»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'p6: say 3;' or rakudo:, std:, or /msg camelia p6: ... | irclog: irc.perl6.org | UTF-8 is our friend!
Set by moritz on 25 December 2014.
adu why can't I access enum members? 00:07
yoleaux 12 Jan 2015 11:38Z <timotimo> adu: perl6 .new() isn't like __init__ or __new__, it's like python's .frob or .blubber or .yoink; also, new calls bless calls BUILDALL calls BUILD unlike what geekosaur said, and skids wrongly wrote bless(*, ...), but the * as first argument is wrong nowadays
12 Jan 2015 11:44Z <moritz> adu: see perlgeek.de/blog-en/perl-6/object-c...ation.html and doc.perl6.org/language/classtut#Constructors
masak adu: indeed, why can't you?
m: enum Foo <a b>; say b
camelia rakudo-moar 050686: OUTPUT«b␤»
masak I can ;)
adu pastie.org/9828545 00:08
masak no idea from just that paste. 00:09
adu the only difference I can think of is that CAST::Op is a class, and CAST::Type is a role 00:10
TimToady so name it with the class it ends up with instead?
adu I can't just s/role/class/ because then I'd have to change 12 files 00:11
TimToady
.oO("We can't fix this--we have 11 users already...")
 00:12
adu ok, it works if CAST::Type is a class 00:13
TimToady enum names work more like sub names than like methods
arguably, it might be possible to make the role names work instead, but I'd like jnthn++'s opionion on that 00:14
adu so I should move the enum somewhere else? 00:15
can I put enums in their own file?
TimToady roles are primarily for abstracting sets of methods, and have not been heavily thought through for other generic purposes
I don't see why not, offhand 00:16
could certainly be put into their own module and exported
assuming we fixed that
00:19 sirdancealot left
TimToady m: module Stuff { enum MyEnum is export <foo bar baz>; }; import Stuff; say baz 00:20
camelia rakudo-moar 050686: OUTPUT«baz␤»
TimToady yes, fixed last Friday
I guess Perl 6 is mature now :P
TimToady kinda thought that whole discussion was a little silly to anyone who has had a 14-year-old in the house 00:21
they generally oscillate in maturity ±10 years or so
so you should not be astonished if Perl 5 acts 24 one day and 4 the next day... 00:23
00:27 colomon left
TimToady s/5/6/ # can't spell today 00:28
00:28 gfldex left
TimToady
.oO(Can you call it spelling if it's numbers? If not, what do you call it?)
 00:29
masak counting. 00:31
TimToady but I wasn't counting, I was just typing a digit
adu uh oh 00:32
Type check failed in assignment to '$!op'; expected 'CAST::OpKind' but got 'OpKind'
masak ooh :)
adu I'm so confused
TimToady if I accidentally typed 555 instead of 666, would that be miscounting the Number of the Beast?
that seems wrongish 00:33
adu TimToady: is it supposed to do that? 00:34
TimToady my last utterance was your answer
exporting a type should not make it a different type 00:35
adu github.com/andydude/p6-c-parser/bl...OpKind.pm6 00:36
japhb
.oO( 668 -- the number of the neighbor of the beast )
 00:37
adu but what's really puzzling me is that it should be CAST::OpKind::OpKind::call that works, but CAST::OpKind::call also works
TimToady CAST::OpKind is the name of the module, not the name of the enum
adu but then why would both work? 00:38
TimToady because enum exports to its immediate surroundings
m: my enum X <a b c>; say X::a: say a
adu ah
camelia rakudo-moar 050686: OUTPUT«a␤Cannot call 'say'; none of these signatures match:␤:(Obsolete:D $: *%_)␤:($: *%_)␤ in block <unit> at /tmp/76vcJh8XGq:1␤␤»
TimToady oh, X is taken 00:39
m: my enum XX <a b c>; say XX::a: say a
camelia rakudo-moar 050686: OUTPUT«a␤Cannot call 'say'; none of these signatures match:␤:(Obsolete:D $: *%_)␤:($: *%_)␤ in block <unit> at /tmp/UjQGzde8PW:1␤␤»
TimToady um...
masak colon.
TimToady ah
m: my enum XX <a b c>; say XX::a; say a
camelia rakudo-moar 050686: OUTPUT«a␤a␤»
TimToady error message is a bit LTA though 00:41
m: my enum XX <a b c>; say XX::a.^methods
camelia rakudo-moar 050686: OUTPUT«enums pair perl pick roll Int invoke Str Numeric gist kv key value Int Num Rat FatRat abs Bridge chr sqrt base expmod is-prime floor ceiling round lsb msb narrow Range sign conj rand sin asin cos acos tan atan atan2 sec asec cosec acosec cotan acotan sinh …»
TimToady m: my enum XX <a b c>; say XX::a.say(a) 00:42
camelia rakudo-moar 050686: OUTPUT«Cannot call 'say'; none of these signatures match:␤:(Obsolete:D $: *%_)␤:($: *%_)␤ in block <unit> at /tmp/1rxt9WMdaV:1␤␤»
adu my enum XX <a b c>; my XX $kind; say $kind;
m: my enum XX <a b c>; my XX $kind; say $kind;
camelia rakudo-moar 050686: OUTPUT«(XX)␤»
adu maybe the error is actually that it's Nil
00:44 dayangkun joined 00:49 ajr joined, ajr is now known as Guest92299 00:51 skids joined 00:52 Guest92299 left 00:53 kaare__ joined, Mso150_e joined 00:54 Mso150 left
masak 'night, #perl6 00:54
00:58 colomon joined 01:01 ajr_ joined 01:02 ajr_ left 01:11 cognominal joined 01:18 Ven joined 01:20 yeahnoob joined
adu TimToady: I figured it out 01:21
I had to declare it "my OpKind" instead of "my CAST::OpKind" 01:22
I still don't quite understand why there is any difference between the two
s/my/has/
TimToady CAST::OpKind is the name of the module, not anything in it or exported from it
adu ah
so CAST::OpKind::OpKind might have worked 01:23
TimToady nod
adu needs to think more
TimToady thinks adu++ needs to give the enum a different name from the module 01:24
adu lol
maybe CAST::Ops::OpKind?
TimToady thinks adu++ needs to give the module a different name from the enum, rather 01:25
adu :)
TimToady ponders the pragmatic ambiguity of "make these things different" 01:26
01:27 broski left 01:29 Rounin joined
raydiak good evening, #perl6 01:37
m: Buf.new.[0, 1] = 2, 3 # this should be reported as a bug, correct? 01:38
camelia rakudo-moar 050686: OUTPUT«Cannot modify an immutable Int␤ in method STORE at src/gen/m-CORE.setting:8022␤ in sub postcircumfix:<[ ]> at src/gen/m-CORE.setting:2829␤ in block <unit> at /tmp/r28g6o8w_a:1␤␤»
TimToady already reported, I think
raydiak thank you
TimToady but you might make sure there's a test :)
01:39 dayangkun left
raydiak looks 01:39
hm...I see buf scattered across several test files, not sure where it would belong, still looking 01:42
adu how do I get a character from a string? 01:47
raydiak! 01:48
raydiak maybe it would go in S09-subscript/slice.t ?
adu!
there's also S03-operators/buf.t , kinda feels like a better place in with other buf stuff, but idk if slicing tests would go with "operators" or not 01:51
01:52 BenGoldberg joined
raydiak adu: if you mean "a character by index", that's .substr 01:52
adu raydiak: thanks 01:53
raydiak you're welcome :)
01:53 Ven left
raydiak bah gotta run for a bit already...back in a while 01:53
02:05 vukcrni left 02:06 vukcrni joined 02:24 cognominal left
adu how do I parse an int? 02:24
m: Int("5")
camelia ( no output )
adu m: say Int("5")
camelia rakudo-moar 050686: OUTPUT«5␤»
adu wow 02:25
japhb adu: Or just +, if you can accept any Numeric value. 02:28
02:28 adu left
colomon note, however 02:28
japhb m: say +"5.24", +"3", +"4+5i";
camelia rakudo-moar 050686: OUTPUT«5.2434+5i␤»
colomon m: say Int("5 fish") 02:29
camelia rakudo-moar 050686: OUTPUT«Cannot convert string to number: trailing characters after number in '5⏏ fish' (indicated by ⏏)␤ in method Int at src/gen/m-CORE.setting:13942␤ in method Int at src/gen/m-CORE.setting:6237␤ in block <unit> at /tmp/4DY6uXgHT9:1␤␤»
japhb m: .say for +"5.24", +"3", +"4+5i";
camelia rakudo-moar 050686: OUTPUT«5.24␤3␤4+5i␤»
japhb m: .WHAT.say for +"5.24", +"3", +"4+5i";
camelia rakudo-moar 050686: OUTPUT«(Rat)␤(Int)␤(Complex)␤»
skids I must say the error messages have become so awesome it's almost easier to code perl6 scripts than perl5. 02:30
They usually tell me exactly where the typo/thinko is and what it is.
