»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'perl6: say 3;' or rakudo:, niecza:, std:, or /msg p6eval perl6: ... | irclog: irc.perl6.org/ | UTF-8 is our friend!
Set by sorear on 4 February 2011.
00:00 Chillance left
jlaire p6: class C { method m() { <a b> } }; my @cs = C.new xx 3; say @cs.map({.m.list}).elems 00:02
p6eval pugs, rakudo d9fd17: OUTPUT«6␤»
..niecza v17-2-gb8fcf85: OUTPUT«3␤»
jlaire :/
00:03 eviltwin_b joined
jlaire how to flatten a list of lists like that in niecza? 00:03
00:03 geekosaur left, eviltwin_b is now known as geekosaur 00:06 dorlamm left
jlaire found a workaround 00:06
00:17 eviltwin_b joined, geekosaur left, eviltwin_b is now known as geekosaur, sisar joined 00:19 eviltwin_b joined, geekosaur left, eviltwin_b is now known as geekosaur 00:22 fgomez left
Radvendii what's the proper way to undefine something? 00:31
00:31 fgomez joined
Radvendii nom: my $a = null; 00:32
p6eval rakudo d9fd17: OUTPUT«===SORRY!===␤CHECK FAILED:␤Undefined routine '&null' called (line 1)␤»
Radvendii nom: my $a = nil;
p6eval rakudo d9fd17: OUTPUT«===SORRY!===␤CHECK FAILED:␤Undefined routine '&nil' called (line 1)␤»
Radvendii nom: my $a.undef
p6eval rakudo d9fd17: OUTPUT«===SORRY!===␤Confused␤at /tmp/8SZTzdweWd:1␤»
idlecuriosity nom: my $a = Nil; 00:33
p6eval rakudo d9fd17: ( no output )
Radvendii ah
thanks
idlecuriosity np, I learnt it just now :)
Radvendii hmm... that doesn't work for removing indicies of an array 00:38
nom: my @a = 1,2,3; @a[1]=Nil; @a;
p6eval rakudo d9fd17: ( no output )
Radvendii nom: my @a = 1,2,3; @a[1]=Nil; say @a;
p6eval rakudo d9fd17: OUTPUT«1 3␤»
Radvendii nom: my @a = 1,2,3; @a[1]=Nil; say @a.perl;
p6eval rakudo d9fd17: OUTPUT«Array.new(1, Nil, 3)␤»
Radvendii nom: my @a = 1,2,3; @a.delete(1); say @a.perl; 00:39
p6eval rakudo d9fd17: OUTPUT«Array.new(1, Any, 3)␤»
Radvendii awww... really?
TimToady are you trying to find splice()?
Radvendii nom: my @a = 1,2,3; @a.splice(1); say @a.perl; 00:40
p6eval rakudo d9fd17: OUTPUT«Array.new(1)␤»
Radvendii huh?
I don't know, what's it do?
jlaire nom: my @a = <a b c d e f>; @a.splice(3); say @a.perl 00:41
p6eval rakudo d9fd17: OUTPUT«Array.new("a", "b", "c")␤»
Radvendii I guess what i'm doing is simpler than i thought
sisar Radvendii: here is splice in perl5 explained: szabgab.com/splice-to-slice-and-dic...-perl.html You can easily figure it out in perl6 :) 00:42
Radvendii i just want to clip off the last element of a list if it's Nil
that's easy enough
sorear Radvendii: go read console_codes(4) or one of the 1e4 equivalent documents on the Web 00:43
Radvendii oh, is that for the clear screen thing? 00:44
sorear n: say <a b> xx 3
p6eval niecza v17-2-gb8fcf85: OUTPUT«a b a b a b␤»
sorear Radvendii: clear screen, cursor positioning, colors, the workd
Radvendii sweet! 00:45
thanks
00:57 icwiener left, scott__ joined 01:01 replore_ joined 01:04 geekosaur left 01:05 geekosaur joined, lestrrat left 01:06 lestrrat joined 01:10 tokuhirom left 01:11 eviltwin_b joined, geekosaur left, eviltwin_b is now known as geekosaur 01:16 sporous left, replore_ left 01:17 eviltwin_b joined, geekosaur left, eviltwin_b is now known as geekosaur 01:21 sporous joined 01:34 raiph joined, eviltwin_b joined, geekosaur left 01:35 eviltwin_b is now known as geekosaur 01:38 scott___ joined 01:41 scott__ left 01:43 scott___ left 01:45 scott__ joined 01:48 geekosaur left, eviltwin_b joined 01:49 eviltwin_b is now known as geekosaur 01:56 whiteknight left 01:59 geekosaur left, eviltwin_b joined 02:00 eviltwin_b is now known as geekosaur 02:01 thou left 02:09 Radvendii left 02:10 eviltwin_b joined, geekosaur left, eviltwin_b is now known as geekosaur 02:21 geekosaur left, eviltwin_b joined 02:22 eviltwin_b is now known as geekosaur 02:23 geekosaur left, eviltwin_b joined 02:24 eviltwin_b is now known as geekosaur 02:32 geekosaur left 02:33 geekosaur joined 02:37 eviltwin_b joined, geekosaur left, eviltwin_b is now known as geekosaur 02:38 orafu left, orafu joined 02:39 eviltwin_b joined, geekosaur left, sftp_ left, eviltwin_b is now known as geekosaur 02:42 eviltwin_b joined, geekosaur left, eviltwin_b is now known as geekosaur 02:44 sftp joined 02:53 eviltwin_b joined, geekosaur left, eviltwin_b is now known as geekosaur 03:04 envi_ joined, crab2313 joined 03:07 wolfman2000 joined 03:11 alvis` joined, eviltwin_b joined, geekosaur left, eviltwin_b is now known as geekosaur 03:12 alvis left 03:15 alc joined 03:37 mikemol left 03:38 mmol_saffron joined 03:41 kaleem joined, geekosaur left, eviltwin_b joined, eviltwin_b is now known as geekosaur 03:45 ponbiki joined 03:48 samurai2 joined 03:49 samurai2 left 03:52 alc left 03:56 geekosaur left 03:57 geekosaur joined 04:07 mmol_saffron left 04:26 crab2313 left 04:38 MContagious joined 04:48 alc joined 04:50 sisar left 04:56 alc left 05:03 thou joined 05:04 eviltwin_b joined, geekosaur left, eviltwin_b is now known as geekosaur 05:08 ponbiki left 05:10 alc joined 05:12 alc left 05:18 eviltwin_b joined, geekosaur left 05:19 geekosaur joined 05:20 eviltwin_b left 05:21 eviltwin_b joined, geekosaur left, eviltwin_b is now known as geekosaur 05:23 fgomez left 05:30 eviltwin_b joined, geekosaur left 05:31 eviltwin_b is now known as geekosaur 05:35 geekosaur left, NamelessTee joined 05:37 geekosaur joined 05:48 eviltwin_b joined, geekosaur left, eviltwin_b is now known as geekosaur 06:06 eviltwin_b joined, geekosaur left, eviltwin_b is now known as geekosaur 06:07 NamelessTee left 06:15 wtw joined 06:17 sisar joined 06:18 sisar left, sisar joined 06:19 eviltwin_b joined, geekosaur left, eviltwin_b is now known as geekosaur 06:21 eviltwin_b joined, geekosaur left, eviltwin_b is now known as geekosaur 06:31 geekosaur left, eviltwin_b joined 06:32 geekosaur joined, fhelmberger joined 06:35 eviltwin_b left, snearch joined 06:38 geekosaur left 06:42 geekosaur joined 07:03 idlecuriosity left 07:04 snearch left 07:07 lestrrat left 07:09 lestrrat joined 07:16 xinming_ joined 07:19 xinming left 07:21 eviltwin_b joined, geekosaur left 07:22 eviltwin_b is now known as geekosaur
moritz \o 07:39
jnthn: new problem: Pair isn't in BOOTSTRAP.pm, so I can't use Pair in Grammar.pm while compiling the setting :/ 07:40
jnthn: and traits.pm does 'use Perl6::BOOTSTRAP' 07:41
hm, I might special-case empty import lists
07:50 thelazydeveloper joined
moritz that way seems simpler anyway 07:50
08:14 fglock joined
jnthn mornin' 08:24
tadzik hello 08:25
jnthn moritz: I don't think you should need import lists while compiling the setting.
It only does whatever the default thing is.
moritz right 08:26
jnthn: oh, and 'use Test :DEFAULT' produces DEFAULT => Mu 08:30
I guess it's the same bug as
r: sub f { return a => 5 }; say f.perl 08:31
p6eval rakudo d9fd17: OUTPUT«"a" => Mu␤»
08:31 fgomez joined 08:32 fgomez left, fgomez joined
jnthn Huh, it shouldn't be producing a named parameter at all... 08:33
That's the point of looking at EXPR...
moritz well, I think it creates a proper Pair 08:34
08:35 LylePerl left
moritz nqp: sub f($x?) { say(nqp::defined($x)) }; f() 08:43
p6eval nqp: OUTPUT«Unrecognized nqp:: opcode 'nqp::defined' at line 1, near ") }; f()"␤current instr.: 'nqp;HLL;Grammar;panic' pc 22008 (src/stage2/gen/NQPHLL.pir:7047) (src/stage2/gen/NQPHLL.pm:329)␤»
moritz nqp: sub f($x?) { say(pir::defined($x)) }; f()
p6eval nqp: OUTPUT«0␤»
moritz nqp: sub f($x?) { say(pir::defined($x)) }; f([])
p6eval nqp: OUTPUT«1␤»
08:48 daxim joined
dalek kudo/use-arglist: 6145974 | moritz++ | src/ (2 files):
allow importing of tags

the default is not imported when an explicit tag is provided		 08:55
08:57 birdwindupbird joined
jnthn yay :) 09:01
tadzik ooo
moritz note that it currently does not complain about nonexisting tags 09:02
so you can write
use Test :NONE;
though I think you can achieve the same with need Test; anyway
tadzik and it won't import anything, right? 09:03
jnthn I'm not sure the spec says whether it should complain or not.
moritz right
jnthn At least, I don't recall S11 indicating as such.
moritz jnthn: to me, importing looks more like a sub call than a method call, so by extension I think it should complain
jnthn Hmm, I'd not thunk of it that way :) 09:04
tadzik ha, clever 09:08
09:09 geekosaur left
tadzik although I see it more like obj.do_something_for_me than obj.?do_something_for_me 09:09
so the conclusion is the same
jnthn Note that there's various other bits of magic in arglist stuff.
moritz well, TimToady++'s ruling was that importing calls a macro in the loaded module 09:10
so until our macros are a bit more capable, we'll just do an approximation anyway
jnthn Of note, :MANDATORY 09:11
09:11 geekosaur joined
jnthn a...macro? 09:11
moritz yes 09:12
because it has the option to do all the intersting stuff that aliasing routines can't do
like, adding a CATCH block
or tweaking slangs
and stuff like that
jnthn ah 09:13
Through COMPILING mostly, I guess.
moritz yes
so no point in doing it now, since we don't hav e COMPILING
jnthn Right
moritz r: class A { sub f() is export { say 'oh shark' } }; import A; f 09:14
p6eval rakudo d9fd17: OUTPUT«===SORRY!===␤CHECK FAILED:␤Undefined routine '&f' called (line 1)␤»
jnthn That doesn't work because we only install in UNIT::EXPORT so far, rather than through all the packages we're nested in. 09:16
moritz so export lists are per package right now? 09:17
erm, sorry, per file?
jnthn Yes, per compilation unit.
