptions/misc.t:
Add test for #122646
synopsebot Link: rt.perl.org/rt3//Public/Bug/Displa...?id=122646
11:36 BenGoldberg joined
FROGGS masak: that's the spectest using ~~: gist.github.com/FROGGS/80b086660badcf365e5a 11:38
running a unmodified now
colomon: btw, v5 should be build- and testable now 11:40
masak FROGGS: only the first and maybe the last two look relevant, methinks. 11:41
FROGGS masak: I think the same
colomon FROGGS: do tests fail in v5 at the moment?
FROGGS colomon: it will just run sanity tests, so it should pass 11:42
masak food &
colomon hmm, they failed in the smoke run last night.
FROGGS colomon: because it did not use Build.pm, but it should now
colomon oh, very recent change, eh?
FROGGS colomon: 5 minutes :o) 11:43
colomon FROGGS++
11:43 cotto joined
jnthn dinner & 11:43
11:43 epochbell joined 11:45 SHODAN joined 11:46 zacts joined
colomon smoke test done: host07.perl6.com:8080/report 11:47
there's an extra batch of build failures there that didn't happen ten hours ago. :\ 11:48
11:48 rindolf left, rindolf joined
nwc10 jnthn: current --profile-comile ccl4.org/~nick/profile-1409312777.79459.html.gz 11:51
(or without .gz if your connection is fine with about 4M
FROGGS masak: only t/spec/integration/advent2010-day12.t fails due to the change 11:54
masak: fails with: Cannot convert string to number: base-10 number must begin with valid digits or '.' in '⏏bar' (indicated by ⏏) 11:55
dalek kudo/nom: b6b3f55 | (Elizabeth Mattijsen)++ | src/core/ (2 files):
A bare 'say' is now an exception
FROGGS lizmat++ # ohh nice!
dalek ast: fd376fc | (Elizabeth Mattijsen)++ | S32-exceptions/misc.t:
Add test for bare say exception
lizmat it's runtime, not compile time yet 11:56
late lunch&
hoelzro_ moritz: I have a fix for the ~Nil problem; I'm just going to test it real quick 11:58
12:00 hoelzro_ is now known as hoelzro
hoelzro ok, my patch fixes perl6/doc 12:06
waiting for spectest
12:07 kaleem joined 12:11 mattp_ joined 12:13 Ben_Goldberg joined, BenGoldberg left, Ben_Goldberg is now known as BenGoldberg, mattp___ left 12:14 smash_ left 12:15 xinming joined 12:18 dayangkun left, arnsholt_ left 12:19 arnsholt joined 12:20 BizarreCake joined
moritz hoelzro: but why does it try to stringify Nil? did it do that before the S17 merge too? 12:23
12:23 smash joined 12:24 xenoterracide joined
hoelzro it's because Mu.Stringy used to unconditionally call .Str 12:24
I changed it yesterday so it doesn't, but I have a fix I'm testing right now for that
12:25 araujo joined 12:26 dayangkun joined 12:28 rindolf left 12:29 rindolf joined 12:32 kaleem left 12:39 xenoterracide left
moritz hoelzro: then my real question is, is it new that htmlify calls Nil.Stringy? 12:42
hoelzro moritz: if you interpolate Nil into a string, it calls Stringy
it probably happens somewhere!
moritz hoelzro: let me rephrase
hoelzro: htmlify makes quite some assumptions regarding the generated Pod tree
hoelzro: and I don't want any regressions due to changed parsing or Pod tree to go unnoticed 12:43
hoelzro: so I wonder if these Nil interpolations point to some new, deeper problem, or if they existed before, and were silent
hoelzro ah ha
moritz hoelzro: did you diff the output before and after the merge, for example? 12:44
hoelzro I did not
moritz s/output/generated files/
hoelzro I can do that later today
moritz that would be cool
also I noticed that it became much slower
it ran in 41m18s today
I remember more like 15 to 20m
hoelzro wow
I'll test timing too 12:45
moritz not sure if that's due to the many new wornings, or something else
hoelzro I just ran it 45 minutes ago or so; it definitely didn't take 40 minutes
12:49 Ulti_ is now known as Ulti
hoelzro ok, my fix for the Stringy stuff is clean 12:53
dalek kudo/nom: 7c8a4a7 | (Rob Hoelz)++ | src/Perl6/Grammar.nqp:
Fix problems with incomplete POD
 12:54
kudo/nom: 187d75c | (Rob Hoelz)++ | src/core/Mu.pm:
Still invoke $.Str from Stringy
hoelzro lizmat++ nice work on the bare say!
I'm seeing a bunch of warnings even before my S26 merge 12:56
12:56 guru joined, xfix joined, anaeem___ left, guru is now known as Guest99782 12:57 anaeem1 joined, Guest99782 is now known as ajr_, anaeem1 left
moritz hoelzro: isn't Mu.Stringy a relatively hot path? 12:59
I'm kinda worried about introducing a dynamic var there
m: my $x := Nil; '' ~ $x 13:00
camelia rakudo-moar b6b3f5: OUTPUT«use of uninitialized value of type Nil in string context in block <unit> at /tmp/CWh1EV7PTJ:1␤␤»
hoelzro hmm, you're probably right 13:01
the dynamic var is only created for undefined invocants, though, if that helps 13:02
FROGGS that's what it was at the release: github.com/rakudo/rakudo/blob/5dd1...Mu.pm#L402
moritz hoelzro: yes, might help 13:04
FROGGS do we need the variable name in that warning at all? 13:07
I mean, this often led to leaking internal variable names
13:10 perlpilot is now known as PerlJam
hoelzro well, we don't *need* it 13:12
it's just a very nice-to-have, especially for people coming from Perl 5 13:13
13:13 rindolf left, rindolf joined
flussence colomon: I've hacked around that Test-ClientServer/t/ hang for the time being, hope that helps. 13:14
BenGoldberg m: my $a = 'abc'; $a R~= 'd'; say $a; # how do I make this produce 'abcd' ? 13:16
camelia rakudo-moar b6b3f5: OUTPUT«Cannot modify an immutable Str␤ in block at src/gen/m-CORE.setting:17800␤ in block at src/gen/m-CORE.setting:17812␤ in block <unit> at /tmp/45fd9tO_tI:1␤␤»
flussence (it's ironic that that problem's in a bit of code that exists entirely to *prevent* test lockups...)
13:16 ivanshmakov left
BenGoldberg m: my $a = 'abc'; $a [R~]= 'd'; say $a; # This doesn't work, either. 13:17
camelia rakudo-moar b6b3f5: OUTPUT«Not enough positional parameters passed; got 0 but expected 2␤ in block at src/gen/m-CORE.setting:17812␤ in block <unit> at /tmp/FFzbNQsN_Y:1␤␤»
FROGGS m: my $a = 'abc'; 'd' R~= $a; say $a;
camelia rakudo-moar b6b3f5: OUTPUT«abcd␤»
BenGoldberg Oops, I mistated my goal: I want 'dabc'
FROGGS BenGoldberg: you have to swap the operands when you use R
ohh
[Coke]_ m: say 13:18
BenGoldberg needs more coffe
camelia rakudo-moar b6b3f5: OUTPUT«Unhandled exception: Unsupported use of bare 'say'; in Perl 6 please use .say if you meant $_, or use an explicit invocant or argument␤ at <unknown>:1 (/home/p6eval/rakudo-inst-2/languages/perl6/runtime/CORE.setting.moarvm:throw:4294967295)␤ from sr…»
[Coke]_ m: say()
camelia rakudo-moar b6b3f5: OUTPUT«Unhandled exception: Unsupported use of bare 'say'; in Perl 6 please use .say if you meant $_, or use an explicit invocant or argument␤ at <unknown>:1 (/home/p6eval/rakudo-inst-2/languages/perl6/runtime/CORE.setting.moarvm:throw:4294967295)␤ from sr…»
13:18 [Coke]_ is now known as [Coke]
PerlJam BenGoldberg: my $a = "abc"; $a = "d$a"; # :-) 13:18
[Coke] std: say
camelia std 53b3ca6: OUTPUT«===SORRY!===␤Unsupported use of bare 'say'; in Perl 6 please use .say if you meant $_, or use an explicit invocant or argument at /tmp/7LcH6qoC9G line 1:␤------> say⏏<EOL>␤Check failed␤FAILED 00:00 120m␤»
[Coke] std: say()
camelia std 53b3ca6: OUTPUT«ok 00:01 121m␤»
BenGoldberg n: say()
camelia niecza v24-109-g48a8de3: OUTPUT«␤»
FROGGS BenGoldberg: I'm afraid, I think you have to do it the 'long' way 13:19
BenGoldberg Oh well :)
m: sub encode($n is copy) { my $m = 0; my $rest = ''; while ($n) { $n = $n.base(10); ++$m; $rest = $n~$rest; $n = $n.chars-1; }; $m.base(36)~$rest; }; say [lt] map { encode($_) }, ^2000 13:20
camelia rakudo-moar b6b3f5: OUTPUT«True␤»
BenGoldberg grondilu, there's a solution for you ^ :)
PerlJam m: my $a = 'abc'; $a=($a.flip ~ 'd').flip; say $a # this is the *long* way :) 13:21
camelia rakudo-moar b6b3f5: OUTPUT«dabc␤»
BenGoldberg m: sub encode($n is copy) { my $m = 0; my $rest = ''; while ($n) { $n = $n.base(36); ++$m; $rest = $n~$rest; $n = $n.chars-1; }; $m.base(36)~$rest; }; say [lt] map { encode($_) }, ^2000 13:22
camelia rakudo-moar b6b3f5: OUTPUT«True␤»
[Coke] n: use Test; eval_dies_ok('say', 'what');
PerlJam BenGoldberg: you can shorten $n = $n.base(10) to $n.=base(10); That'll help mitigate those extra chars you had to use to prepend
camelia niecza v24-109-g48a8de3: OUTPUT«ok 1 - what␤»
[Coke] M: use Test; eval_dies_ok('say', 'what');
13:22 ivanshmakov joined
[Coke] m: use Test; eval_dies_ok('say', 'what'); 13:22
camelia rakudo-moar b6b3f5: OUTPUT«Unsupported use of bare 'say'; in Perl 6 please use .say if you meant $_, or use an explicit invocant or argument␤ in sub eval_exception at lib/Test.pm:319␤ in sub eval_dies_ok at lib/Test.pm:240␤ in sub eval_dies_ok at lib/Test.pm:238␤ in bloc…»
