»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.
00:08 ggoebel left 00:13 ggoebel joined 00:22 thou left
[Coke] phenny: tell ravendii about github.com/petdance/vim-perl 00:28
phenny [Coke]: I'll pass that on when ravendii is around.
00:28 thou joined
benabik [Coke]: I'm pretty sure that's what I'm using, and I see the oddities with $s as well. 00:29
00:30 lestrrat left 00:32 lestrrat joined 00:38 benabik left 00:40 PacoAir left 00:52 Chillance left 00:58 molaf left, thou left 00:59 xeneg joined 01:01 xeneg left
[Coke] b: say 19373 - 19349 # 01/24/2012 - niecza at 99.87% 01:01
p6eval b 1b7dd1: OUTPUT«24␤»
[Coke] ARGH.
01:04 tokuhirom left 01:05 tokuhirom joined 01:09 tokuhirom left 01:11 replore_ joined 01:13 molaf joined
colomon I'm semi-on-the-job tonight, [Coke] 01:23
rjbs- TimToady: /me just sent you an email! 01:26
01:26 rjbs- is now known as rjbs, rjbs left, rjbs joined
colomon rakudo: say "a" ![!eq] "a" 01:32
p6eval rakudo b7cce8: OUTPUT«Bool::True␤»
dalek ast: 9fb6dac | (Solomon Foster)++ | S03-metaops/not.t:
Three tests based on irclog.perlgeek.de/perl6/2012-01-24#i_5045770.
 01:36
[Coke] colomon: sweet. 01:39
sorear good * #perl6 01:40
colomon o/ 01:41
niecza: multi sub infix:<[**]>() { 1; }; say [**] (); say [**] 1, 2, 3; 01:42
p6eval niecza v13-344-gf9b5518: OUTPUT«Potential difficulties:␤ &infix:<[**]> is declared but not used at /tmp/xFO76LU__r line 1:␤------> multi sub infix:<[**]>⏏() { 1; }; say [**] (); say [**] 1, 2, 3␤␤0␤1␤»
colomon niecza: multi sub infix:<[*]>() { 1; }; say [*] (); say [*] 1, 2, 3;
p6eval niecza v13-344-gf9b5518: OUTPUT«Potential difficulties:␤ &infix:<[*]> is declared but not used at /tmp/0GYtNKI_1K line 1:␤------> multi sub infix:<[*]>⏏() { 1; }; say [*] (); say [*] 1, 2, 3;␤␤0␤6␤»
colomon niecza: say [*] 1, 2, 3; multi sub infix:<[*]>() { 1; }; say [*] (); say [*] 1, 2, 3; 01:43
p6eval niecza v13-344-gf9b5518: OUTPUT«Potential difficulties:␤ &infix:<[*]> is declared but not used at /tmp/8DwdhJ2P4O line 1:␤------> say [*] 1, 2, 3; multi sub infix:<[*]>⏏() { 1; }; say [*] (); say [*] 1, 2, 3;␤␤6␤0␤6␤»
colomon niecza: say [*] 1, 2, 3; multi sub prefix:<[*]>() { 1; }; say [*] (); say [*] 1, 2, 3;
p6eval niecza v13-344-gf9b5518: OUTPUT«6␤Unhandled exception: No candidates for dispatch to &prefix:<[*]>; candidates are:␤ ␤ at /tmp/C4IspXeXvK line 1 (mainline @ 4) ␤ at /home/p6eval/niecza/lib/CORE.setting line 3424 (ANON @ 3) ␤ at /home/p6eval/niecza/lib/CORE.setting line 3425 (modul…
colomon niecza: say [*] 1, 2, 3; multi sub prefix:<[*]>() { 1; }; my @a; say [*] @a; say [*] 1, 2, 3; 01:44
p6eval niecza v13-344-gf9b5518: OUTPUT«6␤Unhandled exception: No candidates for dispatch to &prefix:<[*]>; candidates are:␤ ␤ at /tmp/BvsWKc585C line 1 (mainline @ 4) ␤ at /home/p6eval/niecza/lib/CORE.setting line 3424 (ANON @ 3) ␤ at /home/p6eval/niecza/lib/CORE.setting line 3425 (modul…
sorear it's not going to work
[Coke]: Whyfore ARGH? 01:46
rjbs shift @ARGH 01:52
01:59 REPLeffect left
colomon I'm looking at S16-filehandles/chmod.t 02:04
It creates a temp file, then uses chmod to change permissions to 0o000, skips most of the tests because $*EUID is not defined / not true, and then tries to delete the temp file 02:07
but if the file permission was correctly set to 0o000, you can't delete it!
sorear Why not? 02:08
/ What do you mean? 02:09
colomon no write permission on the file
that's what 0o000 means
sorear you don't need write permission on a file to delete it
colomon are you sure?
sorear on UNIX anyway
colomon you do on OS X
sorear you need write permission on the *enclosing directory*
colomon rm frip 02:10
rm: remove write-protected regular empty file `frip'?
sorear rm -f frip
rm stats the file first and prompts you if it's write-protected, unless -f is given 02:11
colomon I'm not saying the file is forever unremovable -- but I can promise you that niecza dies with System.UnauthorizedAccessException: Access to the path "temp_1327456824.112535" is denied. 02:14
on OS X at least, when you try to remove it.
colomon realizes he hasn't tried niecza on Linux in months...
sorear I bet Mono is trying to emulate the Windows semantics here 02:15
colomon seems quite possible 02:16
02:16 Yary joined
Yary Hi all, think I have a small Rakudo issue that's not in RT 02:17
From S06, using "sub MAIN ($directory);" with the semicolon at the end
sorear nobody implements that form yet 02:18
02:18 whiteknight joined
colomon sorear: oh, right, the reason I haven't messed around with niecza on linux is that I had issues getting a decent mono up and running on it. 02:18
abercrombie Hi perl6, a quick question: how do I apply ? to 'something' * 2..** to force minimum match?
Yary OK
sorear You have your stars backward 02:19
'something ** 2..*
'something' ** 2..*
abercrombie yes, my typo, sorry.
sorear I think it's: 'something' **? 2..*
abercrombie ok, I will try it
sorear although this is the reason ** is deprecated
wait, nvm 02:20
abercrombie that's right, thanks. 02:21
02:22 wolfman2000 joined 02:35 thou joined 03:03 whiteknight left 03:24 orafu left, orafu joined 03:25 lutok left 03:36 wolfman2000 left 03:41 Patterner left 03:43 Psyche^ joined, Psyche^ is now known as Patterner
colomon TimToady: ping? 03:45
TimToady yes? 03:48
dalek ecza: 93f71df | sorear++ | src/ (2 files):
Initial pass at importing STD changes

Sigilless \ and | things aren't actually supported yet, and there are still ripple effects of the <initializer> change that need to occur.
colomon TimToady: I'm looking at getting your Set.pm working in niecza, and running it through the existing tests in set.t. 03:49
TimToady: and I'm getting zillions of failures, and I need to work out whether they are code failures, niecza failures, or bad tests.
so I could use a consult. 03:50
sorear TimToady: How bad is it if $a.WHICH eq $b.WHICH occasionally gives false positives?
TimToady: I can make $a.WHICH === $b.WHICH work reliably, but stringification creates problems
by "occasionally" I mean "about 2**-29 chance" 03:51
TimToady well, hopefully we'll have real object keys someday, so it's probably not too bad for a temporary solution till then
the main concern of using the stringification is security; if too many different keys can be forced to hash to the same bucket, you can get a DoS attack 03:53
03:54 kcwu left
TimToady to start with, I'm happy if only the sets of strings work reliably 03:55
colomon TimToady: for instance, given my $s = set <a b foo>; does "is $s<santa>, False" make sense?
TimToady I'm mostly interested in getting the API baked in soonish
yes, it does 03:56
to the extent that a set is treated as a hash, it behaves as a hash returning boolean
colomon the spec says "Hash of True"
TimToady of course a set can only return True
so, I guess it doesn't necessarily make sense 03:57
well, it'll make sense someday when we have defaults
colomon why?
(defaults why?)
TimToady because Sets will default to False
and Bags will default to 0
currently we can only get a default of Any, because it's hardwired 03:58
colomon but you can't add to a Set anyway, so what difference does it make what the default is?
TimToady API
and the guarantee of returning a defined value 03:59
colomon So, should I change postcircumfix:<{ }> with an // False, or should I remove the test?
TimToady we can do // False for now
colomon k 04:00
TimToady it's not like a real Set type isn't going to hardwire that one way or another
there's not even any guarantee that it must be implemented with a hash
colomon Next failure is dies_ok { $s<a> = True } 04:01
TimToady it merely fits into the API of Associative with a defined default value
colomon apparently method postcircumfix:<{ }> ($k) { %!elems{$k} // False } is all you need to modify values in %!elems. 04:02
TimToady and y'know, I don't think we even really need to define a default for every hash or array; it might be sufficient for the default value to be attached to the type 04:03
colomon In this case I'm sure the code is wrong, but don't know whether it's a Set.pm issue or a niecza issue.
TimToady at worst, if you really want different hashes with different values, you have to do a mixin for each
is the proto defined "is rw"? 04:04
colomon no 04:05
TimToady normally %!elems{$k} should decontainerize when it returns so you can't assign to it
colomon oh, that's interesting 04:06
TimToady in the original nom code I stole from masak++, there's a method !STORE that dies
but I think that's only for the set as a shole
colomon I changed it to method postcircumfix:<{ }> (*@k) { %!elems{|@k} // False } 04:07
TimToady *whole
colomon in an effort to handle $set<a b> as a test, and now assignment does fail
but multiple-element access still fails.
TimToady well, it's probably supposed to be a \keys really anyway 04:08
try \@k there maybe
colomon that doesn't work in my copy of niecza (though I haven't upgraded to sorear++'s latest yet)
TimToady \k parses but isn't implemented yet
colomon nope, I get a Unable to resolve method Capture in class Str when I try to use it 04:09
I guess I should be brave and upgrade, I'm not going to manage to finish this before bed anyway. :)
04:10 raiph left
TimToady probably wants something more like map -> $k { %!elems{$k} // False }, @k 04:10
or maybe just map -> $k { so %!elems{$k} }, @k 04:12
ultimately, subscripts want to have something more like ** semantics, but we don't lol much yet 04:13
colomon TimToady++ # map was a noticeable improvement 04:14
TimToady I think just putting each lookup in boolean context should work 04:15
colomon now the only remaining failures I get are the ones that assume type is preserved in the set
which is expected as long as hash keys are always strings, right? 04:16
TimToady which we might possibly work around by storing the real value as part of the value, as long as we can guarantee it always returns true
(in boolen context) 04:17
04:17 DarthGandalf left
TimToady but I'd rather have real === hashes 04:17
hmm 04:18
perl6: say so True => False
p6eval rakudo b7cce8, niecza v13-345-g93f71df: OUTPUT«Bool::True␤»
..pugs b927740: OUTPUT«*** No such subroutine: "&so"␤ at /tmp/bOd4oC7lPZ line 1, column 5 - line 2, column 1␤»
TimToady so if instead of storing True in the value we store True => $object, we could get $object back at need 04:19
it's a hack, of course
sorear TimToady: What I'm trying to ask is, "is WHICH stringification required to be reliable"? 04:21
TimToady: of course === shouldn't use Str
04:23 replore_ left
TimToady is your question really whether WHICH should form the basis of a reliable hash function? 04:24
dalek ast: a2e9f04 | (Solomon Foster)++ | S02-types/set.t:
Fudge for niecza.
