»ö« | perl6.org/ | nopaste: paste.lisp.org/new/perl6 | evalbot usage: 'perl6: say 3;' or rakudo: / pugs: / std: , or /msg p6eval perl6: ... | irclog: irc.pugscode.org/ | UTF-8 is our friend!
Set by diakopter on 25 January 2010.
lue I want to set it as my wallpaper, but then I might not read it :( (only 1024*768) 00:00
bah. I'll crop it later, then I can read it as my wallpaper :) 00:07
ng: my @rray=0..*;say @rray; 00:16
p6eval ng 5635e9: ( no output ) 00:17
lue ng: my @rray=[0,1,2]; say @rray[0..*];
p6eval ng 5635e9: OUTPUT«No applicable candidates found to dispatch to for 'postcircumfix:<[ ]>'␤current instr.: '!postcircumfix:<[ ]>' pc 306755 (src/gen/core.pir:39353)␤»
lue ng: my @rray=0..2; say @rray[0..*];
p6eval ng 5635e9: OUTPUT«No applicable candidates found to dispatch to for 'postcircumfix:<[ ]>'␤current instr.: '!postcircumfix:<[ ]>' pc 306755 (src/gen/core.pir:39353)␤»
lue how does an array store a range? Questions... 00:21
ng: say 0..3.minmax 00:23
p6eval ng 5635e9: OUTPUT«Method 'minmax' not found for invocant of class 'Integer'␤current instr.: '_block14' pc 29 (EVAL_1:0)␤»
lue ng: say (0..3).minmax 00:23
p6eval ng 5635e9: OUTPUT«03␤» 00:24
lue ng: my @rray=0..3;say @rray.minmax
p6eval ng 5635e9: OUTPUT«Method 'minmax' not found for invocant of class 'Array'␤current instr.: '_block14' pc 29 (EVAL_1:0)␤»
lue grrr
mberends lue: ng is probably too incomplete to explore Perl 6 today, You could try rakudo: instead of ng: for many commands.
lue yeah, trying to fix whatever indices. Trying to find where and how to implement it... 00:25
rakudo: my @rray=0..3;say @rray.minmax
p6eval rakudo 1d4928: OUTPUT«Method 'minmax' not found for invocant of class 'Perl6Array'␤in Main (file src/gen_setting.pm, line 324)␤»
lue pugs: my @rray=0..3;say @rray.minmax
p6eval pugs: OUTPUT«*** No such method in class Array: "&minmax"␤ at /tmp/Ell1aSMesK line 1, column 19 - line 2, column 1␤»
diakopter anyone around? hi! 00:48
lue yes! just figuring stuff out. 00:49
whatever indices, mainly.
pugs: @rray=[0,1,2]; say @rray[0..*]; 00:50
p6eval pugs: OUTPUT«*** ␤ Unexpected "=["␤ expecting "::"␤ Variable "@rray" requires predeclaration or explicit package name␤ at /tmp/63rzdlL1du line 1, column 6␤»
lue pugs: @rray=0,1,2; say @rray[0..*];
p6eval pugs: OUTPUT«*** ␤ Unexpected "="␤ expecting "::"␤ Variable "@rray" requires predeclaration or explicit package name␤ at /tmp/mkhmm9hqMV line 1, column 6␤»
lue pugs: my @rray=[0,1,2]; say @rray[0..*]; 00:51
p6eval pugs: OUTPUT«0 1 2␤»
lue pugs: my @rray=[0,1,2]; say @rray[1..*];
p6eval pugs: OUTPUT«␤»
ash_ why are you using pugs? just cuirous
lue to try and see how it should come out. I'd use std, but it only say OK and then a time.
lue so whatever doesn't mean "until the end of the list" (that'd be too easy). 00:54
pugs: my @rray=[0,1,2]; say @rray[*..2]; 00:55
p6eval pugs: OUTPUT«␤»
ash_ you can use rakudo: or ng: if you want an actively developed version of perl6
lue rakudo: my @rray=[0,1,2]; say @rray[*..2];
p6eval rakudo 1d4928: OUTPUT«Use of uninitialized value␤Use of uninitialized value␤␤»
lue rakudo: my @rray=[0,1,2]; say @rray[*..*]; #heh heh
p6eval rakudo 1d4928: OUTPUT«Use of uninitialized value␤␤»
lue pugs: my @rray=[0,1,2]; say @rray[*..*]; #heh heh 00:56
p6eval pugs: OUTPUT«0 1 2␤»
ash_ ng: my @a = 1...5; say ~@a;
p6eval ng 5635e9: OUTPUT«6 6 6 6 6␤»
ash_ hmm
that's not right
colomon ash_: woah.
ash_ rakudo: my @a = 1...5; say ~@a;
p6eval rakudo 1d4928: OUTPUT«1 2 3 4 5␤»
ash_ thats better
colomon: yeah, never seen that problem before... 00:57
lue is that a problem with the array or the range? I'm guessing range.
ash_ ng: my @a = 1,2,3,4,5; say @a;
p6eval ng 5635e9: OUTPUT«12345␤»
ash_ ng: my @a = 1,2,3,4,5; say ~@a;
p6eval ng 5635e9: OUTPUT«1 2 3 4 5␤»
ash_ ng: my @a = 1...6; say ~@a[0];
p6eval ng 5635e9: OUTPUT«7␤»
ash_ lue: doing * in the range is probably not working correctly 00:58
lue for the ... infix, it generates numbers with a "gather loop" (in ng)
ash_ colomon: ... seems to be wrong in most places in ng 00:59
lue how to know if the * indices are working right? Hmmm...
ash_ ng: say 1...6;
p6eval ng 5635e9: OUTPUT«777777␤»
ash_ ng: say 1..6; # infix .. seems alright however 01:00
p6eval ng 5635e9: OUTPUT«123456␤»
lue ng: gather loop(my $calar = 0; $calar <= 6; $calar++){say take $calar;} 01:01
p6eval ng 5635e9: OUTPUT«Whitespace required after 'loop' at line 1, near "(my $calar"␤current instr.: 'perl6;HLL;Grammar;panic' pc 500 (src/stage0/HLL-s0.pir:328)␤»
lue ng: gather loop (my $calar = 0; $calar <= 6; $calar++){say take $calar;}
p6eval ng 5635e9: ( no output )
lue ng: @rray=(gather loop (my $calar = 0; $calar <= 6; $calar++){take $calar;}); say @rray; 01:02
p6eval ng 5635e9: OUTPUT«Symbol '@rray' not predeclared in <anonymous>␤current instr.: 'perl6;PCT;HLLCompiler;panic' pc 137 (src/PCT/HLLCompiler.pir:101)␤»
lue ng: my @rray=(gather loop (my $calar = 0; $calar <= 6; $calar++){take $calar;}); say @rray;
p6eval ng 5635e9: OUTPUT«7777777␤»
lue that gather loop is used in the ... infix.
is ... supposed to be any different from .. ? 01:05
ash_ yes, ... is different than .. 01:06
ash_ ... is a series operator, .. is a range 01:06
lue that loop, then, seems to be broken somewhere. 01:07
colomon ash_: I think it's gather / take that's actually broken. (Sorry for the pause, was backlogging and prepping milk bottles for my boy.) 01:08
lue yeah, that's what it is. Makes most sense.
ash_ no worries colomon 01:09
colomon actually, it looks like the old rakudo master bug where saying take $i yielded a reference to $i rather than $i's value.
ash_ so gather take isn't functioning properly then?
colomon There is some sort of wonkiness with captures, it may be affecting gather / take as well.
most of our tests have looked more like 01:10
lue ng: my $i=5; say $i.VALUE;
p6eval ng 5635e9: OUTPUT«Method 'VALUE' not found for invocant of class 'Integer'␤current instr.: '_block14' pc 29 (EVAL_1:0)␤»
colomon ng: (1...5).map({$_ * 2}).eager.perl.say
lue eh, worth a shot.
p6eval ng 5635e9: OUTPUT«(2, 4, 6, 8, 10)␤»
colomon See, in that case we process each iteration in the gather / take through the map before going on to the next. 01:11
So you wouldn't notice if it was giving you $i instead of the value of $i.
If that's the case, though, there's an easy patch.