The only exception being when Any hits a string or numeric context (line number often wrong) and when a undefined type object is hit (doesn't say what type the undef was). 02:33
02:34 bjz joined 02:39 bjz left
b2gills m: multi sub test ( int \a ){ say a.WHAT; say 'int' }; multi sub test ( Int \a ){ say a.WHAT }; test (my int $ = 1); 02:39
camelia rakudo-moar 050686: OUTPUT«(Int)␤int␤»
b2gills Nevermind, of course it promotes it to an Int if you use a method on it. 02:40
skids m: multi sub test ( int \a ){ say a ~~ int; say 'int' }; multi sub test ( Int \a ){ say a.WHAT }; test (my int $ = 1); 02:43
camelia rakudo-moar 050686: OUTPUT«False␤int␤»
02:43 cognominal joined
raydiak back 02:43
putting my buf slice test in S09-subscript/slice.t unless anyone objects
b2gills The most difficult things in programming is naming things, the second most difficult thing is figuring out where to put the tests 02:47
raydiak and I guess third is finding the ticket number I should attach to the test :) 02:48
ah well, will report just in case, better to have a dup than none 02:51
02:56 adu joined
skids m: my %d; %d.push("b" => "c"); %d.perl.say; %d.push(:a(1)); %d.perl.say; 03:04
camelia rakudo-moar 050686: OUTPUT«("b" => "c").hash␤("b" => "c").hash␤»
_4d47 whats the best way to include ./lib to run a module tests ? ufo, @*INC, export PERL6LIB ? 03:06
colomon prove -e "perl6 -Ilib" t/whatever
dunno if that's best, but it's what I always do. :)
raydiak m: sub foo { @_.perl.say }; foo("b" => "c"); foo(:a(1)); 03:07
camelia rakudo-moar 050686: OUTPUT«Array.new("b" => "c")␤Unexpected named parameter 'a' passed␤ in sub foo at /tmp/PWMdgm4LZB:1␤ in block <unit> at /tmp/PWMdgm4LZB:1␤␤»
03:07 kurahaupo joined
_4d47 cool thanks i like it 03:08
skids I've been staring to move away from @*INC to "use lib './lib';"
raydiak is "#?rakudo skip 'RT #123594'" the right thing to do when adding a test which currently fails? 03:16
synopsebot Link: rt.perl.org/rt3//Public/Bug/Displa...?id=123594
raydiak idk much about those magic little comments we have in roast
raydiak figures out how to read the README and use fudge to answer his own question (the answer seems to be yes) 03:24
dalek ast: b8e8171 | raydiak++ | S09-subscript/slice.t:
Add test for RT #123594
 03:25
synopsebot Link: rt.perl.org/rt3//Public/Bug/Displa...?id=123594
03:28 noganex joined
skids heh. bot chaining win. 03:29
03:31 noganex_ left 03:37 adu left 03:49 telex left 03:50 telex joined 04:02 kurahaupo left 04:14 adu joined 04:25 anaeem1 joined 04:28 yeahnoob left, _4d47 left 04:31 yeahnoob joined
adu lol 04:34
I think I just came up with a name for this 04:36
transformation between c syntax and semantics
04:37 yeahnoob left 04:38 yeahnoob joined, yeahnoob left 04:39 yeahnoob joined, yeahnoob left, yeahnoob joined 04:40 yeahnoob left, yeahnoob joined 04:46 tinyblak joined, tinyblak left 04:47 tinyblak joined, BenGoldberg left 04:50 kaleem joined 04:55 tinyblak_ joined 04:56 tinyblak left 05:01 leont left 05:06 Rounin left 05:12 gcole left 05:13 tinyblak_ left, tinyblak joined 05:14 kurahaupo joined 05:15 yeahnoob left 05:20 tinyblak_ joined, tinyblak left 05:29 vendethiel left 05:30 kaleem left 05:36 vendethiel joined 05:38 kurahaupo left 05:39 tinyblak_ left 05:40 tinyblak joined 05:48 kurahaupo joined 06:00 Rounin joined 06:05 xfix joined 06:08 rindolf joined, sirdancealot joined 06:12 [Sno] left 06:13 yeahnoob joined, yeahnoob left 06:16 yeahnoob joined 06:17 yeahnoob left 06:19 kurahaupo left
moritz m: .say for Signature.^attributes.map: { .^name } 06:27
camelia rakudo-moar 050686: OUTPUT«BOOTSTRAPATTR␤BOOTSTRAPATTR␤BOOTSTRAPATTR␤BOOTSTRAPATTR␤BOOTSTRAPATTR␤»
moritz m: .say for Rat.^attributes.map: { .^name } 06:28
camelia rakudo-moar 050686: OUTPUT«Attribute␤Attribute␤»
06:42 kaleem joined 06:45 mr-foobar left 06:57 szabgab joined 06:58 szabgab left 07:03 [Sno] joined 07:11 Patterner joined 07:15 Psyche^ left 07:21 darutoko joined 07:33 gfldex joined 07:41 FROGGS left 07:42 kaleem left 07:47 FROGGS joined 07:49 FROGGS left 07:50 Sqirrel left 07:51 gfldex left 07:52 FROGGS[mobile] joined 07:59 eiro joined
eiro hello 07:59
FROGGS[mobile] hi 08:01
JimmyZ 😹 08:05
08:10 xfix left 08:16 zakharyas joined 08:21 FROGGS joined 08:25 davido_ left 08:28 FROGGS[mobile] left 08:29 davido_ joined 08:35 Ugator joined, eternaleye left 08:37 eternaleye joined 08:38 silug left 08:55 fhelmberger joined 08:56 slavik left
CurtisOvidPoe m: say 1/(.1 + .2 - .3) # divide by zero error 08:59
camelia rakudo-moar 050686: OUTPUT«Type check failed for return value; expected 'Int' but got 'Failure'␤ in any return_error at src/vm/moar/Perl6/Ops.nqp:649␤ in method floor at src/gen/m-CORE.setting:11400␤ in method Str at src/gen/m-CORE.setting:11421␤ in method gist at src/ge…»
08:59 adu left
CurtisOvidPoe I don’t know how easy it would be, but a more user-friendly error message might help there. 08:59
moritz agreed 09:00
iirc we have a bug report open for that
CurtisOvidPoe m: say 1.1/(.1 + .2 - .3)
camelia rakudo-moar 050686: OUTPUT«Type check failed for return value; expected 'Int' but got 'Failure'␤ in any return_error at src/vm/moar/Perl6/Ops.nqp:649␤ in method floor at src/gen/m-CORE.setting:11400␤ in method Str at src/gen/m-CORE.setting:11421␤ in method gist at src/ge…»
CurtisOvidPoe That’s strange. I would expect that it would at least say “expected ‘Real’”
lizmat m: 1.1 / (my $a = 0)
camelia ( no output )
lizmat m: say 1.1 / (my $a = 0) 09:01
camelia rakudo-moar 050686: OUTPUT«Type check failed for return value; expected 'Int' but got 'Failure'␤ in any return_error at src/vm/moar/Perl6/Ops.nqp:649␤ in method floor at src/gen/m-CORE.setting:11400␤ in method Str at src/gen/m-CORE.setting:11421␤ in method gist at src/ge…»
CurtisOvidPoe moritz: thanks.
moritz CurtisOvidPoe: it's from some internal helper function
lizmat method floor(Rational:D:) returns Int:D {
09:01 FROGGS[mobile] joined
lizmat maybe we need to lose the "returns Int:D ? 09:02
CurtisOvidPoe Unfortunately, that’s a from a key slide in my talk and I can’t omit that example.
jnthn morning, #perl6
CurtisOvidPoe Question: “returns Int” makes sense to me, but what does the :D mean?
moritz lizmat: if it can fail, yes
lizmat tries and runs spectest
moritz CurtisOvidPoe: that it's a defined Int
jnthn lizmat: Sounds like the Int:D is overly restrictive, yes.
moritz CurtisOvidPoe: ie not the type object 09:03
lizmat $ 6 'say 1/(my $a=0)' 09:04
Divide by zero
is that better ?
jnthn Surely... :) 09:05
lizmat CurtisOvidPoe: is that what you expect?
$ 6 'my $a = 1/($=0); say $a.WHAT; say $a' 09:06
(Rat)
Divide by zero
aka, you *do* get a Rat returned, but as soon as you want to do anything with it, it fails 09:07
moritz if that's a feature, we have to give up all type constraints on return values from methods in Rat 09:10
09:11 mvuets joined
lizmat some test failures, but from tests checking the exception type (which clearly has changed) 09:12
moritz: I removed the constraints from Rational.floor/ceiling: they were the only ones with return constraints
09:13 abraxxa left
dalek kudo/nom: 9672c0d | lizmat++ | src/core/Rational.pm:
Rational.floor/ceiling less anal about return vals
 09:13
moritz lizmat++ 09:14
09:14 abraxxa joined 09:15 jluis left
dalek ast: 8da006e | lizmat++ | S03-operators/arith.t:
Fix changed exception type for 3/0
 09:19
jnthn lizmat: Related: rt.perl.org/Ticket/Display.html?id=123077 09:20
09:23 jluis joined 09:24 sqirrel_ joined 09:29 FROGGS[mobile] left
lizmat afk for ~2 hours& 09:33
moritz m: say 1/0 09:35
camelia rakudo-moar 050686: OUTPUT«Type check failed for return value; expected 'Int' but got 'Failure'␤ in any return_error at src/vm/moar/Perl6/Ops.nqp:649␤ in method floor at src/gen/m-CORE.setting:11400␤ in method Str at src/gen/m-CORE.setting:11421␤ in method gist at src/ge…»
moritz still the bad old rakudo
09:36 tinyblak left, tinyblak joined
CurtisOvidPoe lizmat: yes, that’s much better :) 09:38
09:39 kjs_ joined, dakkar joined 09:41 tinyblak left
masak antenoon, #perl6 09:46
FROGGS o/ #perl6
I'll be at the FOSDEM \o/
CurtisOvidPoe Yay! 09:52
09:53 sqirrel_ left
nwc10 with or without the FROGGletS? 09:57
09:58 kaleem joined
FROGGS nwc10: without 10:02
it is too complicated to get all of 'em with me
I mean, we are talking about four hobbits here (wife + three kids) 10:03
I just need to find good topics now and prep my talks :/ 10:04
jnthn FROGGS: I'm also pondering topic :) 10:05
(So you're not alone :))
FROGGS jnthn: :o)
jnthn: if you have spare topics left at the end... don't hesitate to share please 10:06
JimmyZ NSA! :P 10:07
FROGGS that's certainly a jnthnish topic, aye
I could imagine a NativeCall talk 10:08
10:08 rurban joined
moritz that would be nice 10:08
speaking of which: both the sqlite and the postgresql client API are great from a nativecall POV
FROGGS like a walk through...
hmmm
and I bet I can spend more time on NativeCall than on v5 or slangs in general other the other things I usually talk about 10:09
moritz: so you'd favour that somebody extends DBDish::Pg? 10:11
10:11 cognominal left
jnthn FROGGS: A NativeCall talk would be nice; it tends to be well received by many folks who try it. 10:11
moritz FROGGS: no objections to that, but I just wanted to point to useful APIs that can be cleanly nativecalled (as examples for such a talk) 10:12
FROGGS jnthn: exactly my motivation
moritz: my first thought was: "Shout a libname and we'll bind it!" but I dunno how well that would work out :o)
10:13 slavik joined
moritz FROGGS: live demos are always very risky 10:13
jnthn Terribly if somebody shouts "ncurses!" :)
moritz and I guess you don't want to bind something non-trivial in the front of an audience 10:14
Rounin How awful... Next time you should shout "OpenGL"
Then again, repetitive bindings like that can probably be auto-generated to some extent 10:16
raydiak still hopes to make creating bindings a little less laborious when C::Parser is ready
btyler_ also hopes that 10:17
10:17 Alina-malina left 10:22 Alina-malina joined
FROGGS yeah 10:24
10:25 virtualsue joined
FROGGS moritz: I guess that live coding is the extrem of live demos, and we all know the rule of live demos :o) 10:25
but I might stick to libXML 10:27
dalek kudo/nom: 39be5fb | jnthn++ | src/Perl6/World.nqp:
Remove dead code.