TimToady it can be argued that turning an object's WHICH into a string merely to hash it is wasted energy
so maybe we need a HASH that can override the default that hashes WHICH.Str 04:25
so a compound object could merely, say hash together the hashes returned by each of its compotnents
*pone 04:26
I think that WHICH itself must be unique when compared with === or eqv, but the stringification could be (a little) lossy 04:27
a few bits leakage here or there is not a problem for DoS attacks, but if there are a large set of values that can map to the same hash function, you can be attacked 04:29
sorear TimToady: I repeat, I am not planning to hash the stringificaiton in any official code 04:30
TimToady for example, if the WHICH of a 64-bit floater only represented 32-bits worth, you can load 4 billion entries or so into the same hash slot
sorear this is only for the sake of users silly enough to write $a.WHICH eq $b.WHICH
TimToady I'm fine with probabalistic for that
sorear ok
TimToady after all, the whole of git runs on assuming no collisions in the hash space, though it's a pretty big hash space, as these things go... 04:32
04:33 xinming left 04:39 birdwindupbird joined
dalek ecza: 314664d | (Solomon Foster)++ | lib/CORE.setting:
Import version of TimToady++'s Set class. Alas, the operators are all commented out for the moment.
 04:42
ecza: c347cd0 | (Solomon Foster)++ | t/spectest.data:
Turn on S02-types/set.t.
colomon TimToady: thanks for your help 04:46
sorear: you might want to take a look at t/spec/S03-metaops/hyper.t, it's failing after your last patch. (Or maybe because of set, but I cannot imagine how.) 04:47
'night, all.
TimToady o/ 04:48
04:59 thou left 05:01 DarthGandalf joined 05:10 thou joined 05:21 Radvendii joined
Radvendii nom: say <a b>.grep('a' !~~ /$_/) 05:22
p6eval nom b7cce8: OUTPUT«␤»
Radvendii ??
nom: <a b>.grep(say $_) 05:23
p6eval nom b7cce8: OUTPUT«Any()␤»
Radvendii nom: <a b>.grep(say *)
p6eval nom b7cce8: OUTPUT«*␤»
Radvendii nom: <a b>.grep(7say) 05:24
p6eval nom b7cce8: OUTPUT«===SORRY!===␤Confused at line 1, near "<a b>.grep"␤»
Radvendii nom: <a b>.grep(&say)
p6eval nom b7cce8: ( no output )
Radvendii nom: <a b>.grep: {say $_}
p6eval nom b7cce8: ( no output )
sorear Radvendii: you need to delay the evaluation somehow
.grep('a' !~~ /$_/) evaluates the expression before calling grep
Radvendii which evaluation?
why would it do that? 05:25
doesn't it need to know what $_ is?
TimToady it needs to know *when* $_ is
sorear because Perl 6 uses something called "applicative order evaluation", whereby the arguments to functions are evalled before the function is called
as opposed to Haskell, which is pretty much the only "mainstream" language that uses the alternative, "normal order evaluation" 05:26
TimToady I'm a bit surprised that your {say $_} didn't work
Radvendii then how does .map: {} work?
sorear {anything} evaluates to itself
Radvendii ah
TimToady nom: eager <a b>.grep: {say $_} 05:27
p6eval nom b7cce8: OUTPUT«a␤b␤»
TimToady that's your problem
Radvendii it's lazy evaluating?
hwo does that even make sense?
TimToady it's not doing eager evalation by default
nom: eager <a b>.grep: *.say
p6eval nom b7cce8: OUTPUT«a␤b␤»
TimToady nom: sink <a b>.grep: *.say 05:28
cxreg SPUG says hi
p6eval nom b7cce8: OUTPUT«a␤b␤»
Radvendii sink?
sorear hello, SPUG
Who is SPUG?
TimToady waves at SPUG
Radvendii and why would it make a difference if it is lazy evaluating
cxreg seattle.pm, particle is here
TimToady if it never has any demand for the list of values, it never calls say
05:29 xinming joined
sorear particle's cool 05:29
TimToady you guys still on Mercer Island?
Radvendii nom: say <a b>.grep: *.say
p6eval nom b7cce8: OUTPUT«a␤b␤a b␤»
Radvendii ah
cxreg TimToady: no, downtown
TimToady lemme guess, Amazon... 05:30
cxreg nope, marchex
but today we're at a bar
TimToady well that's...traditional... :)
cxreg :-)
sorear wonders if there's any chance cxreg knows the Perl programmer friend of eirs who moved to Seattle years ago 05:31
cxreg good chance of it, who?
TimToady has a number of ancient acquaintances in seattle... 05:32
Radvendii nom: say <a b>.grep: {'a' ~~ *}
TimToady plus 2 daughters and a grandson :)
p6eval nom b7cce8: OUTPUT«a b␤»
Radvendii as far as i know, 'a' ~~ 'b' evaluates to False 05:33
nom: say ?('a' ~~ 'b')
p6eval nom b7cce8: OUTPUT«Bool::False␤»
Radvendii yup
so why is it evaluating to true in the grep?
TimToady because * is special to ~~
it means match anything
cxreg TimToady: well, you can be our next speaker then :D
05:33 koban joined
Radvendii oh. GAH. Splat means too much!! 05:33
TimToady I've sworn off travel for this past while
Radvendii nom: say <a b>.grep: {'a' ~~ $_} 05:34
p6eval nom b7cce8: OUTPUT«a b␤»
sorear cxreg: TimToady does not presently live in Seattle
cxreg TimToady: well, come up soon :)
Radvendii ha. why that?
05:34 koban left
cxreg yeah i know 05:34
TimToady did too much last year, and crashed physically, so I'm taking it easy this year
cxreg cool, good plan
TimToady Radvendii: because the left side of ~~ *sets* $_ as a topicalizer 05:35
nom: say <a b>.grep: -> $x { 'a' ~~ $x }
p6eval nom b7cce8: OUTPUT«a␤»
TimToady that's what you're looking for
05:35 thou left
Radvendii grumble grumble grumble 05:35
:P
thanks
06:09 kaleem joined 06:30 odoacre joined 06:31 fsergot2 left
sorear niecza: <a b> <<~>> <c d> 06:35
p6eval niecza v13-347-gc347cd0: OUTPUT«Use of uninitialized value in string context␤ at /home/p6eval/niecza/boot/lib/CORE.setting line 806 (warn @ 2) ␤ at /home/p6eval/niecza/boot/lib/CORE.setting line 170 (Mu.Str @ 9) ␤ at <unknown> line 0 (ExitRunloop @ 0) ␤ at /home/p6eval/niecza/src/ST… 06:36
sorear std: <a b> <<~>> <c d>
p6eval std b808250: OUTPUT«ok 00:01 109m␤»
06:51 wtw joined
dalek ecza: 6b3961c | sorear++ | src/STD.pm6:
Fix Texas hypers with list-assoc operators (how does this fix work? why did it not fail before?)
 06:53
06:54 kcwu joined 07:10 Khisanth left
sorear niecza: constant @foo = gather { take 1; take $_+1 for @foo }; say @foo[^12] 07:11
p6eval niecza v13-348-g6b3961c: OUTPUT«Rebuild in progress␤»
sorear niecza: constant @foo = gather { take 1; take $_+1 for @foo }; say @foo[^12] 07:14
p6eval niecza v13-348-g6b3961c: OUTPUT«===SORRY!===␤␤Variable @foo is not predeclared at /tmp/d8gOyTNv4P line 1:␤------> t @foo = gather { take 1; take $_+1 for ⏏@foo }; say @foo[^12]␤␤Unhandled exception: Check failed␤␤ at /home/p6eval/niecza/boot/lib/COR…
sorear niecza: my ($x) = $x; 07:15
p6eval niecza v13-348-g6b3961c: ( no output )
07:24 Khisanth joined
sorear std: my \term:<∞> = Inf; say ∞ 07:28
p6eval std b808250: OUTPUT«===SORRY!===␤Term definition requires an initializer at /tmp/psLRf4C6xT line 1:␤------> my \term⏏:<∞> = Inf; say ∞␤Unsupported use of bare 'say'; in Perl 6 please use .say if you meant $_, or use an explicit invocant or argument a…
sorear phenny: tell TimToady my \term:<∞> = Inf; still doesn't work because <identifier> won't parse past the colon. What's the best fix here? 07:29
phenny sorear: I'll pass that on when TimToady is around.
dalek ecza: 552a38b | sorear++ | src/ (2 files):
Fix self-referential constants; add my \foo = ... syntax
 07:30
07:32 Radvendii left 07:38 lutok joined
sorear nom: say ((5 but True) === (5 but False)) 07:44
p6eval nom b7cce8: OUTPUT«Bool::True␤»
sorear ^^^ interesting infelecity
nom: my $a = do { role Moo { method x() { True } }; 5 but Moo }; my $b = do { role Moo { method x() { False } }; 5 but Moo }; say $a.x; say $b.x; say $a.WHICH; say $b.WHICH; say $a === $b 07:45
p6eval nom b7cce8: OUTPUT«None of the parametric role variants for 'Moo' matched the arguments supplied.␤Ambiguous dispatch to multi '_block1400'. Ambiguous candidates had signatures:␤:(Mu )␤:(Mu )␤␤ in specialize at src/gen/Metamodel.pm:1818␤ in <anon> at src/gen/Metamodel.pm:2075␤ in <ano…
sorear nom: my $a = do { my role Moo { method x() { True } }; 5 but Moo }; my $b = do { my role Moo { method x() { False } }; 5 but Moo }; say $a.x; say $b.x; say $a.WHICH; say $b.WHICH; say $a === $b
p6eval nom b7cce8: OUTPUT«Bool::True␤Bool::False␤Int+{Moo}|5␤Int+{Moo}|5␤Bool::True␤»
sorear ^^^ longer version which will thwart jnthn's attempt at an easy fix for the short version 07:46
I think what's really up here is that the WHICH return needs to include .HOW.WHICH, and that needs to incude a WHERE descriminant 07:47
std: :(\foo) 07:50
p6eval std b808250: OUTPUT«Can't call method "Str" on unblessed reference at STD.pm line 40751.␤FAILED 00:01 109m␤»
sorear std: sub (\foo) {}
p6eval std b808250: OUTPUT«Can't call method "Str" on unblessed reference at STD.pm line 40751.␤FAILED 00:01 109m␤»
sorear std: sub (:\foo) {} 07:51
p6eval std b808250: OUTPUT«ok 00:01 109m␤»
sorear std: sub (:cow(\foo)) {}
p6eval std b808250: OUTPUT«===SORRY!===␤Malformed parameter at /tmp/0VT0S9Egof line 1:␤------> sub (:cow(⏏\foo)) {}␤ expecting any of:␤ name␤ named_param␤ parameter␤ routine_def␤ signature␤Parse failed␤FAILED 00:01 109m␤»
sorear phenny: tell TimToady sub (:\foo) is accepted but sub (:foo(\bar)) is rejected. Shouldn't param_var accept \bar ? 07:52
phenny sorear: I'll pass that on when TimToady is around.
sorear phenny: tell TimToady STD croaks on sub (\foo) {} with an unblessed reference error
phenny sorear: I'll pass that on when TimToady is around.