S11 says that in
class A { class B { sub foo() is export { } } } 09:18
It'd be installed in B::EXPORT::DEFAULT::<&foo>, A::EXPORT::DEFAULT::<&foo> and the UNIT one.
Today it only goes in UNIT
dalek kudo/use-arglist: 3c393c6 | moritz++ | src/Perl6/ (2 files):
refactor compile time evaluation into a World method
 09:19
kudo/use-arglist: 5b60299 | moritz++ | src/Perl6/Grammar.pm:
fix import + arglist
09:20 eviltwin_b joined, geekosaur left, kresike joined 09:21 eviltwin_b is now known as geekosaur
moritz meh, it seems that while not importing stuff works, importing custom tags still doesn't install the routines :/ 09:24
but explicitly importing :DEFAULT or :ALL works 09:25
jnthn Does exporting with tags other than the defautl one actually to the right instatllation?
moritz I guess not
jnthn r: sub foo() is export(:BLAH) { }
p6eval rakudo d9fd17: ( no output )
jnthn I don't recall implementing that bit.
moritz I wonder how it dispatches at all 09:26
there's a multi trait_mod:<is>)(Routine:D \$r, :$export!) 09:27
09:27 not_gerd joined
not_gerd hello, #perl6 09:27
some benchmarking fun: gist.github.com/2626839
moritz \o not_gerd
not_gerd the high-level code outperforms the low-level version ;) 09:28
moritz not_gerd: I think if you do subroutines with low-level types, you might be faster
not_gerd: because then rakudo can infer at compile time that there's a native num to pass
and no boxing will happen
jnthn If you use native types in the wrong way, it will cost more due to loads of boxing/unboxing. 09:29
moritz r: multi trait_mod:<is>(Routine:D \$r, :$ext!, $args) { say $args.perl }; sub f is ex(:blah) { } 09:30
p6eval rakudo d9fd17: OUTPUT«===SORRY!===␤Cannot put required parameter after variadic parameters␤at /tmp/lI9rjYYKWy:1␤»
moritz r: multi trait_mod:<is>(Routine:D \$r, $args, :$ext!) { say $args.perl }; sub f is ex(:blah) { }
p6eval rakudo d9fd17: OUTPUT«===SORRY!===␤Cannot call 'trait_mod:<is>'; none of these signatures match:␤:(Attribute:D $attr, :rw(:$rw)!)␤:(Attribute:D $attr, :readonly(:$readonly)!)␤:(Attribute:D $attr, :box_target(:$box_target)!)␤:(Routine:D $r, :rw(:$rw)!)␤:(Routine:D $r, :default(:$default)…
moritz r: multi trait_mod:<is>(Routine:D \$r, $args, :$ext!) { say $args.perl }; sub f is ex('blah') { } 09:31
jnthn moritz: ex vs ext
p6eval rakudo d9fd17: OUTPUT«===SORRY!===␤Cannot call 'trait_mod:<is>'; none of these signatures match:␤:(Attribute:D $attr, :rw(:$rw)!)␤:(Attribute:D $attr, :readonly(:$readonly)!)␤:(Attribute:D $attr, :box_target(:$box_target)!)␤:(Routine:D $r, :rw(:$rw)!)␤:(Routine:D $r, :default(:$default)…
moritz r: multi trait_mod:<is>(Routine:D \$r, $args, :$ex!) { say $args.perl }; sub f is ex('blah') { }
p6eval rakudo d9fd17: OUTPUT«"blah"␤»
moritz r: multi trait_mod:<is>(Routine:D \$r, $args, :$ex!) { say $args.perl }; sub f is ex(:blah) { }
p6eval rakudo d9fd17: OUTPUT«===SORRY!===␤Cannot call 'trait_mod:<is>'; none of these signatures match:␤:(Attribute:D $attr, :rw(:$rw)!)␤:(Attribute:D $attr, :readonly(:$readonly)!)␤:(Attribute:D $attr, :box_target(:$box_target)!)␤:(Routine:D $r, :rw(:$rw)!)␤:(Routine:D $r, :default(:$default)…
moritz jnthn: thanks
r: multi trait_mod:<is>(Routine:D \$r, |$args, :$ex!) { say $args.perl }; sub f is ex(:blah) { }
p6eval rakudo d9fd17: OUTPUT«Capture.new( list => ().list, hash => EnumMap.new("ex", 1, ))␤»
moritz r: multi trait_mod:<is>(Routine:D \$r, *%args, :$ex!) { say %args.perl }; sub f is ex(:blah) { } 09:32
p6eval rakudo d9fd17: OUTPUT«===SORRY!===␤Required named parameter 'ex' not passed␤»
moritz r: multi trait_mod:<is>(Routine:D \$r, :$ex!, *%args) { say %args.perl }; sub f is ex(:blah) { }
p6eval rakudo d9fd17: OUTPUT«().hash␤»
09:36 thou left, eviltwin_b joined, geekosaur left, eviltwin_b is now known as geekosaur
moritz r: class A { sub f is export(:YARR) { } }; say UNIT::EXPORT::.keys 09:40
p6eval rakudo d9fd17: OUTPUT«ALL DEFAULT␤»
09:44 geekosaur left 09:45 geekosaur joined
arnsholt .u ṛ 09:46
phenny U+1E5B LATIN SMALL LETTER R WITH DOT BELOW (ṛ)
arnsholt .u ṝ 09:47
phenny U+1E5D LATIN SMALL LETTER R WITH DOT BELOW AND MACRON (ṝ)
arnsholt .u ḷ
phenny U+1E37 LATIN SMALL LETTER L WITH DOT BELOW (ḷ)
arnsholt .u ḥ
phenny U+1E25 LATIN SMALL LETTER H WITH DOT BELOW (ḥ)
09:55 geekosaur left 09:56 geekosaur joined
moritz jnthn: how do I access named arguments to a trait? 09:58
10:04 brrt joined, eviltwin_b joined, geekosaur left 10:05 eviltwin_b is now known as geekosaur
moritz r: multi trait_mod:<is>(|$a) { say $a.perl }; sub f is foo(:bar) { } 10:05
p6eval rakudo d9fd17: OUTPUT«Capture.new( list => (sub f() { ... },).list, hash => EnumMap.new("foo", 1, ))␤»
moritz no bar in sight :(
(pun intended)
tadzik pun indented 10:13
moritz I guess the reason is that the argument list to the 'is' trait is parsed as a circumfix, which contains a semilist, not an arglist 10:15
so there's no logic to handle named arguments
but, otoh, shouldn't named arguments then become simple pairs?
r: my $x = do { :foo }; say $x
p6eval rakudo d9fd17: OUTPUT«"foo" => Bool::True␤»
moritz r: my $x = (:foo) ; say $x 10:16
p6eval rakudo d9fd17: OUTPUT«"foo" => Bool::True␤»
10:20 crab2313 joined, crab2313 left
moritz ok, importing by tag name works, though you have to set up the export package yourself :-) 10:21
module A { our sub f is export(:YARR) { say 'oh shark' } } 10:22
UNIT::EXPORT::<YARR> := package YARR { our &f = &A::f; }
and then
use A :YARR; f; says 'oh shark'
10:24 not_gerd left 10:28 cogno joined 10:35 cogno left 10:38 brrt left 10:44 JimmyZ joined 10:51 cogno joined
cogno When one watches a commit page on github, how does he know what branch the commit belongs to? 10:58
tadzik I'm not sure if one does 10:59
cogno Too bad. 11:00
tadzik any particular commit you have in mind? 11:03
11:06 brrt joined 11:10 snearch joined, moritz joined 11:11 cogno left
jnthn moritz: I suspect you've found a bug in the handling or args to the trait mod. 11:15
11:17 _Vi joined
_Vi Is Perl6 ready for use cases like "perl -pe 's/something/anything/' -i *.c", but with per6'th numerous regex enhancements? 11:19
11:19 tyatpi joined 11:21 cogno joined 11:26 cogno left
tadzik I don't think we have -i implemented 11:27
brrt would it be possible - and if so, useful - to create a subset Filename of String where the file exists? 11:28
11:28 lestrrat left
tadzik possible? I think so 11:29
11:29 lestrrat joined
tadzik r: subset Filename of Str where { .IO ~~ :e } 11:30
p6eval rakudo d9fd17: ( no output )
jnthn r: subset Filename of Str where { .IO ~~ :e }; say 'README' ~~ Filename; say 'lolno' ~~ Filename
p6eval rakudo d9fd17: OUTPUT«False␤False␤»
tadzik r: subset Filename of Str where { .IO ~~ :e }; Filename $a = "README"
p6eval rakudo d9fd17: OUTPUT«===SORRY!===␤Confused␤at /tmp/s24cf4g888:1␤»
jnthn r: subset Filename of Str where { .IO ~~ :e }; say 'perl6' ~~ Filename; say 'lolno' ~~ Filename
p6eval rakudo d9fd17: OUTPUT«False␤False␤»
tadzik r: subset Filename of Str where { .IO ~~ :e }; my Filename $a = "README"
p6eval rakudo d9fd17: OUTPUT«Type check failed in assignment to '$a'; expected 'Filename' but got 'Str'␤ in block <anon> at /tmp/i_JnSSyRX1:1␤␤»
jnthn r: say "README".IO ~~ :e 11:31
p6eval rakudo d9fd17: OUTPUT«False␤»
jnthn r: say "README" ~~ :e
p6eval rakudo d9fd17: OUTPUT«Method 'e' not found for invocant of class 'Str'␤ in method ACCEPTS at src/gen/CORE.setting:5639␤ in block <anon> at /tmp/Q58SGxK3SP:1␤␤»
moritz brrt: note that type checks are supposed to pure functions
brrt: so if somebody adds or deletes a file while your program is running, that might cause some havoc 11:32
brrt dangerous method 11:33
but still
11:34 geekosaur left 11:36 geekosaur joined 11:45 noam left 11:46 noam joined
_Vi Is (will be) Perl6 in general as good tool for ad-hoc oneliners as Perl5 is? It would be good to be able to use it as "enhanced sed" with advanced Unicode capabilities. 11:46
11:46 noam left, noam joined
moritz _Vi: we haven't reached that level yet, but we certainly intend to do it 11:47
I've been implementing a few command line options these past weeks, maybe I should consider doing some more in the -F or -i direction 11:48
afk&
_Vi (In my head the "soul and spirit of Perl" lies in oneliners and short non-modular text-processing scripts. Perl6 does not have this spirit yet.)