 10:32
kudo/nom: a6ac643 | jnthn++ | src/Perl6/Actions.nqp:
Don't static-inline routine with `is copy` trait.

Fixes RT#118183.
synopsebot Link: rt.perl.org/rt3//Public/Bug/Displa...?id=118183
ast: 3b76c73 | jnthn++ | S06-traits/is-copy.t:
Test for RT#118183.
synopsebot Link: rt.perl.org/rt3//Public/Bug/Displa...?id=118183
10:34 denis_boyun_ joined 10:37 tinyblak joined 10:39 tinyblak left, tinyblak joined, tinyblak left, tinyblak joined
masak today's mini-challenge: a 4x4 grid has 24 internal walls. in the power set of these, some of the members are "mazes", by definition grids with a unique path between every two squares. how many mazes exist? 10:41
a number is fine. (but it should be found with Perl 6, not with oeis.org)
cute ASCII art is even better. 10:42
10:42 tinyblak_ joined, tinyblak_ left, tinyblak_ joined
moritz is that really a *mini* challenge? 10:44
masak solution can fit in ~50 LoC by my estimate. so yes.
10:45 tinyblak left
jnthn 50 lines of Perl 6 can do quite a bit :P 10:48
masak I'm not contesting that. 10:59
10:59 tinyblak joined 11:02 tinyblak left, tinyblak_ left, tinyblak joined
dalek kudo/nom: 0885ec3 | jnthn++ | src/Perl6/Grammar.nqp:
Whine about redeclarations of labels.

Fixes RT#120632.		 11:03
synopsebot Link: rt.perl.org/rt3//Public/Bug/Displa...?id=120632
dalek ast: b57a00b | jnthn++ | S04-statements/label.t:
Test duplicate labels are whined about.

Covers RT#120632.
synopsebot Link: rt.perl.org/rt3//Public/Bug/Displa...?id=120632
11:08 Mso150_e left
dalek p/6pe: 9d6ba70 | jnthn++ | src/vm/moar/QAST/QASTOperationsMAST.nqp:
Map nqp:: ops for parametricity on MoarVM backend.
 11:19
p/6pe: 98b770c | jnthn++ | t/nqp/87-parametric-6model.t:
Some initial tests for 6model parametric ops.
p/6pe: f90dd15 | jnthn++ | t/nqp/87-parametric-6model.t:
Basic tests for parameterization interning.
11:20 sirdancealot left 11:25 Rounin left
dalek ast: f031a89 | lizmat++ | S02-types/ (2 files):
Fudge questionable tests
 11:27
11:33 colomon left
masak hah. I don't even recall submitting RT #120632. :) 11:34
synopsebot Link: rt.perl.org/rt3//Public/Bug/Displa...?id=120632
11:35 colomon joined
jnthn Well, now the redecl error is a-OK :P 11:38
11:48 kaleem left
dalek kudo/nom: 42bb409 | lizmat++ | src/core/ (3 files):
Add custom HashIter constructors

The idea being that we don't need to figure out at run time what type of iteration we need. Hopefully a better target for inlining this way. Shows about a 3% speedup in .keys/kv/values/pairs/invert. Also reads better, I think.		 11:57
lizmat ^^^ also seems to give 1.5% improvement in Text::CSV benchmark 12:01
masak nice!
12:04 uniejo joined 12:06 tinyblak_ joined 12:07 tinyblak_ left, tinyblak_ joined, tinyblak left 12:13 uniejo left 12:18 grondilu joined 12:20 kaare__ left
vendethiel for previous discussion: wouldn't have Failure as a bottom type make sense? 12:25
TuxCM lizmat, I reworked through my CSV_XS changes and got another 10% while implementing options (see test-t.pl) 12:33
lizmat++ 12:34
12:34 tinyblak joined 12:36 tinyblak_ left 12:40 xfix joined, xfix left, xfix joined 12:42 tinyblak left, tinyblak joined
masak vendethiel: TimToady has repeatedly hinted that Failure is effectively our bottom type. 12:42
12:45 kaleem joined 12:46 rhr left 12:54 colomon left, anaeem1 left 12:56 colomon joined 12:58 tinyblak left 12:59 tinyblak joined 13:00 brrt joined
lizmat
.oO( in the end, it is all a mistake )
 13:02
13:02 rurban left 13:03 skids left
moritz in some sense, Failure as a bottom type isn't too far off 13:03
because you can call any method on it
which is a requirement for good bottom type 13:04
psch hi #perl6 \o 13:09
brrt lizmat, in perl6 io, and especially newio, who is responsible for encoding of strings
btw, 'Buf' is our bag-of-bytes type right
13:09 colomon left
moritz Blob is the most general bag-of-thingies type 13:10
and Blob[uint8] the most general bag-of-bytes type 13:11
dalek kudo/nom: fdf18b3 | lizmat++ | src/core/Array.pm:
Streamline Array.at_pos

  - reduce number of boxings/unboxings
  - get better errors on Inf/NaN
Gives about 15% speedup in @a[Int $b] access. Adding a separate Int candidate did not give the desired speedup effect: not sure why that is, maybe the overhead of having another candidate? Unfortunately, not very noticeable in the test-suite.
moritz and Buf is supposed to be writable, but isn't quite yet
masak moritz: also, Failure gets through most typechecks.
13:11 jack_rabbit left
lizmat m: say Failure ~~ Int 13:11
camelia rakudo-moar 42bb40: OUTPUT«False␤»
lizmat m: say 1/($=0) 13:12
camelia rakudo-moar 42bb40: OUTPUT«Divide by zero␤ in method Numeric at src/gen/m-CORE.setting:13949␤ in sub infix:<-> at src/gen/m-CORE.setting:4615␤ in method Str at src/gen/m-CORE.setting:11427␤ in method gist at src/gen/m-CORE.setting:4450␤ in sub say at src/gen/m-CORE.se…»
13:12 sirdancealot joined
lizmat afk for a few hours& 13:12
masak m: sub foo(--> Int) { fail }; my $x = foo; say "alive"
camelia rakudo-moar 42bb40: OUTPUT«Type check failed for return value; expected 'Int' but got 'Failure'␤ in any return_error at src/vm/moar/Perl6/Ops.nqp:649␤ in sub foo at /tmp/WcESj2YqJO:1␤ in block <unit> at /tmp/WcESj2YqJO:1␤␤»
masak huh.
lizmat that's *exactly* the problem! :-) 13:13
13:13 Rounin joined, grondilu left, colomon joined 13:14 rindolf left
jnthn And making it work has a notable cost. 13:14
FROGGS moritz: there is a subbuf-rw fwiw 13:15
moritz FROGGS: yes
FROGGS: my point is that it's not mutable in all the places where it should be
(iirc)
masak jnthn: one that could be optimized away?
brrt but anyway. if in perl6 i have a file handle, and I write a string to it, who is going to do the encoding of the string
FROGGS moritz: correct
jnthn Suppose we have $a + $b + $c and we statically known all are of type Int
moritz brrt: the file handle has an encoding associated
brrt hmm 13:16
what happens if i write something unencodable
FROGGS brrt: assplosion
moritz you get an exception, I believe
jnthn We can not statically inline the second + there, because we cannot rely on return types contributing static knowledge - short of looking deeper.
So we push more things towards dynamic optimization, and mostly leave static to be useful for native types. 13:17
That said, if Int:D rejects failure, we may be able to keep the benefits.
13:20 jack_rabbit joined 13:21 leont joined 13:30 tinyblak left 13:31 tinyblak joined, jack_rabbit left 13:35 rindolf joined 13:36 baest_ joined 13:38 baest left, leont left 13:40 baest_ is now known as baest, synopsebot left 13:41 rjbs left 13:42 rjbs joined, pmichaud left 13:43 pmichaud joined, jack_rabbit joined 13:47 FROGGS left 13:51 leont joined 13:58 kjs_ left 13:59 anaeem1 joined, jack_rabbit left, telex left 14:00 anaeem1 left, telex joined, kjs_ joined 14:01 synopsebot joined 14:02 tinyblak_ joined 14:04 Rounin left 14:05 tinyblak left, tinyblak_ left
[Coke] TimToady: how do you feel about renaming isNaN to is-NaN for is-prime parity? 14:06
14:06 tinyblak joined 14:09 rurban joined
psch p6: my $x = 5; say nqp::isint(nqp::decont($x)) # this still confuses me 14:16
camelia rakudo-{parrot,moar} 42bb40: OUTPUT«0␤»
psch especially in tandem with:
$ ./perl6 -e'my $x = 5; say nqp::isint(nqp::decont($x))'
1
that's perl6-j
14:17 rhr joined
jnthn Int in Perl 6 is a P6opaque with a P6bigint located within it. 14:19
p6: say nqp::isint(Int); say nqp::isint(int); 14:20
camelia rakudo-{parrot,moar} 42bb40: OUTPUT«0␤1␤»
jnthn I'm a bit curious about the -j discrepancy.