BenGoldberg m: sub encode($n is copy) { my $m = 0; my $rest = ''; while ($n) { $n = $n.base(36); ++$m; $rest = $n~$rest; $n = $n.chars-1; }; $m.base(36)~$rest; }; say map { encode($_) }, ^40;
camelia rakudo-moar b6b3f5: OUTPUT«0 11 12 13 14 15 16 17 18 19 1A 1B 1C 1D 1E 1F 1G 1H 1I 1J 1K 1L 1M 1N 1O 1P 1Q 1R 1S 1T 1U 1V 1W 1X 1Y 1Z 2110 2111 2112 2113␤»
BenGoldberg m: sub encode($n is copy) { my $m = 0; my $rest = ''; while ($n) { $n = $n.base(36); ++$m; $rest = $n~$rest; $n = $n.chars-1; }; $m.base(36)~$rest; }; say map { encode($_).chars - $_.Str.chars }, ^100; 13:24
camelia rakudo-moar b6b3f5: OUTPUT«0 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2␤»
BenGoldberg It is a nearly-universal numeric coding, whose outputs are at most 2 letters longer than ordinary base 10. 13:25
13:26 gmunoz joined
BenGoldberg And, unlike ordinary base 10, the results are in lexigraphical order :) 13:26
13:26 gmunoz is now known as zoosha
hoelzro moritz: I found no performance issues, but I did find a regression 13:35
the section links aren't being correctly generated
I have to start getting to work, but I'll check it out tonight 13:36
[Coke] m: my $::x 13:37
13:37 rindolf left
camelia ( no output ) 13:37
[Coke] m: my $::x = 3; say $::x;
camelia rakudo-moar b6b3f5: OUTPUT«===SORRY!=== Error while compiling /tmp/B2xaa_Iy_o␤Variable '$x' is not declared␤at /tmp/B2xaa_Iy_o:1␤------> my $::x = 3; say $::x⏏;␤ expecting any of:␤ postfix␤»
13:37 rindolf joined
dalek ast: 2759035 | (Elizabeth Mattijsen)++ | S32-exceptions/misc.t:
TODO test for bare say being a compile time error
 13:40
[Coke] m: say (3+4i).reals.fmt("%d %d") 13:41
camelia rakudo-moar 187d75: OUTPUT«No such method 'item' for invocant of type 'NQPMu'␤ in method message at src/gen/m-CORE.setting:12938␤␤»
[Coke] lizmat: see also S16-io/bare-say.t
lizmat std: print 13:43
camelia std 53b3ca6: OUTPUT«===SORRY!===␤Unsupported use of bare 'print'; in Perl 6 please use .print if you meant $_, or use an explicit invocant or argument at /tmp/q3iLRLpKId line 1:␤------> print⏏<EOL>␤Check failed␤FAILED 00:00 120m␤»
lizmat std: print()
camelia std 53b3ca6: OUTPUT«ok 00:00 121m␤»
lizmat I'm not sure what one would mean by 'print()' that wouldn't warrant the same error 13:44
[Coke] no clue, just reporting about the old ticket.
carlin m: my $x = "foobar"; say $x.split 13:46
camelia rakudo-moar 187d75: OUTPUT«Cannot call 'split'; none of these signatures match:␤:(Cool: Regex $pat, Any $limit = { ... }, Any :all($all), *%_)␤:(Cool: Cool $pat, Any $limit = { ... }, Any :all($all), *%_)␤:(Str:D: Regex $pat, Any $limit = { ... }, Any :all($all), *%_)␤:(Str:…»
carlin m: my $x = "foobar"; say $x[2]
camelia rakudo-moar 187d75: OUTPUT«Index out of range. Is: 2, should be in 0..0␤ in method gist at src/gen/m-CORE.setting:13255␤ in sub say at src/gen/m-CORE.setting:14193␤ in block <unit> at /tmp/0S0eb1ZMJF:1␤␤»
carlin ah, my local rakudo was busted 13:54
13:55 rindolf left, rindolf joined
BenGoldberg A bare say at least might be expected to produce a newline. A bare print seems useless, unless we want to change the spec to it uses $_ like perl5. 13:57
[Coke] m: my $x = "hi"; say $x[0];
camelia rakudo-moar 187d75: OUTPUT«hi␤»
lizmat fwiw, a bare say in p5 prints a newline
timotimo i would like for say; to generate a newline
lizmat oops, no: in p5 it takes $_ 13:58
and that's what TimToady is against (for good reasons, I might add :-)
timotimo yeah
lizmat $ perl5.20.0 -E '$_=42; say'
42
timotimo well, i'd also accept say doing the newline thing only if you "use idontcareaboutperl5" 13:59
huf no v5; at the top?
timotimo maybe like that, yeah 14:00
grondilu m: sub f returns Int { 1/1 }; say f' # shouldn't there be automatic promotion to Int? 14:02
camelia rakudo-moar 187d75: OUTPUT«===SORRY!=== Error while compiling /tmp/O0w0DKeXje␤Two terms in a row␤at /tmp/O0w0DKeXje:1␤------> sub f returns Int { 1/1 }; say f⏏' # shouldn't there be automatic promot␤ expecting any of:␤ argu…»
timotimo m: try eval "say"; say $!.WHAT; say $!.perl;
camelia rakudo-moar 187d75: OUTPUT«===SORRY!=== Error while compiling /tmp/RLWzU96VhO␤Undeclared routine:␤ eval used at line 1␤␤»
timotimo m: try EVAL "say"; say $!.WHAT; say $!.perl; 14:03
camelia rakudo-moar 187d75: OUTPUT«Unsupported use of bare 'say'; in Perl 6 please use .say if you meant $_, or use an explicit invocant or argument␤ in block <unit> at /tmp/rvw2qgoPEO:1␤␤»
lizmat timotimo [Coke] : working on making it a compile time error
timotimo ... huh?
lizmat m: { try EVAL "say"; say $!.WHAT; say $!.perl; } 14:04
camelia rakudo-moar 187d75: OUTPUT«Unsupported use of bare 'say'; in Perl 6 please use .say if you meant $_, or use an explicit invocant or argument␤ in block <unit> at /tmp/h5PR5W9Fom:1␤␤»
lizmat hmmm...
m: { try EVAL "say"; say $!.WHAT; say $!.perl; }; 1
camelia rakudo-moar 187d75: OUTPUT«Unsupported use of bare 'say'; in Perl 6 please use .say if you meant $_, or use an explicit invocant or argument␤ in block <unit> at /tmp/oNtKmX_AxD:1␤␤»
timotimo i wasn't able to find where that error message is currently implemented
lizmat atm, it is a say MMD candidate
14:04 rindolf left 14:05 rindolf joined 14:06 _thou joined
dalek kudo/nom: 1450e3e | (Elizabeth Mattijsen)++ | src/core/IO.pm:
Remove bare say as a candidate

We need a more comprehensive approach		 14:06
timotimo theoretically, you could have "no v5" implemented in pure perl6 by wrapping the multi sub say() candidate and just { print "\n" } instead 14:07
if it's a compile-time error, that's going to be harder to "fix" 14:08
oh well
it's a step in the right direction.
i'm all for errors exploding at compile time rather than at run time
lizmat if say wouldn't have a 'say (|) candidate', the optimizer would already have been complaining at compile time 14:13
timotimo ah
PerlJam Seems to me you just add a "say" term (just like "now") that calls $*W.throw() 14:14
lizmat PerlJam: but that would be a runtime error, no?
jnthn Oh yay...I have a big bottle of Belgian in the fridge, the AC is cranked up, and I've no work to get up for in the morning. That makes for at lesat a couple of hours hacking time. :)
timotimo yays, jnthn :) 14:15
lizmat: $*W is a compile-time construct
nwc10 the great firewall of China doesn't stop beer getting in? 14:16
[Coke] jnthn: good evening.
PerlJam granted, there's probably some hidden difficulty behind my "just" :) 14:17
jnthn nwc10: No, just googl emaps requests getting in/out :)
o/ [Coke], timotimo
I should probably backlog properly, but - why aren't we stealing precisely the STD solution for say? :)
[Coke] I figure if it was easy, someone would have done it 4 years ago? ;) 14:18
jnthn Ok, fine, I'll do it :P
jnthn looks at STD
timotimo don't waste precious jnthn time for such little details :P
jnthn: i'll have a look 14:19
lizmat wonders where jnthn looks at STD
jnthn Well, this one comes up again and again, and I'd like to see it nails.
std: say
PerlJam ah, STD is more general too in that it fixes this for many things.
camelia std 53b3ca6: OUTPUT«===SORRY!===␤Unsupported use of bare 'say'; in Perl 6 please use .say if you meant $_, or use an explicit invocant or argument at /tmp/ltWTMfGTPT line 1:␤------> say⏏<EOL>␤Check failed␤FAILED 00:01 120m␤»
TimToady except, of course, that it just uses a list
rather than having a trait
PerlJam aye
jnthn Ah, there...
TimToady or a signature
14:19 PotatoGim^Home joined 14:20 PotatoGim^Home left
timotimo ah, yeah, it has the %deftrap thing 14:20
jnthn: OK, you can do it :P
jnthn :P
PerlJam waits for the future to become slightly more distributed ;)
14:20 PotatoGim^Home joined
jnthn has also found a nice Indian restaurant here :) 14:21
lizmat wonders (again) where everybody looks at STD
nwc10 jnthn: that seems to happen to you surprisingly often
jnthn lizmat: github.com/perl6/std/blob/master/STD.pm6
TimToady jnthn: I'm this close to having anonymous rvalue variables working; they successfully declare a variable, but then lose track of it somewhere
timotimo huh? i don't even know what "anonymous rvalue variables" is for :S
jnthn nwc10: Well, it got to the point that eating Chinese/Korean noms each day was starting to overwhelm my stomach, so something more familiar to it is nice to find. :)
TimToady so if I say: $; $ANON_VAR__1 = 42; say $ANON_VAR__1 then it prints 42 14:22
but $ = 42 gets a NQPMu
jnthn A NQPMu where, exactly?