dalek rlito: 6244c95 | (Flavio S. Glock)++ | t5/op/exp.t:
Perlito5 - add t/op/exp.t
 11:53
11:57 brrt left 11:58 _sri joined, skids left, moritz left 11:59 moritz joined 12:05 _Vi left 12:06 mtk joined 12:31 MContagious left 12:34 geekosaur left, eviltwin_b joined, eviltwin_b is now known as geekosaur 12:48 eviltwin_b joined, geekosaur left, eviltwin_b is now known as geekosaur 12:50 PacoAir joined, eviltwin_b joined, geekosaur left 12:51 eviltwin_b is now known as geekosaur 12:52 geekosaur left, geekosaur joined 12:53 snearch left 12:55 PacoAir left 13:04 bluescreen10 joined, PacoAir_ joined 13:06 eviltwin_b joined, geekosaur left, eviltwin_b is now known as geekosaur 13:07 _Vi joined 13:08 bluescreen10 left 13:10 Guest51792 left, eviltwin_b joined, geekosaur left 13:11 eviltwin_b is now known as geekosaur 13:13 eviltwin_b joined, geekosaur left 13:14 bluescreen10 joined, geekosaur joined 13:15 xenu joined, xenu is now known as Guest71264 13:17 eviltwin_b left 13:24 bluescreen10 left 13:25 skids joined, bluescreen10 joined 13:31 eviltwin_b joined, geekosaur left 13:32 eviltwin_b is now known as geekosaur 13:36 tyatpi left 13:40 mynameisbrian joined 13:43 pernatiy left 13:45 araujo left 13:46 araujo joined, pernatiy joined, pernatiy left 13:48 pernatiy joined 13:49 tyatpi joined
colomon n: say '\S\' 13:51
p6eval niecza v17-2-gb8fcf85: 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/BSnyHC3wCm line 1:␤------> say⏏ '\S\'␤␤Confused at /tmp/BSnyHC3wCm line 1:␤------> …
colomon n: say 'A'.ord 13:52
p6eval niecza v17-2-gb8fcf85: OUTPUT«65␤»
13:54 pernatiy_ joined, pernatiy left
colomon what's the p6 equivalent of $' ? 13:54
13:55 kaleem left
colomon ah, $/.postmatch 13:57
13:59 DarthGandalf joined, eviltwin_b joined, geekosaur left 14:00 eviltwin_b is now known as geekosaur
colomon n: say "face".hex 14:00
p6eval niecza v17-2-gb8fcf85: OUTPUT«Unhandled exception: Unable to resolve method hex in class Str␤ at /tmp/r3CZD0vf4z line 1 (mainline @ 2) ␤ at /home/p6eval/niecza/lib/CORE.setting line 3840 (ANON @ 3) ␤ at /home/p6eval/niecza/lib/CORE.setting line 3841 (module-CORE @ 65) ␤ at /home/p6e…
14:00 mynameisbrian left
colomon r: say "face".hex 14:00
p6eval rakudo d9fd17: OUTPUT«Method 'hex' not found for invocant of class 'Str'␤ in block <anon> at /tmp/tC9Oi3Rs_h:1␤␤»
colomon 's brain is starting to hurt from switching back and forth with p5 14:01
jnthn r: say :16('face')
p6eval rakudo d9fd17: OUTPUT«64206␤»
colomon r: my $a = 'face'; say :16($a)
p6eval rakudo d9fd17: OUTPUT«64206␤»
colomon n: my $a = 'face'; say :16($a)
p6eval niecza v17-2-gb8fcf85: OUTPUT«64206␤»
colomon jnthn++ 14:02
14:03 eviltwin_b joined, geekosaur left, eviltwin_b is now known as geekosaur 14:04 kaare_ joined 14:09 Guest71264 left 14:14 xenu joined 14:15 xenu is now known as Guest66045 14:23 eviltwin_b joined, geekosaur left, eviltwin_b is now known as geekosaur, tyatpi left 14:28 kresike left 14:31 PacoAir_ left 14:33 cognominal joined 14:34 cognominal___ left 14:37 tyatpi joined, sporous left, sporous joined 14:39 araujo left, eviltwin_b joined, geekosaur left 14:40 eviltwin_b is now known as geekosaur, araujo joined 14:41 eviltwin_b joined, geekosaur left 14:42 eviltwin_b is now known as geekosaur 14:47 eviltwin_b joined, geekosaur left, eviltwin_b is now known as geekosaur 14:51 Kresike joined 14:53 eviltwin_b joined, geekosaur left 14:54 eviltwin_b is now known as geekosaur 14:55 zby_home_ joined 14:57 fgomez left 14:59 birdwindupbird left 15:01 erkan left 15:05 geekosaur left, eviltwin_b joined, eviltwin_b is now known as geekosaur 15:12 tyatpi left, eviltwin_b joined, geekosaur left 15:13 eviltwin_b is now known as geekosaur 15:23 glass left 15:24 envi_ left 15:25 envi_ joined 15:26 sftp left 15:27 sftp joined 15:28 tyatpi joined, Guest66045 left, tokuhirom joined 15:29 carlin joined 15:30 envi_ left
masak lol shark hai, #perl6 15:30
jnthn fin evening, masak 15:32
15:32 envi_ joined
moritz \o 15:33
15:35 fglock left
masak .oO( have the appropriate amount of fin ) 15:35
15:35 fgomez joined 15:36 MContagious joined 15:38 xenu joined, xenu is now known as Guest74228 15:40 fgomez left 15:49 JimmyZ left, Psyche^ joined
moritz fin ack 15:49
masak .oO( NAK on wood ) 15:50
geekosaur too NAKered 15:51
moritz no details please 15:52
15:53 Patterner left, Psyche^ is now known as Patterner 15:56 geekosaur left, eviltwin_b joined, eviltwin_b is now known as geekosaur 16:09 geekosaur left
raiph p6: say lines("asdf")[0] # expected this to look for a file "asdf" and say its first line 16:13
p6eval pugs: OUTPUT«*** No such subroutine: "&lines"␤ at /tmp/iNzpLXdSwX line 1, column 5 - line 2, column 1␤»
..rakudo d9fd17, niecza v17-2-gb8fcf85: OUTPUT«asdf␤»
jnthn raiph: That's slept lines("asdf".IO)[0]
raiph thx jnthn 16:14
moritz r: say lines("asdf".IO)[0]
p6eval rakudo d9fd17: OUTPUT«Unable to open filehandle from path 'asdf'␤ in method open at src/gen/CORE.setting:6588␤ in method get at src/gen/CORE.setting:6608␤ in code <anon> at src/gen/CORE.setting:6632␤ in sub coro at src/gen/CORE.setting:4926␤ in method reify at src/gen/CORE.setting:…
TimToady p6: say words('the quick brown fox jumped')[3]
p6eval rakudo d9fd17: OUTPUT«===SORRY!===␤CHECK FAILED:␤Undefined routine '&words' called (line 1)␤»
..pugs: OUTPUT«*** No such subroutine: "&words"␤ at /tmp/gtj1gJE0m_ line 1, column 5 - line 2, column 1␤»
..niecza v17-2-gb8fcf85: OUTPUT«===SORRY!===␤␤Undeclared routine:␤ 'words' used at line 1␤␤Unhandled exception: Check failed␤␤ at /home/p6eval/niecza/boot/lib/CORE.setting line 1364 (die @ 3) ␤ at /home/p6eval/niecza/src/STD.pm6 line 1147 (P6.comp_unit @ 33) ␤ at /hom…
jnthn
.oO( I'm not sure if that typo was ironic or Freudian... )
 16:15
bbkr wow, awesome trick with "file".IO
raiph right. (btw i like your spelling of spelled)
16:15 geekosaur joined
TimToady p6: say 'the quick brown fox jumped'.words[3] 16:17
p6eval rakudo d9fd17, niecza v17-2-gb8fcf85: OUTPUT«fox␤»
..pugs: OUTPUT«*** No such method in class Str: "&words"␤ at /tmp/7Jvqdl04jT line 1, column 5 - line 2, column 1␤»
TimToady multi method words ( Str $input: Int $limit = Inf --> List ) is export
supposed to be exported as as a function
16:18 fglock joined, fgomez joined
masak TimToady: could we get 'invert' as a function, too? I'd find that useful. 16:18
16:19 geekosaur left, geekosaur joined
TimToady sure, anything that makes sense both ways and is common enough for it to matter should be considered for export at least 16:20
masak \o/
oh oh
TimToady but I'm sure there are methods that are going to be rare enough that they don't merit such treatement 16:21
masak and List.invert (as being the same as List.kv.hash.invert) -- I'd like that too
TimToady um
why not List.pairs.invert?
.kv is a micro-smell
masak ah. 16:22
16:22 MContagious left
masak yeah. maybe that's enough, then. 16:22
p6: say <a b c>.pairs.invert
p6eval pugs: OUTPUT«*** Not a keyed value: VList [VStr "a",VStr "b",VStr "c"]␤ at /tmp/IrcXokCCwJ line 1, column 5 - line 2, column 1␤»
..niecza v17-2-gb8fcf85: OUTPUT«Unhandled exception: Unable to resolve method invert in class List␤ at /tmp/hLf2h_qa2Z line 1 (mainline @ 3) ␤ at /home/p6eval/niecza/lib/CORE.setting line 3840 (ANON @ 3) ␤ at /home/p6eval/niecza/lib/CORE.setting line 3841 (module-CORE @ 65) ␤ at /home…
..rakudo d9fd17: OUTPUT«Method 'invert' not found for invocant of class 'List'␤ in block <anon> at /tmp/9lmAS5IB4E:1␤␤»
TimToady mind, I dunno if that even works
masak seems not :)
TimToady but we ought to be able to write one that doesn't need an intermediate hash, meseemeth 16:23
that is, a List.invert that assumes pairs the same way a hash initializer does 16:25
masak right.
I keep finding myself in situations when I want that.