psch mvm check repr identity, jvm checks for StorageSpec.CAN_BOX_INT 14:23
...at least if i'm looking at the right spots, which is interp.c:2615 and Ops.java:2350 14:25
14:28 kaleem left
[Coke] imgur.com/gallery/CJfnAL7 - a little nicer than the ones I have! 14:32
tadzik whoa 14:33
14:40 skids joined 14:47 kjs_ left 14:51 kjs_ joined 14:56 anaeem1 joined 15:01 kjs_ left, kjs_ joined 15:02 donaldh joined 15:03 kaleem joined 15:10 tinyblak left 15:27 adu joined, adu left 15:31 kjs_ left, [Sno] left 15:33 treehug88 joined 15:37 treehug88 left 15:40 treehug88 joined
hoelzro o/ #perl6 15:44
psch o/ hoelzro 15:45
jnthn: so jvm is wrong and should also check for REPR id instead of the storagespec? 15:46
which means i'll have to do that myself in RakudoJavaInterop, which isn't bad because it gives me a convenient spot to check for values > Long.MAX_VALUE i guess 15:47
15:47 spider-mario joined, FROGGS joined
[Coke] thanks, all you groovy perl6 people. 15:48
FROGGS o/
jnthn psch: Well, inconsistency is certainly wrong... The JVM way does mean you're asking "can this unbox to a native int"... 15:49
psch o/ FROGGS
jnthn ...which may be useful. 15:50
Mebbe try changing it to be inline with Moar and Parrot, and see if spectest is happy. 15:51
psch right, i'll try that and see what falls out 15:52
got something else right now though, result will come in about 2 hours i think :)
psch will bbl & 15:53
15:58 kaleem left 16:02 kjs_ joined
pyrimidine re: panda, is there a way to run bootstrap/rebootstrap or install w/ different backends? IN other words, maybe have a panda (for the default) and panda-j/panda-p (for JVM or Parrot)? 16:06
*panda-m (for the default moar) # just to clarify 16:07
FROGGS pyrimidine: that is what you will get in a not too distant future for free
hoelzro if two multis exist that have the same signature except for a where condition on a parameter, and both wheres match the incoming value, what determines which is called? 16:08
FROGGS pyrimidine: the "eleven" branches of rakudo+panda will create foo,foo-m,foo-j and foo-p scripts for every binary that gets installed
hoelzro ex. my multi conjugate(Str $verb where /'вить' $/) { ... } ; my multi conjugation(Str $verb where /'ить' $/) { ... } ; say conjugate('язвить') 16:09
FROGGS hoelzro: the first declared IIRC
jnthn hoelzro: design.perl6.org/S12.html#Constrain...Candidates
hoelzro ah, thanks FROGGS + jnthn
jnthn But what FROGGS said
There's an RT recently filed on this by CurtisOvidPoe++, but Rakudo is to spec on this one, so it's not a bug. 16:10
pyrimidine FROGGS: ok, thx.
FROGGS: Just making sure it isn't implemented now (e.g. I'm missing something obvious :)
FROGGS so it is the first hit in the method table
pyrimidine: correct, it is not available yet
jnthn Could be a spec issue, though I already argued this one out with TimToady plenty. :)
So I'm not convinced he's likely to change his mind. 16:11
jnthn was for "ambiguous" also 16:12
FROGGS jnthn: but checking for ambiguity would not be cheaper, right?
CurtisOvidPoe It the issue with that ticket is because runtime checking is so expensive? Would any compile-time checks work (or is that the halting problem?) 16:13
16:14 denis_boyun_ left
jnthn CurtisOvidPoe: Well, the administrative issue with the ticket is that we use RT for implementation bugs, and github.com/perl6/specs/issues for specification issues. :) 16:14
16:15 _4d47 joined
CurtisOvidPoe Ah, I had no idea. 16:15
jnthn CurtisOvidPoe: Compile time resolution quickly ends up being the halting problem.
16:17 mr-foobar joined
jnthn CurtisOvidPoe: I can't personally help much with the issue, though, in so far as I originally implemented it as complaining about ambiguity, and argued for the same semantics you are, but TimToady was insistent on "first match wins" for constraint types (for nominal types, we do care about ambiguity). 16:17
So eventually after a couple of debates on it, I ended up dragging Rakudo in line with spec. 16:18
I think performance was one argument, but "equivalence with given/when semantics" is the other I remember. 16:19
CurtisOvidPoe I wonder about the roles issue then. Maybe there could at least be a warning when there’s potential ambiguity if multiple roles provide mutli-methods with the same name? One of the biggest problems with inheritance and mixins is ordering.
I can buy the equivalence with given/when. It’s roles and inheritance where it bytes. If you have a muli in a parent or child, I assume the child wins? Is that tested? 16:20
And if I have a class and a role and the role is composed into the class, which is considered “first”? 16:21
(sorry if I’m beating a dead horse and being annoying)
16:22 _4d47 left
jnthn The inheritance one has a nice easy answer: the invocant of a method is typed to match the enclosing class, and multi-dispatch on methods doesn't treat the invocant parameter specially in any way, so the first parameter in the child class has a narrower and so wins. 16:23
CurtisOvidPoe That makes perfect sense.
jnthn The roles case has a bit more going on. When you write a method in a role, the invocant's type is the generic type ::?CLASS. That is to say, every Perl 6 role is implicitly parametric on the type it is composed into. Composition thus first instantiates the role, and copies the instantiated method into the method table of the class. So the method the multi-dispatcher actually sees has an invocant typed with the class that composed the method. 16:26
16:26 Sir_Ragnarok_ joined
CurtisOvidPoe And if there’s ambiguity, the “native” class multi is first? 16:27
That might be hard if we’re flattening at compile-time.
CurtisOvidPoe is just guessing.
jnthn In the absence of subset/where that is ambiguous at present. 16:28
Since - as you correctly point out - we flattened things at compile time.
16:29 Sir_Ragnarok left
jnthn You can, if you go introspecting deeply enough, figure out what methods came from roles. 16:29
I *think* 16:30
moritz you can
jnthn moritz: .package?
moritz .^roles
jnthn Oh, I meant from the Method object itself.
moritz and then look into the roles 16:31
jnthn Ah, true.
16:31 zakharyas left
jnthn That'd be more round-about, but yes. 16:31
moritz but .package might work too
jnthn It's probably also worth noting that various very obvious ambiguities in multis could be pointed out at compile time, but are not yet. 16:33
16:33 brrt left 16:34 molaf joined 16:49 kaleem joined 16:50 KCL_ joined 16:52 salv0 left 17:03 anaeem1 left 17:04 anaeem1 joined 17:06 Sir_Ragnarok joined, Sir_Ragnarok_ left 17:09 anaeem1 left, salv0 joined
dalek rl6-roast-data: 2e529d4 | coke++ | / (5 files):
today (automated commit)
 17:11
17:12 virtualsue left, rurban left 17:13 anaeem1_ joined 17:14 mvuets left 17:16 anaeem1_ left
[Coke] no failures in a few days, that's good. 17:16
jnthn Nice
Especially as there's been a flow of patches too :)
17:16 anaeem1 joined
[Coke] mmm, rakudo.moar-jit has gone from 32907 to 33025 passes in 4 days. 17:17
vendethiel nice :D. 17:19
17:20 kaleem left 17:21 anaeem1 left
jnthn Dinner time here soon, and I plan to continue working on the 6model parametric bits this evening. :) 17:24
17:24 fhelmberger left 17:26 abraxxa left
timotimo likes 17:26
bon appetit
17:27 anaeem1 joined 17:28 [Sno] joined
hoelzro [Coke]: regarding Mojo6, are you going to try to preserve the "no dependencies" nature of Mojo5? 17:35
hoelzro may have some tuits tonight 17:36
17:44 dakkar left, anaeem1 left 17:45 treehug88 left 17:46 ab5tract joined
hoelzro I've noticed I often accidentally type a lexical variable (ie. $name) instead of the attribute that I mean ($!name). I was thinking I could add a helpful error message; would it be ok if it were just Variable '$name' is not declared. Did you mean '$!name'? 17:47
or should it be more particular?
[Coke] hoelzro: in general, I don't particular care for "no dependencies", but I've always worked places where installing dependencies was trivial. 17:49
so, probably depends on how hard it is to satisfy.
hoelzro [Coke]: I was thinking of trying my hand at writing Mojo::CSS, and I wouldn't mind avoiding some reinventing of the wheel for the HTML/CSS parsing 17:50
jdv79 ugexe: its not the uri commit - the ssl.close call is hanging. 17:51
PerlJam even mojo5 isn't really "no dependencies" 17:54
17:56 sirdancealot left, gfldex joined 17:58 treehug88 joined
hoelzro PerlJam: fair enough =) 18:00
PerlJam plus, it seems to me that getting working would win on the -Ofun factor initially :) 18:01
*it
arnsholt hoelzro: Suggesting $!foo for $foo sounds reasonable to me. Should be possible to implement as an extension to our existing "name missing" stuff 18:02
hoelzro PerlJam: yeah, I was thinking of using dependencies to get it working; the question is do I then try to make it dependency-less? 18:03
arnsholt: yeah, I don't think it'll be to hard; I'm just wondering if the error should be more clear
something like "There's no lexical '$name', but you have an attribute called '$!name'. Did you mean that?" 18:04
18:04 _4d47 joined
arnsholt Might be nicer. I think your original suggestion is nice as well 18:04
It's a bit more concise
PerlJam aye, me too
18:05 kaare__ joined
hoelzro ok, how about I just leave the copy as-is, and others can improve if desired later? 18:05
18:06 eternaleye left 18:07 _4d47 left 18:08 donaldh left 18:09 eternaleye joined
jnthn hoelzro: About CSS, I think there's already a module in the ecosystem that does CSS parsing, and it looked like a lot of work/thought had gone into it. 18:12
18:13 FROGGS_ joined
jnthn hoelzro: +1 to suggesting an existing $!name if $name does not exist 18:13
hoelzro jnthn: I saw that! I intend to make use of it =)
18:14 KCL_ left, pmurias joined
itz will FOSDEM have a State of the Butterfly in addition to Get Ready to Party? 18:15
lizmat itz: not that I'm aware of (atm)
18:16 FROGGS left 18:18 zakharyas joined 18:19 IllvilJa joined
PerlJam idly wonders what Camelia looked like in her pupal stage 18:21
actually, camelia has been a pupa for about 15 years now, IMHO (Perl 5 was the caterpillar :)
ab5tract that caterpillar's got humps! 18:22
PerlJam
.oO( camelpillar? )
ab5tract has been imagining a Perl 6 camel as having fractal humps :) 18:23
Perl 6: Even the humps have humps. Plenty of places to find a comfortable seat! 18:24
[Coke] hoelzro: please please jump in, even if you steal the repo to do it. :) happy to give out bits there, since I've been so lazy. 18:25
PerlJam perl6.org could have some imagery of a chrysalis slowly opening and Camelia emerging :)
18:27 Mso150 joined
[Coke] hoelzro: you're a collaborator on github.com/coke/mojo6 if you like. 18:27
18:27 dwarring joined
lizmat [Coke]: wrt to isNaN vs Is-NaN, I don't have much of a feeling about that 18:29
PerlJam dwarring: did someone mentioning CSS somehow summon you?
lizmat [Coke]: but with regards to e.g. exists_key vs exists-key (aka anything with an underscore currently)
18:29 virtualsue joined
lizmat [Coke]: I think those should be kebabcased with the underscored version deprecated 18:29
PerlJam lizmat++ (That should be the "p6 standard" IMHO) 18:30
lizmat yup
:-)
dwarring sips caffine & backlogs 18:31
ab5tract lizmat++ 18:32
18:33 anaeem1_ joined
ab5tract fwiw, i think the isNan Is-Nan makes sense too. smells like a YAPE otherwise ;) 18:34
*YAPAE
18:36 pmurias left
[Coke] yapae? 18:36
18:37 mvuets joined
woolfy nine: ping 18:38
dalek ecs: 1ca7203 | coke++ | S99-glossary.pod:
fix link
ab5tract [Coke]: Yet Another Potentially Awkward Explanation - perl6advent.wordpress.com/2014/12/...us-corner/ 18:41
dalek ecs/newio: 8284cda | jnthn++ | S12-objects.pod:
Correct spec of context aspects of the self term.