(while waiting for compile...) 01:13
btw lue, I think your examples of [0..*] and [1..*] back there are misleading.
you said @rray = [1, 2, 3]
I believe that creates a one-element array whose first element is [1, 2, 3]
that would certainly explain your results. 01:14
lue ooh!
colomon try @rray = 1, 2, 3 01:14
lue ng: my @rray = 1, 2, 3; say @rray[0..*];
p6eval ng 5635e9: OUTPUT«No applicable candidates found to dispatch to for 'postcircumfix:<[ ]>'␤current instr.: '!postcircumfix:<[ ]>' pc 306755 (src/gen/core.pir:39427)␤»
lue rakudo: my @rray = 1, 2, 3; say @rray[0..*];
p6eval rakudo 1d4928: OUTPUT«123Use of uninitialized value␤␤»
lue pugs: my @rray = 1, 2, 3; say @rray[0..*];
p6eval pugs: OUTPUT«123␤»
lue pugs: my @rray = 1, 2, 3; say @rray[1..*];
p6eval pugs: OUTPUT«23␤»
lue yes, there we go. (old python habits again) 01:15
colomon > my @a = 1...5; say ~@a;
1 2 3 4 5
That's my local version, ie the patch works.
lue hooray! now, back to my * indices (grumble)... 01:16
rakudo: say (0..3).perl 01:17
p6eval rakudo 1d4928: OUTPUT«0..3␤»
lue rakudo: say (0..3).Str 01:18
p6eval rakudo 1d4928: OUTPUT«0 1 2 3␤»
lue rakudo: say (*..3).Str
p6eval rakudo 1d4928: OUTPUT«WhateverCode()<0x2b768446f740>␤»
lue rakudo: say (0..*).Str
p6eval rakudo 1d4928: OUTPUT«WhateverCode()<0x2ac2c1ec1740>␤»
lue rakudo: say (0..*).perl
p6eval rakudo 1d4928: OUTPUT«{ ... }␤»
lue rakudo: say (*..3).perl
p6eval rakudo 1d4928: OUTPUT«{ ... }␤»
colomon lue: If (*..3).Str "worked properly", you'd get an infinitely long string of numbers ending with 3. 01:19
lue I know :) 01:20
just had to see what happens.
trying to figure out how ranges works, so I can get them to work right with *. 01:22
rakudo: my @rray=0..4; say @rray;
p6eval rakudo 1d4928: OUTPUT«01234␤»
lue rakudo my @rray=0..4; say @rray.WHAT 01:26
rakudo: my @rray=0..4; say @rray.WHAT
p6eval rakudo 1d4928: OUTPUT«Array()␤»
lue rakudo: say (0..4).WHAT 01:27
p6eval rakudo 1d4928: OUTPUT«Range()␤»
lue rakudo: say Array().WHAT 01:27
p6eval rakudo 1d4928: OUTPUT«invoke() not implemented in class 'ResizablePMCArray'␤in Main (file <unknown>, line <unknown>)␤»
colomon lue: pretty sure Ranges do work right with *.
They just don't work in slices.
lue yeah, I'm trying to figure out indices. I think i'm looking in the wrong place. 01:28
colomon Try looking at Seq.
lue thank you, colomon!
colomon probably Seq.pir. 01:29
ash_ colomon: does rakudo/ng have this many precedence levels as perlcabal.org/syn/S06.html#Subroutine_traits (down a bit starts with category:<prefix>) lists?
colomon ng has a lot of those, but as far as I know is tighter/equiv/looser doesn't actually work yet. 01:30
er, I may have just answered completely the wrong thing.
colomon are those actually precedence levels? 01:31
dalek kudo/ng: 8013237 | (Erik Johansen)++ | src/core/ (2 files):
Add .Bool to Rat and Complex
kudo/ng: c0d94fd | (Solomon Foster)++ | src/builtins/Role.pir:
Merge branch 'ng' of [email@hidden.address] into ng
kudo/ng: ae60f8c | (Solomon Foster)++ | src/core/Range.pm:
Workaround in ... to deal with the gather / take bug.
ash_ yeah, thats what I was wondering colomon 01:31
colomon If you look at line 880 of ng's Grammar.pm, that's where the precedence levels for ng are defined. 01:32
ash_ okay, cool 01:33
that helps
lue rakudo: my @rray{1..7};say @rray;
p6eval rakudo 1d4928: OUTPUT«Confused at line 10, near "{1..7};say"␤in Main (file <unknown>, line <unknown>)␤»
lue rakudo: my @rray{1..7}=2..8;say @rray; 01:34
p6eval rakudo 1d4928: OUTPUT«Confused at line 10, near "{1..7}=2.."␤in Main (file <unknown>, line <unknown>)␤»
lue :(
colomon ng: my @rray = 1..7; say @rray
p6eval ng 5635e9: OUTPUT«1234567␤»
colomon rakudo: my @rray; @rray[1..7] = 2..8; say @rray
p6eval rakudo 1d4928: OUTPUT«Use of uninitialized value␤2345678␤»
lue looking at S09 for indices, it uses {}. 01:35
ash_ ng: my @a; @a[1..3] = 2..4;
p6eval ng 5635e9: OUTPUT«No applicable candidates found to dispatch to for 'postcircumfix:<[ ]>'␤current instr.: '!postcircumfix:<[ ]>' pc 306755 (src/gen/core.pir:39269)␤»
colomon lue: I'm guessing that's just not yet implemented in rakudo, either master or ng. 01:37
lue rakudo my @rray{ 1..7 };@rray{1}=0;say @rray{1};
yeah, let me see. 01:38
rakudo: my @rray{ 1..7 };@rray{1}=0;say @rray{1};
p6eval rakudo 1d4928: OUTPUT«Confused at line 10, near "{ 1..7 };@"␤in Main (file <unknown>, line <unknown>)␤»
lue pugs: my @rray{ 1..7 };@rray{1}=0;say @rray{1};
p6eval pugs: OUTPUT«*** Cannot cast into Hash: VRef <Array:0x2abc73ada031>␤ at /tmp/bF3yHkB4xP line 1, column 1-17␤»
lue :(
ash_ my $a, @b, $c = 1, 2, 3, 4, 5; say 'first: ' ~ $a ~ ' last: ' ~ $c;
ng: my $a, @b, $c = 1, 2, 3, 4, 5; say 'first: ' ~ $a ~ ' last: ' ~ $c;
p6eval ng 5635e9: OUTPUT«Symbol '@b' not predeclared in <anonymous>␤current instr.: 'perl6;PCT;HLLCompiler;panic' pc 137 (src/PCT/HLLCompiler.pir:101)␤»
lue nothing seems to agree with the synopsis. 01:39
ash_ i'd guess its not yet implemented
lue so indices are not yet implemented, let alone whatever indices.
ash_ [] work, but thats not a custom index 01:40
ng: my @a = 1, 2, 3; @a[2].perl;
p6eval ng 5635e9: ( no output )
ash_ ng: my @a = 1, 2, 3; @a[2].perl.say;
p6eval ng 5635e9: OUTPUT«3␤»
lue bet that'd be fun to implement. :| 01:41
what you'd have to do is create a {} operator that maps the "invented" index to the "actual" index. 01:43
rakudo: say «1 2 3 4».WHAT
p6eval rakudo 1d4928: OUTPUT«List()␤»
lue rakudo: say (1..4).WHAT 01:44
p6eval rakudo 1d4928: OUTPUT«Range()␤»
lue yeah, indices will be fun :/
ash_ colomon: is that O thing that is used to specify precedence in the docs? i don't see it in S05 01:45
lue rakudo: my enum Days «:Sun(1) Mon Tue Wed Thu Fri Sat»; say Days.WHAT; 01:46
p6eval rakudo 1d4928: OUTPUT«Malformed declaration at line 10, near "enum Days "␤in Main (file <unknown>, line <unknown>)␤»
colomon ash_: I doubt it. 01:47
colomon Is that my $a, @b, $c = 1, 2, 3, 4, 5; actually in the specs? 01:47
I would have thought it would be my ($a, @b, $c) = 1, 2, 3, 4, 5;
and that doesn't work at the moment, unless someone fixed it today while I wasn't looking. 01:48
lue rakudo: my @rray=0..4; say @rray.WHAT #I think I have it!