08:10 sudokode left 08:11 sudokode joined
moritz nom: push @*INC, $@*INC 08:15
p6eval nom b7cce8: OUTPUT«===SORRY!===␤Unsupported use of $@ variable as eval error; in Perl 6 please use $! at line 1, near "*INC"␤»
sorear phenny: tell TimToady There are a few places in Niecza that use method foo(\$: ...) { } to force 'self' to be bound as a parcel. What should be written now? 08:16
phenny sorear: I'll pass that on when TimToady is around.
dalek ecza: 69a1984 | sorear++ | / (2 files):
Implement \foo and |foo parameter forms, use them in setting where possible
 08:17
moritz o/ 08:18
sorear /o
08:23 mikemol left
sorear hmm, perhaps WHICH should only include class info for immutable types 08:24
08:24 mikemol joined
sorear with a mutable type, you can use 'does' on it 08:25
moritz sorear: fwiw I don't believe that WHICH can or should be a macro
sorear and my thought is that WHICH for an object should never change )except in the case where you mix in a role that overrides WHICH)
moritz so you think that (3 but SomeRole) === 3 should be true? 08:26
sorear no
moritz so all roles override WHICH? 08:27
sorear but isn't a mutator
what I'm saying is that my $a = MyClass.new; say $a.WHICH; $a does MyRole; say $a.WHICH; # shouldn't change 08:28
Int.WHICH can get away with including the type because Int subclasses can't change type
3 does MyRole # forbidden!
moritz "subclasses can't change type"? 08:29
sorear on the other hand, Int.WHICH *has* to include the type, because of the example you just gave
moritz subclassing is a type change, no?
sorear no
subclassing creates a new type
subclassing does not magically cause existing objects to change type
moritz right 08:30
sorear [instances of (subclasses of Int)] cannot (change type)
moritz nom: class A { }; role B { }; my $x = A.new; say $x.WHICH; $x does B; say $x.WHICH
p6eval nom b7cce8: OUTPUT«A|-8491906571934514453␤A+{B}|-8491906571934514453␤»
sorear niecza: my $obj = 5; $obj does role { } 08:31
p6eval niecza v13-349-g552a38b: OUTPUT«Unhandled exception: Cannot use 'does' operator with an immutable object␤ at /home/p6eval/niecza/lib/CORE.setting line 1303 (die @ 3) ␤ at /home/p6eval/niecza/lib/CORE.setting line 1242 (infix:<does> @ 6) ␤ at /tmp/xE_E2vA9xi line 1 (mainline @ 2) ␤ a…
08:43 _twitch joined 08:45 robinsmidsrod left
sorear -> sleep 08:45
08:48 rsimoes1 left 08:49 rsimoes1 joined 08:51 robinsmidsrod joined 09:12 grondilu joined
masak morning, #perl6 09:12
moritz \o masak
masak 6-peeps, only four days remain until the final p6cc deadline! 09:13
09:21 grondilu left
dalek kudo/nom: 1cce955 | moritz++ | src/ (2 files):
typed exception for "Cannot use bind operator with this LHS"
 09:32
kudo/nom: 5cf8452 | moritz++ | src/ (2 files):
typed error for binding to native types
kudo/nom: 27a5dd3 | moritz++ | src/Perl6/Actions.pm:
more binding exception fixes
09:32 daxim joined 09:33 Alias left
dalek ast: a824698 | moritz++ | S32-exceptions/misc.t:
typed binding excpetions
 09:33
09:33 Alias joined 09:35 preflex left 09:37 preflex joined, Bucciarati left, pomysl left 09:38 pomysl joined
jnthn mornin' 09:39
moritz \o 09:40
masak jnthn! \o/ 09:41
moritz in arguments to rules, what slang is used?
ie / <.panic: HERE> /
09:42 Bucciarati joined
moritz that's nqp regexes, fwiw 09:43
masak well, S05 has an example <.gather { take foo($¢,1,2,3) }> 09:45
which leads me to think that it's mainline p6 slang.
if that makes any sense.
jnthn moritz: Re-enters <LANG('MAIN', ...> or so iirc 09:46
moritz ok
masak right.
moritz I was thinking about the best way to throw typed exceptions from the grammar
jnthn token arglist {
<arglist=.LANG('MAIN','arglist')>
}
masak though S05 doesn't even mention the <.foo: |@arglist> syntax; only the <.foo(|@arglist)> syntax. 09:47
jnthn masak: Sure I saw that syntax on a recent pass through S05...
masak then I missed it.
jnthn moritz: $*W is available in the grammar.
moritz replacing <.panic: 'Malformed Regex'> with { $*W.throw($/, ['X', 'Syntax', 'Regex']) } is possible 09:48
but a bit ugly
I could make another panicy wrapper that looks like
<.typed_panic: 'X::Syntax::Regex'>
jnthn moritz: That's be preferable
moritz when then simply redispatches the latter
s/when/which/ 09:49
would that simply be a method in Perl6::Grammar? 09:50
nom: /=/
p6eval nom b7cce8: OUTPUT«===SORRY!===␤Unrecognized regex metacharacter (must be quoted to match literally) at line 1, near "/"␤»
jnthn moritz: Writing a method for it would work, yes
moritz jnthn: and in that method, how do I get a $/ ? 09:51
jnthn moritz: Why do you need $/?
moritz: Or better asked: what info do you really need?
moritz: If you do $/.CURSOR.pos() or so, then you can just do self.pos() in your method 09:52
tadzik 'morning
moritz jnthn: well, I want to pass it to Perl6::World.throw, which calls $/.from, $/.orig and $/.CUROSR.panic
(the latter as a fallback while compiling the setting) 09:53
jnthn moritz: Ah. If you were passing the cursor rather than the match object to Perl6::World.throw it woulda been easier here :) 09:54
moritz: Well, you can always get a $/ by writing it as a token instead of a method, then putting a codeblock inside.
moritz: oh
moritz jnthn: well, then I have to write $/.CURSOR in Actions.pm everywhere
how does the token get the match object? 09:55
something as simple as self.MATCH?
jnthn oh, yeah
You can do that :)
moritz or self.CURSOR.MATCH?
jnthn If you write a method foo() { } in a grammar, self is a Cursor
moritz ok 09:56
jnthn So just self.MATCH()
09:57 dakkar joined 10:18 _twitch left 10:29 Yary left 10:36 _twitch joined 10:37 xinming left 10:41 GlitchMr joined 10:45 _twitch left 10:51 xinming joined
bbkr rakudo: say so \() ~~ :(); # bug or NYI? IMO it should return true 11:07
p6eval rakudo 27a5dd: OUTPUT«Bool::False␤»
11:07 Chillance joined
jnthn NYI 11:10
moritz nom: for (1; 1; 1) { } 11:17
p6eval nom 27a5dd: OUTPUT«===SORRY!===␤Unsupported use of C-style "for (;;)" loop; in Perl 6 please use "loop (;;)" at line 1, near "(1; 1; 1) "␤»
11:45 Alias_ joined 11:46 Alias left 11:49 bbkr2 left 11:55 xinming left 12:06 Grrrr joined 12:13 bkolera joined 12:22 mtk joined, mtk left, mtk joined 12:31 MayDaniel joined
daxim perlmonks.org/index.pl?node_id=949890 12:31
12:32 MayDaniel left 12:35 mj41 left 12:37 bluescreen10 joined 12:40 kaleem left 12:44 benabik joined
masak on the Perl course I gave these last two days, I finished off with a Perl 6 section. 12:49
12:49 skids joined
masak reaction seems to have been a mix of "ok, so I'm definitely gonna learn Perl 6 now" and "hm, this is nothing that I think I need for what I do" 12:49
which is fine.
12:50 mj41 joined
masak got the question "so when will it be, you know, here?" to which I answered "both major compilers make releases each month. each one more stable and featureful than the last." 12:50
to which the reaction was something like "oh!"
then I added "the the speed, memory leanness, and stability we deliver may not be within your required range yet. there's no conduit to CPAN beyond Blizkost." 12:51
to which the reaction was something like "oh :/"
12:54 am0c joined
skids looking forward to trying out Zavolaj when he gets the time, psyched to see work in this area. jnthn++ 12:56
Was wondering if endian-reversed representations were on the TODO list. 12:58
moritz why would you want such a thing? 12:59
for NCI? 13:00
skids moritz: network packet buffer processing 13:02
moritz skids: sounds more like a job for pack/unpack than for reprs
skids I beg to differ. 13:03
moritz how so?
skids No reason an IP header should not be a viable object.
moritz well, for IP headers you'd need a custom, bit-packed repr anyway, no? 13:04
masak right. let's keep separate "values of the attributes" of an IP header object, and "serialization format" of same. 13:05
that's what reprs are for :)
moritz well, not exactly
that's what serialization formats are for 13:06
masak hm, yes.
it's three different things.
moritz of course you *can* create a custom REPR that happens to have the same storage format as the serialization
masak nod
13:07 _twitch joined
skids I kinda cringe when I hear "serialization format" because it sort of implies bounce-buffer semantics, when I could very well want to work on a live register set mmio region. 13:07
flussence wants a struct() or something, that takes a single pack() format and returns an object with .pack(*@values) and .unpack(Buf) methods
skids wants same, but regex-based :-) 13:08
Anyway I'm sure it's all doable, just need to find the tiime to learn enough guts to do it. 13:10
benabik Is it possible to send information from the class definition to the REPR? I.E. class IPPacket is REPR(EndianREPR) { ...; has int16 $.length is endianness(NetworkByteOrder); ... }
jnthn Well, the point of doing NCI with REPR stuff is that it stores the data as it needs to be in native land.
masak a repr would be perfect for something like libxml, where the objects coming in from C already have a given format. so we want to pass those objects around in Perl 6 without converting them, but they should behave like real Perl 6 obejcts.
...jnthn said it much shorter than I did :) 13:11
jnthn masak: No, you don't need a custom REPR for that. You just declare a class XmlThingy is repr('Struct') { }
And then in that class declare the attributes as the elements of the structure
A REPR tends to correspond to a world view on arranging stuff in memory. C structures *in general* therefore are a candidate for being a representation. 13:12
masak jnthn: oh! 13:16
even better.
benabik jnthn: Do REPRs receive any information on the annotations attached to attributes? Could you say `has int32 $!val is NetworkByteOrder` and get it to DTRT?
masak still, reprs are involved somehow... :)
13:16 GlitchMr left
[Coke] sorear: argh because only down by 25 tests yesterday. so close. 13:16
masak benabik: an excellent question.
jnthn masak: Yeah, people just have a tendancy to think "oh, I need a custom REPR for X" way too often :) 13:17
skids benabik: more or less precisely what I was wondering.
jnthn benabik, skids: When you declare a new type, the REPR's type_object_for is called. It receives the meta-object for the type in question. 13:18
That can (though you'd best wait until declaration is done with) be used to get at the meta-attributes, which in turn mean you can get at the attribute type, plus if any traits were mixed in you can get at the values from those too.
So basically, yes, though there's no explicit passing mechanism, it's just that all such information about a type is reachable through its meta-object somehow. :) 13:19
That's how CStruct and P6opaque compute memory layouts.
13:20 _twitch left
[Coke] (4 days left until p6cc deadline). Guess I'm losing! ;) 13:22
masak depends how you spend those 4 days... :) 13:24
benabik jnthn: (poking through src/6model) It looks like P6opaque is lazy about calculating its size, and only does so when it really needs it (via compute_allocation_strategy), right?
jnthn benabik: Correct, though there is a little more to it than that (more)
type_object_for is called too early for you to do the compuation. 13:25
colomon perl6: my $a := 1, 2; say $a
p6eval pugs b927740, niecza v13-350-g69a1984: OUTPUT«1␤»
..rakudo 27a5dd: OUTPUT«1 2␤»
13:26 fdgbch456 joined
masak colomon: "The binding operator parses as a list assignment, so it is reasonable to 13:27
generate a list on the right without parens"
moritz rakudo++
masak S03:2299
fdgbch456 hi
masak fdgbch456: hello!
colomon masak: niecza's handling of that changed sometime in the last two days. 13:28
benabik jnthn: The REPR is initialized (more or less) at the beginning of the class, right? (NQPClassHOW.new_type). Is there a reason it couldn't be delayed until compose()? It seems like it would be very difficult to determine how much memory you need before then.
fdgbch456 i'm following the instructions on rakudo.org/how-to-get-rakudo/ and it says i need to install "the ICU library." So i'm downloading "icu4c-4_8_1_1-src.tgz", is it correct ? 13:29
moritz fdgbch456: what operating system are you using?
fdgbch456: and how do your pronounce your nick name? 13:30
jnthn benabik: We need the type object right away
fdgbch456 i figured i don't need icu4J (?)
i'm using Ubuntu 11.10
moritz fdgbch456: there should be a libicu-dev package that you can install with your package manager
jnthn nom: class A { method m() { say 'oh lol' }; BEGIN { A.m } }
p6eval nom 27a5dd: OUTPUT«oh lol␤»
moritz fdgbch456: which is much easier for you
fdgbch456 lol...oh, its just a random ncik
ok, will try ap-get then... 13:31
jnthn benabik: See ^
moritz fdgbch456: at leat it's called libicu-dev on debian, and usually the package names are the same
benabik jnthn: I don't (immediately) see how that works. Is it composing it mid-class? 13:32
13:33 xinming joined
jnthn benabik: It's not composed yet 13:34
(at the time the method is called)
13:36 timotimo joined
timotimo hello, i'm trying to build rakudo-nom, but at the Configure.pl --gen-parrot step i get a failure with no obvious error message 13:37
masak timotimo: could you nopaste it to gist.github.com? 13:38
benabik jnthn: Oh. An NQPClassHOW class couldn't do that because it uses @!mro which is calculated at compose time... But Perl6::ClassHOW is smarter.
timotimo yes, i will
benabik jnthn: Sorry if I'm bothering you, I just find this stuff fascinating. 13:39
jnthn benabik: Yeah, NQPClassHOW is decidedly simpler. That's deliberate. I wanted classes for implementing Perl 6's classes, but they didn't need to do everything Perl 6 classes do :)
timotimo sprunge.us/VBJH <- there you go 13:41
benabik jnthn: It seems like if you try to create an object before declaring all of it's attributes the REPR would get confused. Also, many (most?) REPRs will need to be lazy about allocation information because they'll want the attribute information. Shouldn't they be notified on compose so they have a chance to do it without all the `if(!alloc) calc_alloc()` ?
moritz timotimo: that's weird 13:42
timotimo: what platform is this?
timotimo i agree :)
debian sid on a 64 bit intel
jnthn benabik: I have considered giving the REPR API a compose hook. 13:43
benabik: Just means every meta-object needs to remember to call it, but maybe that's not such a burden. 13:44
moritz timotimo: do you have any env variables set that could affect the compilation?