 18:42
ecs/newio: 1ca7203 | coke++ | S99-glossary.pod:
fix link
ecs/newio: f907cb8 | lizmat++ | S (2 files):
Merge branch 'master' into newio
PerlJam ab5tract: might want to add that to S99 18:44
[Coke] btw, not "Is-NaN" but "is-NaN". also, I wonder if "is-nan" would be ok. 18:47
18:49 anaeem1_ left
ab5tract PerlJam: i would be honored :D 18:50
18:51 sirdancealot joined
dwarring hoelzro: CSS::Grammar implements a basic selector parser - based on www.w3.org/TR/2011/REC-css3-selectors-20110929/ 18:54
CSS::Module extends this - its based on www.w3.org/TR/css3-selectors/ 18:55
18:55 anaeem1 joined, beastd joined
dwarring which is what Mojo::CSS implements 18:56
18:57 virtualsue left
dwarring these modules do parsing only no attempt at implementing a CSS selector engine 18:58
18:58 anaeem1 left
dwarring ...so wide open 18:58
18:59 anaeem1 joined
dwarring Mojo::DOM and Mojo::CSS look interesting 19:00
itz I hadn't noticed perltricks.com/article/144/2015/1/1...-in-Perl-6 19:01
19:01 Sqirrel joined
dalek kudo/nom: b4a936e | lizmat++ | src/core/List.pm:
Streamline List.at_pos for 15/20% improvement
 19:02
kudo/nom: ac75b0a | lizmat++ | src/core/List.pm:
Handle List.at_pos(Inf/NaN) better
kudo/nom: 389326a | lizmat++ | src/core/ (2 files):
Make sure X::Item exceptions are thrown properly

Stubbing X::Item too late, would cause a "Cannot find &Item" error, probably because X *was* already stubbed, and X::Item was interpreted as a sub in X.
19:03 anaeem1 left
dalek ecs: 08e5b67 | PerlJam++ | S99-glossary.pod:
[S99] Add YAPAE
 19:03
19:04 Mso150_b joined 19:05 Mso150 left, telex left 19:06 telex joined, eternaleye left 19:07 eternaleye joined
ab5tract [Coke]: ah, the capital 'I' did seem suspicious to me 19:08
lizmat jnthn: how would you feel of about only making int/Int candidates for at_pos in classes, and a catchall at_pos in Any that would the Any->int conversion ? 19:09
ab5tract keeping the casing of the type seems important to me
lizmat perhaps a Numeric catcher as well, which would do the Inf/Nan check 19:10
19:12 rjbs left
ab5tract PerlJam: ah, thanks for adding :) i just finished filing a PR when i saw it announced here :) 19:12
jnthn Why do we have an is-NaN method, ooc? 19:18
Is ~~ NaN not workable
?
lizmat: That would probably save some duplication... 19:19
lizmat yes, quite some, and reduce the number of candidates significantly
jnthn lizmat: It'd imply a level of delegation, but that's probably not a big problem. 19:20
lizmat which would also be good, right ?
it would be punishing the programs that don't use int/Int for []
jnthn Well, the dispatch cache is keyed on the types that actually show up.
So if a program hits just a selection of candidates, the cache is quickly primed and the number of candidates matters rather less
lizmat types that actually show up where? 19:21
jnthn In calls 19:22
lizmat but if in the setting, we don't see a call with a certain type, how would a program than handle that?
dwarring hoelzro: example CSS selector parse...
perl6-m -I blib/lib -MCSS::Module::CSS3 -e'my $actions = CSS::Module::CSS3::Actions.new; say CSS::Module::CSS3.parse( q<input[class~="foo"]>, :rule<selectors>); say $/.ast.perl'
jnthn lizmat: The cache is built dynamically at runtime.
lizmat or is the cache outside of the setting
ok
jnthn lizmat: It depends on what types actually show up at runtime, like much of our optimization :) 19:23
lizmat yeah, ok
jnthn lizmat: So if somebody never does @foo[$a_str] then that candidate won't be in the multi dispatch cache.
19:23 kurahaupo joined
dwarring perl6-m -I blib/lib -MCSS::Module::CSS3 -e'my $actions = CSS::Module::CSS3::Actions.new; say CSS::Module::CSS3.parse( q<input[class~="foo"]>, :rule<selectors>, :$actions); say $/.ast.perl' 19:23
lizmat well, I'm streamlining at_pos candidates now for all classes
jnthn The number of candidates being less is good of course for overall code size in memory.
lizmat yup
and when that's done, I'll see about reducing the number of candidates (except in Any) 19:24
and see what that brings us :)
jnthn And good if the program uses so many of the candidates or so crazy range of types on a candidate that it overflows the multi cache and we fall back to the slower path.
lizmat ok 19:25
jnthn Thankfully, megamorphic things ain't hugely common.
lizmat so reducing candidates is good :-)
19:25 CurtisOvidPoe left
jnthn lizmat: Yes, to a degree 19:27
ugexe jdv79: those messages you got with my examples are also known to be produced by older versions of openssl with AES-ni (and from what you posted your openssl is not up to date)
jnthn lizmat: For example, turning two multi candidates into a single one that does a type check is unhelpful. 19:28
lizmat yeah, got that :-)
this is exactly what I'm splitting in a lot of cases now
jnthn OK. Then sounds like you're doing very agreeable things. :) 19:29
lizmat yeah :-)
jnthn ponders whether to work on 6pe JVM port, or on figuring out the bounded serialization interaction with it... 19:30
Mebbe I should do the port so I can remember how this code I wrote a month ago works. :)
ab5tract JVM is a huge "production ready" checkbox, imo 19:32
jercos I presume 6pe is not IPv6 over MPLS? >.> 19:33
TimToady datapoint: I had a helper doing quiz scorekeeping this past weekend, and he duplicated the dataentry using the JVM while I used MoarVM to run our quiz editor; both performed flawlessly 19:34
jnthn \o/
TimToady and he was running on a Mac 19:35
(which is why he had the JVM version from December, because MoarVM was still flakey on Mac back then)
19:36 kjs_ left
jnthn Yeah, OSX is a good canary for some reason or toher. 19:36
*other
lizmat m: say (1,2,3,4.5)[-1] 19:38
camelia rakudo-moar 389326: OUTPUT«===SORRY!=== Error while compiling /tmp/xXEOxdv6nA␤Unsupported use of a negative -1 subscript to index from the end; in Perl 6 please use a function such as *-1␤at /tmp/xXEOxdv6nA:1␤------> say (1,2,3,4.5)[-1]⏏<EOL>…»
TimToady I'm fine with hypehening 'is' functions in general
El_Che TimToady: you heard the Language track at fosdem is at Sunday?
lizmat m: my $a = (1,2,3,4.5); say $a[-1]
camelia rakudo-moar 389326: OUTPUT«===SORRY!=== Error while compiling /tmp/4G6lTlWbXZ␤Unsupported use of a negative -1 subscript to index from the end; in Perl 6 please use a function such as *-1␤at /tmp/4G6lTlWbXZ:1␤------> my $a = (1,2,3,4.5); say $a[-1]…»
lizmat hmmm....
TimToady El_Che: yes
TimToady gets to have the 'last word' :)
lizmat there *are* other presentations in that room after that, though
but not in the "Language" track 19:39
El_Che "And now, I declare all other languages obsolete" :)
19:39 bartolin joined
vendethiel please tell me fosdem will have some kind of recording 19:39
TimToady nah, just obsolescent :)
El_Che vendethiel: it has
to my surprise there is an other Perl talk just before TimToady's
TimToady "There's More Than One Way To Talk About It!" 19:42
FROGGS_ who's speaking? rjbs?
El_Che fosdem.org/2015/schedule/event/des...ry_module/
the perl talks arein the dev room saturday
hehe 19:43
lizmat Dana
FROGGS_ "Fastest bigint probable prime testing to 10k digits".... I wonder how we do compete 19:44
19:44 uniejo joined
TimToady well, we do is-prime with libtommath 19:45
woolfy El_Che++ 19:46
El_Che woolfy++
(that's how we say hi nowadays?) :)
woolfy oO(let's ++ eachother more often!) 19:47
El_Che Is that a hint for Larry for Presenting Perl++ at Fosdem? :)
19:47 uniejo left
woolfy :-) 19:50
dalek p: 1200c79 | peschwa++ | src/vm/jvm/NQP/Ops.nqp:
Implement missing falsey(num) for JVM backend.