p6eval rakudo 1d4928: OUTPUT«Array()␤»
lue rakudo: my @rray=<0 1 2 3 4>; say @rray.WHAT; 01:49
p6eval rakudo 1d4928: OUTPUT«Array()␤»
lue I've got it! I just need to know where to put it! 01:50
pugs_svn r29659 | colomon++ | [t/spec] Defudge test fixed by the gather / take workaround in infix:<...>. 01:51
colomon lue: my @rray=<0 1 2 3 4>; actually is a one-element array whose one element is an array of strings, AFAIK. 01:51
lue just figuring out if some defines an indice as <Summer Autumn Spring Winter>. 01:53
ash_ colomon: that $a, @b, $c wasn't in the spec that I can find, i was just guessing, it might not be allowed at all
lue ex. my @rray { <Sum Aut Spr Win> };
colomon ash_: for sure my ($a, @b) is allowed. 01:54
sorry, I mean, my ($a, @b) = @c.
ash_ can't you do that with params? like sub foo($a, @b, $c) ?
colomon ash_: yes, but that's different. 01:55
because then you have to pass three arguments, the middle of which is an array.
my ($a, @b) = list puts the first element of the list in $a and the rest in @b.
lue ng: my $calar=<0 1 2>; say $calar; 01:58
p6eval ng ae60f8: OUTPUT«0 1 2␤»
lue ng: my $calar=<0 1 2>; say $calar.WHAT;
p6eval ng ae60f8: OUTPUT«Null PMC access in elements()␤current instr.: 'perl6;SeqIter;get' pc 11981 (src/builtins/Positional.pir:16)␤»
lue rakudo: my $calar=<0 1 2>; say $calar.WHAT;
p6eval rakudo 1d4928: OUTPUT«Array()␤»
lue pugs: my $calar=<0 1 2>; say $calar.WHAT;
p6eval pugs: OUTPUT«Array␤»
lue ah. Is that legal? Array in a scalar? 01:59
ash_ yea
lue hm, I think I'm close to the how on indices. 02:00
ash_ you can override how postcircumfix:{ } works, for a special index on an array, i imagine thats how it will be done eventually 02:01
ash_ [ ] for normal access, { } for special access 02:01
lue how can you see if a scalar ends up holding an array (ie variable.WHAT==Array, or something like that) 02:04
colomon lue: $var ~~ Array
lue ah. 02:05
lue thinks about how he REALLY REALLY should learn perl before fixing it.
colomon Though I think $var ~~ Seq and $var ~~ Positional should both work as well, and be more inclusive.
ash_ the specs are a great place to start, they change a lot though
colomon (Seq is an immutable Array, and Positional is the role both Seq and Array do.) 02:06
Not sure if Positional is actually in place in ng at the moment or not.
ash_ ng: say Array ~~ Positional
p6eval ng ae60f8: OUTPUT«0␤»
ash_ ng: say Positional ~~ Array; # oops had it backwards
p6eval ng ae60f8: OUTPUT«0␤»
colomon No, doesn't look like Positional is in Seq yet (just checked the source). 02:07
ash_ ng: my $foo = (1, 2, 3); say $foo.WHAT; for @($foo) { say $_.perl; }
p6eval ng ae60f8: OUTPUT«Null PMC access in elements()␤current instr.: 'perl6;SeqIter;get' pc 11981 (src/builtins/Positional.pir:16)␤» 02:08
colomon Positional is defined, though... 02:08
ash_ ng: my $foo = (1, 2, 3); say $foo.WHAT; 02:09
p6eval ng ae60f8: OUTPUT«Null PMC access in elements()␤current instr.: 'perl6;SeqIter;get' pc 11981 (src/builtins/Positional.pir:16)␤»
ash_ hmmm thats not right i think
pugs_svn r29660 | colomon++ | [t/spec] Unfudge tests that now work in ng. 02:11
lue rakudo: my $foo = (1, 2, 3); say $foo.WHAT;
p6eval rakudo 1d4928: OUTPUT«Array()␤»
ash_ ng: my $foo = (1, 2, 3); say $foo.perl; 02:12
p6eval ng ae60f8: OUTPUT«Method 'perl' not found for invocant of class 'Seq'␤current instr.: '_block14' pc 29 (EVAL_1:0)␤»
ash_ thats why its doing that
lue I'm looking for a way to take an unnamed array (<1 2 3>) and put it into a variable as its original 1d list. 02:13
rakudo: my @rray=\<1 2 3>; say @rray[1];
p6eval rakudo 1d4928: OUTPUT«Use of uninitialized value␤␤»
lue rakudo: my $calar=<1 2 3>; my @rray=$calar; say @rray[1]; 02:14
p6eval rakudo 1d4928: OUTPUT«Use of uninitialized value␤␤»
lue hmmm
ash_ my $a = <1 2 3>; my @a = @($a); say ~@a; ? 02:15
oops
ng: my $a = <1 2 3>; my @a = @($a); say ~@a; # like that?
p6eval ng ae60f8: OUTPUT«1 2 3␤»
ash_ or if you want to check that it did it right you can try:
ng: my $a = <1 2 3>; my @a = @($a); for @a { say $_.perl } 02:16
lue ng: my $a = <1 2 3>; my @a = @($a); say @a[1];
p6eval ng ae60f8: OUTPUT«"1"␤"2"␤"3"␤»
ng ae60f8: OUTPUT«2␤»
lue yes! Thank you!
ash_ hehe, i have bad variable names
lue ng: my $a = 1..3; my @a = @($a); say @a[1];
p6eval ng ae60f8: OUTPUT«2␤»
ash_ perl 6 lets you get away with re-using variables with different sigils even though they never point to the same thing
lue yay, I don't need any ifs! 02:17
ash_ $a is not related to @a at all
so you know
lue yes, same way Iterator.pm and iterator.pm are two different files in *nix.
ash_ @() returns something in list context, so does the list op if its implemented (but might want to double check that)
list $a == @($a) # is what that means 02:19
lue I now have the how of _storing_ custom indices. 02:21
The {} postcircumfix should intercept the given values, and store them in an array. 02:22
ash_ a translation table basically right? 02:22
how to translate @a{1} into @a[0] for instance
lue Yeah. You intercept, interpret, and send.
yes. Exactly. Or @rray{January} into @rray[0]. 02:23
ash_ yeah
are you just calling Int or something on the params of { } ?
lue I'll post some basic code of storing custom indices when I have it all worked out. 02:24
(on paste.lisp.org of course.)
ash_ hmm or does the param of { } need to have 'succ' implemented?
colomon ng: my ($a, $b) = 1, 2; say $a; say $b 02:26
p6eval ng ae60f8: OUTPUT«Mu()␤Mu()␤»
ash_ ah, i see i just read over the section on user defined indices and it talks about what you need to do when you encounter an index not specified 02:27
lue if you get it implemented, why not write some tests for pugscode? :-D 02:28
lue I've got just the how of storing it. Nowhere near usable yet :) 02:29
ash_ are you augmenting Array? 02:31
lisppaste3 lue pasted "basic howto store indices" at paste.lisp.org/display/94551
lue heh, that's all I coded so far :\ 02:32
pugs: @rray{1..3}=1,2,3;say @rray; 02:33
p6eval pugs: OUTPUT«*** ␤ Unexpected "{"␤ expecting "::"␤ Variable "@rray" requires predeclaration or explicit package name␤ at /tmp/WA30Grp57C line 1, column 6␤»
lue I'll find a place to put this stuff and start putting it in properly! 02:34
colomon ng: say 0i.Bool 02:37
p6eval ng ae60f8: OUTPUT«0␤»
colomon ng: say 1i.Bool
p6eval ng ae60f8: OUTPUT«1␤»
ash_ lue: i'd do it something like: gist.github.com/297833 but thats not complete right now
lue yeah, I was thinking of using the index of where the custom value is, but a hash makes more sense. 02:41
ash_ i am not sure how you would do my @a{2, 3} = 1, 2 though. I know how to override { } but making the array only accept 2 values and return them at the given keys seems a bit beyond me, like how would i know to map 2 to 1 and 3 to 2, that makes me think the { } needs to be aware of the other side of the = 02:42
i was thinking of a hash that takes the param of {} and gets a key returned from the hash to input into the array 02:43
so really that could be as simple as, return self.[%!key_map{*@slice}]; i think eventually 02:44
is there a way to tell if your doing $a = $foo{'a'} vs $foo{'a'} = 5 ? 02:45
lue store custom index elements as keys, actual index elements as values, 02:46
use value for array I/O.