(something like concurrency options or so)
jnthn moritz: Do you even see the actual error in that output? I'm missing it if it's there...
benabik jnthn: Depends on if you want every meta-object to remember to call repr_compose or every repr to remember to calculate the allocation any place they may need it. :-D
timotimo apparently i have MAKEFLAGS= j2 j2 j2 j2, interesting
moritz timotimo: please try clearing it 13:45
jnthn: make[1]: *** read jobs pipe: Is a directory. Stop.
jnthn: a quick google tells me that that happens when one launches a sub-make in a parallalized build with 'make' instead of $(MAKE)
jnthn moritz: hm. Note:
cd 3rdparty/dyncall && BUILD_DIR=. make
moritz jnthn: it seems that GNU make wants some kind of control over its child processes
jnthn cd 3rdparty/dyncall && BUILD_DIR=. make
moritz: It does it twice or some such... 13:46
moritz weird
that's why I asked about concurrency options
timotimo the environment was at fault, indeed
thank you. i hopefully also identified and removed the culprit
moritz timotimo: curiously, a simple 'make -j3' works for me in nqp 13:47
timotimo another question: i have never written any perl5 and am new to perl6. is there a tutorial written for people like me? 13:48
moritz timotimo: the "Using Perl 6" book is meant for people who know other programming languages, but not necessarily perl 5 13:49
timotimo i've drooled over a couple of examples from perl6advent, but i'd like a more thorough approach
moritz github.com/perl6/book/downloads 13:50
timotimo sounds good. my prior experience consists mostly of c++ and python
moritz if you find anything unclear in the book (it's under development), feel free to ask 13:51
most of the authors hang out here
timotimo sure, thanks!
benabik jnthn: Oh. A.new doesn't work in BEGIN inside a class (method 'new' not found), so the allocation strategy can't get goofed. Is that deliberate?
moritz well, the .new comes in through inheritance, which is resolved at compose time, iirc 13:52
and compose time is when the class' final } is parsed
benabik moritz: Makes sense. 13:53
jnthn moritz: Yeah, that does offer some protection here. :)
moritz one can still nqp::create :-)
jnthn I'm tempted to do the compose hook anyway at some point 13:54
Well, time to dig back into $dayjob stuffs... 13:55
benabik I'm still somewhat worried about more general usage of 6model... Having a repr compose would let it at least know if what's going on is a bad idea...
Although I guess the HOW should yell about creating an uncomposed class.
jnthn benabik: Yeah, the more I ponder it, the more I think it's worth adding. 13:57
benabik digs into how objects are created.
Yeah, looks like creation bypasses the HOW. So the REPR really has no way to check if the class is composed or not. 14:02
ANYway...
moritz on what level? :-)
jnthn nqp::create is a REPR op, not a HOW-related one, yes. 14:03
benabik should probably spend this time on $work or $thesis, not random 6model digging.
jnthn: BTW, 6model is an excellent piece of work. And thanks for helping me understand it. 14:06
jnthn++
masak timotimo: if you go to strangelyconsistent.org/blog/list-of-posts and search on that page for "June 1 2011", you'll get to the start of a 30-day intro course to Perl 6. that might be something for you. 14:09
benabik It seems like we have more and more people showing up going "how do I get this working" and "how do I use this". I like this trend. 14:12
14:13 dhs227 joined, dhs227 left
PerlJam guten Morgen meine Freunde 14:14
moritz \o PerlJam 14:15
fdgbch456 so i have reached the step "To run a Perl 6 program with Rakudo, include the install directory in your system PATH variable", but there are two "perl6" executables, one in ~/rakudo/ and one in ~/rakudo/install/bin/ 14:27
which one do i include?
flussence the install/ one
moritz it even says "include the *install* directory" :-) 14:28
fdgbch456 the page rakudo.org/how-to-get-rakudo/ says, "This will build and install a “perl6″ or “perl6.exe” executable, by default in the parrot_install/bin/ subdirectory of the rakudo build directory."
moritz oh, that's outdated
fdgbch456 & there is no "parrot_install" directory
PerlJam fdgbch456: you could nuke everything else, as long as you keep that install directory around, you can run rakudo
fdgbch456 moritz: oh 14:29
moritz it's now just named 'install/', not 'parrot_install/'
fdgbch456: rakudo.org/how-to-get-rakudo-nom/ is probably more interesting for you
once we have release a nom-based star, which should remove the old instruction page 14:30
PerlJam watches some really dark clouds move in and hopes that "tornado watch" doesn't turn into a "tornado warning" 14:41
jnthn That'd suck. 14:42
TimToady wonders if OZ has any internetses... 14:43
phenny TimToady: 07:29Z <sorear> tell TimToady my \term:<∞> = Inf; still doesn't work because <identifier> won't parse past the colon. What's the best fix here?
TimToady: 07:52Z <sorear> tell TimToady sub (:\foo) is accepted but sub (:foo(\bar)) is rejected. Shouldn't param_var accept \bar ?
TimToady: 07:52Z <sorear> tell TimToady STD croaks on sub (\foo) {} with an unblessed reference error
TimToady: 08:16Z <sorear> tell TimToady There are a few places in Niecza that use method foo(\$: ...) { } to force 'self' to be bound as a parcel. What should be written now?
TimToady oh, and jnthn++ for the terrible pun 14:44
masak we let jnthn work on Rakudo just so he can provide us with bad puns all the time. ;) 14:46
TimToady gah, I really hate it when ^O takes me to a different file in vim. Is there a way to reverse and/or block that? 14:47
masak sounds like Unity's new behavior of Alt-Tabbing to a different workspace :( 14:49
jasonmay ^I is the reverse
benabik TimToady: Jumplists are per window, so I think you'd have to open a new window for each file. <Tab> should also bring you forward again quickly.
14:50 PacoAir joined 14:52 fdgbch456 left
TimToady thanks 14:53
14:55 birdwindupbird left
dalek d: 655cc68 | larry++ | STD.pm6:
:= and ::= must parse RHS as list assignment prec

The operator forms do that, but the initializer forms falsely generalized from = parsing.		 15:00
TimToady phenny: ask sorear Would (\self:) work? 15:04
phenny TimToady: I'll pass that on when sorear is around.
moritz eeks. 15:06
jnthn hmm...in Rakudo I'd have to take care to suppress auto-generation/binding of self. 15:07
Oh
It'd cause a further problem.
I rely on self being decontainerized.
For attribute lookup
So a naive patch to make the above work would break attribute lookups in any method that did that. 15:08
Which I guess I could work around but ew.
moritz also wouldn't that bind self to a parcel rather than to an object of the method you're currently in?
jnthn moritz: "is parcel" doesn't mean that.
moritz: It means "leave the thing exactly at is without re-containerization" 15:09
*as
Or context.
(The context-ignoring part being why we don't want the sigils on there.)
Hmm, what if you do method foo(\self) { ... } also. Then you get a self which is the wrong thing. 15:13
moritz then you get a redeclaration error?
because there's one declaration for the implicit self, and another for one you just declared 15:14
and if you do method foo($other_self: \self), well, then screw you.
jnthn moritz: Yes, but iiuc TimToady was suggesting we might end up suppressing declaration for implicit self 15:15
moritz: Writing method foo($other_self:) { } does not mean you don't get self too
moritz ah right 15:16
15:19 kcwu left 15:24 xinming left, cognominal_ left 15:25 xinming joined, cognominal joined 15:32 wtw left 15:34 mj41 left, phenny left, krunen_ left, arnsholt left, timotimo left, bkolera left, Alias_ left, lestrrat left, charsbar___ left, felher left, huf left, Yappoko___ left, drbean left, dukeleto left, sirhc left, tokuhirom3 left, Khisanth left, kshannon left, aloha left, rsimoes left, Maddingue left, groky left, [Coke] left, jnthn left, bacek_at_work left, pomysl left, ruoso left, Trashlord left, yeltzooo left, araujo left, Vlavv` left, mls left, c1sung left, LoRe left, silug left, Lorn left 15:37 drbean joined 15:38 kaare_ joined, xinming left 15:39 ranguard joined, krunen joined, timotimo joined, arnsholt joined, kshannon joined 15:40 [Coke] joined 15:42 xinming joined, aloha joined, bacek_at_work joined 15:44 doy joined, Maddingue joined
masak <Radvendii> ǧis al čiuj! 15:47
Radvendii: your funny hats are upside down. 15:48
15:48 TimToady joined
TimToady whew, just discovered how to recover from a bad netsplit with /reconnect 15:49
15:50 phenny joined, jnthn joined
TimToady a /reconnect is a lot easier than restarting irssi... 15:50
rjbs restarting irssi always feels like an admission of failure to me
and I always forget to rejoin at least 1d6 channels
moritz doesn't irssi reconnect automatically? 15:51
rjbs It tries to, but it's imperfect.
jnthn It did for me, but I ended up on a server with, er, 6 of the channel's inhabitants.
A /reconnect got me back to a better connected one :)
TimToady hah, my orphan server had 7 inhabitants, nyaa nyaa nyaa 15:52
15:52 am0c left
TimToady blames solar storms 15:52
15:53 mj41 joined, kcwu joined, bkolera joined, Alias_ joined, lestrrat joined, charsbar___ joined, felher joined, huf joined, Timbus joined 15:54 rsimoes joined, cxreg joined, pmurias joined, Yappoko___ joined, dukeleto joined, sirhc joined, tokuhirom3 joined 15:55 kaleem joined, amkrankruleuen joined, pomysl joined, ruoso joined, Trashlord joined, yeltzooo joined, araujo joined, Vlavv` joined, mls joined, c1sung joined, LoRe joined, silug joined, Lorn joined, hillu joined, awwaiid joined, mathw joined
moritz TimToady: that does sound a bit like an BOfH excuse :-) 15:55
timotimo when i find spelling mistakes etc in the perl6 book, is it ok to fork+edit+pullrequest them? 15:56
moritz timotimo: or just commit them directly; what's your gitub ID?
timotimo of course i could do that, too :p - it's "timo" 15:57
moritz timotimo: you have commit access now. Have fun!
timotimo: note that the PDF might be slightly out of sync with the sources, so some typos might already be corrected
timotimo wow, i hope i won't disappoint you by not working hard enough ;)
yes, i've made sure to look into the code first 15:58
pmichaud good morning, #perl6
phenny pmichaud: 24 Jan 19:32Z <sorear> tell pmichaud At some point I'd like an overview of how LoL, $obj.lol, **@foo, etc work and fit together in Rakudo
moritz good am, pm
colomon pmichaud: good morning!
moritz (I've missed that bad joke)
dalek d: afb775b | larry++ | STD.pm6:
allow bare \ and \term:<your-ad-here> forms
 15:59
15:59 groky joined 16:00 Khisanth joined
masak pmichaud! \o/ 16:00
jnthn o/ pmichaud
TimToady you just missed the great netsplits this morning 16:01
masak \term:<your-ad-here> -- oh man.
masak imagines term:<this spec commit was brought to you by Pepsi>
jnthn hoped that the solar storm might mean he saw the aurora this far south, but I didn't see much... 16:03
TimToady evalbot rebuild std
p6eval OK (started asynchronously)
[Coke] glares at masak.