Before this the new test in nqp/13-op.t killed NQP with a VerifyError due to invalid bytecode.		 19:51
psch got sidetracked by that when running tests for the nqp::isint change... :) 19:52
dalek kudo/nom: 958cbe7 | lizmat++ | src/core/Parcel.pm:
Streamline Parcel.at_pos, for about 10% improvement
19:57 kjs_ joined
jnthn psch++ 19:58
ab5tract speaking of ++, i owe psch++ # now the mountains of perl6-processing are visible enough to plot a base camp 20:00
\
dalek kudo/nom: 28e93b6 | lizmat++ | src/core/Range.pm:
Streamline Range.at_pos: not much to win here
 20:01
20:01 cognominal joined 20:03 kurahaupo left
hoelzro [Coke]: cool, thanks! 20:08
dwarring: thanks for the example!
dwarring hoelzro: np 20:09
20:12 cognominal left, kurahaupo joined
[Coke] jnthn: (remove isNaN) works for me. 20:19
20:19 kjs_ left 20:22 zakharyas left 20:26 darutoko left, beastd left 20:35 amoquin joined
amoquin masak: Is the mini-challenge orientation-dependent? 20:36
Heh
20:36 amoquin is now known as Mouq
Mouq better 20:36
20:41 rindolf left, ab5tract left 20:44 kjs_ joined 20:48 virtualsue joined 20:51 ab5tract joined, telex left 20:52 telex joined 20:56 rurban joined 21:02 zakharyas joined 21:05 synopsebot left, eternaleye left 21:06 Util left, eternaleye joined 21:08 kurahaupo left
dalek p/6pe: 648d0ca | jnthn++ | src/vm/jvm/runtime/org/perl6/nqp/runtime/IndyBootstrap.java:
Quiet warnings about variable name _.
 21:09
p/6pe: c194b3a | jnthn++ | src/vm/jvm/runtime/org/perl6/nqp/sixmodel/ (4 files):
Stub in STable parametricity extensions on JVM.
p/6pe: 7661100 | jnthn++ | src/vm/jvm/QAST/Compiler.nqp:
Map parametric 6model nqp:: ops on JVM.

They're not yet implemented, but it gets the test file compiling.
ab5tract sub prune_array( $list, $indices ) { [ for $list.kv { $^v unless $indices (cont) $^k } ] } 21:10
21:10 telex left
ab5tract ^^ this solution makes me happy 21:10
21:10 zakharyas1 joined 21:11 zakharyas left, hugme joined, ChanServ sets mode: +v hugme 21:12 synopsebot joined 21:13 Util joined
ab5tract not that there was really a problem floating around, but that showed up out of the mists today. 21:13
it worked well as an example to explain the $^ twigil (or is it a stwigil when the sigil and twigil are presented in combination) 21:15
21:16 Mso150_b_g joined, Mso150_b left 21:17 telex joined
moritz lizmat: + multi method at_pos(Parcel:D: Int \pos) is rw { 21:18
+ my $pos = nqp::unbox_i(pos);
lizmat: shouldn't it be Int:D \pos 21:19
21:19 rhr left
moritz and then my int $pos = nqp::unbox_i(pos) ? 21:19
21:19 Juerd left 21:21 rhr joined, synopsebot left 21:22 pmichaud left, Juerd joined 21:23 Util left, Util joined, synopsebot joined 21:25 Mso150_b_g_d joined
psch jnthn: replacing the check for StorageSpec.CAN_BOX_INT with REPR.equals(P6int.class) (and similarly for the other is$type ops) doesn't seem to cause any roast failures 21:25
21:26 Mso150_b_g left
psch all i got was what looks like a Supply-flap, in S17-supply/unique.t 21:26
jnthn: i'm wondering if we want an op along the lines of what isint on jvm does currently though.
jnthn psch: Only if we have a use case ;) 21:27
21:27 pmichaud joined
lizmat moritz: entirely correct :-) 21:28
psch jnthn: i have exactly on use case, for unmarshaling towards java types, which is jvm-specific
*one
i can definitely make do with isbig_I and checking StorageSpec myself though 21:29
jnthn Yeah, if you're already in Java land just do it...
lizmat m: my $a; $a[Int] = 1
jnthn (With the StorageSpec)
camelia rakudo-moar 28e93b: OUTPUT«Invocant requires an instance, but a type object was passed␤ in method Bridge at src/gen/m-CORE.setting:4905␤ in sub infix:<<> at src/gen/m-CORE.setting:4828␤ in sub postcircumfix:<[ ]> at src/gen/m-CORE.setting:2736␤ in block <unit> at /tmp/HU…»
21:30 xfix left 21:32 synopsebot left, Util left, Mouq left, Alina-malina left, vukcrni left, amkrankruleuen left, dj_goku left, jakesyl_insomnia left, prime left, sven_123 left, broquaint left, ribasushi left, Hor|zon left, dagurval_ left, sergot left, atta left, telex left, hugme left, sirdancealot left, FROGGS_ left, vike left, dylanwh left, araujo left, go|dfish left, xinming left, frew left, pyrimidine left, Mso150_b_g_d left, kaare__ left, treehug88 left, berekuk left, TuxCM left, ggoebel111111112 left, avuserow_ left, tadzik left, Woodi left, mst left, pochi left, ka05 left, ingy left 21:33 Sqirrel left, Sir_Ragnarok left, mr-foobar left, zoosha left, yakudza left, Khisanth left
dwarring r: say "abc".substr(0,4) 21:34
camelia rakudo-{parrot,moar} 28e93b: OUTPUT«abc␤»
dwarring r: say "abc".substr(*-4)
camelia rakudo-moar 28e93b: OUTPUT«===SORRY!===␤Start argument to substr out of range. Is: -1, should be in 0..3␤»
..rakudo-parrot 28e93b: OUTPUT«Start argument to substr out of range. Is: -1, should be in 0..3␤ in method gist at gen/parrot/CORE.setting:13906␤ in method gist at gen/parrot/CORE.setting:1165␤ in sub say at gen/parrot/CORE.setting:16640␤ in sub say at gen/parrot/CORE.sett…»
dalek kudo/nom: 6306593 | moritz++ | src/core/Parcel.pm:
micro-optimize at_pos a bit more
21:34 skaufman1 joined, ingy joined, daxim joined, jnthn joined, cxreg joined, mtj- joined, mathw joined, El_Che joined, Timbus joined, tinita_ joined, leedo_ joined, muraiki joined, robinsmidsrod joined, Guest22354 joined, nine joined, cibs joined, sivoais joined, MilkmanDan joined, atta joined, ka05 joined, sergot joined, dagurval_ joined, Hor|zon joined, pochi joined, ribasushi joined, broquaint joined, mst joined, sven_123 joined, Woodi joined, pyrimidine joined, tadzik joined, avuserow_ joined, SABFOSP joined, ggoebel111111112 joined, frew joined, TuxCM joined, dj_goku joined, xinming joined, go|dfish joined, berekuk joined, araujo joined, amkrankruleuen joined, dylanwh joined, vike joined, vukcrni joined, treehug88 joined, kaare__ joined, FROGGS_ joined, sirdancealot joined, Mouq joined, hugme joined, telex joined, Util joined, synopsebot joined, jakesyl_insomnia joined, Sqirrel joined, Sir_Ragnarok joined, mr-foobar joined, zoosha joined, yakudza joined, Khisanth joined, sinisalo.freenode.net sets mode: +v hugme, Alina-malina joined
dwarring ^^ I wonder if that usage of .substr should be more forgiving 21:35
21:35 telex left, sivoais left, vukcrni left, sivoais_ joined
Mouq I'm surprised I'm blanking on how to write "all but the last element"... Surely there's a better way than @a[^(*-1)] 21:35
dwarring a lot can happen in a minute
[Coke] moritz: so now we cannot have a bigints worth of things in a parcel? 21:36
TimToady 0..*-2 mebbe
21:36 frew left, xinming left, Alina-malina left, sivoais_ left, jakesyl_insomnia left
dwarring just a[*-1] ? 21:36
21:36 [Sno] left, leont left, Ugator left, davido_ left, bbkr left, nyuszika7h left, Grrrr left, masak left, jantore_ left, woshty left, wtw left, avar left, xinming joined, frew joined, telex joined 21:37 timotimo left
[Coke] dwarring: @a[0..^*-1] 21:37
lizmat Mouq: @a[0..*-2]
21:37 Alina-malina joined, vukcrni joined, [Sno] joined, leont joined, Ugator joined, davido_ joined, bbkr joined, Grrrr joined, masak joined, jantore_ joined, woshty joined, wtw joined, avar joined
[Coke] ... TimToady's is saner, I think. 21:37
21:38 brrt joined, jakesyl_insomnia joined, skarn joined, rhr left, ab5tract left, spider-mario left, clkao left, rmgk left, lue left, mls left, Pleiades` left, torbjorn left, lestrrat left, labster left, felher left, ponbiki left, epochbell left, ajr joined 21:39 ajr is now known as Guest58559, Guest58559 is now known as ajr_ 21:40 gfldex left, gfldex joined
dwarring all BUT last element 8-) 21:41
21:41 nyuszika7h joined
lizmat m: my @a = ^10; say @a[0..*-2] 21:41
camelia rakudo-moar 28e93b: OUTPUT«0 1 2 3 4 5 6 7 8␤»
TimToady bad netweather today... 21:42
21:42 rhr joined, ab5tract joined, spider-mario joined, clkao joined, rmgk joined, lue joined, mls joined, Pleiades` joined, torbjorn joined, lestrrat joined, labster joined, felher joined, ponbiki joined, epochbell joined 21:43 timotimo joined, jakesyl_insomnia left, esaym153 left, lizmat left, FlipBill left, tony-o left, awwaiid left, psch left, nwc10 left
TimToady ab5tract: your (cont) code might be more efficient if you guarantee $indices is a set outside the loop 21:43
21:43 esaym153 joined, lizmat joined, FlipBill joined, tony-o joined, awwaiid joined, psch joined, nwc10 joined, perlfan joined
TimToady otherwise might rebuild the set every iteration 21:43