ash_ ruby had [] and []= for telling if your doing an access vs an assignment, i wonder how you tell the difference in perl 6 02:49
lue maybe ternary operators for assignment? 02:50
hm, interesting question
lue thinks this place is deserted because of the Super Bowl, which should be over about now (or soon) 02:51
ash_ that and a lot of developers are in europe so its extra late (or early depending on your perspective) plus its the weekened
colomon naw, probably most of the European contingent is asleep right now.
lue tomorrow's a weekday, and I can't be on until *gasp* 15:30 PST. 02:53
ash_ colomon do you have any idea if there is a way to tell?
i am sure he's busy with his kids, no worries if you don't know 02:54
colomon ash_: you mean if there's a way to tell if you are an lvalue or an rvalue? 02:56
ash_ if those are the right terms for $a = $foo{'a'} vs $foo{'a'} = 5 then yes 02:57
colomon I asked pmichaud a similar question the other day, and didn't understand the answer. :(
lue Would that tell you the difference between assignment and access? Seems like it would. 02:58
ash_ perlcabal.org/syn/S06.html#Lvalue_subroutines seems sorta to answer my question but i can't tell if multi dispatch will tell the difference between is rw and not is rw
only way to know is to try it out i suppose 02:59
colomon That seems like a sensible approach. 03:00
ash_ i think perlcabal.org/syn/S06.html#Subroutine_traits 'is rw' has the wrong test associated with it. that test is using a param 'is rw' not a sub 'is rw' 03:01
lue they must have done it for standard indices. 03:03
Just use what they used for [] :)
colomon ng: my @a = 1..5; sub foo() is rw { @a[0] }; say @a[0]; foo() = 2332; say @a[0]; 03:03
p6eval ng ae60f8: OUTPUT«No applicable candidates found to dispatch to for '&trait_mod:<is>'␤current instr.: '' pc -1 ((unknown file):-1)␤»
ash_ yeah, it seems is rw is not implemented currently 03:04
colomon You want to look in Array.pir for how postcircumfix:<[ ]> is currently implemented on Array.
ash_ i think ill ask pmichaud about is rw next time i see him, just to see if it does make a difference in the multi dispatcher or if there is a way inside the function to see if its being assigned to 03:05
lue yeah, {} could, at its simplest, be a sort of middleman, letting [] methods do all the legwork.
ash_ thats what i'd do 03:06
make [] do the work of finding the real value
lue I'm thinking two ways: 03:12
1. Create a middleman and do tricky on the left or on the right stuff. 03:13
2. Add to the pir code the {} operator.
I understand about 80% of Array.pir 03:14
ash_ well, you could do have { } return a reference to the array it self and override the assignment = for array's to check if you have defined your set of keys
lue You could have {} look in a hash, find the appropriate index, then call the [] operator, all through pir code. 03:17
Ah, the possibilities...
lue thinks of TIMTOWTDI
colomon looking at postcircumfix:<[ ]>, it looks to me like the fail path is handling setting up an lvalue using the Proxy class. 03:18
ash_ colomon: where are you looking at? just curious? 03:19
colomon Array.pir and Proxy.pir in src/builtins in ng 03:20
colomon ng: my @a = 1..3; @[-1] = 10; say @a; 03:20
p6eval ng ae60f8: OUTPUT«sh: ./perl6: No such file or directory␤»
lue Maybe the {} pir code could be simply a copy'n'paste of the [] pir code, with appropriate modifications.
eh, middleman still seems easiest to me, especially when implemented as pir code. 03:21
ash_ i think it would be better if {} used [] to cut down on repetition
colomon I don't see why it particularly should be pir code, either. 03:22
lue Just spewing ideas... where would the hash be stored, for middleman stuff?
colomon lue: I'm not sure. I've got to say, my instinct would be that you wouldn't do these things in the Array class at all. 03:25
It's not clear to me what the spec intends there.
lue questions, questions... 03:26
colomon But essentially what you are doing there is adding the Associative role to an Array type.
colomon I don't have a feel whether it makes more sense to add the Associative role to the particular variable you're looking at, or to create a new Array class which also does Associative. 03:28
(if you're wondering, @a means "does Positional", not "is Array".) 03:29
lue look in src/core/operators.pm. There is the {} circumfix, and it seems to create a hash... 03:33
colomon all right, I've got to get to bed here. happy hacking...
lue ng: %h=<1 2 3>;say %h;
p6eval ng ae60f8: OUTPUT«Symbol '%h' not predeclared in <anonymous>␤current instr.: 'perl6;PCT;HLLCompiler;panic' pc 137 (src/PCT/HLLCompiler.pir:101)␤»
ash_ ttyl
lue ng: my %h=<1 2 3>;say %h;
p6eval ng ae60f8: OUTPUT«Odd number of elements found where hash expected␤current instr.: '&die' pc 15000 (src/builtins/Junction.pir:165)␤»
ash_ ng: my %h = :K<4>; 03:34
p6eval ng ae60f8: ( no output )
ash_ ng: my %h = :K<4>; say %h;
p6eval ng ae60f8: ( no output )
lue I'll mull over how to do it. I have to go get things done.
ash_ ng: my %h = :K(4); say %h;
p6eval ng ae60f8: ( no output )
ash_ hmm
ng: my %h = k => 5; say %h; 03:35
p6eval ng ae60f8: ( no output )
ash_ ummm i must be doing something wrong
ng: my %h; %h<k> = 5; say %h<k>;
p6eval ng ae60f8: OUTPUT«5␤»
ash_ ng: my %h; %h<k> = 5; say %h<k> ~ ' is the same as' ~ %h{'k'}; 03:36
p6eval ng ae60f8: OUTPUT«5 is the same as5␤» 03:37
lue I'm back and ready to continue pondering. 04:11
rakudo: my @rray=1..5; say @rray[2..*]; 04:12
p6eval rakudo 1d4928: OUTPUT«345Use of uninitialized value␤␤»
lue pugs: my @rray=1..5; say @rray[2..*];
p6eval pugs: OUTPUT«345␤»
lue ng: my $a=4; my $b=1; $a=(5 if $b==1); say $a; 04:20
p6eval ng ae60f8: OUTPUT«sh: ./perl6: No such file or directory␤»
lue awww!
ng: my $a=4; my $b=1; $a=(5 if $b==1); say $a;
p6eval ng ae60f8: OUTPUT«sh: ./perl6: No such file or directory␤»
lue :/
ng: my $a=4; my $b=1; $a=(5 if $b==1); say $a; 04:24
p6eval ng ae60f8: OUTPUT«5␤»
lue ng: my $a=4; my $b=2; $a=(5 if $b==1); say $a;
p6eval ng ae60f8: OUTPUT«␤»
lue ng: my $a=4; my $b=2; $a=($b==1 ?? 5 !! 4); say $a; 04:25
p6eval ng ae60f8: OUTPUT«4␤»
lue ng: my $a=4; my $b=1; $a=($b==1 ?? 5 !! 4); say $a;
p6eval ng ae60f8: OUTPUT«5␤»
lue hello, anyone there? 04:25
lue The Hexclock time is D_9E in roughly the PST area 04:26
lue is broadcasting a welcome signal. If you see this, please respond. 04:34
rakudo: say [1,2,3].WHAT 04:35
p6eval rakudo 1d4928: OUTPUT«Array()␤»
lue rakudo: say {1 2 3}.WHAT
p6eval rakudo 1d4928: OUTPUT«Confused at line 10, near "2 3}.WHAT"␤in Main (file <unknown>, line <unknown>)␤»
lue rakudo: say {1 2 3 4}.WHAT
p6eval rakudo 1d4928: OUTPUT«Confused at line 10, near "2 3 4}.WHA"␤in Main (file <unknown>, line <unknown>)␤» 04:36
lue hello. 04:37
ng:say 5.FETCH 04:48
ng: say 5.FETCH
p6eval ng ae60f8: OUTPUT«sh: ./perl6: No such file or directory␤»
lue AA
rakudo: say 5.FETCH
p6eval rakudo 1d4928: OUTPUT«Method 'FETCH' not found for invocant of class 'Int'␤in Main (file src/gen_setting.pm, line 324)␤»
lue rakudo: my Parcel $a=5; say a.FETCH 04:49
p6eval rakudo 1d4928: OUTPUT«Malformed declaration at line 10, near "Parcel $a="␤in Main (file <unknown>, line <unknown>)␤»
ash_ hi lue 05:00
lue there you are! 05:08
I'm looking through STD.pm for a potential clue to custom indices.