TimToady I don't know how to answer that, which is part of the problem :) 14:23
14:23 aborazmeh left
jnthn Does it leak into Perl 6 land? 14:23
Or does it crop up during compilation? 14:24
It may be easiest for me to look at a patch; working on the death trap first. :)
TimToady m: $ = 42 14:25
camelia rakudo-moar 187d75: OUTPUT«===SORRY!===␤Unknown QAST node type NQPMu␤»
TimToady an older version, but shows the issue 14:26
nwc10 jnthn: it was more the "nice" part. You seem to have consistent good luck in finding reasonable Indians
timotimo hm, so maybe it expects to build a QAST::Var somewhere, but if the name isn't set it doesn't end up assigning to the $var that would hold it
and then it puts $var into the qast tree without checking?
jnthn nwc10: Research. ;) 14:29
nwc10: I've found the odd bad one, but online reviews are good for avoiding many of the sucky ones.
dalek kudo/nom: c711b1e | TimToady++ | src/Perl6/Actions.nqp:
latest version of rvalue $
 14:30
nwc10 do you have a co-worker who always seens to end up going to gigs where all the Indians are LTA, and he/she has no idea why that happens? :-)
jnthn TimToady: That often means you incorporated something into the QAST tree that was undefined; most often due to .ast on something make wasn't called on :)
nwc10: No, actually I have many co-workers who also like Indian noms, so we share tips :) 14:31
TimToady but it's in the same path as a normal token variable, so it oughta return the ast the same way from there upward
jnthn At one place we got it down to the level of "don't eat at X on a Monday 'cus the usual chef is off that day and it sucks, but any other day is wonderful" :)
nwc10 jnthn: so if everyone is busy, clients who are in places with good Indians seem to get priority for gigs? :-)
14:32 mberends joined
nwc10 oh gosh 14:32
mberends: MoarVM is fast!
TimToady well, faster :)
nwc10 I suppose that's true too. 14:33
timotimo nwc10: that seems like a response to something, but i don't see what it's in response to ... !?
nwc10 partly, that I've not seen mberends in ages
mberends nwc10: yes! I'm currently at DebConf14 and one of my goals here is to become Debian Maintainer for MoarVM :-) 14:34
dalek kudo/nom: dd580df | (Elizabeth Mattijsen)++ | src/core/Exception.pm:
We don't need no bare exception either
nwc10 mberends: cool
mberends: I don't know how many architectures it's currently portable to
lizmat one of these days I hope to be able doing some commits that I don't need to revoke a few hours later
[Coke] nwc10: nobody sees mberends. 14:35
nwc10 at one point it has worked on x86_64, x86, arm and ppc
PerlJam lizmat++ I was just looking at that. Apparently we already have X::TypeCheck::Argument which does that job already
mberends nwc10: surely just amd64
[Coke] you're just talking to yourself, it seems. :)
There he is! whew.
timotimo the first bunch of yapceu talk recordings have been uploaded: www.youtube.com/user/yapceu/videos
AFAICT no perl6-related videos yet
[Coke] I even see him in backscroll a bit. weird! :)
hugme: hug mberends
hugme hugs mberends
tadzik mberends: it runs on my ARM phone just fine :) 14:36
nwc10 mberends: I fail to think of an actually funny response to that, so all I can think of is "isn't it heresy in debian to just think of x86_64?"
mberends hugs [Coke]
nwc10 the crazy bit was that ppc (big endian) was fairly easy, and ARM wasn't
I don't have access to anything else useful, and haven't had time to check them recently. 14:37
14:37 Alula_ joined
masak mberends! \o/ 14:37
mberends: I've *missed* you! :D
mberends tadzik: that's awesome, must try it on my Nexus 4
nine moritz: I managed to write a failing test with only inline Perl 5: github.com/niner/Inline-Perl5/comm...6a7895dcc4
mberends hugs masak extra hard :D
masak FROGGS: t/spec/integration/advent2010-day12.t is the only file that fails? that's telling me we're onto something here.
masak feels extra hug'd :)
nwc10 mberends: getting moarvm on debian (and hence rakudo available on more-than-just-parrot) would be really $expletive useful. 14:38
tadzik mberends: well, "just fine" is maybe a bit of an overstatement. It did run, but running nqp on top of it resulted in a mysterious sigpipe
japhb_ jnthn: re: the gist I sent you yesterday ... is it expected that if I create promises in a tree (starting tasks that themselves make promises), I need a fair amount of threads just to make any progress at all under r-j?
14:38 japhb_ is now known as japhb
nwc10 mberends: please ask for help if stuck, although I can't promise that I can answer 14:38
mberends nwc10: thanks, I make no promises about timescales either 14:39
masak m: say "bar" ~~ (15..25)
camelia rakudo-moar 187d75: OUTPUT«False␤»
nwc10 mberends: so, in the big mushy window after "this week" but "before Christmas"? :-/
jnthn japhb: If you use "await" a bunch it's not smart enough (anywhere) to give the thread back to the scheduler for the time being.
japhb (anywhere) meaning "any back end"?
Also: ah, hmmm. 14:40
jnthn japhb: That'll likely happen on MoarVM at some point in the not too distant future. On the JVM...continuation barrier stuff...mumble...argh...so yeah :)
japhb heh
14:40 rindolf left 14:41 rindolf joined
jnthn (I really want to do it on JVM too, it's just I tried it once, and know there's some work to do) 14:41
japhb Is there a better construction for my "outer loops" than what I gist'ed?
(Meaning, one that doesn't trip over that problem?)
14:42 donaldh left
lizmat I'm considering adding an "abspath" attribute to IO::Handle 14:42
japhb lizmat: as a cache, to avoid having to compute it if someone asks? 14:43
lizmat which would contain the absolute path
14:43 smash left
lizmat yes, because now we get a "rel2abs" call for every file test operation 14:43
14:43 guru joined, ajr_ left
lizmat and that's expensive 14:44
14:44 guru is now known as Guest53698, Guest53698 is now known as ajr_
lizmat method abspath { $!abspath //= rel2abs($!path) } # something like that 14:44
japhb Yeah, it is. I haven't been able to follow the discussion -- was there any progress in making rel2abs any less expensive, or has it just hit a wall that we can't improve right now? (And if so, what is that wall?) 14:45
14:45 SamuraiJack joined
lizmat for Unix we can simplify, but not for Win 14:45
14:45 BizarreCake left
lizmat in the best case, rel2abs is: return self.canonpath($path) if self.is-absolute($path); 14:46
the canonpath is rather extensive and involved 14:47
mberends nwc10: yes, definitely between next week^Wmonth and Christmas :-/
japhb nodnod
lizmat I guess I could make the canonpath a multi 14:48
japhb lizmat: But is it *correct*? There are so many edge cases in path canonicalization, I'm wondering if we even handle it properly (where "as well as Perl 5" would be a good start)
14:48 smash joined
lizmat one with and one withhout :$parent 14:48
japhb: I only know that if a change the smallest thing, I get spectest errors :-)
*I 14:49
14:49 denis_boyun_ joined
japhb nods in appreciation of that particular pain 14:50
lizmat I guess another thing making this slow, is the forced runtime lookup of methods 14:51
method rel2abs( |c ) { $SPEC.rel2abs( |c ) }
14:52 rindolf left 14:55 ajr joined, ajr_ left
masak m: set(1, 2) ~~ set(1, 2, 3) 14:56
camelia ( no output )
14:56 ajr is now known as Guest66751
masak m: say set(1, 2) ~~ set(1, 2, 3) 14:56
camelia rakudo-moar 1450e3: OUTPUT«False␤»
14:56 Guest66751 is now known as ajr_
masak m: say set(1, 2) (<) set(1, 2, 3) 14:56
camelia rakudo-moar 1450e3: OUTPUT«True␤»
14:57 bjz joined
lizmat m: say set(1, 2) (<) bag(1, 2, 3) 14:58
camelia rakudo-moar 1450e3: OUTPUT«True␤»
lizmat m: say set(1, 2) (<) bag(1, 1, 2, 3)
camelia rakudo-moar 1450e3: OUTPUT«True␤»
lizmat m: say set(1, 2) (<) bag(2, 3)
camelia rakudo-moar 1450e3: OUTPUT«False␤»
lizmat seems to work?
masak yes :) 14:59
14:59 fhelmberger_ joined, fhelmberger_ left
masak is thinking about smartmatching 14:59
14:59 rindolf joined
masak here is the draft of my change so far: gist.github.com/masak/1322886ed7092c522bd8 15:00
a. smartmatch on lists/parcels becomes structural; that is, it does ~~ on its elements. the fallout is that `(1, 2, 3) ~~ (1, 2, Int)` does what I expect.
b. the funky behavior with whatevers in lists/arrays/parcels goes away. whatevers now mean only "whatever". tadzik++'s example in the advent post then actually becomes correct. :) 15:01
lizmat and how would that work for (1,2,3,4) ~~ (1,2,anyothernumber of ints) 15:02
15:02 H2O1 joined, PotatoGim^Home left
masak m: say (1, 2, 3, 4) ~~ :(1, 2, *@ints where { .all ~~ Int }) 15:02
camelia rakudo-moar 1450e3: OUTPUT«False␤»
masak hrm. 15:03
jnthn std: my $s = set;
camelia std 53b3ca6: OUTPUT«===SORRY!===␤The 'set' listop may not be called without arguments (please use () or whitespace to clarify) at /tmp/4Znrem_HlD line 1:␤------> my $s = set⏏;␤Check failed␤FAILED 00:01 123m␤»
jnthn S02-types\set.rakudo.moar fails like that
15:03 kaare_ joined
lizmat actually, I just realize an abspath wouldn't speed up Perl6 start time 15:03
masak m: say (1, 2, 3, 4) ~~ :(1, 2, *@ints where { .all >>~~>> Int })
camelia rakudo-moar 1450e3: OUTPUT«True␤»
masak \o/
15:03 H2O1 left
lizmat as each path (except one) is only ever checked once during startup anyway 15:03
masak lizmat: when you want to do fancy stuff like "any number of X", what you want is not a smartmatch against a simple sequence. making sequences do that fragile/weird logic in the first place was crazy. 15:04
doubly so after signatures turned out to do all that better.
lizmat ok, just checking :-)
jnthn is $s, set⏏.new(5), '(|) has correct precedence.'; 15:05
bah :)
masak jnthn: Set :)
jnthn Right :)
15:05 donaldh joined
masak lizmat: it's OK, I want this proposal to be challenged. that's why I didn't just commit it directly :) 15:05
lizmat the coercer has the same name as the type
jnthn TimToady: I get a failure in S32-str/substr-rw.t claiming "Unsupported use of $$ variable; in Perl 6 please use $*PID" over is($$r⏏, "gloop", '$r referent is eq to the sub 15:06
masak m: say (1, 2, 3, 4, 5) >>~~>> Int
camelia rakudo-moar 1450e3: OUTPUT«True True True True True␤»
jnthn TimToady: Related to your recent patches? I don't see right off how it could be my one... 15:07
TimToady: Also another test has a "can't find method returns"
Bah, I'll just push this thing then we can distribute triage :) 15:08
dalek kudo-star-daily: 94f4ab8 | coke++ | log/ (14 files):
today (automated commit)
kudo/nom: e616060 | jonathan++ | src/Perl6/Grammar.nqp:
Add deftrap mechanism from STD.