21:43 eiro left, yoleaux left, jakesyl_insomnia joined, jakesyl_insomnia left, eiro joined, yoleaux joined, DarthGandalf joined, sinisalo.freenode.net sets mode: +v yoleaux 21:44 DarthGandalf left, timotimo left, skarn left, vukcrni left, Juerd left, zakharyas1 left, eternaleye left, rurban left, kjs_ left, baest left, jluis left, Patterner left, vendethiel left, Vlavv left, geekosaur left, dalek left, lumimies left, danaj left, haroldwu left, PerlJam left, jferrero left, JimmyZ left, avalenn left, hobbs left, Celelibi left, ggherdov left, nebuchadnezzar left, gtodd left, btyler_ left
TimToady in the absence of a "sufficiently smart optimizer" 21:44
21:45 DarthGandalf joined, sivoais joined 21:46 nyuszika7h left
ab5tract TimToady: thanks for the suggestion 21:47
21:48 ajr_ left, brrt left, pmichaud left, virtualsue left, mvuets left, sunnavy left, ugexe left, ruoso left, mattp_ left, eMBee left, apejens left, kshannon left, vukcrni joined, nyuszika7h joined, jakesyl_insomnia joined
ab5tract was actually assuming the presence of such a sufficiently smart optimizer 21:49
sub prune_array( $list, $indices is copy ) { $indices .= Set; [ for $list.kv { $^v unless $indices (cont) $^k } ] }
moritz [Coke]: that's going to be hard :-) 21:50
21:50 ggherdov joined, Celelibi joined, dalek joined, timotimo joined, skarn joined, Juerd joined, zakharyas1 joined, rurban joined, kjs_ joined, baest joined, jluis joined, Patterner joined, vendethiel joined, Vlavv joined, geekosaur joined, lumimies joined, danaj joined, haroldwu joined, jferrero joined, PerlJam joined, JimmyZ joined, avalenn joined, hobbs joined, nebuchadnezzar joined, gtodd joined, btyler_ joined, Spot__ joined, xiaomiao joined, jakesyl_insomnia left, jakesyl_insomnia joined, brrt joined
moritz [Coke]: if you have 8 billion GB of RAM, maybe 21:50
TimToady that's a heck of a lot of ram 21:51
21:51 skarn left, ChanServ sets mode: +v dalek 21:52 ggherdov left, ggherdov joined, mattp__ joined, colomon left, eternaleye joined, skarn joined 21:53 molaf left, slavik left, mephinet left, integral left, woolfy left, [Tux] left, flussence left, lsm-desktop left, yogan left, bcode left, fwilson left, obra left, renormalist left, kst left, krunen left, Exodist left, osfameron left, huf left, [Coke] left, sjn left, yeltzooo9 left, kshannon_ joined, molaf joined, slavik joined, mephinet joined, integral joined, woolfy joined, [Tux] joined, flussence joined, lsm-desktop joined, yogan joined, bcode joined, fwilson joined, obra joined, renormalist joined, kst joined, krunen joined, Exodist joined, huf joined, osfameron joined, [Coke] joined, sjn joined, yeltzooo9 joined, dg joined, cosimo joined, charsbar_____ joined, pnu joined
psch marshaling of listy types to jvm is limited to 2**31 - 6 already, so small parcels don't hurt there at least :) 21:54
21:54 simcop2387 joined, PZt left, ilbot3 left, hoelzro left
psch s/jvm/java/ 21:54
fsvo "small"
21:54 PZt joined, isacloud joined, camelia joined, profan joined, BooK joined, ivan`` joined, arnsholt joined, sorear joined, SHODAN joined, ilbot3 joined, hoelzro joined, sinisalo.freenode.net sets mode: +v camelia, PZt left, mattp__ left, sivoais left, IllvilJa left, salv0 left, petercommand left, breinbaas left, itz left, moritz left, Maddingue left, djanatyn left, Bucciarati left, BinGOs left, Gothmog_ left, TimToady left, anocelot left, jtpalmer left, jdv79 left, salparadise left 21:55 ajr_ joined, WAAWSLC joined, pmichaud joined, virtualsue joined, mvuets joined, sunnavy joined, ugexe joined, ruoso joined, mattp_ joined, eMBee joined, apejens joined, kshannon joined, raydiak joined, prammer joined, sunnavy_ joined, mattp_ left, kshannon left, sunnavy left, mvuets left 21:57 dwarring left, vladtz left, stux|RC-only joined, sivoais joined, dwarring joined, vladtz joined
masak amoquin: it's up to you whether you want to count different flips/rotations as equivalent or not. 21:57
lizmat m: say (0,1,2,3,4)[5]; say (0,1,2,3,4).list[5] # this seems inconsistent 21:58
camelia rakudo-moar 28e93b: OUTPUT«(Mu)␤Nil␤»
21:58 nyuszika7h left
lizmat m: my @a=^5; say @a[5] 21:58
camelia rakudo-moar 28e93b: OUTPUT«(Any)␤»
21:59 mattp__ joined, IllvilJa joined, salv0 joined, petercommand joined, breinbaas joined, itz joined, moritz joined, Maddingue joined, mtj_- joined, orevdiabl joined, erdic joined, oetiker joined, retupmoca joined, djanatyn joined, salparadise joined, Bucciarati joined, BinGOs joined, Gothmog_ joined, TimToady joined, anocelot joined, jdv79 joined, jtpalmer joined 22:00 nyuszika7h joined
TimToady ah, there's a global notice about the netsplits, which are not unexpected 22:01
dalek p/6pe: a2814dc | jnthn++ | src/vm/jvm/runtime/org/perl6/nqp/runtime/Ops.java:
Port parametric interning to point MoarVM does it.

With this, JVM passes all the parametric 6model tests so far, just as MoarVM does.
Heuristic branch merge: pushed 36 commits to rakudo/newio by lizmat
22:02 skarn left, eternaleye left, simcop2387 left, ggherdov left, Celelibi left, dalek left, timotimo left, Juerd left, zakharyas1 left, rurban left, kjs_ left, baest left, jluis left, Patterner left, vendethiel left, Vlavv left, geekosaur left, lumimies left, danaj left, haroldwu left, PerlJam left, jferrero left, JimmyZ left, avalenn left, hobbs left, nebuchadnezzar left, gtodd left, btyler_ left, Spot__ left, xiaomiao left, telex left
lizmat is going to sleep out the splits 22:02
moritz good night lizmat 22:03
jnthn 'night, lizmat
masak good night, lizmat
22:05 sivoais left, baest joined 22:06 skids left, Spot__ joined, simcop2387 joined, skarn joined, eternaleye joined, ggherdov joined, timotimo joined, Juerd joined, zakharyas1 joined, rurban joined, kjs_ joined, jluis joined, Patterner joined, vendethiel joined, Vlavv joined, geekosaur joined, lumimies joined, danaj joined, haroldwu joined, jferrero joined, PerlJam joined, JimmyZ joined, avalenn joined, hobbs joined, nebuchadnezzar joined, gtodd joined, btyler_ joined, xiaomiao joined, Celelibi joined 22:07 treehug88 left, Spot__ left, Spot__ joined, sivoais joined, Celelibi left, ggherdov left, zakharyas1 left 22:08 skarn left, dalek joined, ChanServ sets mode: +v dalek, simcop2387 left, Celelibi joined, skarn joined 22:09 donaldh joined, simcop2387 joined 22:10 telex joined, ggherdov joined 22:11 molaf left, timotimo left 22:14 telex left 22:19 timotimo joined 22:22 telex joined 22:24 ajr_ left 22:27 colomon joined, donaldh left, sirdancealot left
jnthn sleep & 22:30
japhb o/
22:32 sirdancealot joined 22:37 andreoss joined
brrt yes, me too 22:38
22:38 brrt left
andreoss may two lazy lists be mutually recursive in perl6? 22:43
japhb m: my @a; my @b; @a := { @b * 2 } ... *; @b := { @a + 1 } ... *; say @a[^5]; say @b[^5]; 22:46
camelia rakudo-moar e2c18c: OUTPUT«Inf Inf Inf Inf Inf␤Inf Inf Inf Inf Inf␤»
22:46 sirdancealot left
Mouq m: my @a; my @b; @a := { @b[(state$)++] * 2 } ... *; @b := 0,{ @a[(state$)++] + 1 } ... *; say @a[^5]; say @b[^5]; 22:47
camelia rakudo-moar e2c18c: OUTPUT«0 2 6 14 30␤0 1 3 7 15␤»
japhb m: my @a; my @b; @a := 0, { @b[$^i] * 2 } ... *; @b := 0, { @a[$^i] + 1 } ... *; say @a[^5]; say @b[^5];
camelia rakudo-moar e2c18c: OUTPUT«0 0 0 0 0␤0 1 1 1 1␤»
japhb m: my @a; my @b; @a := 0, { @b[$^i + 1] * 2 } ... *; @b := 0, { @a[$^i] + 1 } ... *; say @a[^5]; say @b[^5]; 22:48
camelia rakudo-moar e2c18c: OUTPUT«use of uninitialized value of type Nil in numeric context in block <unit> at /tmp/Vr5yN8SlDa:1␤␤This type does not support positional operations␤ in block at src/gen/m-CORE.setting:8408␤ in method reify at src/gen/m-CORE.setting:8384␤ in met…»
psch m: my @a; my @b; @a := 0, 1, { @b[$^i] * 2 } ... *; @b := 0, 1, { @a[$^i] + 1 } ... *; say @a[^5]; say @b[^5]; 22:50
camelia rakudo-moar e2c18c: OUTPUT«0 1 2 4 10␤0 1 2 3 5␤»
andreoss $^i has special meaning here?
psch m: say 1, 2, { $^a } ... 10 22:51
camelia rakudo-moar e2c18c: OUTPUT«1 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 ...␤»
andreoss m: my @a; my @b; @a := 0, 1, { @b[$_] * 2 } ... *; @b := 0, 1, {
@a[$_] + 1 } ... *; say @a[^5]; say @b[^5];
camelia rakudo-moar e2c18c: OUTPUT«===SORRY!=== Error while compiling /tmp/TvDxbQZ2Oz␤Unable to parse expression in block; couldn't find final '}' ␤at /tmp/TvDxbQZ2Oz:1␤------> , 1, { @b[$_] * 2 } ... *; @b := 0, 1, {⏏<EOL>␤»
andreoss m: my @a; my @b; @a := 0, 1, { @b[$_] * 2 } ... *; @b := 0, 1, { @a[$_] + 1 } ... *; say @a[^5]; say @b[^5];
camelia rakudo-moar e2c18c: OUTPUT«0 1 2 4 10␤0 1 2 3 5␤»
andreoss ok 22:52
psch andreoss: the ^-twigil syntax is useful e.g. when you want more than argument to the block
m: say 1, 2, { $^a } ... 10
camelia rakudo-moar e2c18c: OUTPUT«1 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 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 ...␤»
psch err
m: say 1, 2, { $^a + $^b } ... 10
camelia rakudo-moar e2c18c: OUTPUT«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 …»
psch m: say 1, 2, { $^a + $^b } ... 55 # that's how it stops...