ash_ i kinda agree with colomon, custom indices seem like hashes basically IMO 05:11
lue Yeah, just that which side of the equation deal...
lue ng: say 'hi'; 05:17
p6eval ng ae60f8: OUTPUT«hi␤»
lue yes, done compiling!
lue rakudo: $a=5; $b=\$a; $b=3; say $a 05:18
p6eval rakudo 1d4928: OUTPUT«Symbol '$a' not predeclared in <anonymous> (/tmp/DtErD9Iioy:10)␤in Main (file <unknown>, line <unknown>)␤»
lue rakudo: my $a=5; my $b=\$a; $b=3; say $a
p6eval rakudo 1d4928: OUTPUT«5␤»
lue rakudo: my $a=5; my $b=\$a; $b.WHERE=3; say $a 05:19
p6eval rakudo 1d4928: OUTPUT«5␤»
lue rakudo: my $a=5; my $b=\$a; say $b.WHAT
p6eval rakudo 1d4928: OUTPUT«Int()␤»
lue rakudo: my $a=5; my $b=\$a; say $b.VALUE
p6eval rakudo 1d4928: OUTPUT«Method 'VALUE' not found for invocant of class 'Int'␤in Main (file src/gen_setting.pm, line 324)␤»
lue rakudo: my $a=5; my $b=\$a; say $b.WHERE
p6eval rakudo 1d4928: OUTPUT«47501651931336␤»
lue hm... I'm thinking of using references to avoid guessing which side of the equation, but... 05:20
ash_ ng: my $a = 5; my $b := $a; say $b; $b = 10; say $a;
p6eval ng ae60f8: OUTPUT«sh: ./perl6: No such file or directory␤»
ash_ ng: my $a = 5; my $b := $a; say $b; $b = 10; say $a;
p6eval ng ae60f8: OUTPUT«sh: ./perl6: No such file or directory␤»
lue rakudo: my $a = 5; my $b := $a; say $b; $b = 10; say $a; 05:21
p6eval rakudo 1d4928: OUTPUT«5␤10␤»
lue hooray! This keeps us from having to worry about which side of the equation!
ash_ hmmm 05:22
how so?
lue if we want to access, then the reference will automagically give us the number. If we want to assign, then the reference takes care of that!
ash_ so i assume in my @a{2, 3} = 1, 2; you'd return 2 captures 05:25
and they'd update indices 0 and 1 to 1, 2
yeah, i guess that should work fine
lue ah, now the initialization code and retrieval code just needs to be written and put in the right place. 05:26
gah, where to put it, that's the one thing! 05:28
ash_ i'd do it in the src/core/Array.pm 05:31
its really late here, so ima head to bed 05:32
cya tomorrow
lue It's 9:30 PM here. 'Night. 05:33
...to you, I'm going to work on the indices
ash_ good luck
lue rakudo: my %ash; say %ash; 05:47
p6eval rakudo 1d4928: OUTPUT«␤»
lue rakudo: my %ash; say %ash~~"";
p6eval rakudo 1d4928: OUTPUT«1␤»
lue rakudo: my %ash; say %ash~~undefined; 05:49
p6eval rakudo 1d4928: OUTPUT«Could not find non-existent sub undefined␤in Main (file src/gen_setting.pm, line 324)␤»
lue rakudo: say @(3); 05:53
p6eval rakudo 1d4928: OUTPUT«3␤»
lue pugs: @rray{1..3}=0,1,2; say @rray; 05:57
p6eval pugs: OUTPUT«*** ␤ Unexpected "{"␤ expecting "::"␤ Variable "@rray" requires predeclaration or explicit package name␤ at /tmp/vr5tR7SWgO line 1, column 6␤»
lue std: @rray{1..3}=0,1,2; say @rray;
p6eval std 29660: OUTPUT«Potential difficulties:␤ Variable @rray is not predeclared at /tmp/WHkDoXH6Mb line 1:␤------> @rray⏏{1..3}=0,1,2; say @rray;␤ Variable @rray is not predeclared at /tmp/WHkDoXH6Mb line 1:␤------> @rray{1..3}=0,1,2; say @rray⏏;␤ok 00:01 109m␤»
lue pugs: my @rray{1..3}=0,1,2; say @rray;
p6eval pugs: OUTPUT«*** Cannot cast into Hash: VRef <Array:0x2b881a87fe61>␤ at /tmp/Zkm2VN4AaI line 1, column 1-21␤»
lue std: my @rray{1..3}=0,1,2; say @rray;
p6eval std 29660: OUTPUT«ok 00:01 107m␤»
lue rakudo: my @rray=1,2,3; say @rray.max; 06:10
p6eval rakudo 1d4928: OUTPUT«3␤»
lue rakudo: my @rray=1,2,3; say @rray.end;
p6eval rakudo 1d4928: OUTPUT«2␤»
lue rakudo: my %ash=("hi"=>3);say %ash; %ash{"oi"}=4; say %ash; 06:14
p6eval rakudo 1d4928: ( no output )
lue rakudo: my %ash=("hi"=>3);say %ash;
p6eval rakudo 1d4928: ( no output )
lue ng: my %ash=("hi"=>3);say %ash;
p6eval ng ae60f8: ( no output )
lue pugs: my %ash=("hi"=>3);say %ash;
p6eval pugs: OUTPUT«hi 3␤»
lue pugs: my %ash=("hi"=>3);say %ash; %ash{"oi"}=4; say %ash;
p6eval pugs: OUTPUT«hi 3␤hi 3oi 4␤» 06:15
lue pugs: my %ash;say %ash; %ash{"oi"}=4; say %ash;
p6eval pugs: OUTPUT«␤oi 4␤»
lue pugs: my %ash;say %ash; %ash{"oi"}=4; say %ash{"oi"}; 06:17
p6eval pugs: OUTPUT«␤4␤»
lue pugs: say <1 2 3>.end 06:18
p6eval pugs: OUTPUT«2␤»
lue pugs: say <1>.end
p6eval pugs: OUTPUT«0␤»
lue rakudo: my @rray=1 2 3 4; say @rray; @rray[0..2]=5 6 7; say @rray; 06:20
p6eval rakudo 1d4928: OUTPUT«Confused at line 10, near "2 3 4; say"␤in Main (file <unknown>, line <unknown>)␤»
lue pugs: my @rray=1 2 3 4; say @rray; @rray[0..2]=5 6 7; say @rray;
p6eval pugs: OUTPUT«*** ␤ Unexpected "2"␤ expecting operator or ","␤ at /tmp/QWUbAVjbtR line 1, column 12␤»
lue rakudo: my @rray=1,2,3,4; say @rray; @rray[0..2]=5,6,7; say @rray;
p6eval rakudo 1d4928: OUTPUT«1234␤5674␤»
lue pugs: say <1>.end ~~ 1 06:28
p6eval pugs: OUTPUT«␤»
lue pugs: say <1>.end ~~ 0
p6eval pugs: OUTPUT«1␤»
Su-Shee good morning 07:41
mathw Morning 09:11
pugs_svn r29661 | mberends++ | [MiniPerl6] add a perlito launcher script and some docs 09:47
jnthn_ ohhai 10:14
JimmyZ hello 10:20
colomon o/ 10:22
colomon So, by my count ng had 19 commits yesterday, by five different people. 10:45
jnthn Nice :-) 10:48
colomon now if we could do that every day for the next three months.... :) 10:53
moritz_ good morning 11:05
jnthn hi moritz_ 11:08
colomon \o
colomon jnthn: turns out our lazy gather / take has that old bug where saying take $i actually stashes a reference to $i rather than its value. 11:12
dalek ok: 41d351c | moritz++ | src/ (3 files):
There is no undef anymore
jnthn colomon: Aww. 11:13
colomon ng: my @rray=(gather loop (my $calar = 0; $calar <= 6; $calar++){take $calar;}); say @r 11:14
p6eval ng ae60f8: OUTPUT«Symbol '@r' not predeclared in <anonymous>␤current instr.: 'perl6;PCT;HLLCompiler;panic' pc 137 (src/PCT/HLLCompiler.pir:101)␤»
colomon ng: my @rray=(gather loop (my $calar = 0; $calar <= 6; $calar++){take $calar;}); say @rray
p6eval ng ae60f8: OUTPUT«7777777␤»
colomon lue++ # for the example 11:15
colomon rakudo: say 1, 3, 5 ... 9 11:33
p6eval rakudo 1d4928: OUTPUT«Multiple Dispatch: No suitable candidate found for 'cmp', with signature 'PP->I'␤in Main (file <unknown>, line <unknown>)␤»
colomon jnthn: any notion why moving infix:<...> from Range.pm to operators.pm would break it? 11:51
mathw morning 11:52
colomon o/
moritz_ colomon: perhaps different order of declared classes?