TimToady then it doesn't matter whether whether you .kv or .pairs
masak :)
TimToady but it would require one or the other to introduce the .keys 16:26
16:26 fglock left
TimToady it might be a little surprising to someone who expects @array.invert to behave exactly like @hash.invert though 16:28
sorear good * #perl6 16:29
TimToady \o
16:29 thou joined
jnthn masak: Curious - what kind of situation? 16:29
16:31 ponbiki joined
TimToady emulating content-addressable memory 16:31
16:31 eviltwin_b joined, geekosaur left
masak jnthn: hex puzzle. mapping forwards from j2-like coords to 1-d board indices, and then backwards to 2-d coordinate pairs. 16:31
16:31 eviltwin_b is now known as geekosaur 16:32 wtw left
jnthn ah, ok 16:32
masak jnthn: the forwards mapping is done most easily as 'my %horizontals = <i1 i2 ...>.kv.hash.invert'
TimToady basically, any 2-way mapping where one of them is naturally positional 16:34
16:34 mspaulding joined
masak right. 16:34
16:34 mspaulding left 16:35 wolfman2000 left
sisar um, installing LWP::Simple fails for me with failure in dependence resolution: gist.github.com/2628829. 16:36
TimToady I could probably be argued into List.invert assuming it should invert on the indexes, and leave out the hashlist stuff
16:36 NamelessTee joined
TimToady while less powerful, it's probably more POLS 16:37
the other way is to distinguish Array.invert from List.invert, but that bothers me
masak oh, I meant inverting on the indices. 16:38
sisar: panda seems to complain because the URI tests don't pass -- so it doesn't go on and install LWP::Simple. 16:39
TimToady perl6: say (a => 42).invert.perl
p6eval rakudo d9fd17, niecza v17-2-gb8fcf85: OUTPUT«42 => "a"␤»
..pugs: OUTPUT«*** No such method in class Pair: "&invert"␤ at /tmp/ifgTGz1xwR line 1, column 5 - line 2, column 1␤»
TimToady we can always hyperinvert a pairlist
masak that still doesn't feel Huffmanized enough to me. 16:40
TimToady I mean that in addition to List.invert
masak ah. 16:41
TimToady because that would do the wrong thing with a list of pairs
sisar masak: ok. So I need to wait till URI is fixed ? (i know, i should fix it myself)
:) 16:42
TimToady you'd end up with 0 => (a => 42)
perl6: my %hash = (:a(1), :b(2), :c(3)).pairs».invert; say %hash<c> 16:43
p6eval pugs: OUTPUT«*** Not a keyed value: VList [VRef <Pair:0xf6d85bb9>,VRef <Pair:0xf6d85c31>,VRef <Pair:0xf6d85c8d>]␤ at /tmp/bzrU0kk1FC line 1, column 12-47␤»
..rakudo d9fd17, niecza v17-2-gb8fcf85: OUTPUT«Any()␤»
TimToady oops
perl6: my %hash = (:a(1), :b(2), :c(3))».invert; say %hash<c> 16:44
p6eval pugs: OUTPUT«*** No such method in class Array: "&>>invert"␤ at /tmp/eZZ5NPmqIS line 1, column 12-41␤»
..rakudo d9fd17, niecza v17-2-gb8fcf85: OUTPUT«Any()␤»
16:44 libertyprime left
masak sisar: yes, please do. or at least figure out more about the test failures. 16:44
sisar ok
TimToady perl6: my %hash = <a b c d e>.pairs».invert; say %hash<c>
p6eval rakudo d9fd17, niecza v17-2-gb8fcf85: OUTPUT«2␤»
..pugs: OUTPUT«*** Not a keyed value: VList [VStr "a",VStr "b",VStr "c",VStr "d",VStr "e"]␤ at /tmp/2FMJ66De4g line 1, column 12-37␤»
TimToady that at least avoids creating a temp hash 16:45
16:45 libertyprime joined 16:48 eviltwin_b joined, geekosaur left 16:49 eviltwin_b is now known as geekosaur 16:55 erkan joined, erkan left, erkan joined
masak perl6: my %hash = <a b c d e> Z=> 0..*; say %hash<c> 16:55
p6eval rakudo d9fd17, niecza v17-2-gb8fcf85: OUTPUT«2␤»
..pugs: OUTPUT«*** unknown parse error␤ at /tmp/sxmYSABSen line 1, column 27␤»
masak so does this. but I don't much like the end weight of it.
16:55 eviltwin_b joined, geekosaur left 16:56 eviltwin_b is now known as geekosaur
masak that's why I'd like a sub form of 'invert', too. because it places the transformation up front. 16:57
16:58 geekosaur left, geekosaur joined
sorear p6: &invert 16:59
p6eval pugs: OUTPUT«*** ␤ Unexpected end of input␤ expecting "::" or signature value␤ Variable "&invert" requires predeclaration or explicit package name␤ at /tmp/3ubh2sCbl8 line 1, column 8␤»
..rakudo d9fd17, niecza v17-2-gb8fcf85: ( no output )
sorear n: say invert [<a i u e o>] 17:00
p6eval niecza v17-2-gb8fcf85: OUTPUT«Unhandled exception: Nominal type check failed in binding '%h' in 'invert'; got Array, needed Associative␤ at /home/p6eval/niecza/lib/CORE.setting line 0 (invert @ 1) ␤ at /tmp/OFg3FW4C_I line 1 (mainline @ 2) ␤ at /home/p6eval/niecza/lib/CORE.setting li…
sorear n: say invert {<a i u e o> Z=> 0..*}
p6eval niecza v17-2-gb8fcf85: OUTPUT«Unhandled exception: Nominal type check failed in binding '%h' in 'invert'; got Block, needed Associative␤ at /home/p6eval/niecza/lib/CORE.setting line 0 (invert @ 1) ␤ at /tmp/rU3cw1ymAB line 1 (mainline @ 2) ␤ at /home/p6eval/niecza/lib/CORE.setting li…
sorear n: say invert {%(), <a i u e o> Z=> 0..*}
p6eval niecza v17-2-gb8fcf85: OUTPUT«Unhandled exception: Nominal type check failed in binding '%h' in 'invert'; got Block, needed Associative␤ at /home/p6eval/niecza/lib/CORE.setting line 0 (invert @ 1) ␤ at /tmp/FvJjCHG074 line 1 (mainline @ 2) ␤ at /home/p6eval/niecza/lib/CORE.setting li…
sorear n: say invert %(<a i u e o> Z=> 0..*)
p6eval niecza v17-2-gb8fcf85: OUTPUT«0 => "a" 1 => "i" 2 => "u" 3 => "e" 4 => "o"␤» 17:01
masak \o/ 17:02
17:02 vlixes joined
masak any good reason *not* to make 'invert <a i u e o>' work, too? 17:02
17:04 PacoAir joined
TimToady I already said I was fine with that part, I think...my wonderings were more about the types involved 17:04
17:06 carlin left, eviltwin_b joined, geekosaur left 17:07 carlin joined, eviltwin_b is now known as geekosaur
tadzik 'afternoon, #eplr6 17:07
or whatever channel I'm on
17:10 simcop2387 joined 17:11 Chillance joined
Woodi hallo 17:12
perl6: my %hash = (:a(1), :b(2), :c(3)); say %hash
p6eval niecza v17-2-gb8fcf85: OUTPUT«{"a" => 1, "b" => 2, "c" => 3}␤»
..rakudo d9fd17: OUTPUT«("a" => 1, "b" => 2, "c" => 3).hash␤»
..pugs: OUTPUT«a 1b 2c 3␤»
masak .oO( Perl 6 was delayed many years, because the community members kept logging into anagrams of the channel name ) 17:13
Woodi it's strange that compilers have different output...
masak Woodi: no, it's not.
Woodi: .gist defaults to .perl, which is compiler-dependent.
Woodi ah, right. so "internal dumper" is standarized ? 17:14
I think about it as kind of code serialization... for RPC or something... 17:15
17:19 Guest74228 left 17:26 geekosaur left, eviltwin_b joined, eviltwin_b is now known as geekosaur 17:34 geekosaur left, eviltwin_b joined, eviltwin_b is now known as geekosaur 17:35 ocharles joined 17:40 xenu joined, eviltwin_b joined, geekosaur left 17:41 xenu is now known as Guest40378, eviltwin_b is now known as geekosaur 17:42 not_gerd joined
not_gerd hello, #perl6 17:42
google fail - is it possible to define your own ternary operators?
17:42 geekosaur is now known as allbery_b 17:45 geekosaur joined
moritz not yet, i fear 17:45
17:46 eviltwin_b joined, geekosaur left 17:47 geekosaur joined
masak what's missing? 17:48
jnthn Something that does the right thing with the definition, I suspect.
masak r: sub infix:<!? !?>($a, $b, $c) { "$a, $b and $c went into a bar..." }; say 1 !? 2 !? 3
p6eval rakudo d9fd17: OUTPUT«===SORRY!===␤Confused␤at /tmp/rLuIegwVIH:1␤»
masak aww
moritz ttiar 17:49
hm, no
moritz can't parse
masak ;)
moritz r: sub infix:<FOO BAR>(*@a) { say @a.perl }; 1 FOO 2 BAR 3
p6eval rakudo d9fd17: OUTPUT«===SORRY!===␤Confused␤at /tmp/UU5finyUCs:1␤»
17:49 allbery_b left
moritz r: sub infix:<FOO BAR>(*@a) { say @a.perl }; 1 FOO BAR 3 17:49
p6eval rakudo d9fd17: OUTPUT«Array.new(1, 3)␤»
moritz r: sub infix:<FOO BAR>(*@a) { say @a.perl }; 1 FOO BAR 3 17:50
p6eval rakudo d9fd17: OUTPUT«===SORRY!===␤Confused␤at /tmp/qHyav8EPpc:1␤»
not_gerd anyway, I had some more fun with the n-body benchmark
if anyone wants to take a look: gist.github.com/2626839
jnthn moritz: Heh. Was just glancing the code and realized...it just...installs the token, including the space. :) 17:51
moritz jnthn: I figured :-) 17:52
jnthn I'm gonna redo that whole load of stuff soonish anyways. 17:53
17:53 geekosaur left
jnthn To fix the serialization issues, and also to enable the precedence traits. 17:53
tadzik jnthn: did you have any chance to look at that non-deterministic precompilation fun? 17:54
[Coke] jnthn: are you using win xp or win 7? 17:55
jnthn tadzik: There was a non-deterministic one?
[Coke]: Win 7.
tadzik jnthn: yeah, URI and panda suffer from it. Basically, loading sometimes fails with "No object at STable <n>", and sometimes not
jlaire .u ’
phenny U+2019 RIGHT SINGLE QUOTATION MARK (’)
jnthn tadzik: "sometimes"? :/ 17:56
tadzik jnthn: yeah, that's the point :/
jnthn: I can provide a shell if you can't reproduce
jnthn What do you have to do to make it suceed? Just run it again? Or re-compile? 17:57
17:57 eviltwin_b is now known as geekosaur
tadzik jnthn: you have a 50% chance. Recompilation triggers it 17:57
if it's compiled with luck, then it always works
if without luck, then it never does
afair, will recheck
jnthn: yes, it's a matter of recompiling it 18:00
18:03 tokuhirom left, tokuhirom joined 18:07 pernatiy_ left 18:08 pernatiy joined, tokuhirom left 18:10 alester joined
masak tadzik: so that makes it deterministic, then? 18:23
18:28 wolfman2000 joined
tadzik no 18:28
you have approx. 50% chance it'd work 18:29
masak :/
grep the code for any 1d2 die rolls... 18:30
18:30 eviltwin_b joined, geekosaur left 18:31 eviltwin_b is now known as geekosaur 18:33 zby_home_ left
tadzik my only guess is the Parrot's hash randomization 18:37
I know of no other things that are random
18:37 fhelmberger left
TimToady parallel make could be random 18:47
jnthn Ah 18:48
And parallel make would potentially cause issues like this...