camelia rakudo-moar e2c18c: OUTPUT«1 2 3 5 8 13 21 34 55␤»
psch *more than *one* argument... :) 22:53
andreoss {@a[$^i]} ... * will be depending on index than
so will work only for equally sized lists, not good 22:54
psch it takes the index from the current list and uses it to look up the next value from the other list, yes
lazy lists are potentially infinite
22:54 spider-mario left, kurahaupo joined
japhb Takes the previous value from the current list, not the index. 22:55
andreoss s/equally sized/eqaully fast growing/
psch japhb: "takes the previous value from the current list as the index for the other list" was my intended meaning. that didn't get clear i think 22:56
japhb andreoss, when the term on the left side of ... is a block, that block is inspected for parameter count, and that many previous values from the list are given to the block
m: my @a; my @b; @a := 0, 1, { @b[$_] * 2 } ... *; @b := 0, 1, { @a[$_] + 1 } ... *; say @a[^10]; say @b[^10];
camelia rakudo-moar e2c18c: OUTPUT«use of uninitialized value of type Nil in numeric context in block <unit> at /tmp/trfuVUGpsI:1␤␤use of uninitialized value of type Nil in numeric context in block <unit> at /tmp/trfuVUGpsI:1␤␤This type does not support positional operations␤ i…»
japhb Wah ...?
m: my @a; my @b; @a := 0, 1, { @b[$_] * 2 } ... *; @b := 0, 1, { @a[$_] + 1 } ... *; say @a[^5]; say @b[^5]; 22:57
camelia rakudo-moar e2c18c: OUTPUT«0 1 2 4 10␤0 1 2 3 5␤»
japhb m: my @a; my @b; @a := 0, 1, { @b[$_] * 2 } ... *; @b := 0, 1, { @a[$_] + 1 } ... *; say @a[^6]; say @b[^6];
camelia rakudo-moar e2c18c: OUTPUT«use of uninitialized value of type Nil in numeric context in block <unit> at /tmp/vyvOwdeX6T:1␤␤use of uninitialized value of type Nil in numeric context in block <unit> at /tmp/vyvOwdeX6T:1␤␤This type does not support positional operations␤ i…»
japhb Weird.
m: my @a; my @b; @a := 0, 1, { @b[$_] * 2 } ... *; @b := 0, 1, { @a[$_] + 1 } ... *; say @a[^5]; say @b[^10];
camelia rakudo-moar e2c18c: OUTPUT«0 1 2 4 10␤use of uninitialized value of type Nil in numeric context in block <unit> at /tmp/3q6kE3V1wI:1␤␤This type does not support positional operations␤ in block at src/gen/m-CORE.setting:8408␤ in method reify at src/gen/m-CORE.setting:83…»
japhb Oh ah, yes, running off each other's buffers. 22:58
I guess the point is it's *possible* to have mutually recursive infinite lists without too much work. Getting them really *right* is more work. 22:59
And of course, you can just do it a pile of other ways.
andreoss my @odds; my @evens; @odd = 1,3, @evens.map(* + 1); @evens = 2,@odd.map(* + 1); say @odd[^3]; 23:00
Mouq andreoss: assignment is eager 23:01
23:01 virtualsue left
Mouq my @odds; my @evens; @odd := 1, @evens.map(* + 1); @evens := 2,@odd.map(* + 1); say @odd[^3]; 23:01
andreoss m: my @odds; my @evens; @odd := 1,3, @evens.map(* + 1); @evens := 2,@odd.map(* + 1); say @odd[^3];
camelia rakudo-moar e2c18c: OUTPUT«===SORRY!=== Error while compiling /tmp/SQOFUgkij6␤Variable '@odd' is not declared. Did you mean '@odds'?␤at /tmp/SQOFUgkij6:1␤------> my @odds; my @evens; @odd⏏ := 1,3, @evens.map(* + 1); @evens := 2,␤ exp…»
andreoss m: my @odds; my @evens; @odds := 1,3, @evens.map(* + 1); @evens := 2,@odd.map(* + 1); say @odd[^3];
camelia rakudo-moar e2c18c: OUTPUT«===SORRY!=== Error while compiling /tmp/VD5MuuA5OL␤Variable '@odd' is not declared. Did you mean '@odds'?␤at /tmp/VD5MuuA5OL:1␤------> .map(* + 1); @evens := 2,@odd.map(* + 1)⏏; say @odd[^3];␤»
andreoss i'm writing code here because perl6 doesn't work inside Emacs btw 23:03
ab5tract andreoss: you can start an individual chat with camelia 23:04
andreoss m: my @odds; my @evens; @odds := 1,3, @evens.map(* + 1); @evens := 2, @odds.map(1 + *); say @odds[^5];
camelia rakudo-moar e2c18c: OUTPUT«1 3␤»
psch m: my @odds; my @evens; @odds := 1,3, @evens.map(* + 1) ... *; @evens := 2, @odds.map(1 + *) ... *; say @odds[^7]; 23:05
camelia rakudo-moar e2c18c: OUTPUT«1 3 5 7 9 11 13␤»
psch andreoss: you didn't make the lists lazy
err
infinite
lazy they couldve been, but they didn't generate more elements
andreoss isn't ...* stuff a range? 23:06
psch m: (1...*).WHAT.say
camelia rakudo-moar e2c18c: OUTPUT«(List)␤»
psch m: (1..*).WHAT.say
camelia rakudo-moar e2c18c: OUTPUT«(Range)␤»
andreoss foo...* = foo, foo+1, etc i mean
vendethiel m: my $a = 5; say ($a, *+1 ... *)[^10]; 23:07
camelia rakudo-moar e2c18c: OUTPUT«5 6 7 8 9 10 11 12 13 14␤»
vendethiel andreoss: this?
andreoss vendethiel: i wanted mutual recursion between lists 23:08
vendethiel oooh, backlogging again
andreoss vendethiel: what your code was supposed to show? 23:09
vendethiel that you can "iterate off" a value
psch andreoss: you need ... * to have the lists continue to generate values 23:16
m: my @a := 1,2,3; @a.infinite.say; @a := 1,2,3...*; @a.infinite.say 23:18
camelia rakudo-moar e2c18c: OUTPUT«Nil␤1␤»
23:19 kjs_ left 23:20 jack_rabbit joined 23:21 kjs_ joined 23:25 kjs_ left
andreoss m: multi car2([$x, *@xs], [$y, *@ys] where $x == $y) { $x, $y }; say car2([1,2,3], [1,3,5]); 23:27
camelia rakudo-moar e2c18c: OUTPUT«Cannot call 'infix:<==>'; none of these signatures match:␤:(Any $?)␤:(Any \a, Any \b)␤:(Real \a, Real \b)␤:(Int:D \a, Int:D \b)␤:(int $a, int $b)␤:(Num:D \a, Num:D \b --> Bool)␤:(num $a, num $b --> Bool)␤:(Rational:D \a, Rational:D \b)␤:(…»
andreoss m: multi car2([$x, *@xs], [$y, *@ys]) { $x, $y }; say car2([1,2,3], [1,3,5]);
camelia rakudo-moar e2c18c: OUTPUT«1 1␤»
andreoss a bug? 23:28
psch m: sub f ($a, $b where $a == $b) { 1 }; say f 1, 1 23:29
camelia rakudo-moar e2c18c: OUTPUT«1␤»
psch considering that ^^^ works i'd say "maybe". i'm not sure subsignature can be used in post-constraints outside of themselves
andreoss: maybe check the specs 23:30
i'll go to bed o/
23:30 telex left 23:32 telex joined 23:33 clkao left, Spot__ left
Mouq has been trying to do the lazy @odd/@even snippet with Supplies and finds it surprisingly difficult 23:44
Dang immutability and all :P
lue found she somehow gets duplicate entries in a Set while trying to solve masak's mini-challenge; hopefully updating rakudo will fix it :) .
Mouq lue: What kind of entities are they? I.e. are you sure they have the same .WHICH ? 23:45
lue Mouq: the set I get in the end is set($(0, 0),$(0, 1),$(0, 2),$(2, 0),$(2, 1),$(2, 2),$(0, 0),$(1, 0),$(2, 0),$(0, 2),$(1, 2))
masak has a solution to his own mini-challenge which is 33 LoC, but which counts rotations/flips as distinct 23:46
lue The weird thing is that I _do_ eliminate some duplicates when I convert it to a set, just not all of them, apparently :/
Mouq m: say $(0,0) === $(0,0)
camelia rakudo-moar e2c18c: OUTPUT«True␤»
Mouq Oh, huh
lue m: Array.new($(0, 0), $(0, 1), $(0, 1), $(0, 2), $(2, 0), $(2, 1), $(2, 1), $(2, 2), $(0, 0), $(1, 0), $(1, 0), $(2, 0), $(0, 2), $(1, 2), $(1, 2), $(2, 2)).Set.perl.say
camelia rakudo-moar e2c18c: OUTPUT«set($(0, 0),$(0, 1),$(0, 2),$(2, 0),$(2, 1),$(2, 2),$(1, 0),$(1, 2))␤»
lue if I take the array and convert it to a set in a snippet like that, it works. Somehow. 23:47
23:47 skids joined
Mouq m: say $(0,0) === (0,0) 23:47
camelia rakudo-moar e2c18c: OUTPUT«True␤»
Mouq m: say $(0,0) === (0;0)
camelia rakudo-moar e2c18c: OUTPUT«False␤»
lue That's why I'm hoping a recent Rakudo will solve it, 'cos otherwise I don't know. 23:49
23:50 bbkr left
lue masak: my solution is 179 lines so far (granted there are plenty of comments). I think I might be just a bit out of shape on this whole programming thing :P . 23:51
andreoss /part 23:52
23:52 andreoss left
masak lue: don't fret. I'm creating majorly. 23:52
cheating*
(and I learned a thing or two about spanning trees on grids in the process.)
lue I suppose if updating rakudo doesn't work, I could switch to Complex numbers and hope that solves it :P . 23:55
masak I don't really like using Complex as a 2-tuple. 23:56
IMO, it's a code smell to use Complex as a pair of reals without using any aspects of its multiplication. 23:57
Mouq loved to use them as LoLs before he implemented LoLs :P
23:58 adu joined
lue masak: Of course Complex numbers aren't the best choice here, but if I find lists don't want to behave properly (yet again), I'll use complex numbers instead. 23:59

Powered by Cro and the Raku® Programming Language.

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

Thank you!