colomon: how does it break?
colomon everything compiles, but you get Could not find non-existent sub &infix:<...> 11:53
when you try to use it.
moritz_ in master there was some magic which would export routines in the setting to the CORE namespace (or whatever it was caused) 11:54
colomon oh hey, good guess, other ops in operators.pm are declared "our" 11:55
I'll try that 11:56
that did it! 11:57
moritz_++
moritz_ "educated guessing" :-) 11:59
colomon it does make me wonder why there was a difference.... 12:01
jnthn colomon: Sorry, phone-meeting...anyways, subs are lexical by default in Perl 6, so need exporting or marking "our". 12:20
colomon jnthn: so, why did it work in Range.pm without the our?
moritz_ so rakudo doesn't consider the setting yet as an outer scope of the program? 12:21
jnthn moritz_: Not yet. 12:22
colomon: I'm not sure how it worked without the "our". :-/
dalek kudo/ng: 157750d | (Solomon Foster)++ | src/core/ (2 files):
Move infix:<...> to operators.pm.
 12:30
kudo/ng: 2312138 | (Solomon Foster)++ | src/core/operators.pm:
Replace our primitive infix:<...> implementation with a more sophisticated one from master.
masak o/ 12:56
moritz_ \o 12:58
masak I just blogged a slightly acerbic post: use.perl.org/~masak/journal/40167 13:03
I hope I managed to walk the fine line of making fun of the present situation and not misrepresenting anyone or anything.
jnthn reads nervously 13:04
masak if not, here's your chance to help me fix it early. :) 13:05
takadonet morning all 13:06
masak takadonet: \o
jnthn *sigh* Patches welcome, or I'll add it to my queue of 10000000 other things we also need to do. 13:07
masak jnthn++ # my hero
jnthn masak: I think I worked out why Parrot started getting annotations so wrong, btw. 13:08
masak that's good news.
jnthn masak: The bad news is that I'm a tad lost on how to fix it.
masak ok.
jnthn masak: Essentially, though, we end up not having the program counter to hand any more when there's a non-user-generated exception. 13:09
masak aha.
masak tries hard not to sound like the Q guy in the FAQ 13:10
jnthn Which is...unhelpful. Of course, keeping track of it does slow things down a bit... :-/ 13:10
jnthn Anyway, that's as far as I dug into it. 13:10
frettled masak++ - nice blog post, again.
masak frettled: thanks. :) 13:11
it was fun to write.
frettled I can imagine. It was also quite informative and enjoyable. 13:12
masak \o/
moritz_ masak++ 13:18
masak clearly, this format is a winner. :) 13:19
colomon :) 13:29
colomon So, if we say May ... *, should we detect that Dec is the end of the series by checking to see if $x eq $x.succ ? 13:33
masak colomon: why would Dec be its own successor? 13:35
colomon because it's the last one in the (presumed) enum.
masak sure but the .succ of the last element could be any number of things, on the design level. 13:37
it could be Mu.
or it could throw an exception.
I was just wondering if you were reading from the spec and deduced that Dec.succ === Dec
colomon No, intepolating. 13:38
Though in ng right now, True.succ == True, I believe.
jnthn I don't think it's throw
colomon ng: say True.succ
p6eval ng 231213: OUTPUT«1␤»
jnthn but it might fail
colomon I've just been trying to think of how to support:
May..* # May through December
jnthn Well, there is that behavior in Bool, yeah...
colomon: Yeah..hmm. 13:39
masak oh yes, whatever the answer is, it has to congrue with Bool.
jnthn Tssk.
That means range needs to be aware of enum elements. :-)
Let's see how enums fall out in the implementation, anyways.
jnthn The spec improved on those. 13:40
colomon I originally was thinking that we needed another function or two to go with the .succ interface
but if we simply followed the Bool model, it would be pretty easy to detect end of range.
jnthn
.oO( If it returns identity at the end, doesn't that mean it'll succ less? )
colomon: Yes, it does make it easy in that snese. 13:41
*sense
colomon The other thing I'm not sure is if we want to allow *..May
jnthn Hmm
masak I'm actually quite eager to implement those new enums. 13:41
jnthn How to find the start
Yeah
colomon: otoh 13:42
both become easy-ish if you can go from an element back to the enum.
colomon jnthn: ?
jnthn masak: Feel free. :-)
masak thanks. :)
jnthn colomon: If you can get the full set of values...
colomon oooh.
jnthn ...then you can make a range from start to end right off
Without having to care about how it behaves at the end. 13:43
'cus you know where the end is.
:-)
colomon so something like ::T $enum where T ~~ Enum
and then say T.keys?
jnthn Not quite sure it wants to be that.
I'd have to check the spec
Maybe Enum ::T $enum would work... 13:44
oh, wait
No, that won't
Well, maybe it should
Gah
jnthn looks at what S12 akshually says
masak not S09? 13:46
colomon (I'm just trying to figure out an approach, not necessarily the actual code. :)
jnthn masak: Enums are in S12 13:47
Alas, I gotta look at some other stuff for a bit now. 13:48
But I figure that you can (a) know when you have an enum element, not just a plain old value, (b) multi-dispatch on that for the range constructors and (3) get the full set of enum values so you'll be able to know the start and the end. 13:50
jnthn masak: btw, a good place to start on enum is to implement the anonymous enum constructor. 13:50
masak thanks for the tip :)
colomon go masak!
jnthn It shouldn't be too hard at all. And you should be able to do it in Perl 6. ;-) 13:51
masak colomon: sure, once I finish my day job. :)
jnthn: that sounds wonderful.
jnthn goes to spend a while in the less wonderful world of Java web services 13:52
pmichaud good morning, #perl6 14:37
phenny pmichaud: 07 Feb 17:05Z <jnthn> tell pmichaud seems there's some issues with parsing <-> lambdas, I think the xblock's EXPR swallows up < as an infix operator, and then -> { } as a pointy block. Stared at STD a bit, don't immediately notice the solution. LTM? Any thoughts?
pmichaud did someone do the ng1->master merge?
I can't find ng1 on github
moritz_ or maybe ng1 -> ng? 14:38
ng is younger
pmichaud hmmmm
takadonet morning pmichaud
pmichaud I only have a few minutes to try to sync up before leaving (to return back home).... but I can't seem to sync anything because the branches aren't where i expected them to be :-) 14:39
colomon pmichaud: we did ng1 to ng.