18:52 wolf2k_away joined 18:55 wolfman2000 left
sisar tadzik: now i remember, i got the exact same error "No STable at index 8" when compiling panda today. Any info I can provide? I still have the terminal output of that try. 18:57
18:57 _Vi left
Celelibi About the >> and << metaoperators: what is an operator? 18:59
masak Celelibi: how do you mean?
an operator is anything that S03 talks about. to a first approximation :P
Celelibi How can I make my own operator to put between >> and << ?
masak Celelibi: just define it, as a sub. 19:00
colomon Celelibi: sub infix:<whatever>($a, $b) { ... }
masak r: sub infix:<!!!>($a, $b) { "A $a and a $b walk into a bar..." }; say "platypus" !!! "philosopher"
p6eval rakudo d9fd17: OUTPUT«A platypus and a philosopher walk into a bar...␤»
Celelibi What are they talking about ? 19:01
:)
masak Plato :)
colomon n: sub infix:<whatever>($a, $b) { $a ~ " blue" x $b }; say <go blue> >>whatever<< [2, 3, 4]
p6eval niecza v17-2-gb8fcf85: OUTPUT«Unhandled exception: Ran off end of non-dwimmy left␤ at /home/p6eval/niecza/lib/CORE.setting line 1364 (die @ 3) ␤ at /home/p6eval/niecza/lib/CORE.setting line 3019 (ANON @ 18) ␤ at /home/p6eval/niecza/lib/CORE.setting line 3026 (_hyper_posi @ 6) ␤ at /…
19:02 GlitchMr joined
Celelibi what is <go blue> ? 19:02
masak Celelibi: ('go', 'blue')
19:02 wolf2k_bahhumbug joined
GlitchMr perl6: print sqrt -1 19:02
Celelibi like qw ?
p6eval pugs, rakudo d9fd17, niecza v17-2-gb8fcf85: OUTPUT«NaN»
colomon yes
masak Celelibi: to a first approximation, yes.
GlitchMr perl6: print sqrt(-1 + 0i)
Celelibi ok
p6eval rakudo d9fd17: OUTPUT«6.12303176911189e-17+1i»
..niecza v17-2-gb8fcf85: OUTPUT«6.1230317691118863E-17+1i»
..pugs: OUTPUT«0 + 1i»
masak Pugs++
19:03 eviltwin_b joined, geekosaur left, eviltwin_b is now known as geekosaur 19:04 wolf2k_bahhumbug left 19:05 wolf2k_away left
sisar aye, Pugs++ 19:06
19:09 eviltwin_b joined, geekosaur left, eviltwin_b is now known as geekosaur
Celelibi masak: and what's the difference between <> and qw ? 19:09
cognominal none, I think 19:10
Celelibi And can I, in some way use the meta operators >> and/or << on more that two arguments? 19:12
cognominal r: say << a b > < c >> # except that you can do it multilevel 19:13
p6eval rakudo d9fd17: OUTPUT«a b > < c␤»
19:13 geekosaur left
Celelibi I don't understand that line. 19:14
cognominal r: say << a b > < c >>.perl # except that you can do it multilevel
p6eval rakudo d9fd17: OUTPUT«("a", "b", ">", "<", "c")␤»
19:15 geekosaur joined
masak Celelibi: <> has val() semantics. (see S02). qw doesn't. 19:15
19:16 birdwindupbird joined
Celelibi cognominal: this is just a list... 19:16
I don't get it.
masak Celelibi: cognominal wants to show off that if you have a '<' inside the <> list, you can just delimit with <<>> instead.
Celelibi ok 19:17
masak and so on.
cognominal you can't do that with qw
masak cognominal: in Perl 5, you can't.
cognominal: this is #perl6 :P
19:17 _dolmen_ joined
cognominal I am watching the Almadovar, so I am not very attentive 19:17
masak n: .say for qw<<< > >> >>> 19:18
p6eval niecza v17-2-gb8fcf85: OUTPUT«>␤>>␤»
Celelibi ok ok
masak .oO( I thought he was called the Hollande... )
cognominal …in spanish to make it more challenging
masak n: .say for qw<[< a b c >]> 19:19
p6eval niecza v17-2-gb8fcf85: OUTPUT«[<␤a␤b␤c␤>]␤»
19:20 Mescale joined
Celelibi oh oh, was the ">>." in "Str.methods(:all)>>.name" the metaoperator >> applied on "."? 19:22
masak conceptually, yes.
Celelibi which means "actually no"? :)
masak "." isn't really an operator.
or rather, if it's an oprator, it's a rather strange one. 19:23
Celelibi hum... Is >>. the operator as whole?
masak in the sense that >>+<< is the operator as a whole, yes.
Celelibi ok so this >>. is actually part of the language while >>+<< is not. right? 19:24
tadzik no, it's just a combo
gfldex r: my @a = 1,2,3; say @a>>.?foo; 19:25
p6eval rakudo d9fd17: OUTPUT«Nil Nil Nil␤»
masak Celelibi: it's all just parsing. there's no real "actually part of the language" distinction.
Celelibi what is .?
masak Celelibi: it's a dottyop
Celelibi: see STD.pm6 :)
Celelibi: or did you ask 'what is C<.?>?' :) 19:26
Celelibi: it's "call method, but don't mind if dispatch fails".
cognominal everything is more or less dotty in Perl 6 :)
Celelibi I mean fucking connect my brain to that computer so that I know everything!
:p
gfldex Celelibi: perlcabal.org/syn/S12.html#Calling_...of_methods
19:27 geekosaur left
gfldex it's like a twigil but not :-> 19:28
Celelibi what means "call all the methods"? 19:29
masak Celelibi: it means call all the methods.
Celelibi How can there be several appliable methods?
19:29 geekosaur joined
masak Celelibi: inheritance. 19:29
Celelibi ok
gfldex how does that play with MMD? 19:30
masak r: class A { method foo { say "A" } }; class B is A { method foo { say "B" } }; B.new.*foo
p6eval rakudo d9fd17: OUTPUT«B␤A␤»
Celelibi So $obj.*methods will call the methods of $obj and its parents.
masak gfldex: very well.
Celelibi: $obj doesn't have parents. its class does. and it's not just the parents, it's all applicable methods, any level up.
Celelibi parents 19:31
ok
masak that's not the point :)
Celelibi I fogot about the multiple inheritance.
masak "parents" only indicates the case of MI.
you meant "ancestors" or something.
Celelibi yes.
masak don't blame me for your being inexact :P
TimToady parent's parents are parents too :P
Celelibi English is not my native language. :)
masak TimToady: yes, but not $obj's. 19:32
Celelibi: neither is it mine.
I suspect that there's more non-native English speakers here than native ones.
19:32 geekosaur is now known as allbery_b
Celelibi possibly. :) 19:32
TimToady regarding whether . is an operator, not by itself; rather it's considered to introduce a postfix, so the whole method call is the postfix operator 19:33
19:33 eviltwin_b joined
masak ah. 19:33
learn something every day ;)
Celelibi And can I declare several methods that could match the same arguments?
TimToady and some of our native English speakers have studied linguistics, which means they don't know how the language works anymore either 19:34
cognominal a spread postfix then.
masak Celelibi: yes, if you declare them 'multi'.
19:34 allbery_b left
Celelibi ok 19:34
and I guess .* will call them all.
19:34 Mescale left
masak hm, with .* you don't need the 'multi'. 19:34
as I showed above.
Celelibi r: class A { method foo { say "A" } }; class B is A { method foo { say "B" } method foo {say "C"} }; B.new.*foo 19:35
p6eval rakudo d9fd17: OUTPUT«===SORRY!===␤Unable to parse blockoid, couldn't find final '}' at line 2␤»
Celelibi r: class A { method foo { say "A" } }; class B is A { method foo { say "B" }; method foo {say "C"} }; B.new.*foo
p6eval rakudo d9fd17: OUTPUT«===SORRY!===␤Package 'B' already has a method 'foo' (did you mean to declare a multi-method?)␤»
moritz note that classes usually have to be explicitly designed for .* use, otherwise .* won't work as expected
Celelibi r: class A { method foo { say "A" } }; class B is A { multi method foo { say "B" }; multi method foo {say "C"} }; B.new.*foo
p6eval rakudo d9fd17: OUTPUT«Ambiguous call to 'foo'; these signatures all match:␤:(B , Mu *%_)␤:(B , Mu *%_)␤␤ in method foo at src/gen/CORE.setting:293␤ in method dispatch:<.*> at src/gen/CORE.setting:821␤ in block <anon> at /tmp/Ebas_rsdm9:1␤␤»
TimToady .* is kind of a power tool; it can cut your arm off really quick 19:36
add in >> and it'll gladly cut off all your limbs 19:37
19:37 Mescale joined
Celelibi I guess from the last error that .* cannot call several methods in the same class. 19:37
TimToady nope
you can only do that if you add 'multi' 19:38
and then distinguish the type sigs enough so that's not ambiguous
Celelibi didn't I put multi in the right place?
masak Celelibi: your problem wasn't .* at all
TimToady you didn't distinguish the signatures
masak Celelibi: your problem was an ambiguous dispatch. 19:39
Celelibi r: class A { method foo { say "A" } }; class B is A { multi method foo { say "B" }; multi method foo (int $a = 0) {say "C"} }; B.new.*foo
p6eval rakudo d9fd17: OUTPUT«Ambiguous call to 'foo'; these signatures all match:␤:(B , Mu *%_)␤:(B , int $a = { ... }, Mu *%_)␤␤ in method foo at src/gen/CORE.setting:293␤ in method dispatch:<.*> at src/gen/CORE.setting:821␤ in block <anon> at /tmp/u2sOxmc5rI:1␤␤»
masak Celelibi: generally, if you make your method signatures exactly the same, you'll get one.
TimToady .* isn't meant to help with that
jnthn Remember that .* is *only* about the MRO
masak Celelibi: 'multi' really only makes sense if you also make sure to make your signatures different enough.
jnthn The multi-dispatch is then a secondary dispatch *after* normal method dispatch. 19:40
Celelibi ok
jnthn And the "call all the things" semantics is not applied recursviely into the proto's dispatch.
Celelibi MRO ?
jnthn Method Resolution Order. Basically, "what order do we visit classes in to find the method"?
TimToady if you really need to call a set of indentical signatures, just shove 'em in a data structure and call 'em with >>.
jnthn When you have multiple inheritance, there's more than one way you could walk the tree.
Celelibi ok 19:41
TimToady s/indent/ident/
Celelibi I think I get it.
TimToady good questions though
it's not something we've always "got" ourselves 19:42
moritz regularly confuses how inheritance and multi methods interact 19:43
and I think we've been back and forth a few times about how it should be
masak I also think I get it now.
well, nowadays, I should say. since about mid-2011. 19:44
it's really quite pretty.
19:44 not_gerd left
masak and yeah, keeping the MRO and the proto/multi bits apart is kinda important ;) 19:44
jnthn wishes it had been so obvious from the start, so he didn't have to re-implement all the things... 19:45
TimToady we'll all be smart by the time this is done :) 19:46
jnthn ;)
TimToady otoh, I hope we're all dead before this is done too, in the 100-year-language sense :) 19:47
masak I hope we're all alive by the time this is done, in the 100-year language sense... :) 19:48
TimToady the Plurality is Near!
I suppose it doesn't have to be me, as long as it's somebody who thinks he's me... 19:49
The Strong Turing Test: emulating you so well that you don't know whether you're the real you or not... 19:51
masak TimToady: you reminded me of blog.plover.com//brain/immortality.html 19:52
19:52 jaldhar left
TimToady yes, I've always liked that one. 19:53
masak something about all that bothers me; can't quite say what. 19:54
TimToady s/you don't know/neither of you knows/
well, the whole issue rests on the extent to which your identity includes the reality in which you're embedded 19:55
Celelibi <TimToady> The Strong Turing Test: emulating you so well that you don't know whether you're the real you or not... <-- please stop messing with my mind! :p
moritz Celelibi: you'll have to get used to that if you hang around here for longer :-) 19:56
TimToady um...no? :P
eviltwin_b shatterday?