Catches say; and similar mistakes.		 15:09
timotimo hurm. i cannot build rakudo-moar on a vserver i'm on because there's a 300 seconds cpu time ulimit 15:10
TimToady jnthn: found the problem, kinda
Ulti >>~~>> just need a <3 operator to go with that
jnthn japhb: I can't see any issues in the code you sent me right away...
TimToady in Action variable, there's a make $past at the end, but it doesn't seem to work
if I put an explicit make $past in my branch and return, it works 15:11
timotimo wait, 300 seconds should be enough
TimToady hang on, I'll test and push
jnthn timotimo: Dunno what the hardware you have is like but on my box I can build all of Moar/NQP/Rakudo in under 300s :)
timotimo after 1:11 minutes it gets "Terminated"
jnthn
.oO( A server *so* fast 300s time-dilates to 71s! )
 15:13
timotimo 2009216maxresidentk 15:14
m: say 2009216 / 1024; say 2009216 / 1024 / 1024;
camelia rakudo-moar 1450e3: OUTPUT«1962.125␤1.916138␤»
timotimo 2gb of ram
could be the cause for the kill
stack size (kbytes, -s) 10240 - but we shouldn't get past that limit, right? 15:15
dalek kudo/nom: 3d233ed | TimToady++ | src/Perl6/ (2 files):
oddly, need explicit make $past; return;

It should fall through and do the 'make $past' at the end of the function but doesn't.
  (Also, remove a useless line from check_variable.)
TimToady there, that version works, but I don't know why it's different
timotimo TimToady: you have a "my $past" in the inner scope 15:16
TimToady ah!
timotimo that's why the $past at the end of the function ends up NQPMu
dalek ast: 42e5824 | jonathan++ | S0 (2 files):
A couple of fixes for the deftrap rule.

STD rejected these two before; now Rakudo does also.
TimToady timotimo++
timotimo is pleased with himself
jnthn heh, nice, does that mean I needn't look at it? :) 15:17
timotimo i hope so :)
TimToady feels appropriately stupid :)
timotimo blame your recent eye troubles
TimToady jnthn: yes, that's preciselyu what it means
well, and that I'm a lousy programmer :) 15:18
jnthn std: say()
camelia std 53b3ca6: OUTPUT«ok 00:01 121m␤»
jnthn That isn't meant to be an error, right?
('cus we have a spectest that things it is, and I'm about to nuke it)
*thinks
dalek kudo/nom: 114ba1d | TimToady++ | src/Perl6/Actions.nqp:
remove stupid 'my' spotted by timotimo++
TimToady say() and say '' are both fine, I think, and likewise set(), to indicate emptiness explicitly 15:19
timotimo std: say() 15:20
camelia std 53b3ca6: OUTPUT«ok 00:01 121m␤»
timotimo m: say()
camelia rakudo-moar 1450e3: OUTPUT«␤»
timotimo m: say
camelia rakudo-moar 1450e3: OUTPUT«␤»
timotimo i'm fine with say() working and say complaining
it's a bit less pretty, but it'll still get the point across
tadzik masak: which example?
masak tadzik: irclog.perlgeek.de/perl6/2014-08-29#i_9263844 15:22
tadzik: FROGGS was spectesting making ~~ structural, and it had exactly one failure in that file.
FROGGS++
[Coke] m: say for 1 15:23
camelia rakudo-moar 1450e3: OUTPUT«␤»
[Coke] m: say for 1
camelia rakudo-moar 1450e3: OUTPUT«␤»
[Coke] oh, are we not caught up here?
timotimo exactly
it should work as expected after the next update
[Coke] jnthn: "say for 1" dies as expected. "say for 1" does not. 15:24
std: say for 1
masak TimToady: any thoughts on gist.github.com/masak/1322886ed7092c522bd8 ?
camelia std 53b3ca6: OUTPUT«===SORRY!===␤Unsupported use of bare 'say'; in Perl 6 please use .say if you meant $_, or use an explicit invocant or argument at /tmp/Ceno6bWhot line 1:␤------> say⏏ for 1␤Check failed␤FAILED 00:01 121m␤»
[Coke] std: say for 1
camelia std 53b3ca6: OUTPUT«===SORRY!===␤Unsupported use of bare 'say'; in Perl 6 please use .say if you meant $_, or use an explicit invocant or argument at /tmp/gbC6IdRsb7 line 1:␤------> say⏏ for 1␤Check failed␤FAILED 00:01 121m␤»
jnthn [Coke]: Yeah, just noticed that. wtf. 15:26
dalek ast: 06159ea | jonathan++ | S16-io/bare-say.t:
Unfudge passing tests; toss a bogus one.
 15:27
15:29 rindolf left, rindolf joined
jnthn [Coke]: Fixed it locally, so now bare-say.t can pass in full 15:30
timotimo looking forward to seeing what caused that behavior %) 15:31
15:31 rindolf left
jnthn Well, I kinda cheated on the fix. 15:32
:)
lizmat m: do { return 42 }
camelia ( no output )
timotimo does that mean say for 1 will still fail now? :P
lizmat hmmm.. that segfaults locally
timotimo and after that say for 1? :D
masak m: do { return 42 }; say "I will survive" 15:33
camelia ( no output )
masak lizmat: well, *something* is wrong.
(and it's in RT at least once already)
lizmat yeah, I know: I was just wondering about the segfault
easy segfaults are the best food of the debugger :-) 15:34
cxreg after all these years i still <3 electricfence 15:36
15:43 tphilipp joined
dalek kudo/nom: a98d264 | jonathan++ | src/Perl6/Grammar.nqp:
Make "say for 1;" complain also.
 15:53
jnthn m: say;
camelia rakudo-moar dd580d: OUTPUT«␤»
jnthn ENOTYET...
dalek ast: 927f3e1 | jonathan++ | S16-io/bare-say.t:
Unfudge "say for 1;" test.
 15:54
15:54 tphilipp left
dalek ast: 1059587 | TimToady++ | S03-operators/context.t:
update for anonyvars
 15:55
jnthn read that as "annoyvars" :)
TimToady m: say ++$ 15:57
camelia rakudo-moar dd580d: OUTPUT«===SORRY!===␤Cannot find method 'returns'␤»
TimToady not there yet
so, who owned "planeteria"? 15:59
15:59 MilkmanDan joined 16:01 jack_rabbit joined
TimToady it's kind of a pity we can't just upload our newly compiled version to run under camelia immediately... 16:01
or run camelia on a machine with enough cores to recompile as soon as a change comes it 16:04
16:06 virtualsue joined 16:09 virtualsue left 16:21 xinming left 16:24 cognome_ left
lizmat not ok 3 - Calling sub if without parens parsefails 16:24
jnthn TimToady ^^^
16:25 cognome joined
lizmat from t/spec/S02-lexical-conventions/one-pass-parsing.t 16:25
16:25 raiph joined
lizmat also: not ok 253 - alternation and conjunction (&|) - parse error 16:26
from t/spec/S05-mass/rx.t
TimToady probably anonyvar collision there 16:28
jnthn TimToady: Which one? Or both?
I can't imagine the latter is mine; first one may be... 16:29
TimToady the latter
probably parsing & as a variable somehow
16:29 cognome left
TimToady but I'll look at the &| one 16:29
m: say ++$ 16:30
camelia rakudo-moar dd580d: OUTPUT«===SORRY!===␤Cannot find method 'returns'␤»
lizmat $ 6 'say ++$'
1
jnthn std: sub if() { "#foo" }; say if;
camelia std 53b3ca6: OUTPUT«===SORRY!===␤Unsupported use of bare 'say'; in Perl 6 please use .say if you meant $_, or use an explicit invocant or argument at /tmp/kG_B89Ln4H line 1:␤------> sub if() { "#foo" }; say⏏ if;␤Confused at /tmp/kG_B89Ln4…»
jnthn Rakudo now matches that error 16:31
TimToady note that 'if' is a kind of terminator
jnthn But the test wants it to be Confused.
So, the test wants an update
TimToady it is confused, in std
std: sub if() { "#foo" }; say if(); 16:32
camelia std 53b3ca6: OUTPUT«ok 00:01 126m␤»
lizmat all current fails: gist.github.com/lizmat/d0968a1791e605f52ba1
jnthn I guess it does both
But throws_like now sees the first...
TimToady yes, std is being more lazily panicky
lizmat throws_like just makes sure it dies for the right reason
dalek line-Perl5: 2e80595 | nine++ | / (3 files):
Allow passing Perl 6 objects through Perl 5 space.

Create Perl 5 objects for Perl 6 objects passed to Perl 5 methods. Use an "unwrap" callback to get back the Perl 6 object if a wrapped object is returned by a Perl 5 method.		 16:33
jnthn lizmat: If you want an easy patch to do, making the patch I did throw a typed exception could be nice...
lizmat: And then update that S02 test to look for it
TimToady it's not necessary to follow STD slavishly there
lizmat if it shouldn't die, or differently from the test, the test should be adapted
jnthn Oh, wait...