16:04 groky left
dalek Rebuild of std complete. 16:05
TimToady std: my \term:<∞> = 42; say ∞;
p6eval std afb775b: OUTPUT«ok 00:01 113m␤»
[Coke] why is dalek announcing the rebuild is complete?
TimToady so we can know when to torment it again? 16:06
[Coke] (as opposed to p6eval)
TimToady less noise the failed tries
*than
oh
dalek ok: eb5b12d | timo++ | src/basics.pod:
Removed stale references to " vs " in table tennis data
TimToady std: constant term:<∅> = set(); 16:07
p6eval std afb775b: OUTPUT«ok 00:01 109m␤»
TimToady \o/
std: constant \term:<∅> = set();
p6eval std afb775b: OUTPUT«ok 00:01 109m␤»
16:07 Khisanth is now known as Guest20598
TimToady (the \ is optional in constants) 16:07
masak timotimo++ 16:08
TimToady std: sub (\foo) {} 16:11
p6eval std afb775b: OUTPUT«ok 00:01 109m␤»
skids jnthn: WRT github.com/perl6/specs/pull/10 now that I read S02:1776 and S02:1917 it appears :: is both en-passant type capture and class predeclarator simultaneously. So I should undo that pull request, or have there been decisions in this area not reflected in S02? Or am I totally misunderstanding something? 16:14
dalek ecza: 7bfcbe7 | (Solomon Foster)++ | lib/CORE.setting:
Add Bag modeled after Set.
ecza: 6694950 | (Solomon Foster)++ | t/spectest.data:
Turn on S02-types/bag.t.
timotimo masak: :) 16:15
colomon notes that both set.t and bag.t could use quite a few more tests....
masak skids: no, you're correct.
skids: it's a way to refer to undeclared classes outside of signatures, and en-passant capture inside of them. 16:16
jnthn Yup. Looks like a good clarification.
skids But the S02 fragments say it predeclares the class as well inside a signature... ? 16:17
jnthn :S 16:18
It never pre-declares anything. It just defers lookup until runtime.
Thus why it can't be used in declarative contexts, which need to know stuff at compile time.
16:19 am0c joined
skids Well, OK, my vocabulary is off... S02 says: "Within a signature or other declaration, the :: pseudo-sigil followed by an identifier marks a type variable that also declares the visibility of a package/type name without the sigil within the scope of the declaration" 16:20
So that's what's re-confusing me.
16:21 rsimoes left
masak means you do ::Foo in a signature, and then you can talk about Foo in that scope. 16:22
jnthn It just means...what masak said. Could be clearer though 16:23
16:25 rsimoes joined
pmichaud Rakudo 2012.01 is using Parrot 3.11.0 ? (just checking) 16:25
masak skids: however, Foo is not a class, or even really a type. it's just a type variable. 16:26
16:28 colomon left
lumi___ Does that mean that "my (::T $x) = 5" (or so) means that T is Int for the containing scope? 16:28
jnthn pmichaud: Yeah. The Rakudo compiler release went out with that, more by accident than anything. But next to nothing happened between 3.11.0 and 4.0.0 so it didn't seem worth doing a point release to correct.
PerlJam nom: my (::T $x, T $y) = (1,"foo"); # and should this fail?
p6eval nom 27a5dd: ( no output )
pmichaud okay. One minor difference may be that packagers won't package a 3.11 Parrot release
PerlJam std: my (::T $x, T $y) = (1,"foo");
p6eval std afb775b: OUTPUT«ok 00:01 115m␤»
skids Ah, comprehension dawns. Maybe I'll try to work on that S02 verbiage as well. 16:29
jnthn I suspect it should moan about the signature being too complex to turn that into a parcel of ($x, $y) and do list assignment in that case.
benabik hopes packagers would build rakudo with their packaged parrot, rather than --gen-parrot
pmichaud benabik: oh, that's a good point. 16:30
jnthn pmichaud: Hm, and a release that requires a 3.11 won't build against and installed 4.0?
pmichaud it should work against a 4.0, yes.
jnthn forgets the exact rules...I know a later commit is OK.
OK.
benabik has had no problem with NQP/Rakudo on 4.0.0-devel
pmichaud it just seemed odd to me
masak PerlJam: of course, STD allows that because STD doesn't typecheck. 16:31
jnthn pmichaud: We could fix it, it just means issuing point compiler releases of NQP and Rakudo.
PerlJam yeah, I just wanted to see if there was something std noticed that I didn't
16:31 kaleem_ joined
masak nom: sub foo(::T $a, T $b) { say "OH HAI" }; foo 5, 42 16:31
p6eval nom 27a5dd: OUTPUT«OH HAI␤»
pmichaud jnthn: I'm fine with leaving it as-is for this release, if you are.
masak nom: sub foo(::T $a, T $b) { say "OH HAI" }; foo 5, "wrong"
p6eval nom 27a5dd: OUTPUT«Nominal type check failed for parameter '$b'; expected Int but got Str instead␤ in sub foo at /tmp/4fWEnqFA3Y:1␤ in block <anon> at /tmp/4fWEnqFA3Y:1␤ in <anon> at /tmp/4fWEnqFA3Y:1␤»
jnthn pmichaud: Yeah, I'm fine with it.
16:32 am0c left
masak jnthn: will we ever be able to catch the above at compile-time? :) 16:32
16:32 colomon joined
dalek d: b5dc271 | larry++ | STD.pm6:
new rule to handle what can go inside param :x()
 16:32
[Coke] tries to find more Set & Bag tests to unfudge for colomon.
masak jnthn: with one of those "will never work" errors...
jnthn masak: Yeah, it'd need patches to compile time bind analysis in bind.c. 16:33
TimToady evalbot rebuild std
p6eval OK (started asynchronously)
jnthn masak: It's not particularly tricky to catch the above case.
masak: Oh, wait...it does get a bit tricky as compile time binding analysis doesn't currently have something to hand to serve as the type environment. 16:34
masak: But even that's not hard to solve.
16:34 kaleem left
dalek Rebuild of std complete. 16:34
masak jnthn: cool. 16:35
I think it's kinda cute that p6eval says "OK" and dalek comes back and says "complete".
dalek ast: a45df5f | coke++ | S0 (3 files):
niecza autounfudge
masak makes it look like the bots are one big family :) 16:36
[Coke] colomon: there's 5 more on top. ;)
tadzik it could say LOL DONE
[Coke] er, 8! 16:37
TimToady std: sub foo (:x(:y(\z))) { ... }
p6eval std b5dc271: OUTPUT«ok 00:01 110m␤» 16:38
benabik std: sub foo (:\z) { ... }
p6eval std b5dc271: OUTPUT«ok 00:01 109m␤»
TimToady phenny: tell sorear I think all your concerns are addressed now; bare \ is now allowed for the invocant
phenny TimToady: I'll pass that on when sorear is around.
TimToady std: method (\:) {} 16:39
p6eval std b5dc271: OUTPUT«ok 00:01 109m␤»
TimToady std: sub foo (:x(:y(\term:<※>))) { say ※ } 16:40
p6eval std b5dc271: OUTPUT«ok 00:01 114m␤»
16:43 kaleem_ left 16:48 wamba joined 16:49 wamba left
masak decommutes 17:00
17:08 MayDaniel joined
pmichaud time for lunch here, and I need to think a bit more about star configure/build 17:11
17:12 MayDaniel left
pmichaud bbiaw 17:12
flussence yay! github.com/flussence/specgraphs/bl.../impls.png 17:16
flussence &
17:20 mj41 left
timotimo in the perl6 book, I'm about to change "say @a === @a; # 1" to "say @a === @a; # Bool:True", is that OK? 17:30
uh, Bool::True, of course
moritz timotimo: yes 17:31
dalek ok: 850c094 | timo++ | src/operators.pod:
say @a === @a outputs Bool::True rather than 1.
timotimo oh, there's more
moritz timotimo++ 17:32
timotimo for the example say 3 == '3b'; i get "No applicable candidates found to dispatch to for 'Numeric'." rather than a 1. what to do with that? 17:33
moritz hm, that's a failure by new numifying spec 17:34
(and rakudo doesn't do well in propagating the failure through ==)
timotimo: either add '# fails' or leave a TODO comment 17:35
dalek ok: e8cc91e | timo++ | src/operators.pod:
changed more 0 and 1 into True/False. Note, that 3 == '3b' fails.
 17:36
timotimo why does it say "timo++" in there? 17:37
17:39 jaldhar left
timotimo oh, also in the text it claims, that perl will use 0 if there's no way to turn the string into a numeric value. is that also false as of the new numifying spec? 17:40
Util #ps time, 10m ago 17:41
timotimo: It is using your GitHub ID. 17:42
timotimo but my github id is timo, not timo++
Util The ++ is kept track of by a "karma" bot on this channel. You get a ++ when you do something good. 17:43
timotimo ah, all right
Util Like a few mintes ago, when moritz did it to you.
karma timo
aloha timo has karma of 0.
Util karma timotimo
aloha timotimo has karma of 2.
timotimo yup, i was aware of the karma thing, it just confused me to see it in the bots output 17:44
where would i put feedback like "in chapter 3, the sort example drops new syntax onto the reader out of nowhere without a deeper explanation"?
Util Commits are "something good" :)
timotimo would an issue in the tracker be the right thing?
Util Yes.
and that is an excellent issue to raise 17:45
sorear good * #perl6 17:46
phenny sorear: 15:04Z <TimToady> ask sorear Would (\self:) work?