I'm still uncomfortable making it master, but making it ng was a no-brainer. 14:40
pmichaud okay. what about list assignment (that was in the ng branch?)
is it still there, or ... ?
colomon we did a merge, not a replace.
jnthn pmichaud: I think still there but not functioning properly. 14:41
colomon as far as I know, though, list assignment is broken.
jnthn pmichaud: Likely just in need of updates since The World Changed.
colomon ng: my ($a, $b); ($a, $b) = 1, 2; say $a;
p6eval ng 231213: OUTPUT«Mu()␤»
pmichaud I'm sure it's still broken, yes
mathw pmichaud: and they rewrote Rakudo in Java
pmichaud because it now goes into Parcel instead of List
jnthn It runs faster now! 14:42
;-)
pmichaud: I got much of Hash in place, and colomon++ has done Many Cool Things too.
pmichaud looks like list assignment got lost. I'll have to look it up from the history.
colomon jnthn++ got file I/O working again. 14:43
pmichaud excellent work on all of the other stuff! jnthn++ colomon++
jnthn mberends++ provoked me into it!
pmichaud sorry I didn't get much time to work on things this weekend, but it happens
jnthn pmichaud: Yes, we had quite a commit fest at the weekend. :-) 14:43
colomon we actually had five different patch authors yesterday. :)
pmichaud anyway, I just wanted to sync things before getting on the plane today, so I can work on things on the plane
jnthn \o/ 14:44
colomon :D
pmichaud: where are you off to?
pmichaud back home
my wife and I decided we needed an "emergency vacation" :-) 14:45
colomon :)
mathw sounds good
pmichaud anything else important I should work on besides list assignment? 14:47
jnthn pmichaud: wiki.github.com/rakudo/rakudo/ng-ma...res-needed 14:48
pmichaud: Slices would be my choice.
pmichaud slices, whatever, and grammars look like good hits
jnthn pmichaud: Also array init, so my @array; @array[0] = 42; works
pmichaud okay, got it 14:49
I'll work on those
okay, time to go. see you all tomorrowish (maybe later tonight from airport)
colomon safe home!
pmichaud bbl 14:51
colomon afk # storytime 15:00
masak have people been replying off-list to jonasbn++ about his request for information? I know moritz_++ and jnthn++ have replied on-list, but also listed are TimToady, mberends, and pmichaud, who haven't replied on-list. 16:14
diakopter g'morn 17:02
takadonet morning 17:07
colomon ng: my @a = gather for 1..3 { my $a = $_; take $a, $a * 2; }; say @a.perl; 17:53
p6eval ng 231213: OUTPUT«too many positional arguments: 2 passed, 1 expected␤current instr.: '&take' pc 15322 (src/builtins/Junction.pir:347)␤»
jnthn OK...need food and then I'll do some Rakudo bits. :-) 18:05
colomon \o/ 18:06
jnthn ah...food turns out to have a dependency on shopping.
takadonet hehe
order pizza?
jnthn takadonet: Nah, I need some other bits from the store anyway, I may as well go now. :-) 18:08
jnthn back soon 18:09
spinclad $cupboard.grep: {when Food}>>.nom 18:19
ohno, sorry: ($cupboard.grep: {when Food})>>.nom # not a whole lot more natural, i know 18:20
if EMPTY { shop } 18:21
here would be a good place to use the new E operator, if my brane would only engage 18:23
dalek kudo/ng: 7e4f20e | (Solomon Foster)++ | src/core/operators.pm:
Add simplest Whatever case to infix:<...>.
 18:36
kudo/ng: 4eb9894 | (Solomon Foster)++ | src/core/operators.pm:
Implement stupid first stab at lazy series operator with code blocks.
kudo/ng: f3e62dc | (Solomon Foster)++ | src/core/Iterator.pm:
Add Iterator.GrabAndSay method -- most definitely NOT in the spec -- in an effort to simplify lazy iterator implementation and debugging.
colomon for anyone looking, .GrabAndSay(5) (for instance) says the first five elements of an iterator. Basically I got sick of writing "last if $_ > 5" and .eager when testing infinite iterators. :) 18:41
colomon Though now that I look at it, I'm wondering if I should have stuck it on Iterable instead of Iterator. 18:44
k23z__ hey check this out 18:50
point checked if interior in polygon in c++ codepad.org/MmuDEHvO
can you do better than that in Perl6 ? :D
diakopter k23z__: what do you mean by 'better' 19:16
ash_ does nqp support multi methods? 19:27
jnthn ash_: no 19:45
ash_ does rakudo rely on parrots multi dispatch? or does it implement its own? 19:46
ash_ in my nq-nqp i have it parsing some basic variable declarations and function declarations, so now i am wondering how dispatching works 19:51
diakopter what's nq-nqp 19:56
diakopter ash_: I mean, in what are you writing it, is it public, etc 19:56
ash_ i am writing it, its not quite nqp
diakopter in what language(s)? 19:57
and is it public?
jnthn ash_: Implements its own.
ash_ i am taking a class on compilers, so i was going to try to apply what i have been learning and implementing part of nqp seemed like it might be worth taking a whack at
ash_ its on github if you wana see it or help i can give you a commit bit, its in C++ because i am using the llvm for actual code generation 19:58
C++ was mainly chosen because the llvm is written in C++, it does have other bindings but the C++ seem like the easiest for me to work with for now 19:59
jnthn: where is that in rakudo? i am curious about how you'd go about implementing multi dispatch 20:00
diakopter: my repo is github.com/ashgti/nq-nqp-rx on github
jnthn ash_: src/pmc/perl6multisub.pmc 20:01
ash_ cool, thanks 20:02
jnthn ng: say min(1,2)
p6eval ng f3e62d: OUTPUT«Could not find non-existent sub &min␤current instr.: '_block14' pc 29 (EVAL_1:0)␤»
jnthn ng: my @x = 1,2,3; say @x.min 20:04
p6eval ng f3e62d: OUTPUT«1␤»
ash_ jnthn: do you have any idea how you could implement the custom array indicies in ng? it seems like an odd feature IMO since it makes the array nearly act like a hash but its in S09 20:12
mberends ash_: I'm interested in nq-nqp-rx, so you might be interested in svn.pugscode.org/pugs/src/perl6/vill in which case we could contribute to each other's projects 20:13
jnthn ash_: Not thought much about them.
ash_ yeah, i have seen vill, i do think it would be cool if we had nqp-rx implemented in llvm and if there was a way to extend it in some way so rakudo could do more work in nqp and less in pir, but that may not be possible/practical 20:14
mberends llvm looks a lot dumber than pir, unfortunately. no hashes, for example. 20:15
both projects coulod do with a library of similar primitive building blocks 20:16
*could
ash_ std::map<std::string *, p6opaque *> is what i was planning on using but i am having doubts about how i'll construct a C++ based p6opaque object, i am wondering if i need to build some infastructure in the llvm to support what I am doing, but i still haven't finished the dragon book, so i'll probably focus on that, my projects really more for me to learn, but i hope it might be useful to someone eventually 20:18
well, the map is for a basic hash object 20:19
mberends the map template seems to be available to the compiler, but not to the generated IR :(
ash_ you can call it from IR if you know how to mangle the name right 20:20
which is the other thing i have been having problems with, C++ name mangling is so annoying. C names don't get mangled, but C++ ones do
mberends that's always the way with those two languages 20:21
it's because the underlying linkers and loaders are not namespace aware 20:22
ash_ for instance, i have been trying to figure out how to make some sort of std lib to add things to like say, but the C++ name gets mangled into _GLOBAL__I__Z3sayi when i compiled my version of say. so thats why i am trying to figure out what kind of infastructure i'd need to build in the llvm to support what i want to do
ash_ i'll probably see how other projects like unladen swallow or rubinius do it before i do much more with my project 20:24
pugs_svn r29662 | jnthn++ | [t/spec] Corrections and an unfudging for a multi test. 20:30
jnthn Another 30. :-) 20:41
mberends \o/ ** 30 20:42
dalek kudo/ng: 4d6f164 | jonathan++ | src/builtins/Seq.pir:
Seq (and thus Array) should do Positional.
 20:43
kudo/ng: 6c7f9ec | jonathan++ | t/spectest.data:
S06-multi/syntax.t now passes again.
takadonet how many test are we passing now in ng branch? 20:45
ash_ ng: my @a = 1, 2; say @a.WHAT;
p6eval ng f3e62d: OUTPUT«Null PMC access in elements()␤current instr.: 'perl6;SeqIter;get' pc 11981 (src/builtins/Positional.pir:16)␤»
ash_ jnthn do you know whats up with that?
jnthn ash_: I've got an idea, yeah.
takadonet: I think we must be around the 3000 mark or so. 20:46
takadonet: The rate at which we're winning them back seems to be picking up though.
ash_: It's on my "things to try and look at this evening", but quite a few things are. :-)
takadonet jnthn: Good to hear. Waiting for you the ng branch to land before doing any perl6 coding
ash_ ah, okay, i was just wondering thats all ^_^
takadonet s =~ /you// 20:47
dalek kudo/ng: 7459139 | jonathan++ | src/builtins/ (3 files):
Avoid some NPMCAs and make sure Positional and Associative roles know their names.