TimToady sorry, I mess regularly with my own mind, and some of it naturally wanders out into the reality in which I am embedded 19:57
masak *lol*
jnthn moritz: Latest on use-malice is...import works and I need to look at export? 19:58
masak .oO( we have a level-3 TimToady containment breach... all personel to section 4A of the reality in which TimToady is embedded... )
19:58 eviltwin_b is now known as geekosaur
moritz jnthn: correct 19:59
Celelibi TimToady: I already do this myself. I don't really need more. :)
jnthn Pulled...building. :)
masak personnel*
TimToady
.oO(the Gray Geek Catastrophe)
masak TimToady: :D
moritz jnthn: it works when I build the export tag myself :-)
jnthn moritz: OK :) 20:00
masak TimToady: weirdest mental image today: Guido van Rossum, riding the Googleplex, trying to fend of an ocean of your brain with a pitchfork.
jnthn So, design question
masak off* 20:01
jnthn sub foo is export(:lol, :wut) { }
What gets passed to the trait_mod there?
Clearly :export is passed.
Then do we get: a) two named args, :lol and :wut, or b) one positional arg, which is a Parcel containing two Pairs?
TimToady: ^^ :)
cognominal sometimes, #perl6 sounds like inception, the movie.
20:02 GlitchMr left
TimToady 'is export' is 'long shorthand' for :export so the rules of :export(:lol, :wut) apply, which is to say it's the same as export => (:lol, :wut) 20:02
Celelibi inception is a fary tale. :)
moritz jnthn: how about a a Capture containing two pairs?
Celelibi (Well, I don't thing about Alice in wonderland)
moritz carlin: as is #perl6. Oh, wait...
erm, meant Celelibi 20:03
Celelibi: intellectual milk and honey flow in generally friendly atmosphere
jnthn TimToady: Oh. :)
TimToady: Not the answer I expected, but I can implement that one :)
masak moritz: though milk flows better than honey. 20:04
moritz masak: yes, but having two different viscosities makes for much more interesting flow patterns
masak ooh, yes. 20:05
(different viscosities)++
TimToady suspects that honey stays liquid-ish at a temperature where milk freezes...
[Coke] , in backscroll, wonders if we can have a Modula-3 TimToady containment breach instead.
TimToady [Coke]: the breach is Oberon the other side from where you're looking... 20:06
masak if I ever form a post-anarchist unplugged punk duo with someone, I'm gonna name it "Two Different Viscosities".
20:06 fhelmberger joined
masak TimToady: When you look into an ABSYS, the ABSYS also looks into you. 20:08
ooh, ABSYS was a Prolog precursor... 20:09
TimToady well, so? Neutrinos are into physicists...
geekosaur and everyone else
Celelibi masak: this may sound quite sexual in some ways. :) 20:10
TimToady and everything else
masak Celelibi: I have such an innocent mind that the thought didn't even occur to me. 20:12
20:13 Kresike left
Celelibi I wish I would think a bit less sometime. 20:13
masak that's an unusual wish. 20:14
[Coke] TimToady: *groan*
masak though not entirely unreasonable.
Celelibi: tried meditation?
tadzik that's one of the reasons for drinking alcohol, I guess
one once said "I drink so I tolerate people I spend my time with" 20:15
[Coke] masak: I often complain that my brain is full. That's almost the same complaint.
Celelibi masak: actually it's not the "quantity" it's what I think to.
masak Celelibi: nevertheless. 20:16
Celelibi tadzik: that's right, I'm a bit more open minded when I'm drunk. :) 20:17
masak notes for the first time that Celelibi's nick could be seen as a portmanteau of "celibate" and "libido"
tadzik :>
masak tadzik: I try to seek out friends that I'm able to stand when sober.
Celelibi Tell me more about your innocence masak. :) 20:18
tadzik masak: same :)
masak Celelibi: I was totally not serious. my father is the King of innuendo, srsly.
TimToady as opposed to friends who will help you stand when you're not sober?
tadzik :)
masak TimToady: incidentally, those overlap quite well :) 20:19
TimToady overlap!?!?
masak *lol*
Celelibi I learned a new word today: "innuendo".
masak .oO( must be something they put in the #perl6 welcome punch tonight... ) 20:20
sisar tadzik: read my earlier comment? re panda build fails
tadzik sisar: yeah, seen that, but that's the same thing here 20:21
20:21 Mescale left, tyatpi left
sisar does make parallize on its own or do i need to tell it to ? 20:22
Celelibi is "hyperoperator" synonymous for "meta operator" ?
make -j 8
tadzik sisar: that's not the thing, sadly :/ 20:23
sisar oh, ok, fine :)
20:23 ponbiki left
cognominal <tadzik> one once said "I drink so I tolerate people I spend my time with". Paul Valéry said "un homme seul est toujours en mauvaise companie" : a man alone is aways in bad company, but I suppose you did not mean it that way :) 20:23
Celelibi If I understand correctly in perl6 there are less {} and more <>. :) 20:24
tadzik cognominal: nope :)
masak Celelibi: no, the hyperops are just a type of metaop.
Celelibi: reduction ops are another type.
Celelibi hyperoperators are only >> and << ? 20:25
masak aye.
Celelibi ok
masak and the things that go between them :)
Celelibi ok
so >>+<< as a whole is an hyperoperator.
TimToady yes, we use the term both ways 20:26
Celelibi is >>+>> also one?
TimToady yes, but that one's dwimmy on the right
Celelibi dwimmy? 20:27
cognominal I said that Perl 6 is dotty, I forgot to said it was loopy too
TimToady perl6: say (1,2,3) »+« (1,2)
p6eval niecza v17-2-gb8fcf85: OUTPUT«Unhandled exception: Ran off end of non-dwimmy right␤ at /home/p6eval/niecza/lib/CORE.setting line 1364 (die @ 3) ␤ at /home/p6eval/niecza/lib/CORE.setting line 3020 (ANON @ 19) ␤ at /home/p6eval/niecza/lib/CORE.setting line 3026 (_hyper_posi @ 6) ␤ at …
..pugs: OUTPUT«243␤»
..rakudo d9fd17: OUTPUT«Sorry, lists on both sides of non-dwimmy hyperop are not of same length:␤ left: 3 elements, right: 2 elements␤␤ in sub hyper at src/gen/CORE.setting:10792␤ in block <anon> at src/gen/CORE.setting:10773␤ in block <anon> at /tmp/BygWuXgsSN:1␤␤»
TimToady perl6: say (1,2,3) »+« (1,2,3)
p6eval pugs: OUTPUT«246␤»
..rakudo d9fd17, niecza v17-2-gb8fcf85: OUTPUT«2 4 6␤»
TimToady perl6: say (1,2,3) »+» (1,2)
p6eval pugs: OUTPUT«*** ␤ Unexpected "\187+\187"␤ expecting operator, ":" or ","␤ at /tmp/3kODkyzBcV line 1, column 13␤»
..rakudo d9fd17, niecza v17-2-gb8fcf85: OUTPUT«2 4 4␤»
TimToady perl6: say (1..10) »+» (1,2) 20:28
p6eval rakudo d9fd17, niecza v17-2-gb8fcf85: OUTPUT«2 4 4 6 6 8 8 10 10 12␤»
..pugs: OUTPUT«*** ␤ Unexpected "\187+\187"␤ expecting operator, ":" or ","␤ at /tmp/S1kzO_7N9V line 1, column 13␤»
TimToady other metaops have different policies, for instance
masak Celelibi: DWIM means "Do What I Mean", i.e. employ some context-sensitive logic to figure out some thing that is not explicitly stated. in the case of >>+>>, make the right-hand list as long or short as makes sense.
TimToady perl6: say 1..10 Z+ 1..3 20:29
p6eval pugs: OUTPUT«12345678910␤»
..rakudo d9fd17, niecza v17-2-gb8fcf85: OUTPUT«2 4 6␤»
masak Celelibi: Perl 5 uses a lot of DWIM that (from the perspective of Perl 6) is considered deprecated.
TimToady a 'zip' metaop just trims to the shorter of the two lists
Celelibi ok
TimToady since you can't have a zipper with only one side
masak was gonna say
Celelibi And yes, that was a question I had last time. What does mean "make the list as long as needed"? 20:30
By repeating its elements?
TimToady depends on what is making the list as long as needed
masak .oO( innuendo )
TimToady with hypers, it repeats the short side
20:31 phenny left, phenny joined
TimToady with a degenerate (!?!) case that a single element just repeats 20:31
Celelibi ok, and the list extended/shortened is the one on the "small side" of the >> 20:33
TimToady if it's the xx operator, it just repeats the list on the left
Celelibi what's the equivalent of x operator in perl5? 20:34
masak x
unless you want the list-context variant, then it's xx
TimToady r: say (1..10) »+» (1..20)
p6eval rakudo d9fd17: OUTPUT«2 4 6 8 10 12 14 16 18 20␤»
TimToady yes, the dwimmy side can shorten as well as repeat
Celelibi r: say (1 .. 10)x2
p6eval rakudo d9fd17: OUTPUT«===SORRY!===␤Confused␤at /tmp/ToHwjimFfT:1␤»
Celelibi what's wrong? 20:35
TimToady std: say (1 .. 10)x2
p6eval std fab64fc: OUTPUT«===SORRY!===␤Whitespace is required between alphanumeric tokens at /tmp/biESrIXzTw line 1:␤------> say (1 .. 10)x⏏2␤Check failed␤FAILED 00:00 42m␤»
TimToady n: say (1 .. 10)x2
p6eval niecza v17-2-gb8fcf85: OUTPUT«===SORRY!===␤␤Whitespace is required between alphanumeric tokens at /tmp/5UXVcpE1LS line 1:␤------> say (1 .. 10)x⏏2␤␤Unhandled exception: Check failed␤␤ at /home/p6eval/niecza/boot/lib/CORE.setting line 1364 (die @ 3) …
tadzik masak: autopun? "Hypochondria is the only illnes I don't have"
TimToady what's wrong is that rakudo isn't quite using the standard parser yet
jnthn moritz: Think I've got it workin' :) 20:37
Celelibi but it don't seem to work either with niecza 20:38
jnthn moritz: But it turned out a bunch of other traits were relying on the arg being passed int he wrong way, so been clearing up that.
masak tadzik: yeah :P
TimToady n: say (1..10) x 2
p6eval niecza v17-2-gb8fcf85: OUTPUT«1 2 3 4 5 6 7 8 9 101 2 3 4 5 6 7 8 9 10␤»
masak tadzik: though I would turn it around for maximum punch.
TimToady n: say (1..10) xx 2
p6eval niecza v17-2-gb8fcf85: OUTPUT«1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10␤»
tadzik masak: I saw it turned around in one film I think 20:39
TimToady Celelibi: it works if you read the error message :P
tadzik masak: "Hypochondria?" "Yeah, it's the illnes that (...describes...)" "Oh my god. I have that too!"
masak tadzik: :P
20:40 sivoais left
Celelibi TimToady: they are hard to read with these escape codes. 20:40
masak tadzik: yeah, I'm probably silly to want to turn it around. the usual way with autopuns is mention + use, in that order.