sorear: 16:38Z <TimToady> tell sorear I think all your concerns are addressed now; bare \ is now allowed for the invocant
moritz it's #phasers time 17:49
sorear [Coke]: (dalek reporting end) Convenience, mostly
17:56 dakkar left 18:04 tarski joined 18:06 Radvendii joined
Radvendii hey, guys? Y'know how the zip operator truncates the smaller list? is there a way to make it instead repeat the smaller list? 18:07
TimToady (list) xx * 18:08
or use the dwimmy hyper 18:09
@a »op» @b # repeats @b
18:10 daxim left
TimToady perl6: say ('a'..'z') »~» (1,2,3) 18:10
p6eval rakudo 27a5dd, niecza v13-352-g6694950: OUTPUT«a1 b2 c3 d1 e2 f3 g1 h2 i3 j1 k2 l3 m1 n2 o3 p1 q2 r3 s1 t2 u3 v1 w2 x3 y1 z2␤»
..pugs b927740: OUTPUT«decodeUTF8': bad data: '\187'␤decodeUTF8': bad data: '\187'␤*** ␤ Unexpected "\187~\187"␤ expecting operator, ":" or ","␤ at /tmp/_Cx0CnH2Ml line 1, column 16␤»
TimToady perl6: say 'a'..'z' X~ (1,2,3) xx * 18:11
p6eval rakudo 27a5dd, niecza v13-352-g6694950: OUTPUT«(timeout)»
..pugs b927740: OUTPUT«abcdefghijklmnopqrstuvwxyz␤»
TimToady perl6: say 'a'..'z' Z~ (1,2,3) xx *
p6eval rakudo 27a5dd, niecza v13-352-g6694950: OUTPUT«a1 b2 c3 d1 e2 f3 g1 h2 i3 j1 k2 l3 m1 n2 o3 p1 q2 r3 s1 t2 u3 v1 w2 x3 y1 z2␤»
..pugs b927740: OUTPUT«abcdefghijklmnopqrstuvwxyz␤»
TimToady perl6: say 'a'...* Z~ (1,2,3) xx * 18:12
p6eval rakudo 27a5dd: OUTPUT«Decrement out of range␤ in method Str at src/gen/CORE.setting:7212␤ in method Stringy at src/gen/CORE.setting:678␤ in sub infix:<eq> at src/gen/CORE.setting:977␤ in method ACCEPTS at src/gen/CORE.setting:3230␤ in sub coro at src/gen/CORE.setting:4618␤ in meth… 18:13
..niecza v13-352-g6694950: OUTPUT«===SORRY!===␤␤"..." and "Z~" are non-associative and require parens at /tmp/3lZ7_j1rh_ line 1:␤------> say 'a'...* Z~ ⏏(1,2,3) xx *␤␤Unhandled exception: Check failed␤␤ at /home/p6eval/niecza/boot/lib/CORE.setting lin…
..pugs b927740: OUTPUT«*** No such subroutine: "&Z"␤ at /tmp/HbJ57Pupn4 line 1, column 5 - line 2, column 1␤»
TimToady perl6: say ('a'...*) Z~ (1,2,3) xx *
p6eval niecza v13-352-g6694950: OUTPUT«(timeout)»
..rakudo 27a5dd: OUTPUT«Decrement out of range␤ in method Str at src/gen/CORE.setting:7212␤ in method Stringy at src/gen/CORE.setting:678␤ in sub infix:<~> at src/gen/CORE.setting:965␤ in sub <anon> at src/gen/CORE.setting:9024␤ in sub coro at src/gen/CORE.setting:4618␤ in method re…
..pugs b927740: OUTPUT«*** ␤ Unexpected " Z"␤ at /tmp/BsbWzeI54C line 1, column 14␤»
TimToady perl6: .say for ('a'...*) Z~ (1,2,3) xx *
jnthn ...decrement? :)
p6eval rakudo 27a5dd: OUTPUT«Decrement out of range␤ in method Str at src/gen/CORE.setting:7212␤ in method Stringy at src/gen/CORE.setting:678␤ in sub infix:<~> at src/gen/CORE.setting:965␤ in sub <anon> at src/gen/CORE.setting:9024␤ in sub coro at src/gen/CORE.setting:4618␤ in method re… 18:14
..niecza v13-352-g6694950: OUTPUT«(timeout)a1␤b2␤c3␤d1␤e2␤f3␤g1␤h2␤i3␤j1␤k2␤l3␤m1␤n2␤o3␤p1␤q2␤r3␤s1␤t2␤u3␤v1␤w2␤x3␤y1␤z2␤aa3␤ab1␤ac2␤ad3␤ae1␤af2␤ag3␤ah1␤ai2␤aj3␤ak1␤al2␤am3␤an1␤ao2␤ap3␤aq1␤ar2␤as3␤at1␤au2␤av3␤aw1␤ax2␤ay3␤az1␤ba2␤bb3␤bc1␤bd2␤be3␤bf1␤bg2␤bh3␤bi1␤bj2␤bk3␤bl1␤bm2␤bn3␤bo1␤bp2␤􏿽xE2􏿽x80
..pugs b927740: OUTPUT«*** ␤ Unexpected " Z"␤ at /tmp/9eNfrh7inL line 1, column 19␤»
TimToady rakudo is suffering from increment whether
jnthn *groan* 18:15
18:15 fsergot joined
Radvendii isn't »op» act like X? 18:15
dalek d: 47c76b3 | larry++ | STD.pm6:
change <initializer> to category for extensibility
 18:16
TimToady no, only if there's exactly one thing on the right
in which case, either construct repeats it
well, multiple things can repeat too, I guess 18:17
perl6: say 'a'..'z' X~ 1,2,3
p6eval rakudo 27a5dd, niecza v13-352-g6694950: OUTPUT«a1 a2 a3 b1 b2 b3 c1 c2 c3 d1 d2 d3 e1 e2 e3 f1 f2 f3 g1 g2 g3 h1 h2 h3 i1 i2 i3 j1 j2 j3 k1 k2 k3 l1 l2 l3 m1 m2 m3 n1 n2 n3 o1 o2 o3 p1 p2 p3 q1 q2 q3 r1 r2 r3 s1 s2 s3 t1 t2 t3 u1 u2 u3 v1 v2 v3 w1 w2 w3 x1 x2 x3 y1 y2 y3 z1 z2 z3␤»
..pugs b927740: OUTPUT«abcdefghijklmnopqrstuvwxyz␤»
TimToady so yes, you are correct about that
[Coke] flussence++ 18:18
TimToady note, however, that X is lazy, while hyper is eager, so only X or Z can deal with infinite lists
Radvendii right...
nom: say <a b> Z 1,2,3
p6eval nom 27a5dd: OUTPUT«a 1 b 2␤»
TimToady and X can only deal with an infinite list on its left 18:19
Radvendii nom: say <a b> »~» 1,2,3
p6eval nom 27a5dd: OUTPUT«a1 b123␤»
TimToady precedence
hypers are also scalar ops, not list ops
hypers expect to run on an object, not a list
Radvendii nom: say <a b> »~» (1,2,3) 18:20
p6eval nom 27a5dd: OUTPUT«a1 b2␤»
sorear reads the last few STD commits
Radvendii ah. the issue is i have a list on either side and i dont know which one is going to be infinite.
sorry
which one is going to be bigger
TimToady if you want the shorter to repeat in any case, use «op» 18:21
Radvendii nom: say <a b> «~» (1,2,3)
p6eval nom 27a5dd: OUTPUT«a1 b2 a3␤»
Radvendii i see
benabik ponders what infix:<op> should do...
sorear I was wondering earlier if anyone had any use whatsoever for the "repeat either" semantics of «op»
Radvendii haha. i do! 18:22
oh wait. i dont want them concatenated though
[Coke] wonders if we think perl6 will suffer from the same issues being discussed on p5p today in regards to distributions and "what is perl". /me suspects he will just use perlbrew if it comes to that. (esp. nice to have a perl6brew given we have different perls to start with.)
TimToady then use a different op
Radvendii nom: say <a b> «,» (1,2,3)
p6eval nom 27a5dd: OUTPUT«a 1 b 2 a 3␤»
Radvendii there we go
sorear minor problem being that niecza doesn't use 'add_name' and patch isn't thrilled by the fact that niecza's context lines are lacking { $ċ.add_name... 18:24
Radvendii [Coke]: here's my question: why are we splitting our efforts between different (compilers?) of perl (e.g. niecza, rakudo) 18:25
sorear because I'm a pathetic wretch who doesn't work well with others 18:26
TimToady because efforts are not a zero-sum game
Radvendii sorear: :P
TimToady because when two different implementations agree on how something should work, it's a pretty good indication that it's probably how it should be in the spec 18:27
Radvendii that seems like a pretty secondary goal but that's just IMO, and i'm pretty new to this. 18:28
TimToady the primary goal is to produce a better language, however long it takes
the secondary goal is that we want it now :)
Radvendii i suppose...
sorear There isn't actually a *splitting* of effort involved. 18:29
Radvendii sorear: no?
sorear After I determined that I was mental-health-ily incapable of contributing to Rakudo, my choices were starting Niecza or not contributing at all
std: sub (\foo) { } 18:30
p6eval std b5dc271: OUTPUT«ok 00:01 109m␤»
Radvendii Okay...
TimToady Radvendii: I strongly suggest reading S01 for all about this meta-project-stuff
Radvendii TimToady: okay
18:31 bluescreen10 left
Radvendii nom: sub a {say "in"; return True}; say [||] (a xx 3) 18:33
p6eval nom 27a5dd: OUTPUT«===SORRY!===␤CHECK FAILED:␤Undefined routine '&xx' called (line 1)␤»
Radvendii nom: say 1 xx 3 18:34
p6eval nom 27a5dd: OUTPUT«1 1 1␤»
sorear decides to try editing the patch before applying it
jnthn std: sub a {say "in"; return True}; say [||] (a xx 3)
p6eval std b5dc271: OUTPUT«===SORRY!===␤Undeclared routine:␤ 'xx' used at line 1␤Check failed␤FAILED 00:01 111m␤»
sorear perl6: say rand xx 3
jnthn Radvendii: I think you need a() there
p6eval pugs b927740: OUTPUT«*** No such subroutine: "&xx"␤ at /tmp/QT0oWEDtGH line 1, column 5 - line 2, column 1␤»
..niecza v13-352-g6694950: OUTPUT«0.11041774186790816 0.80917691244239776 0.39022010303578342␤»
..rakudo 27a5dd: OUTPUT«0.36751254841602 0.36751254841602 0.36751254841602␤»
Radvendii nom: sub a {say "in"; return True}; say [||] (a() xx 3)
p6eval nom 27a5dd: OUTPUT«in␤Bool::True␤»
jnthn goes to find something to eat
18:34 bluescreen10 joined
Radvendii nom: sub a {say "in"; return True}; say [||] <a b c>.map: {a()} 18:35
p6eval nom 27a5dd: OUTPUT«in␤in␤in␤Bool::True␤»
Radvendii so .map isn't lazy
[Coke] Radvendii: same question for "why perl6? why not just work on perl5!"
volunteer time isn't completely fungible.
moritz Radvendii: wrong conclusion
Radvendii [Coke]: good question. well, actually my question would be why perl5?
moritz: how so? 18:36
moritz nom: sub a {say "in"; return True}; say <a b c>.map(&in).[0]
p6eval nom 27a5dd: OUTPUT«===SORRY!===␤Symbol '&in' not predeclared in <anonymous> (/tmp/XG9pl3TIhD:1)␤»
sorear patch: **** malformed patch at line 36
TimToady Radvendii: that's a polite way of saying you're wrong
moritz nom: sub a {say "in"; return True}; say <a b c>.map(&a).[0]
p6eval nom 27a5dd: OUTPUT«Too many positional parameters passed; got 1 but expected 0␤ in sub a at /tmp/5a_WhVnDb5:1␤ in method reify at src/gen/CORE.setting:4471␤ in method reify at src/gen/CORE.setting:4376␤ in method gimme at src/gen/CORE.setting:4740␤ in method exists at src/gen/CORE.…
moritz nom: sub a($) {say "in"; return True}; say <a b c>.map(&a).[0]
p6eval nom 27a5dd: OUTPUT«in␤Bool::True␤»
moritz Radvendii: see? if map were eager, you'd see three 'in'
Radvendii but when i try to [||] it... 18:37
nom: sub a {say "in"; return True}; say any( <a b c>.map: {a()} )
p6eval nom 27a5dd: OUTPUT«in␤in␤in␤any(Bool::True, Bool::True, Bool::True)␤»
18:37 bluescreen10 left
Radvendii oh right, any does something else 18:37
nom: sub a {say "in"; return True}; say ?any( <a b c>.map: {a()} )
p6eval nom 27a5dd: OUTPUT«in␤in␤in␤Bool::True␤»
timotimo huh, in the perl6 book, out of nowhere it says :type<<double-beef>>. my rakudo-nom shows no difference between :type<double-beef> and the one with <</>> 18:39
sorear hmm 18:41
Radvendii nom: say <a b c>.map: {$_}.map: {$_}
p6eval nom 27a5dd: OUTPUT«Method 'count' not found for invocant of class 'List'␤ in method reify at src/gen/CORE.setting:4471␤ in method reify at src/gen/CORE.setting:4376␤ in method reify at src/gen/CORE.setting:4376␤ in method gimme at src/gen/CORE.setting:4740␤ in method eager at src/g…
sorear timotimo: Double angles interpolate like double quotes
Radvendii nom: say (<a b c>.map: {$_}).map: {$_}
p6eval nom 27a5dd: OUTPUT«a b c␤»
sorear patch doesn't seem to like it when I delete individual lines from a unidiff 18:42
18:42 GlitchMr joined
timotimo sorear: so i could do things like :types<<rakudo-$currentbranch niecza>> and it would, for instance, give me "rakudo-nom", "niecza"? 18:43
sorear yes
timotimo well, "types" => (rak...)
neat
doy sorear: you also need to update the line counts in the hunk header
sorear wonders how the line counts are defined 18:44
doy it's a pain, and usually less effort to just regenerate the patch after fixing it in the actual source 18:45
timotimo another thing that confused me (maybe i should write a ticket for that as well?) is that in the dance moves example, the subs in the moves hash are called like this: %moves{$move}.(); - why is the . needed there?