It *is* a typed already...
TimToady though the Confused is perhaps useful information as to why say has no argument
lizmat if it is X::AdHoc, it isn't :-)
TimToady Any is a type too :P 16:34
16:34 denis_boyun_ left
TimToady maybe we should rename Any to AdHoc :) 16:34
dalek ast: 7910209 | jonathan++ | S02-lexical-conventions/one-pass-parsing.t:
Sync test to expect same error STD gives.
jnthn Well, STD calls sorryobs in this case. 16:35
TimToady which is lazyish
jnthn So we get an X::Obselete
nine It really makes me nervous that this Inline::Perl5 stuff works with so little hassle. This ought to be much more complicated.
lizmat actually, renaming X::AdHoc to X::Die would make sense to me
16:35 SAA2TAD joined
lizmat or X::Died 16:35
jnthn or X::pired :P
TimToady nine: well, have you looked at diakopter's design to see where your design is going to break?
nine TimToady: not at all. Where can I find it? 16:36
TimToady well, it's in his grant proposal
jnthn Oh heck, I need to implement DESTROY stuff for nine++ too, don't I... :)
lizmat I thought most of the design was in diakopter's head ?
nine jnthn: yes, pretty please :)
TimToady dunno if there's a separate document somewhere offhand
lizmat jnthn: well, *some* code to be executed at GC expulsion would be nice
jnthn lizmat: Yeah, the GC changes are not too bad 16:37
lizmat: It's where to run the code that's tricky.
TimToady I'm sure diakopter++ would be glad to foist some of the work off on nine++
16:37 BizarreCake joined
carlin using NativeCall, I have a CStruct that `has int $.v`, how do I use that to pass to a C function? `some_struct.new(v => 42)` doesn't work 16:37
lizmat TimToady: I hope so too 16:38
16:38 jdv79 joined
nine TimToady: what I can find in his grant proposal is completely different from my implementation. 16:38
jdv79 where is the planet site now?
i wanted to find the link to that json lib integration so i could test the speed with the new rakudo but that's the only place i know to find that link 16:39
jnthn carlin: Dunno off hand but maybe check the nativecall README, or failing that the cstruct tests...
nwc10 jnthn: Josette is already busy that weekend, so can't make the APW in Salzburg :-(
It's also Pycon Ireland 16:40
TimToady would prefer to attend Pyron Iceland 16:41
jnthn nwc10: Aww
TimToady but I guess they've closed their airspace or so 16:42
dalek kudo/nom: 4887ae5 | (Elizabeth Mattijsen)++ | src/core/IO.pm:
Make filetests only check for existence once
btyler_ jdv79: github.com/kanatohodets/p6-json-jansson ? I'm still working on it, but I'd love feedback if you end up playing with it
jnthn Only over the immediate surroundings afaik 16:43
It's "only" a fissure eruption, last I checked...
jdv79 jansson - that's it
jnthn Probably quite pretty from a distance. :)
TimToady well, but that's where the Pyron is :)
jnthn But not ash spewing...yet. :)
jnthn got to see Eyjafjallajokull errupting and since then has kinda hoped to have chane to see more volcanic stuff at some point :) 16:45
*chance
TimToady saw Mt St Helens erupt...from 150 miles away... 16:47
16:48 domidumont joined
lizmat is getting lost in the maze that is IO IO::Handle IO::Path and IO::Spec 16:49
16:50 dayangkun left
jnthn lizmat: You and me both... :) 16:50
16:52 ajr_ left 16:53 donaldh left 16:54 rindolf joined
tony-o btyler_: looks p cool 16:54
btyler_ tony-o: thanks! I think I'm going to change the api a bit, "from-json(<foo>, True)" isn't a great way to declare that you want a handle on a jansson object instead of a core p6 data structure 16:56
lizmat hmmm.... according to spec, IO::Handle should *not* be doing IO::FileTestable
btyler_ but otherwise I feel ok about it. colomon++ spotted some bugs, the worst of which I fixed, will probably pull from his repo at some point this weekend
lizmat is that a problem in the spec? or a problem in the implementation ?
also: according to spec (S32/Str), a bare say or print should warn, not die 16:58
"As with C<print>, the compiler will warn you if you use a bare sub C<say>
without arguments."
About IO::FileTestable: the spec states: "If you apply that role to a class, that class must provide a C<Str> method which returns the full path." 17:00
TimToady: is "full path" the same as "absolute path", and is therefore the .Str of IO faulty atm ?
TimToady lost track of everything IO about 87,000 years ago 17:01
lizmat so "slushy" would be a good description of the current state of IO 17:02
tony-o lol
17:04 rindolf left, BizarreCake left
TimToady m: say lol 17:04
camelia rakudo-moar a98d26: OUTPUT«␤»
lizmat m: lol
camelia ( no output )
17:05 rindolf joined
lizmat
.oO( silently laughing out load )
 17:05
TimToady well, that's why people invented "lol"
cuz you can't hear it
17:05 dakkar left
lizmat for Dutch people, "lol" just means "fun" 17:06
with "lollig" being "funny"
huf yeah, that's how we use it too, pretty much
it's a word.
TimToady wonders if it's etymologically related to "droll"
huf hmm.... i think our "lol" is a bit more schadenfreude-liek 17:07
has the connotations of "what has this idiot done again?"
lizmat you can have "lol" on your own... 17:08
17:08 itz_ joined 17:12 kaare_ left
huf you can lol if you want to 17:13
you can leave your friends behind... 17:14
17:17 gfldex joined 17:19 donaldh joined 17:22 donaldh left 17:23 rindolf left, rindolf joined 17:25 itz joined 17:27 itz_ left
[Coke] (planeteria) site admin is working on the issue, please be patient. 17:28
17:28 zacts left
dalek kudo/nom: 6a6b5d0 | TimToady++ | src/core/List.pm:
gist should, like, give you the tl;dr of the list
 17:29
rl6-roast-data: d4ca3cc | coke++ | / (6 files):
today (automated commit)
17:30 BizarreCake joined
[Coke] hey, someone borked the spectests. 17:30
TimToady yes, there's still a bit of fallout from the anonyvars
[Coke] jvm: 49, moar: 68; moar-jit: 67; parrot: 731
TimToady but it's not all my fault 17:31
[Coke] (as of rakudo 54cd8fb, spectest 7fc8efd)
grondilu p6: subset LinComb of Hash where *.values.all ~~ Numeric; say (my % = a => 1, b => -1) ~~ LinComb
[Coke] likely story. :)
17:31 itz_ joined
TimToady m: say ++$ 17:31
camelia rakudo-jvm 4887ae: OUTPUT«(timeout)»
..rakudo-{parrot,moar} 4887ae, niecza v24-109-g48a8de3: OUTPUT«False␤»
rakudo-moar 4887ae: OUTPUT«1␤»
17:31 itz left
TimToady say what? 17:31
oh, that's you 17:32
or, rather, not you
[Coke] ponders a script that does a roast blame on people. 17:33
"200 new failures: possible culprits include:"
hoelzro [Coke]++ # knowing full well I will probably often appear in that listing
[Coke] Seriously, though, I'll see if I can clean those for tomororw. 17:34
nwc10 0) Jon Orwant, for throwing coffee mugs in the first place
etc
17:34 zacts joined
grondilu mayb camelia should reply with username: 17:34
I mean nick:
TimToady some of them may be transients 17:35
lizmat jnthn: from a performance point of view, would it make sense to just add a fallback to IO::Spec that would install the right IO::Spec::xxx method at runtime 17:37
rather than the class method -> variable class.method approach right now?
17:37 rindolf left 17:38 rindolf joined
lizmat to prevent this kind of mapping: method rel2abs( |c ) { $SPEC.rel2abs( |c ) } 17:38
TimToady m: say 1...* 17:39
lizmat possibly doing this in the optmizer?
camelia rakudo-moar 4887ae: OUTPUT«(timeout)»
nwc10 Rakudo 2014.05 doesn't build on NQP 2014.05 on MoarVM 2014.05 on PPC :-(
grondilu so with this new $++ thing I remembered there is a (state $)++ in RC's evolutionary algorithm. So I try it before I change it and I got:
Internal error: invalid thread ID in GC work pass
nwc10 so, didn't nail that bug back then
:-(
grondilu wonders if there was not some test for RC regressions 17:40
issue was in rosettacode.org/wiki/Evolutionary_a...thm#Perl_6
zacts hi
TimToady zacts: lo
zacts lo 17:41
TimToady and behold
17:41 shlomif joined, rindolf left, shlomif is now known as rindolf
grondilu (the $++ seems to work, though. No idea why) 17:43
grondilu meant the $++ version of the same code 17:44
zacts $++ looks kind of scary to me, coming from perl5. is that $+ with a +? or is $ the default variable? I have much to learn. 17:46
17:46 telex left
moritz zacts: an anonymous variable $ being ++ed 17:46
zacts ah ok
moritz m: say ++(state $) for ^10 17:47
camelia rakudo-moar 4887ae: OUTPUT«1␤2␤3␤4␤5␤6␤7␤8␤9␤10␤»
BenGoldberg Say grondilu, did you see my solution for your problem?
m: sub encode($n is copy) { my $m = 0; my $rest = ''; while ($n) { $n = $n.base(36); ++$m; $rest = $n~$rest; $n = $n.chars-1; }; $m.base(36)~$rest; }; say [lt] map { encode($_) }, ^2000 17:48
camelia rakudo-moar 4887ae: OUTPUT«True␤»
17:48 telex joined
grondilu m: sub encode($n is copy) { my $m = 0; my $rest = ''; while ($n) { $n = $n.base(36); ++$m; $rest = $n~$rest; $n = $n.chars-1; }; $m.base(36)~$rest; }; say map { encode($_).chars }, ^10 17:49
camelia rakudo-moar 4887ae: OUTPUT«1 2 2 2 2 2 2 2 2 2␤»
grondilu m: sub encode($n is copy) { my $m = 0; my $rest = ''; while ($n) { $n = $n.base(36); ++$m; $rest = $n~$rest; $n = $n.chars-1; }; $m.base(36)~$rest; }; say uniq map { encode($_).chars }, ^1000
camelia rakudo-moar 4887ae: OUTPUT«1 2 4␤»
grondilu hum not bad
BenGoldberg++
BenGoldberg :)
It does have a limitation: if $m exceeds 36, then it won't be lexigraphically sorted. On the other hand, you would need truly gargantuan numbers to do that, so it's not a practical concern. 17:52
grondilu ok
is decoding easy? 17:54
BenGoldberg I think so.