Celelibi: check the IRC logs.
20:40 fhelmberger left
TimToady Celelibi: which IRC client are you using? irssi just shows things in colors, since those are terminal color codes 20:41
Celelibi masak: they're stored in the logs. :)
TimToady: xchat.
20:41 sivoais joined
Celelibi these are color codes for a terminal, not for irc. :) 20:41
TimToady I'm on a terminal
tadzik weechat doesn't show them
20:41 birdwindupbird left
geekosaur mmm, yes, I need a script to do something about the broken ansi x3.64 codes in those 20:42
(also xchat)
Celelibi the best fix would be that p6eval translate them to irc color codes. :)
geekosaur indeed 20:45
20:45 fhelmberger joined
sisar is rosettacode.org/wiki/Sorting_algori...ort#Perl_6 using a feature which is NYI in rakudo? coz it hangs Rakudo, but works fine in Niecza. 20:45
Celelibi and so... <<+>> is just like a regular + ? 20:46
diakopter sisar: how long did you wait while it was hanging?
sisar 2-3 mins...
should i wait moar ? 20:47
diakopter how long did niecza take
sisar diakopter: 3 seconds.
diakopter oh; hm
hard to tell (to me) 20:48
I guess it's worth trying again and waiting longer until someone replies with more knowledge
sisar diakopter: when people ask the time, should i report the real, user, or sys time ? 20:49
diakopter wall clock, I guess
eh 20:50
TimToady Celelibi: it's only like regular + if you feed it two numbers. Otherwise + will try to coerce to numbers, so lists return the number of elements with bare +, whereas <<+>> will try to apply across lists
sisar i mean "time perl6 <script-nae>" reports three times, right?
dalek kudo/use-arglist: 6fa9034 | jnthn++ | src/ (3 files):
Refactor handling of traits taking named arguments. The argument is now passed along with the name. Updated various of the built-in traits along with this.
kudo/use-arglist: f9a94d8 | jnthn++ | src/core/traits.pm:
Handle export tags.
jnthn moritz: There you go. :) A test in advent2009-day12.t fails, but looking at it, I think it's wrong. 20:51
Celelibi TimToady: yep I saw. It expand the shortest list.
r: say ((1, 2, 3) X (4, 5, 6)).perl
p6eval rakudo d9fd17: OUTPUT«((ListIter.new(),), (ListIter.new(),), (ListIter.new(),), (ListIter.new(),), (ListIter.new(),), (ListIter.new(),), (ListIter.new(),), (ListIter.new(),), (ListIter.new(),)).list␤»
jnthn moritz: Also a trait one, but should be an easy fix when merging. 20:52
Celelibi what does that mean? :)
jnthn r: say ((1, 2, 3) X (4, 5, 6))
masak Celelibi: means you're in over your head :P
p6eval rakudo d9fd17: OUTPUT«1 4 1 5 1 6 2 4 2 5 2 6 3 4 3 5 3 6␤»
jnthn Celelibi: Just Rakudo being lazy :)
(That is, not evaluating stuff until it has to.)
Celelibi ok, these are lazy lists.
jnthn Looks like... 20:53
Celelibi How do I force the evaluation of an expression?
jnthn .eager should do it.
r: say ((1, 2, 3) X (4, 5, 6)).eager.perl
p6eval rakudo d9fd17: OUTPUT«((ListIter.new(),), (ListIter.new(),), (ListIter.new(),), (ListIter.new(),), (ListIter.new(),), (ListIter.new(),), (ListIter.new(),), (ListIter.new(),), (ListIter.new(),)).list␤»
jnthn Hm...
[Coke] jnthn: at this rate, all the advent tests will eventually break.
jnthn r: say ((1, 2, 3) X (4, 5, 6)).flat.perl
p6eval rakudo d9fd17: OUTPUT«(1, 4, 1, 5, 1, 6, 2, 4, 2, 5, 2, 6, 3, 4, 3, 5, 3, 6).list␤»
diakopter sisar: 60x slower for some operations wouldn't be surprising a while ago; I don't know about recently
[Coke] as long as we do so intentionally and not accidentally, that's the best I can hopef or.
jnthn That flattens it out at least.
Celelibi why flat? 20:54
I do not want it flat. :)
jnthn Celelibi: Well, .tree it or something then :)
TimToady r: say ((1, 2, 3) X (4, 5, 6)).tree.perl
p6eval rakudo d9fd17: OUTPUT«([1, 4], [1, 5], [1, 6], [2, 4], [2, 5], [2, 6], [3, 4], [3, 5], [3, 6]).list␤»
jnthn Celelibi: Things don't commit to flat or structured until you put them in a context that makes them do so.
[Coke]: I don't see how this one could ever have worked on an implementation that does import tags correctly. 20:55
Celelibi ok
jnthn [Coke]: It relies on something being in :DEFAULT, when the exporting module is putting it elsewhere.
Celelibi then flat method appies a list context in the data.
and tree? 20:56
20:56 envi_ left
TimToady doesn't :) 20:56
jnthn LoL context, or whatever we call it... :)
[Coke] jnthn: it was 2009. We didn't know any better!
TimToady lol only goes one level down
jnthn Troo.
[Coke] no, *tree* 20:57
20:57 _Vi joined
TimToady and X should be returning Parcels 20:57
so it can come out (1,4),(1,5),(1,6) etc...
jnthn [Coke]: :P
TimToady: Well, that's just how it comes out once everything gets reified...
TimToady: I guess you're saying it's being a bit too lazy? 20:58
TimToady seems so
Celelibi Hum... My version of rakudo do not evaluate the lazy lists with .tree
TimToady x knows perfectly well how many elements it's returning each time, and that they're scalars
X*
it's only the outer list of parcels needs to be lazy 20:59
jnthn TimToady: OK. Maybe masak++ would be so good as to rakudobug this; pmichaud++ may be trolled into fixing it or commenting "yes, just needs X" :)
masak r: say 1, 2, 3 X* 4, 5, 6
p6eval rakudo d9fd17: OUTPUT«4 5 6 8 10 12 12 15 18␤»
masak jnthn: can I do it tomorrow. my mana is too low right now... :) 21:00
or someone else could, too...
sisar "mana" ?
jnthn masak: Sure :)
TimToady polynesian god power
jnthn sisar: masak is fed by food falling from the sky, but not much fell today, depleting his energy...
[Coke] r: (1..6 X* 1..6).uniq.sort.say 21:01
p6eval rakudo d9fd17: OUTPUT«1 2 3 4 5 6 8 9 10 12 15 16 18 20 24 25 30 36␤»
TimToady wonders if masak has played Populace...
21:01 kaare_ left
[Coke] masak: coding taks mana and not health? hurm. 21:02
TimToady er, Populous...
[Coke] *takes
Celelibi How could I create a lazy list myself? 21:03
tadzik using gather take
Celelibi To create an infinite list for instance.
tadzik r: my @a = gather for 1..* { take $_ * 2 }; say @a[^30] 21:04
eek, eager assignment :/ 21:05
p6eval rakudo d9fd17: OUTPUT«(timeout)»
masak [Coke]: compiler hacking takes mana: strangelyconsistent.org/blog/perl-6-is-my-mmorpg
tadzik r: my @a := gather for 1..* { take $_ * 2 }; say @a[^30]
p6eval rakudo d9fd17: OUTPUT«2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56 58 60␤»
Celelibi := ?
TimToady there are a number of easier ways to make infinite lists 21:06
sisar zz.. 21:07
masak Celelibi: binding, as opposed to assignment.
TimToady it depends on how fancy you need the list to be
note that tadzik's example starts with the infinite list 1..*
Celelibi .oO(Y U SO SLOW TO LEARN BRAIN?!)
21:08 skids left
TimToady (well, techinically, it's an infinite range that gets turning into an infinite list...) 21:08
and the even numbers are more easily written just as 2,4,6...*
Celelibi How heavy is the CPU-cost of making a lazy list? 21:09
TimToady so generally you don't need to write your own gather most of the time
Celelibi: a lot lighter than generating the whole list
</rimshot>
masak unless the whole list is empty...
TimToady serious answer: heavier currently than we hope it will be in the future
Celelibi I hope so, indeed. :) 21:10
masak .oO( native lazy lists o.O )
Celelibi But does it make sens to make every list lazy?
I think not.
TimToady the compiler is free to optimize when it can determine there's an eager context 21:11
Celelibi I think lists of tenth of elements are not worth lazying.
TimToady and the lazy implementation is allowed to batch things up for efficiency 21:12
Celelibi what does mean eager actually?
and the gather keyword? And take?
I think "take" means "stop execution here, take this value" 21:13
TimToady gather/take is essentially our coroutine interface, but it's orthogonal to function calls
Celelibi (and we'll continue execution for the next value)
TimToady yes, take, is a 'yield'
Celelibi ok.
TimToady but easier to spell
Celelibi indeed. 21:14
diakopter beavers can be eager. gather is a gospel singer/songwriter.
Celelibi and gather means "lazy expression follow here"?
or lazy list only?
TimToady it means to set up a coroutine and/or thread 21:15
with a thread, the insides could work ahead until the pipeline stalls
Celelibi ok.
TimToady again, up to the compiler
Celelibi but the programmer then have to handle the synchronization on its own (if any). 21:16
I mean, mutex and stuff.
TimToady no, the compiler would have to handle the handshaking if it implements it that way
Celelibi can it? 21:17
TimToady if a gather's code depends on side-effects from outside, then it's likely erroneous
but if it depends on other lazy lists, that's okay, since those can block automatically till there are values available 21:18
Celelibi I thought about that, and cannot figure out a use case where this would make sens. But that does not mean it doesn't make sens at all.
TimToady we're trying to make Perl 6 very multicore/manycore friendly
so a lot of the constructs are promising what you don't care about, to give room for the compiler to do whatever is best 21:19
whether that's mapping to manycores, or pipeline processors, or GPUs...
our current compilers don't try to do any of that, so they're necessarily a little slow to emulate parallelism 21:21
Celelibi I think these places where parallelism can occur will have to be well documented since it will be the source of numerous insidious bugs. 21:22
If you did some parallel coding you should know how there bugs are hard to track. :) 21:23
and randomly appears. 21:24
and disappear...
TimToady certainly, but generally the more you stick to the FP way of doing things, the safer you'll be, since it's usually side effects that get you into trouble 21:25
Celelibi FP? Functionnal programming? 21:26
TimToady OO also helps, insofar as side-effects are encapsulated into objects
yes
Celelibi k
ow ow ow. Does "gather" create a closure?
TimToady we recognize that we are multiplying the possibilities for people to write erroneous code, but so far we haven't run into many problems 21:27
the block form does
Celelibi nice.
TimToady but usually we just use 'gather' as a prefix for a loop that has its own closure
Celelibi ok, then this lower the risk of unpredictable behavior. 21:29
TimToady unpredictable behavior is not always completely evil, either, as long as the rocketship stays on course 21:31
Celelibi indeed.