moritz timotimo: feel free to ticket it
(it's not needed, just a matter of style preference)
timotimo well, the example has a "todo: find a different example", so maybe it's not needed after all 18:47
Radvendii nom: say ?('wK' !~~ 'b')
p6eval nom 27a5dd: OUTPUT«Bool::True␤»
Radvendii nom: say ?('wK' !~~ 'w') 18:48
p6eval nom 27a5dd: OUTPUT«Bool::True␤»
Radvendii nom: say ?('wK' !~~ /w/)
p6eval nom 27a5dd: OUTPUT«Bool::False␤»
18:49 lichtkind joined
lichtkind cheers o/ 18:51
moritz nom: say pir::perl6ize_type(1).WHAT
p6eval nom 27a5dd: OUTPUT«use of uninitialized value of type Mu in string context␤use of uninitialized value of type Mu in string context␤===SORRY!===␤error:imcc:The opcode 'perl6ize_type_p' (perl6ize_type<1>) was not found. Check the type and number of the arguments␤ in file '(file unknown)' …
sorear o/ lichtkind
moritz nom: say pir::perl6ize_type__(1).WHAT 18:52
p6eval nom 27a5dd: OUTPUT«===SORRY!===␤rtype not set␤»
moritz nom: say pir::perl6ize_type__PP(1).WHAT
p6eval nom 27a5dd: OUTPUT«Int()␤»
moritz nom: say pir::perl6ize_type__PP('1').WHAT
p6eval nom 27a5dd: OUTPUT«Str()␤»
dalek ecza: 32dcbb0 | coke++ | t/spectest.data:
run S03-smartmatch/any-type.t
jnthn moritz: Not convinced that's testing what you want it to :) 18:53
moritz nom: say pir::perl6ize_type__PP(nqp::unbox_s '1').WHAT
p6eval nom 27a5dd: OUTPUT«Str()␤»
jnthn That's more like it :) 18:54
jnthn goes to cook stuff
dalek ok: 40b1d5d | timo++ | src/subs-n-sigs.pod:
in MAIN, limit sides to Numeric, update the usage output to also include --sum
 18:55
lichtkind jnthn: goof om
i just come from it
i mean nom
Radvendii nom: $_ = 'wK'; say "yay" if 1== /w/ ?? 1 !! 2 19:00
p6eval nom 27a5dd: OUTPUT«No applicable candidates found to dispatch to for 'Numeric'. Available candidates are:␤:(Mu:U \$v, Mu %_!)␤␤ in method Numeric at src/gen/CORE.setting:649␤ in sub infix:<==> at src/gen/CORE.setting:2276␤ in block <anon> at /tmp/3qPRgCHe_G:1␤ in <anon> at /tmp/3qPR…
Radvendii nom: $_ = 'wK'; say "yay" if 1== (/w/ ?? 1 !! 2)
p6eval nom 27a5dd: OUTPUT«yay␤»
Radvendii nom: say "yup" if 1==2|1 19:01
p6eval nom 27a5dd: OUTPUT«yup␤»
Radvendii is there a Bool !! else ?? if syntax? 19:02
reversing the if and the else, kind of like an "unless" 19:03
moritz no
Radvendii okay
moritz jnthn: gist.github.com/1677947 that's where I'm currently stuck (includes patch and backtrace) 19:08
OOOH, I know what's wrong
unstuck
don't call an attribute $.new
Radvendii haha
moritz it generates an accessor which overrwrites the constructor
Radvendii good idea
[Coke] moritz, masak: can you tell me how many people have submitted anything? (I know i haven't submitted anything yet.) 19:09
Radvendii is there a way to overwrite variables as they are being passed into a subroutine? 19:10
like a default value, except that it is applied to the value they pass in 19:11
[Coke] ... you mean, can you ignore the value they pass in?
that seems dubious.
Radvendii nom: sub a($a=$a+1){say $a}; a(3) 19:12
p6eval nom 27a5dd: OUTPUT«3␤»
Radvendii except in what i am talking about that would be 4...
nom: sub a($a+1){say $a}; a(3)
p6eval nom 27a5dd: OUTPUT«===SORRY!===␤Malformed block at line 1, near "($a+1){say"␤»
Radvendii nom: sub a($a+=1){say $a}; a(3) 19:13
p6eval nom 27a5dd: OUTPUT«===SORRY!===␤Malformed block at line 1, near "($a+=1){sa"␤»
moritz nom: sub a($a is copy) { $a++; say $a }
p6eval nom 27a5dd: ( no output )
moritz nom: sub a($a is copy) { $a++; say $a }; a(3)
p6eval nom 27a5dd: OUTPUT«4␤»
19:13 alester joined
moritz signatures are for matching and binding, not for computations 19:13
Radvendii i guess...
[Coke] there's already a perfectly good place to write code, in the body. ;) 19:14
Radvendii haha. i know. it just seems more clear to have all of the variable processing in the signature
moritz no.
the signature is part of the public API
Radvendii nom: sub a($tmp, :$a=$tmp+1){say $a}; a(3) 19:15
moritz how you treat it internally isn't part of the API
p6eval nom 27a5dd: OUTPUT«4␤»
TimToady sorear: often a context diff will do better than a unidiff, if there is mismatched context in the middle of a unidiff
Radvendii moritz: understood.
skids nom: sub a($a, :$b? = $a + 1) { say $b; }; a(3); # but then you have the :b named parameter publicly available even if you do not want it.
p6eval nom 27a5dd: OUTPUT«4␤»
benabik Radvendii: Would you want to write `my $a + 1 = 3` and have $a = 4?
moritz
.oO( let's integrate an algebraic equation solver in p6... )
 19:16
Radvendii no...
moritz: agreed
moritz in the multi dispatcher! jnthn, we look at you! (jk)
Radvendii what's the perl6 equivalent to Javas toString() method? 19:21
benabik .Str
Radvendii and that's what say will call?
moritz .gist
benabik IIUC, "$a" and print $a will call $a.Str. say $a will call $a.gist 19:22
[Coke] (that is going to cause so much trouble. :P
Radvendii why are they different?
shouldn't they do exactly the same thing? 19:23
moritz then we wouldn't need different methods :-)
skids say is for programmers. Print is for users.
moritz aye
Radvendii oh.
but for thingsl like ints, it's going to do the same 19:24
moritz aye
Radvendii haha. i've been pronouncing gist in my head with a hard g, so i was confused as to it's meaning. now i'm confused as to how it relates to string conversions... 19:25
TimToady and I used to pronounce zavolaj with an affricate on the end until I thought about it... 19:27
Radvendii . . . 19:29
pmichaud back from lunch
huf what's an affricate?
i mean in this context
TimToady a stop plus its associated fricative, in this case j as in judge
huf oh, so zavoladj when it should be zavolay? 19:30
19:31 birdwindupbird joined
timotimo is foo() and foo always equivalent? 19:31
TimToady well, zah voe lie in English
pmichaud timotimo: no
TimToady (I suspect)
moritz nom: sub f(*@a) { say @a.join }; f() + 1; f + 1
p6eval nom 27a5dd: OUTPUT«␤1␤»
moritz a case where it's not equivalent 19:32
huf TimToady: got it.
moritz gah, tim<tab> can now result in 3 different nicks :/ 19:33
huf TimToady: btw, what was the mnemonic (or the idea behind) calling bless bless in p5?
benabik moritz: It's parsing that as f ( +1 ) ?
TimToady it's asking your boss to bless your proposed project
skids
.oO("fricative" sounds like it should mean "an almost dirty word")
huf well _now_ i'm confused. 19:34
Radvendii moritz: why did your gah, ... comment make the bloop sound that usually asociates someone "Radvendii:"ing me
TimToady huf: it's extending Official Recognition to the fact that a particular thing is in a particular category 19:35
Radvendii nom: $a = "hi"; say $a.length;
p6eval nom 27a5dd: OUTPUT«===SORRY!===␤Variable $a is not predeclared at line 1, near " = \"hi\"; s"␤»
Radvendii nom: my $a = "hi"; say $a.length;
p6eval nom 27a5dd: OUTPUT«Method 'length' not found for invocant of class 'Str'␤ in block <anon> at /tmp/lBerofVIYp:1␤ in <anon> at /tmp/lBerofVIYp:1␤»
skids nom: my $a = "hi"; say $a.chars; 19:36
p6eval nom 27a5dd: OUTPUT«2␤»
Radvendii ah
nom: my $a = "hi"; say $a.bytes; 19:37
skids length is now ambiguous what with unicode/charset support.
p6eval nom 27a5dd: OUTPUT«2␤»
huf TimToady: so bless as in sanctify? (as in "admit to the list of saints")
Radvendii so what's the difference betwixt .chars and .bytes?
skids if your charset is 8-bit, nothing.
benabik Radvendii: One returns the number of characters, the other the number of bytes. ;-P
TimToady huf: yes, there's some of that meaning too
Radvendii nom: my $a = "āł"; say $a.bytes; 19:38
p6eval nom 27a5dd: OUTPUT«4␤»
TimToady a constructor looks at the arguments and decides whether they make a sheep or a goat
Radvendii i see
huf TimToady: i am not familiar with this meaning of bless
should've spent more time in church, i guess
or reading bookses or somethjing
skids Actually it's a secret conspiracy to rhyme with "dress", which will be revealed in an exceptionally bad limerick in 2015. 19:39
huf that'd be really really great
pmichaud skids: does that make it a revealing dress? ;-)
Radvendii groan... 19:40
TimToady I doubt we can make skids fess
Radvendii meant that in the way-you're-supposed-to-react-to-puns way
dalek kudo/nom: 3670b71 | moritz++ | src/ (2 files):
<.obs> in the grammar now throws X::Obsolete objects

also provides a small piece of infrastructure for easily throwing typed exceptions from the grammar
pmichaud TimToady: You're probably right, I guess. 19:41
TimToady no less! :)
dalek ast: dc765eb | moritz++ | S32-exceptions/misc.t:
tests for typed exceptions
huf Radvendii: this channel would sound like the zombie apocalypse if everyone groaned everytime someone made a pun :D
Radvendii huf: haha
pmichaud phenny: tell jnthn I'm working on the star configure, but will likely need at least one more day to finish it. 19:42
phenny pmichaud: I'll pass that on when jnthn is around.
Radvendii nom: $a = 1,2,3; .say for $a 19:43
p6eval nom 27a5dd: OUTPUT«===SORRY!===␤Variable $a is not predeclared at line 1, near " = 1,2,3; "␤»
Radvendii nom: my $a = 1,2,3; .say for $a
p6eval nom 27a5dd: OUTPUT«1␤»
Radvendii nom: my $a = (1,2,3); .say for $a
p6eval nom 27a5dd: OUTPUT«1 2 3␤»
Radvendii nom: my $a = (1,2,3); .say for @$a
p6eval nom 27a5dd: OUTPUT«1␤2␤3␤»
Radvendii nom: my $a=1; my $b=$a.clone; $b=2; say $a; 19:46
p6eval nom 27a5dd: OUTPUT«1␤»
benabik nom: my $a = 1; my $b = $a; $b = 2; say $a
p6eval nom 27a5dd: OUTPUT«1␤»
benabik nom: my $a = 1; my $b := $a; $b = 2; say $a
p6eval nom 27a5dd: OUTPUT«2␤»
benabik nom: my $a = 1; my $b := $a.clone; $b = 2; say $a
p6eval nom 27a5dd: OUTPUT«Cannot assign to a non-container␤ in block <anon> at /tmp/ufOuNPBTWL:1␤ in <anon> at /tmp/ufOuNPBTWL:1␤»
benabik awh.