Split off the first character, decode that char in base 36; that's the number of times to go through the inner loop.
dalek ecs: 6ab1396 | TimToady++ | S02-bits.pod:
.gist only prints 100 entries of long lists
 17:56
BenGoldberg Then... hmm.....
TimToady m: say 1...*
ENOTYET
camelia rakudo-moar 4887ae: OUTPUT«(timeout)»
hoelzro are dynamic variables specific to a single compilation unit? 17:58
ex. if I have $*VAR_NAME in my module, and someone else uses that same variable name in another module, will they interfere?
nwc10 mberends: sorry, looks like I ran out of time in May and never figured out how to get Rakudo itself building on PPC
mberends: sorry, part 2, NQP used to work, but seems to be bust
bisecting 17:59
dalek kudo/nom: 32cb14d | TimToady++ | src/core/List.pm:
off-by-one on gist's tl;dr
18:00 lizmat_ joined
moritz hoelzro: perl6/doc htmlify.p6 is back to < 10min with latest rakudo 18:01
18:01 rindolf left 18:02 rindolf joined
TimToady guesses that's de-pessimizing push 18:02
moritz possible
hoelzro moritz: yeah, performance wasn't an issue when I tried it this morning 18:03
but the generated Pod doesn't match up yet
I'll fix that after work
jnthn TimToady: About push: the reason it became pessimized wasn't so much the special case for item and avoiding the *@foo, afaict
TimToady: It wsa the callsame and the fact that ruins the optimizer's ability to avoid building *%_ 18:04
TimToady: I *think* if we can do the special case and avoid he callsame we may come out ahead...
TimToady well, the old n-ary push was not as efficient as the new one, especially for many values 18:06
but it was certainly beating the unary one in this case
jnthn Oh, there's a new n-ary push? 18:07
that may well have shifted the balance since I did the original opt
I don't commit opts on guesswork; I certainly woulda measured that one at the time. 18:08
TimToady yes, it just does a splice rather than pushing one at a time
jnthn If you want to see an interesting profile about push, though, see the while push one in perl6-bench
It spent - last I checked - > 30% of its time in sink, somehow...
TimToady that's just one-arg
jnthn I think 'cus it was sinking the list after the push, since push returns the list itself... 18:09
TimToady I haven't got around to adding @a.push(@a) yet
which is blazingly faster than it used to be
for larger N
18:09 molaf__ joined
jnthn :) 18:09
18:11 lizmat_ is now known as lizmat
moritz tadzik++ # panda, rebootstrap.pl 18:12
BenGoldberg grondilu, here's a decoding function written in perl5: 18:14
eval: use warnings; sub unencode { local $_ = shift; s/(.)//; my $m = $1 or return 0; my $n = 1; for my $mm ( 1 .. $m ) { warn "[$m;$n;$mm;$_]\n"; s/(.{$n})// or die; $n = $1 }; $_ }; [ unencode("3114562949953421312") ];
I'm sure it can be easily be rewritten in perl6... but I'm not good at golfing, yet :) 18:15
lizmat the more I look at it, I think IO::Path is overenginieered and mostly useless
a lot of effort is spent converting IO::Path to Str and vice-versa 18:16
grondilu p6: role Foo {}; my %h = ("foo" but Foo) => 1; say %h.keys[0] ~~ Foo;
camelia rakudo-{parrot,jvm,moar} 4887ae, niecza v24-109-g48a8de3: OUTPUT«False␤»
18:16 perfectredneck joined
grondilu so Hash keys can not have a role attached to them? 18:17
lizmat typed hashes can contain anything, including objects with roles attached to them?
or is that not what you mean?
TimToady m: role Foo {}; my %h{Any} = ("foo" but Foo) => 1; say %h.keys[0] ~~ Foo;
camelia rakudo-moar 4887ae: OUTPUT«True␤»
TimToady grondilu: ^^ 18:18
grondilu oh, ok
grondilu is not quite familiar with typed structures
18:18 perfectredneck left
lizmat some subs even have a Str candidate in the spec: 18:18
multi sub visitdir(Str:D)
multi sub visitdir(IO::Path:D)
nine moritz: do you have some time to look at my strange method call problem?
jnthn nine: What is the problem, ooc? 18:19
nine: Is it a MOP-related thing?
jnthn might be able to give a hint if so... 18:20
nine jnthn: yes, you might. I managed to write a test with only inline Perl 5: github.com/niner/Inline-Perl5/comm...6a7895dcc4
jnthn: we're using add_fallback to pass through method calls on Perl5Objects to the Perl 5 interpreter: github.com/niner/Inline-Perl5/blob...5.pm6#L247 18:22
jnthn: but there's strange calls to "sink" and in the test, $!perl5 seems to end up being a Perl5Object instead of a PerlInterpreter
jnthn nine: Yes, that's because things in void context do .can('sink') or so 18:23
Perl5Object.^add_fallback(-> $, $ { True },
That means you're claiming you support *every* method.
You migh try adding a method sink() { self } in Perl5Object 18:24
Eventually, though, we really should write a Perl5HOW meta-object...
That way we can even have Perl 5 object dispatches go through the method cache rather than the slow path. 18:25
18:27 rindolf left 18:28 rindolf joined
nine jnthn: ok that takes care of the "sink" problem. But there's still the following error: paste.scsys.co.uk/420296 caused by $!perl5 being a Perl5Object instead of a PerlInterpreter. I really have no idea why. The code seems to be very straight forward. 18:29
jnthn My latest talk slides are now linked from jnthn.net/articles.shtml 18:30
18:31 brrt joined
brrt \o 18:31
18:31 guru joined 18:32 guru is now known as ajr_
dalek line-Perl5/transparent_method_call_regression: 71dcdeb | nine++ | lib/Inline/Perl5.pm6:
Add a sink method to Perl5Object as suggested by jnthn++

Apparently can('sink') may be called on Perl5Objects in some situations. Since we claim to support _any_ method, we have to provide a somewhat sane implementation.		 18:32
ajr_ 150 miles away sounds like a good distance at which to view Mt. St. Helens erupt. 18:33
dalek ast: 48160f2 | TimToady++ | S03-operators/context.t:
anonymous & doesn't make a lot of sense
BenGoldberg If I could, I'd watch Mt. St. Helens erupt from Iceland, and vice versa. ;) 18:34
dalek kudo/nom: 9a9da9f | TimToady++ | src/Perl6/Grammar.nqp:
anonymous & doesn't make a lot of sense

  (and causes bugs in rx parsing of &)
jnthn nine: Hmm....what on earth...
nine: You're sure it's actually that the value is changing? 18:35
TimToady m: say 1...* 18:36
camelia rakudo-moar 32cb14: OUTPUT«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 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 8…»
lizmat given a file "foo" with "foo" in it, and "bar" with "bar in it:
this seems *very* wrong to me:
$ 6 'my $h = "foo".IO.open; $h.open("bar"); say $h.lines'
bar
TimToady m: say 1,1...*
camelia rakudo-moar 32cb14: OUTPUT«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 53 54 55 56 57 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 ...␤»
18:36 denis_boyun_ joined
dalek line-Perl5/call_backs: a563d94 | nine++ | / (3 files):
Failing call backs
 18:36
TimToady notice that's rakudo's ... there
lizmat in other words, you can open another file on an open file handle, and it will read from the new file from there one 18:37
*on
is this intended as a feature? or is this a bug?
TimToady but what kind of output is that....
BenGoldberg lizmat, That's not very different from how perl5 does it.
Or stdio.
Reusing a handle isn't that uncommon. 18:38
brrt ooh TimToady i meant to bug you about the recent div spec change
lizmat that it changes the file handle *itself* ?
BenGoldberg ?
jnthn TimToady: wtf, why the 1 1 1...
brrt that's a spesh or jit error 18:39
TimToady m: say (1,1...*)
camelia rakudo-moar 32cb14: OUTPUT«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 53 54 55 56 57 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 ...␤»
TimToady cool!
brrt bug!
TimToady m: say (1,1.1...*)
camelia rakudo-moar 32cb14: OUTPUT«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 53 54 55 56 57 6.7 6.8 6.9 7 7.1 7.2 7.3 7.4 7.5 7.6 7.7 7.8 7.9 8 8.1 8.2 8.3 8.4 8.5 8.6 8.7 8.8 8.9 9 9.1 …»
brrt :-)
TimToady wow 18:40
nine jnthn: I put a warn $!perl5 in there and I get a line with: Perl5Object<139780065295616> in block at /home/nine/Inline-Perl5/lib/Inline/Perl5.pm6:237
jnthn 57 is a magical number
BenGoldberg m: say ('a',1.1...*);
camelia rakudo-moar 32cb14: OUTPUT«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 53 54 55 56 57 -54.9 -55.9 -56.9 -57.9 -58.9 -59.9 -60.9 -61.9 -62.9 -63.9 -64.9 -65.9 -66.9 -67.9 -68.9 -69.…»
BenGoldberg m: say (3,'a'...*);
camelia rakudo-moar 32cb14: OUTPUT«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 53 54 55 56 57 be bf bg bh bi bj bk bl bm bn bo bp bq br bs bt bu bv bw bx by bz ca cb cc cd ce cf cg ch ci c…»
brrt what the
TimToady m: say (0, 1/10 ... 1)
camelia rakudo-moar 32cb14: OUTPUT«1 2 3 4 5 6 7 8 9 10 11␤»
18:40 SamuraiJack left
jnthn 100 = number of iterations of a loop for OSR to trigger + 8 = number of trace iterations... 18:40
TimToady there's a spectest failing from this one
brrt i can imagine 18:41
TimToady specifically that 0, 1/10 one
brrt what i wanted to bug about: integer division rounds to <negative infinity | zero >
which one?