TimToady the trick is to know when you care, and when you don't
so all of the parallel constructs in Perl 6 tend to make you swear that you don't care :) 21:32
apart from the guarantess of the particular constructs
hypers still preserve order, even though they don't care what order the computation proceeds 21:33
likewise gather will return elements in the order that take sends them
Celelibi One thing I was thinking about was a lazy/infinite list that rely upon I/O to construct continuous flow of elements. And when the input change, the list elements will eventyally change. You don't know when, you only know this will happen.
-y+u
But well, the lists may not be well suitied for a flow. 21:35
(suited?)
TimToady r: for '/dev/random'.IO.comb -> { say .ord } 21:36
p6eval rakudo d9fd17: OUTPUT«Method 'comb' not found for invocant of class 'IO'␤ in block <anon> at /tmp/rzUIx934xc:1␤␤»
Celelibi I'm sure there are use case for things like this.
21:37 jferrero joined
TimToady r: for '/dev/random'.IO.lines».comb -> { say .ord } 21:37
p6eval rakudo d9fd17: OUTPUT«Malformed UTF-8 string␤␤ in method get at src/gen/CORE.setting:6611␤ in code <anon> at src/gen/CORE.setting:6632␤ in sub coro at src/gen/CORE.setting:4926␤ in method reify at src/gen/CORE.setting:4907␤ in method reify at src/gen/CORE.setting:4678␤ in method r…
TimToady r: for open('/dev/random', :raw).lines».comb -> { say .ord }
p6eval rakudo d9fd17: OUTPUT«open is disallowed in safe mode␤ in sub restricted at src/SAFE.setting:2␤ in sub open at src/SAFE.setting:5␤ in block <anon> at /tmp/QEoAmVvyp6:1␤␤»
TimToady aww
we don't seem to have any iterating method defined on binary files yet; read just reads a single buf, not a sequence of them 21:40
and you can't use slurp on an infinite file 21:41
Celelibi thought /dev/random is not infinite /dev/urandom is;
TimToady anyway, that should certainly be possible, mapping a binary file to a lazy list 21:42
currently we can do it with .lines on a text file
Celelibi I think "take" forces the evaluation of its expression. 21:43
TimToady well, on some level that does not include eager-izing the list 21:44
n: my @a := gather loop { take 2,4,6...* }; say @a[^10] 21:45
p6eval niecza v17-2-gb8fcf85: OUTPUT«(timeout)»
TimToady innersting
r: my @a := gather loop { take 2,4,6...* }; say @a[^10]
p6eval rakudo d9fd17: OUTPUT«2 4 6 8 10 12 14 16 18 20␤»
TimToady looks like a nieczabug
Celelibi how does perl infer the list I want?
r: say 1, 2, 3 ... 10 21:46
p6eval rakudo d9fd17: OUTPUT«1 2 3 4 5 6 7 8 9 10␤»
Celelibi r: say 1, 3, 5 ... 10
r: say 1, 3, 5 ... 11
p6eval rakudo d9fd17: OUTPUT«(timeout)»
rakudo d9fd17: OUTPUT«1 3 5 7 9 11␤»
TimToady see S03:1806
Celelibi r: say 1, 3, 5 ... 10
21:46 Guest40378 left
Celelibi interesting. 21:47
p6eval rakudo d9fd17: OUTPUT«(timeout)»
TimToady that's an infinite loop
21:48 _dolmen_ left
TimToady bbl & 21:51
21:52 tyatpi joined 21:53 xenu joined
Celelibi oh, are closures and thread nice to combine in perl6? 21:53
21:53 xenu is now known as Guest21759
Celelibi I think I made perl5 crash with things like returning a function (with a closure from inside a map coderef or something like that) as a result of a thread. 21:54
22:06 jferrero left 22:12 jferrero joined 22:26 whiteknight joined 22:33 James joined 22:34 James is now known as Guest69870 22:36 fhelmberger left
jnthn 'night, #perl6 22:39
22:41 tokuhirom joined, fgomez left 22:42 PacoAir left 22:49 bluescreen10 left 22:51 fgomez joined 22:54 Guest69870 left 23:04 Timbus left
gfldex p6: class A { class Str {}; method new() { return Str.new }; }; say A.new.WHAT; 23:06
p6eval pugs: OUTPUT«Str␤»
..rakudo d9fd17, niecza v17-2-gb8fcf85: OUTPUT«Str()␤»
23:06 Timbus joined
gfldex is there any way to get hold of the outer class name? 23:06
23:07 skids joined
gfldex p6: package A { class Str {}; method foo() { return Str.new }; }; say foo.WHAT; 23:08
p6eval pugs: OUTPUT«*** No such subroutine: "&foo"␤ at /tmp/IziyrsP17y line 1, column 67-75␤»
..niecza v17-2-gb8fcf85: OUTPUT«===SORRY!===␤␤A package cannot have methods added at /tmp/3bLD3APrKG line 1:␤------> package A { class Str {}; method foo⏏() { return Str.new }; }; say foo.WHAT;␤␤Undeclared routine:␤ 'foo' used at line 1␤␤Potential diff…
..rakudo d9fd17: OUTPUT«Useless declaration of a has-scoped method in package␤===SORRY!===␤CHECK FAILED:␤Undefined routine '&foo' called (line 1)␤»
gfldex p6: package A { class Str {}; method foo() { return Str.new }; }; say (A::foo).WHAT;
p6eval pugs: OUTPUT«*** No compatible multi variant found: "&A::foo"␤ at /tmp/g4zjF335HE line 1, column 68-74␤»
..niecza v17-2-gb8fcf85: OUTPUT«===SORRY!===␤␤A package cannot have methods added at /tmp/eQ4ydRPDaL line 1:␤------> package A { class Str {}; method foo⏏() { return Str.new }; }; say (A::foo).W␤␤Potential difficulties:␤ 'method' declaration outside o…
..rakudo d9fd17: OUTPUT«Useless declaration of a has-scoped method in package␤Could not find symbol 'A::&foo'␤ in block <anon> at /tmp/x_cIUUBbOz:1␤␤»
gfldex p6: package A { class Str {}; sub foo() { return Str.new }; }; say (A::foo).WHAT;
p6eval rakudo d9fd17: OUTPUT«Could not find symbol 'A::&foo'␤ in block <anon> at /tmp/P_yi5zLoUR:1␤␤»
..niecza v17-2-gb8fcf85: OUTPUT«Potential difficulties:␤ &foo is declared but not used at /tmp/iQmOxjP6q0 line 1:␤------> package A { class Str {}; sub foo⏏() { return Str.new }; }; say (A::foo).W␤␤Unhandled exception: Unable to resolve method postcircumfix:<( )> in cl…
..pugs: OUTPUT«Str␤»
gfldex p6: package A { class Str {}; our sub foo() { return Str.new }; }; say (A::foo).WHAT; 23:09
p6eval pugs: OUTPUT«Str␤»
..niecza v17-2-gb8fcf85: OUTPUT«Str()␤»
..rakudo d9fd17: OUTPUT«===SORRY!===␤This type cannot box a native string␤»
gfldex :)
n: say Str.^methods; 23:10
p6eval niecza v17-2-gb8fcf85: OUTPUT«Unhandled exception: Unable to resolve method methods in class ClassHOW␤ at /tmp/pHIchPEjyY line 1 (mainline @ 3) ␤ at /home/p6eval/niecza/lib/CORE.setting line 3840 (ANON @ 3) ␤ at /home/p6eval/niecza/lib/CORE.setting line 3841 (module-CORE @ 65) ␤ at …
gfldex n: say Str.new.^methods;
p6eval niecza v17-2-gb8fcf85: OUTPUT«Unhandled exception: Unable to resolve method methods in class ClassHOW␤ at /tmp/jlOmCxTZcI line 1 (mainline @ 4) ␤ at /home/p6eval/niecza/lib/CORE.setting line 3840 (ANON @ 3) ␤ at /home/p6eval/niecza/lib/CORE.setting line 3841 (module-CORE @ 65) ␤ at …
sorear don't bother
gfldex p6: package A { class Str {}; our sub foo() { return Str.new }; }; say (A::foo).chomp; 23:11
p6eval pugs: OUTPUT«<obj:Str>␤»
..niecza v17-2-gb8fcf85: OUTPUT«Unhandled exception: Unable to resolve method chomp in class Str␤ at /tmp/BusWEUMYFX line 1 (mainline @ 4) ␤ at /home/p6eval/niecza/lib/CORE.setting line 3840 (ANON @ 3) ␤ at /home/p6eval/niecza/lib/CORE.setting line 3841 (module-CORE @ 65) ␤ at /home/p…
..rakudo d9fd17: OUTPUT«===SORRY!===␤This type cannot box a native string␤»
gfldex niecza is doing the right thing as it seams
gfldex finds his rakudobug pencil 23:12
23:13 REPLeffect left, REPLeffect joined
moritz r: my $f = open(:bin, '/dev/urandom'); say $f.read(20) 23:14
p6eval rakudo d9fd17: OUTPUT«open is disallowed in safe mode␤ in sub restricted at src/SAFE.setting:2␤ in sub open at src/SAFE.setting:5␤ in block <anon> at /tmp/CslPnRpCyt:1␤␤»
moritz r: my $f = '/dev/urandom'.IO; say $f.read(20)
p6eval rakudo d9fd17: OUTPUT«Method 'read_bytes' not found for invocant of class 'Any'␤ in method read at src/gen/CORE.setting:6638␤ in block <anon> at /tmp/KwbBSL_MsL:1␤␤»
23:15 _ilbot left, _ilbot joined
gfldex std: package A { class Str {} our sub foo() } 23:16
p6eval std fab64fc: OUTPUT«===SORRY!===␤Strange text after block (missing comma, semicolon, comment marker?) at /tmp/oUZxkkwmjC line 1:␤------> package A { class Str {}⏏ our sub foo() }␤ expecting any of:␤ infix or meta-infix␤ infixed function␤ statement…
gfldex std: package A { class Str {} sub foo() }
p6eval std fab64fc: OUTPUT«===SORRY!===␤Strange text after block (missing comma, semicolon, comment marker?) at /tmp/QgsC1Eytse line 1:␤------> package A { class Str {}⏏ sub foo() }␤ expecting any of:␤ infix or meta-infix␤ infixed function␤ statement
..mod…
gfldex std: package A { class Str {} sub foo(){} }
p6eval std fab64fc: OUTPUT«===SORRY!===␤Strange text after block (missing comma, semicolon, comment marker?) at /tmp/QFqrtjHfGG line 1:␤------> package A { class Str {}⏏ sub foo(){} }␤ expecting any of:␤ infix or meta-infix␤ infixed function␤ statement
..m…
gfldex std: package A { class Str {}; sub foo(){} }
p6eval std fab64fc: OUTPUT«ok 00:00 41m␤»
dalek blets: 62f4516 | (Herbert Breunung)++ | docs/tablet-2-basic-syntax.txt:
start working on tablet 2
 23:17
23:18 fgomez left 23:20 fgomez joined 23:25 Timbus left, Timbus joined
raiph re: "The Strong Turing Test: emulating you so well that you don't know whether you're the real you or not..."; www.newbanner.com/SecHumSCM/WhereAmI.html 23:33
23:45 jferrero left 23:49 Chillance left 23:53 NamelessTee 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!