 20:49
mberends aborts a spectest run to git pull again...
mberends need not have, pulls were up to date, (slow dalek)-- 20:53
jnthn ash_: Aww. It's not quite as easy to fix as I was hoping. Or at lesat, my first attempt failed.
ash_ why is it doing that?
jnthn ash_: Well, I had thought it was just hitting the .Str of the class rather than the one that a proto-object has. 20:54
(Well, I still think it is...I'm just not sure why exactly.)
mberends jnthn: do variables need to know their own names too, or just classes and roles? 20:55
jnthn mberends: Variables need not know their names, no. 20:56
mberends: For trait handlers, those know 'em, but they get passed a ContainerDeclarand object, not just the actual container. 20:57
subs/methods etc should know their names too, btw.
mberends jnthn: thought so, thanks. It's like filesystem hard links afaics
jnthn Oh eww 21:03
We hadn't got a proper impl of prefix:<~> for one thing...
jnthn ng: my $x = 0.5; say $x.WHAT 21:10
p6eval ng 745913: OUTPUT«Rat()␤»
ash_ ng doesn't have infix ~ yet?
jnthn I think it does 21:11
ng: say "a" ~ "advark"
ash_ s/infix/prefix/
p6eval ng 745913: OUTPUT«aadvark␤»
jnthn ash_: It has it, just not done right.
ash_ gotcha
jnthn I figure it's meant to be like prefix:<?>, which just calls .Bool, apart from it calls .Str.
ash_: Oh! I think I know what's up. :-/ 21:12
ash_: It sees the proto-object for Array and is like "oh, that's an Array, I should flatten that!" 21:13
mberends ng passes 3334 spectests on linux/amd64 21:16
jnthn 10% to master then... ;-) 21:27
PerlJam anyone familiar with code_swarm? I was thinking about putting a code_swarm configuration file in the Rakudo repo so that others could hack on the visualization for Rakudo. Anyone see any problems with that or have any comments?
mberends correction, ng passes 3037, todos 66 and skips 231 out of 3334
jnthn ng: say <a b c> 21:32
p6eval ng 745913: OUTPUT«Method 'returns' not found for invocant of class 'String'␤current instr.: 'perl6;Perl6;Actions;handle_named_parameter' pc 149464 (src/gen/perl6-grammar.pir:30472)␤»
jnthn ng: <a b c>
p6eval ng 745913: ( no output )
jnthn ng: <a b c>.perl.say
p6eval ng 745913: OUTPUT«("a", "b", "c")␤»
jnthn hmm 21:33
ash_ thats an interesting bug
jnthn Yeah, a tad perplexed on that one.
ash_: Finding a couple of curiosities when stumbling through S06-multi/type-based.t
Tene something is make-ing a literal string instead of a PAST::Val node in Actions.pm
PerlJam (anyone care to see a code_swarm video for rakudo?)
Tene and something else tries to set or get a type on it. 21:34
ash_ i don't know what code_swarm is, or i might?
Tene notice where the exception is thrown from.
jnthn PerlJam: Code swarm? :-)
mberends it would look pretty vis.cs.ucdavis.edu/~ogawa/codeswarm/
jnthn Tene: Oh, it's doing a get on it, yeah.
PerlJam code_swarm is a repo visualization tool
jnthn Tene: Just not sure why it ain't making a decent node. :-)
oh, I see it. 21:35
Tene jnthn: that's certainly one argument against allowing literals in PAST.
PerlJam Here's a run I did for Rakudo earlier today: feather.perl6.nl/~duff/rakudo-swarm.avi
dalek kudo/ng: 2d24344 | jonathan++ | (3 files):
The role-in-PIR glue should wrap things in Perl 6 Methods as it adds them to the role. This means composition works out rather better, and we can make EnumMap and thus Hash do Associative, so they'll beind the %foo style parameters.
jnthn Tene: Hmm...they are allowed though. :-)
I'll fix handle_named_parameters to deal with it. :-)
Tene jnthn: has that ever stopped people from arguing? ;)
Glad I could help you find it, though. 21:36
jnthn PerlJam: That's quite surreal. :-) 21:40
ng: sub foo($a) { }; foo(<a b c>) 21:41
p6eval ng 745913: OUTPUT«Method 'returns' not found for invocant of class 'String'␤current instr.: 'perl6;Perl6;Actions;handle_named_parameter' pc 149464 (src/gen/perl6-grammar.pir:30472)␤»
jnthn It actually things it's looking through a list of parameters there. :-/ 21:42
dalek kudo/ng: c5340f2 | jonathan++ | src/Perl6/Actions.pm:
Make sure handle_named_parameter doesn't blow up on literals in PAST trees.
 22:10
mberends trying to use proto on ng, the first blocker is qqx{}; any volunteers to add it? 22:41
jnthn mberends: You? ;-) 22:43
mberends: In Grammar.pm see
token quote:sym<qq> { 'qq' <![(]> <.ws> <quote_EXPR: ':qq'> }
You can probably do something very similar.
mberends was afraid that would be the answer ;) 22:44
jnthn Then add an action for it that builds a PAST::Op( :name('!qx'), :pasttype('call'), $<quote_EXPR>.ast ) node or something like that.
Then find the !qx from master.
:-)
mberends ok-ish :-) 22:45
jnthn (.sub '!qx' or so, I think)
jnthn mberends: Just yelp if you get stuck. :-) 22:53
mberends just finishing a proto bisect first... some pattern matching problem now... 22:54
rakudo: my $foo="abcde"; $foo ~~ / b (.) d /; say $0 22:56
p6eval rakudo 1d4928: OUTPUT«c␤»
mberends ng: my $foo="abcde"; $foo ~~ / b (.) d /; say $0
p6eval ng c5340f: OUTPUT«Confused at line 1, near "say $0"␤current instr.: 'perl6;HLL;Grammar;panic' pc 500 (src/stage0/HLL-s0.pir:328)␤»
mberends $0 $1 etc are broken :( 22:57
jnthn mberends: More likely just nyri
mberends ok, time to look at qqx{}
ash_ ng: "test" ~~ / t (.*) /; say $/; 22:59
p6eval ng c5340f: OUTPUT«test␤»
ash_ ng: "test" ~~ / t (.*) /; say $/.0;
p6eval ng c5340f: OUTPUT«Confused at line 1, near "say $/.0;"␤current instr.: 'perl6;HLL;Grammar;panic' pc 500 (src/stage0/HLL-s0.pir:328)␤»
ash_ ng: "test" ~~ / t (.*) /; say $/[0];
p6eval ng c5340f: OUTPUT«Can't postcircumfix:<[ ]> foreign objects yet.␤current instr.: '!postcircumfix:<[ ]>' pc 309458 (src/gen/core.pir:40328)␤»
ash_ hmmm no good 23:01
it looks like $/ might have the right values, but they might be accessible
dalek kudo/ng: 8faeb9e | jonathan++ | src/ (2 files):
Pave the way for lexical classes and even anonymous ones with names (er... :-)) to respond correctly to .WHAT; also make type objects of puns of roles report for .WHAT the name of the role.
 23:07
jnthn Well, there's the fix masak++ asked for. :-) 23:09
mberends heh, testing Reini's B::C on the eeePC eated almost all memory and froze the GUI 23:46
jnthn
.oO( you gave me a ram but I eeeted it )
 23:47
mberends gimme moar memry! 23:48
mberends perlcc viv 23:52
/usr/local/bin/perlcc: viv did not compile, which can't happen:
OH YES IT CAN!
jnthn lol! 23:53
Talk about a stress test. :) 23:54
mberends 'twas a good plan, theoretically 23:55
jnthn :-)
mberends if TPF gives Reini $1000, perlcc will compile viv 23:56
diakopter hee
rurban is on magnet 23:57
mberends then please send him my best wishes and appreciation :) 23:58
mberends +1 to that grant application, still writing a comment :) 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!