TimToady integration/advent2013-day15.t
lizmat BenGoldberg: indeed, I'm surprised... 18:42
and this is a good thing because ?
I mean, in Perl 6 the old handle at the OS level will probably stay open, as we don't DESTROY
TimToady m: say ~(0, 1/10 ... 1)
camelia rakudo-moar 32cb14: OUTPUT«0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1␤»
nine jnthn: $!perl5 is actually the same as self in that case.
jnthn lizmat: Feels like a wtf to me...
TimToady ooh, it's something about my gist change 18:43
BenGoldberg I'm not saying it's a good thing, just not a surprising thing.
TimToady m: say (0, 1/10 ... 1).gist
camelia rakudo-moar 32cb14: OUTPUT«1 2 3 4 5 6 7 8 9 10 11␤»
TimToady huh
lizmat it's surprising to me.. :-)
jnthn lizmat: Well, the VM GCs IOHandles and will close them...
lizmat: So we're not *too* bad from that angle.
lizmat: Unless the original handle makes gen2.
From data I'm seeing at present, full collection frequency wants to becme a function of bytes promoted 18:44
lizmat ok, but still, you run the risk of running out of OS handles pretty soon in some applications that way
BenGoldberg If you reopen the same handle to a different file, the old c-level filedescriptor gets closed right away, instead of having to wait for GC.
TimToady oh, duh, I'm an idgit
BenGoldberg Or at least, it should.
lizmat pretty sure it doesn't in rakudo atm 18:45
jnthn BenGoldberg: That's now how things actually work unless we make it so
And what lizmat said
lizmat it's just doing a nqp::bindattr
and that doesn't know about filehandles
jnthn To me, trying to do .open on a dhnel a second time is likely to be a bug.
*handle
So better to whine, I think...
lizmat feels "DOS"sy to me
TimToady testing fix 18:46
lizmat as in MS-DOSsy
BenGoldberg If we run out of OS-level filehandles, does it trigger a GC?
nwc10 BenGoldberg: no. (I know that already)
TimToady though I don't understand why it worked around my nested $_ binding bug halfway through 18:47
probably indicates a different bug
18:47 ggherdov left
brrt we have a known inlining-related bug in moar with $_ 18:48
jnthn brrt: We...do? 18:49
brrt aye
dalek kudo/nom: 1ff1d30 | TimToady++ | src/core/List.pm:
don't use $_ two different ways

previous .gist patch did just that, TimToady--
jnthn Ticket?
brrt you probably have missed that since flight
let me see that
TimToady would that account for my buggy code getting the right (wrong) $_ after 57? 18:50
brrt github.com/MoarVM/MoarVM/issues/127 jnthn
jnthn TimToady: As I said, 57 is a *very* suspect number
lizmat TimToady: not sure whether that fix isn't just circumventing a spec bug
*spesh / jit
TimToady what I just said
18:51 ggherdov joined
TimToady it just showed the other bug by accident 18:51
brrt i'd note that the old behaviour (div int rounding to zero) never caused one spectest to fail :-)
TimToady by compensating for my mistake partway through :)
18:52 elfphiltrum joined, elfphiltrum left
TimToady unfortunately, I tested the sequence 1...* so I didn't see the error, and wrote off the failure in integration as some Rat problem 18:53
wasn't until I tried the 1,1...* here to get the gist to fit into camelia's limit that I spotted the goofup 18:54
lizmat finally sees the goofup 18:55
TimToady yes, it sort of doesn't register at first 18:56
m: say 1,1,*+*...* 18:57
camelia rakudo-moar 32cb14: OUTPUT«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 53 54 55 56 57 591286729879 956722026041 1548008755920 2504730781961 4052739537881 6557470319842 106102098577…»
TimToady we now get an answer to that from camelia
jnthn brrt: Thanks, got a fix for #127 locally
brrt jnthn++
TimToady even if it's still the wrong answer in the first 57 spots
jnthn Nemozem inline le getlexouter... :) 18:58
TimToady wow, and that's the exact bug I hit...how lucky I am! :)
brrt :-) 18:59
TimToady escape analysis should be really easy on anonymous variables 19:00
jnthn Pushed the fix.
brrt jnthn++ 19:01
19:01 rindolf left
jnthn And yeah, guess that one got lost in my travel blind spot :) 19:01
FROGGS o/
jnthn o/ FROGGS 19:02
19:02 rindolf joined
TimToady jnthn: you weren't living in interesting times 19:03
brrt \o FROGGS
TimToady good grief, how did it get to be lunchtime already? 19:04
jnthn You think you got it bad. It's got to be 3am here... :P
TimToady which city? 19:05
jnthn Shenzhen
TimToady wow, I even know where that is now...
jnthn :)
FROGGS m: [&baz]
camelia rakudo-moar 32cb14: OUTPUT«===SORRY!=== Error while compiling /tmp/bDTNoZdkZU␤Undeclared routine:␤ &baz used at line 1. Did you mean '&bag'?␤␤» 19:06
jnthn My last time in China was in Beijing. My weather app told me the sky was clear every day. I saw a blue sky one day out of 7 thanks to smog.
FROGGS m: [[&baz]]
camelia rakudo-moar 32cb14: OUTPUT«===SORRY!=== Error while compiling /tmp/NBXzhwwfJT␤Undeclared routine:␤ &baz used at line 1. Did you mean '&bag'?␤␤»
FROGGS ohh, somebody fixed that
jnthn Here I've seen it every day. :)
So at least the air here isn't so bad. :)
19:08 Ven joined
Ven chuckles at colomon++'s IRC client inserting "…" instead of "..." 19:11
colomon Ven: it's downright evil. 19:12
Ven makes backlogging a bit funier :-)
colomon distressingly, I think it may be general OS X behavior instead of just my IRC client. :\ 19:13
19:13 mberends left
TimToady I suppose .gist could use … too 19:15
Ven colomon: it is, but it doesn't happen on ,ine 19:16
TimToady your l has gone commatose 19:17
Ven stackoverflow.com/questions/1917898...able-types <- oh, Perl6 can do that ;) 19:18
Ven is writing an answer
19:19 rindolf left 19:20 rindolf joined
jnthn Time for some sleep, before I accidentally get myself back onto European time... & 19:20
brrt sleep well :-)
TimToady o/ 19:21
now we can Break All The Things
actually, if .gist is supposed to be human readable, it should use etc. instead of ... :) 19:23
well, ... translates better to other languages, except maybe Latin 19:24
Ven stackoverflow.com/questions/1917898...6#25574936 <- answered with PErl6 :)
synopsebot Link: rt.perl.org/rt3//Public/Bug/Displa...d=25574936
Ven calm down, synopsebot
we're not there yet.
brrt m: my int $i = 8; my int $y = -3; say $i div $y; 19:25
camelia rakudo-moar 32cb14: OUTPUT«-3␤»
brrt m: my int $i = 8; my int $y = 3; say $i div $y;
camelia rakudo-moar 32cb14: OUTPUT«2␤»
Ven ^ if anybody has to fix my answer, feel free to :)
brrt p6: my int $i = 8; my int $y = -3; say $i div $y; 19:26
camelia rakudo-jvm 32cb14: OUTPUT«(timeout)»
..rakudo-{parrot,moar} 32cb14: OUTPUT«-3␤»
..niecza v24-109-g48a8de3: OUTPUT«===SORRY!===␤␤Malformed my at /tmp/tmpfile line 1:␤------> my⏏ int $i = 8; my int $y = -3; say $i div ␤␤Parse failed␤␤»
19:27 domidumont left
brrt ok, i won't be complaining anymore :-) 19:30
TimToady p6: my int $i = 8; my int $y = 0; say $i div $y; 19:32
camelia rakudo-jvm 1ff1d3: OUTPUT«(timeout)»
..niecza v24-109-g48a8de3: OUTPUT«===SORRY!===␤␤Malformed my at /tmp/tmpfile line 1:␤------> my⏏ int $i = 8; my int $y = 0; say $i div $␤␤Parse failed␤␤»
..rakudo-moar 1ff1d3: OUTPUT«Division by zero␤ in block <unit> at /tmp/tmpfile:1␤␤»
..rakudo-parrot 1ff1d3: OUTPUT«(signal )»
brrt why does that cause a timeout on the jvm? 19:33
or is that the jvm's way of dying
TimToady j: say "hi" 19:34
camelia rakudo-jvm 1ff1d3: OUTPUT«hi␤»
TimToady beats me
Ven TimToady: camelia-r-j only is able to say. 19:35
19:35 Ven left
TimToady m: say 1,1,*+*...* 19:36
camelia rakudo-moar 1ff1d3: OUTPUT«1 1 2 3 5 8 13 21 34 55 89 144 233 377 610 987 1597 2584 4181 6765 10946 17711 28657 46368 75025 121393 196418 317811 514229 832040 1346269 2178309 3524578 5702887 9227465 14930352 24157817 39088169 63245986 102334155 165580141 267914296 433494437 70140873…»
TimToady that's better
j: say 1,1,*+*...*
camelia rakudo-jvm 1ff1d3: OUTPUT«1 1 2 3 5 8 13 21 34 55 89 144 233 377 610 987 1597 2584 4181 6765 10946 17711 28657 46368 75025 121393 196418 317811 514229 832040 1346269 2178309 3524578 5702887 9227465 14930352 24157817 39088169 63245986 102334155 165580141 267914296 433494437 701408733…»
19:37 brrt left 19:38 dwarring joined
moritz mm: say (1, 1, *+* ... *).gist.chars 19:41
m: say (1, 1, *+* ... *).gist.chars
camelia rakudo-moar 1ff1d3: OUTPUT«1174␤»
TimToady 100 seems about the right limit to me; unlikely to interfere with printing small lists, and not so long it scrolls off your screen 19:45
I also don't bother to check .infinite 19:46
if it's not infinite, I'd have to check .elems, and that could be huge, or effectively infinite
m: my @a = 1..*; say @a 19:47
camelia rakudo-moar 1ff1d3: OUTPUT«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 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 8…»
TimToady works there too
19:56 mberends joined 19:57 Ven joined, Ven left 20:02 mberends left 20:04 Ven joined 20:05 cognome joined
nwc10 jnthn: other PPC "bug" was solved with fadd779... Update to simpler MAST::Label API. 20:13
ie was trying to use too old an NQP 20:14
raiph
.oO( subset Probability of Num where 0.0 <= * <= 1.0; ??? .oO( % of Probability always total 1.0; ??? ) )
20:16 beastd joined, rindolf left
nwc10 botherit, Rakudo still unhappy 20:17
Ven sometimes, I see posts about TCL, and I really laugh. 20:19
Like, well, you may badmouth php, but when you see how terrible tcl is... hahaha
20:21 [Sno] joined
[Coke] I would say that tcl has a consistency that php lacks. 20:21
20:22 molaf__ left
Ven [Coke]: it's consistently offering the worst the world has to offer 20:24
tadzik naah
zacts I've decided to specifically put on my portfolio/resume that I don't know php, and I specifically won't do anything php related. 20:25
they'll have to use someone else for that 20:26
20:26 MilkmanDan left
Ven zacts: I put that I know it, but the last time a recruiter called for a php job, I said "sorry, I'd rather be programming" 20:27
20:31 MilkmanDan joined
[Coke] Ven: I respectfully disagree. 20:31
Ven [Coke]: let's :)
20:32 Ven left
nine How can a typed private attribute of an object get a value of a different type?! 20:44
20:45 ivanshmakov left
TimToady binding? 20:46
20:46 Ven joined
TimToady m: my Int $x; $x := "foo"; say $x.WHAT 20:46
camelia rakudo-moar 1ff1d3: OUTPUT«Type check failed in binding; expected 'Int' but got 'Str'␤ in any bind_error at src/vm/moar/Perl6/Ops.nqp:221␤ in block <unit> at /tmp/2vMWNk6dgO:1␤␤»
TimToady nope, checks there
nine TimToady: this is the complete source of the class: gist.github.com/niner/11490afd74ede2af7b8b 20:47
20:49 ivanshmakov joined
TimToady you aren't deriving Perl5Object from Perl5Interpreter or some such? 20:51
nine TimToady: not that I know of
TimToady maybe the Opaque is overflowing somehow?
you on 32 or 64-bit? 20:52
nine 64
TimToady if something is assuming 32 and something else assuming 64, you could get that sort of damage
nine Changeing the order of attributes in that class changes the error message. I now get: No such method 'call' for invocant of type 'Perl6::Metamodel::ContainerDescriptor'
diakopte1 o_O 20:53
nine And if I put a dummy Int in between those two attributes I get Cannot call method 'isa' on a null object 20:54
I suspect, my code is not at fault here ;) 20:55
diakopte1 not solely, anyway ;)
20:57 nbrown__ joined
nine If I change the private attributes to public ones and remove the BUILD method the errors disappear. 20:59
21:00 tphilipp joined 21:03 itz joined
moritz nine: if you make them public, but keep the BUILD, what happens? 21:03
nine moritz: Cannot find method 'STORE' in submethod BUILD at /home/nine/Inline-Perl5/lib/Inline/Perl5.pm6:262 21:05
Ven no cont? 21:06
raiph m: subset Probability of Num where 0.0 <= * <= 1.0; # Is this a solution to question part `1` in SO question (as answered by ven++ at stackoverflow.com/a/25574936/1077672)? What about a solution to part `2`?
camelia ( no output )
21:06 itz_ left, dolmen joined
Ven hahaha someone downvoted my perl6 answer on that SO question 21:07
nine Ven: cont?
Ven nine: container
21:08 nbrown__ left 21:09 amkrankr1leuen left 21:10 amkrankruleuen joined, amkrankruleuen left, amkrankruleuen joined
moritz nine: this is getting weirder and weirder 21:10
21:11 mj41 left
Ven TimToady: subsets can't get a type parameter, can they? So, we can't subset a role[::T], right? 21:11
nine moritz: jnthn gave some advice: irclog.perlgeek.de/perl6/2014-08-29#i_9266406 21:13
21:13 avuserow_ left
Ven 's trying to find the second part to raiph++'s answer 21:14
raiph
.oO( raiph thought so, in a bubbly kinda way )
 21:15
nine moritz: I fear I know just too little about Perl 6 to make real progress here :/
Ven role DPD[::T] { has @.values where [+] == 1; } 21:16
raiph: ^ how does that sound?
raiph m: role DPD[::T] { has @.values where [+] == 1; } # ven: really? 21:17
camelia rakudo-moar 1ff1d3: OUTPUT«===SORRY!=== Error while compiling /tmp/da9kDOkt2L␤Preceding context expects a term, but found infix == instead␤at /tmp/da9kDOkt2L:1␤------> ole DPD[::T] { has @.values where [+] ==⏏ 1; } # ven: really?␤»
Ven [+] *, whoops
raiph m: role DPD[::T] { has @.values where [+] * == 1; } # ven: really!? 21:18
TimToady * doesn't usually work in argument position
camelia rakudo-moar 1ff1d3: OUTPUT«Unhandled exception: Could not instantiate role 'DPD':␤Not enough positional parameters passed; got 1 but expected 2␤ at <unknown>:1 (/home/p6eval/rakudo-inst-2/languages/nqp/lib/Perl6/Metamodel.moarvm:specialize:4294967295)␤ from src/gen/m-Metamo…»
Ven true.
raiph m: role DPD { has @.values where [+] * == 1; } 21:19
camelia ( no output )
TimToady * won't work there
21:19 tphilipp left
Ven TimToady: I'm using *.values now anyway 21:19
raiph: updated
TimToady you still have a proecedence problem 21:20
Ven 1 ==, then !
TimToady or [+](@.values)
moritz nine: well, the part about method sink { self } is worth trying, at least 21:21
TimToady or { ([+] @$_ ) == 1 } or so
Ven @.values?
it's a Map, just [+](.values) I think
TimToady okay 21:22
Ven because, well, since it's a "Map"(whatever *that* is), I switched back to $.values
I'll rename it...
dalek line-Perl5: 67dfcac | moritz++ | lib/Inline/Perl5.pm6:
give Perl5Object a no-op sink method
moritz nine: if it doesn't work, just revert the commit that made the attributes private 21:23
21:25 mberends joined 21:30 BizarreCake left 21:37 dwarring left 21:38 Ven left 21:41 ggoebel111119 joined 21:46 guru joined, guru is now known as Guest59771 21:47 jdv79_ joined 21:48 itz_ joined 21:51 yeltzooo9 joined 21:52 itz left, ajr_ left, jdv79 left 22:01 rindolf joined 22:02 cognome left 22:03 cognome joined 22:07 cognome left 22:11 _sri joined, silug joined, \00 joined, ingy joined, sjohnson joined
dalek line-Perl5/call_backs: 1baec29 | nine++ | / (3 files):
Implement parameterless callbacks
 22:14
line-Perl5/call_backs: 5ea1c61 | nine++ | / (2 files):
Implement passing parameters to call backs

We now can pass Perl 6 objects to Perl 5 code and have that code call methods on the Perl 6 objects and pass parameters to them.		 22:16
22:16 Guest59771 left
dalek line-Perl5/call_backs: 10aa036 | nine++ | / (3 files):
Implement parameterless callbacks
 22:16
line-Perl5/call_backs: 770eb5c | nine++ | / (2 files):
Implement passing parameters to call backs

We now can pass Perl 6 objects to Perl 5 code and have that code call methods on the Perl 6 objects and pass parameters to them.
line-Perl5: 10aa036 | nine++ | / (3 files):
Implement parameterless callbacks
line-Perl5: 770eb5c | nine++ | / (2 files):
Implement passing parameters to call backs

We now can pass Perl 6 objects to Perl 5 code and have that code call methods on the Perl 6 objects and pass parameters to them.
diakopte1 dalek--
22:19 mberends left
spider-mario nqp on parrot builds with various -Werror=… parameters 22:21
this causes problems, as it also builds dyncall with those
but dyncall doesn’t try to avoid those warnings 22:22
22:25 skarn joined 22:26 skarn is now known as Guest21595
spider-mario and once built, it doesn’t appear to know its version number 22:28
$ nqp-p --version
This is nqp version built on parrot 6.7.0 revision 0
22:35 BenGoldberg left 22:36 cognome joined 22:39 mberends joined 22:41 donaldh joined
donaldh noticed some spam in RT that needs deleted rt.perl.org/Ticket/Display.html?id=122651 22:41
22:42 Guest21595 left, Guest21595 joined, Guest21595 is now known as skarn 22:45 donaldh left 22:46 cognome_ joined, cognome left
colomon nine++ 22:52
spider-mario even after patching nqp so it prints “This is nqp version 2014.08 built on parrot 6.7.0 revision 0”, Rakudo’s Configure.pl thinks it’s too old… what’s going on? :/ 22:55
dalek kudo/nom: 7467ee8 | (Rob Hoelz)++ | src/core/Pod.pm:
Add :content in Pod::Block constructor, but warn about deprecation
 22:57
c: 0a5229d | (Rob Hoelz)++ | lib/Pod/Convenience.pm6:
Don't use deprecated contents attribute
 22:58
22:58 xragnar_ joined, xragnar_ is now known as xragnar
hoelzro .tell moritz I fixed up and checked the docs; they're fine now! 22:58
yoleaux hoelzro: I'll pass your message to moritz.
23:00 denis_boyun_ left 23:16 avuserow_ joined 23:20 mberends left 23:24 gfldex left
dalek kudo/nom: d8c8342 | TimToady++ | src/core/List.pm:
first whack at implementing .plan
 23:25
23:26 mbaby joined
TimToady > perl6 -e 'my @a = 1,2,3; @a.plan(4..*); say @a' 23:26
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 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 ...
23:27 mbaby left
TimToady unfortunately, hangs if you try to subscript into the lazy part 23:28
23:28 mbaby joined
TimToady but it's a start 23:28
23:29 mbaby left, je joined 23:37 dolmen left, itz joined 23:38 nbrown__ joined, cognome_ is now known as cognome, mberends joined 23:39 itz_ left 23:51 nbrown__ left, itz_ joined 23:52 itz left 23:58 itz joined 23:59 itz_ 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!