Radvendii ;P almost 19:47
nom: my $a=\1; my $b=$a.clone; $b=\2; say $a;
p6eval nom 27a5dd: OUTPUT«1␤»
Radvendii \ is still the way to create references right?
nom: my $a=(1,2,3); my $b=$a.clone; $b[0]='a'; say $a;
p6eval nom 27a5dd: OUTPUT«Cannot assign to a non-container␤ in block <anon> at /tmp/BdgBWrzT0O:1␤ in <anon> at /tmp/BdgBWrzT0O:1␤»
Radvendii nom: my @a=(1,2,3); my $b=@a.clone; $b[0]='a'; say @a; 19:48
p6eval nom 27a5dd: OUTPUT«elements() not implemented in class 'Mu'␤ in method REIFY at src/gen/CORE.setting:4952␤ in method REIFY at src/gen/CORE.setting:5186␤ in method reify at src/gen/CORE.setting:4376␤ in method gimme at src/gen/CORE.setting:4740␤ in method exists at src/gen/CORE.sett…
benabik Radvendii: I do not think that's doing what you think it's doing. 19:49
Radvendii it appears not.
timotimo raw.github.com/perl6/book/master/s...bjects.pod ← the line $some-task === any(@!dependencies) in "method depends-on" seems useless. its return value isn't used, is it? i suppose the intent was to make it the implicit return value, so adding an or in between the lines would fix it?
Radvendii what is .clone doing?
moritz Radvendii: see S12 19:50
Radvendii clone should create an identical copy right? but it should be it's own variable, so when i modify it the original is untouched? 19:51
pmichaud clone is typically for values, not containers 19:52
so "it should be its own variable" reads kind of wrongish there
nom: my @a = 1,2,3; my $b = @a.clone; say $b.perl; 19:53
p6eval nom 27a5dd: OUTPUT«elements() not implemented in class 'Mu'␤ in method REIFY at src/gen/CORE.setting:4952␤ in method REIFY at src/gen/CORE.setting:5186␤ in method reify at src/gen/CORE.setting:4376␤ in method reify at src/gen/CORE.setting:4376␤ in method gimme at src/gen/CORE.setti…
benabik I think that Array.clone has problems.
pmichaud that looks buggy-ish
I agree; I think Array.clone is broken
jnthn Array.clone is just doing whatever Mu.clone is doing. That's almost certainly inappropriate. 19:54
phenny jnthn: 19:42Z <pmichaud> tell jnthn I'm working on the star configure, but will likely need at least one more day to finish it.
pmichaud it's probably defaulting to Mu..... what jnthn++ said
jnthn Array likely needs its own .clone but I dunno what it should look like
(esp. with regard to what it does with the unreified bits)
pmichaud Whoever is responsible for the Array and List implementation really nee.... er. Hmm. 19:55
Radvendii okay. but I am cloning an object i made, modifying the clone, and the original is modified as well...
jnthn pmichaud: ;)
moritz Radvendii: it's not doing a deep clone
jnthn Mu.clone does something sane with scalar containers. If you've got an array or hash in there, it won't be doing much about that at the moment. 19:56
TimToady unreified bits should generally be assumed to be idempotent, I think
Radvendii so .clone is not very well implimented yet...
understood
pmichaud since iterators/cursors are considered immutable, the cloned array probably shares that part
19:57 bluescreen10 joined
jnthn Sounds sane. 19:57
pmichaud: (one more day) as in, might be done tomorrow, or at earliest Friday?
pmichaud might be done tomorrow 19:58
jnthn OK.
jnthn crosses fingers for that :)
pmichaud I could push it through today, but I'd be taking a couple of shortcuts that we'd have to clean up later 19:59
having an extra evening to reflect on it a bit will give a better impl
jnthn pmichaud: So long as we get it released this week, I'm happy enough.
pmichaud (this week) as in "before Saturday" or ...? 20:02
jnthn Yeah 20:03
Radvendii something about .gist is printing out TrueTrueTrueTrueTrueTrue at the end of what i print
well, 8 true's...
which is the size of my array 20:04
pmichaud I suspect you may be printing the result of say() 20:05
nom: say say say "what you want"
p6eval nom 3670b7: OUTPUT«what you want␤Bool::True␤Bool::True␤»
Radvendii oh
yes i am
thanks
20:06 GlitchMr left
pmichaud nom: take take take "what you need" 20:07
p6eval nom 3670b7: OUTPUT«stray take statement␤ in block <anon> at /tmp/1VT3DvTK9K:1␤ in <anon> at /tmp/1VT3DvTK9K:1␤»
pmichaud wonders if anyone recognizes the reference. :-)
skids Steppenwolf?
pmichaud en.wikipedia.org/wiki/Say_Say_Say 20:08
Radvendii got that there was a reference 20:09
pmichaud do we have a "pray" function yet? ;-) ( pray pray pray "everyday" ) 20:10
Radvendii we should... and make it have something to do with bless... 20:11
jnthn pmichaud: I didn't, but having looked at the page I realized that the reference is older than I am. ;-)
skids considers golfing Janis Joplin "Piece of My Heart" lyrics until they parse.
pmichaud jnthn: yeah, I was afraid I now fall into the "Oldies" genre. :-) 20:12
Radvendii this is quite odd...
jnthn pmichaud: Having just listened to a clip of the song, I am pretty sure I've heard it before, however. :)
20:13 Guest20598 is now known as Khisanth
skids
.oO(it isn't that you're an oldie if you like songs made before you were born. In that case you're an archivist, not an oldie.)
 20:13
Radvendii i run $a = ThisClass.new; $a.b = self.b; $a[0]=1; and suddenly self[0] is 1! (@.b handles array)
20:14 ksi joined
pmichaud I was in high school when that song came out, so I don't fall in the "song made before I was born" category, though. :-/ 20:14
timotimo niecza: multi barf($) { "meep" }; only sub barf(Int $a) { $a + 5 }; barf("a") 20:17
p6eval niecza v13-352-g6694950: OUTPUT«===SORRY!===␤␤Illegal redeclaration of routine 'barf' (see line 1) at /tmp/S5VBAvFAVi line 1:␤------> multi barf($) { "meep" }; only sub barf⏏(Int $a) { $a + 5 }; barf("a")␤␤Unhandled exception: Check failed␤␤ at /hom…
timotimo pugs: multi barf($) { "meep" }; only sub barf(Int $a) { $a + 5 }; barf("a")
p6eval pugs b927740: OUTPUT«*** ␤ Unexpected "$)"␤ expecting formal parameter or ")"␤ at /tmp/5TBV_NF__4 line 1, column 12␤»
20:19 birdwindupbird left
Radvendii nom: Class A {has @.b=1,2,3; method c { my $a = A.new; $a.b=self.b; $a.b[0]=4; self.b.say;}}; my $x = A.new; A.c; 20:20
p6eval nom 3670b7: OUTPUT«===SORRY!===␤You can't declare an attribute here; maybe you'd like a class or a role? at line 1, near "=1,2,3; me"␤»
Radvendii nom: Class A {has @.b = (1,2,3); method c { my $a = A.new; $a.b=self.b; $a.b[0]=4; self.b.say;}}; my $x = A.new; A.c;
p6eval nom 3670b7: OUTPUT«===SORRY!===␤You can't declare an attribute here; maybe you'd like a class or a role? at line 1, near " = (1,2,3)"␤»
jnthn Radvendii: wRONG CASE
dalek ok: c6ccbe3 | timo++ | src/multi-dispatch.pod:
Rakudo implements "only sub YourSub", niecza and pugs still don't.
 20:21
Radvendii nom: class A {has @.b = (1,2,3); method c { my $a = A.new; $a.b=self.b; $a.b[0]=4; self.b.say;}}; my $x = A.new; A.c;
p6eval nom 3670b7: OUTPUT«Cannot look up attributes in a type object␤ in method b at src/gen/CORE.setting:1384␤ in method c at /tmp/gWyM3qRy6m:1␤ in block <anon> at /tmp/gWyM3qRy6m:1␤ in <anon> at /tmp/gWyM3qRy6m:1␤»
Radvendii jnthn: thanks
i dont get the error... 20:23
what attribute am i look up in which type object?
pmichaud self.b from inside of A.c 20:24
Radvendii why is that a problem?
pmichaud because A is a type object
Radvendii nom: class A {has @.b = (1,2,3); method c { my $a = A.new; $a.b=@.b; $a.b[0]=4; @.b.say;}}; my $x = A.new; A.c;
p6eval nom 3670b7: OUTPUT«Cannot look up attributes in a type object␤ in method b at src/gen/CORE.setting:1384␤ in method c at /tmp/LgYglK6jPj:1␤ in block <anon> at /tmp/LgYglK6jPj:1␤ in <anon> at /tmp/LgYglK6jPj:1␤»
Radvendii oh
that doesnt help
what does type object mean? 20:25
pmichaud A is the object that represents the class
jnthn Radvendii: Did you meant to call $x.c at the end?
Radvendii ah
yes
nom: class A {has @.b = (1,2,3); method c { my $a = A.new; $a.b=@.b; $a.b[0]=4; @.b.say;}}; my $x = A.new; $x.c;
p6eval nom 3670b7: OUTPUT«1 2 3␤»
pmichaud so when you have A.c at the end, you're invoking the 'c' method on the type object, instead of an instance of the type object
Radvendii right
that makes sense
i thought it was something i was doing within the c method
nom: class A {has @.b = (1,2,3); method c { my $a = A.new; $a.b=@.b; $a.b[0]=4; @.b.say; $a.b.say}}; my $x = A.new; $x.c; 20:26
p6eval nom 3670b7: OUTPUT«1 2 3␤4 2 3␤»
20:27 tokuhirom joined
pmichaud afk again for a while... probably won't be back until tomorrow morning 20:28
Radvendii o/
jnthn o/
Radvendii nom: say <a b c>.map({$_}) 20:33
p6eval nom 3670b7: OUTPUT«a b c␤»
20:34 skids left 20:42 bbkr1 joined
masak nom: class A { our method foo { say "OH HAI" unless self.^name eq "A"; return &foo } }; class B is A { method bar(&c) { self.&c } }; B.bar(A.foo) 20:45
p6eval nom 3670b7: OUTPUT«OH HAI␤»
masak \o/
that's... sick :)
niecza: class A { our method foo { say "OH HAI" unless self.^name eq "A"; return &foo } }; class B is A { method bar(&c) { self.&c } }; B.bar(A.foo)
p6eval niecza v13-352-g6694950: OUTPUT«OH HAI␤» 20:46
masak \o/
niecza: class A { our method foo { say "OH HAI" unless self.^name eq "A"; return &foo } }; class B { method bar(&c) { self.&c } }; B.bar(A.foo)
p6eval niecza v13-352-g6694950: OUTPUT«Unhandled exception: Nominal type check failed in binding self in A.foo; got B, needed A␤ at /tmp/cS5Pj7J_Fl line 0 (A.foo @ 1) ␤ at /tmp/cS5Pj7J_Fl line 1 (B.bar @ 3) ␤ at /tmp/cS5Pj7J_Fl line 1 (mainline @ 5) ␤ at /home/p6eval/niecza/lib/CORE.settin…
masak indeed.
jnthn masak: Why the "our"? :) 20:47
bbkr1 What is the best practice to define attribute default value in child class? Let's say I have class A { has Str $.id }. Now I can write class B is A { has $.id = 8 } which is not good as it redeclares type. Or I can write class B is A { submethod BUILD { $.id = 8 } } which is also not good as this is slow runtime assignment and it has to be repeated in every BUILD method. 20:48
jnthn You shouldn't be doing $.id in BUILD anyway
20:48 ggoebel left
jnthn BUILD submethod should always refer directly to the storage slot ($!id) 20:49
bbkr1 I know, that's why i search for better solution
masak jnthn: so that the method gets registered in the namespace of class A, so that I can return &foo from it :) 20:51
jnthn masak: ah, I see
bbkr1 for example P5-ism from Moose was: has '+id' => ('default' => 8 );
masak jnthn: I bet there's an easier way to return a routine from itself... :)
jnthn Guess you could return &?ROUTINE
nom: sub foo() { say "hi"; return &?ROUTINE; }; my $x = foo(); $x() 20:52
p6eval nom 3670b7: OUTPUT«hi␤hi␤»
20:52 fsergot left 20:53 ggoebel joined 20:54 MayDaniel joined
timotimo rakudo: my %foo = 1 => "one", 2 => "two"; say %foo[1] 21:01
p6eval rakudo 3670b7: OUTPUT«Failure.new(exception => Exception.new())␤»
timotimo i must be very confused. why doesn't that work?
jnthn Because you're doing array indexing, not hash indexing. :)
rakudo: my %foo = 1 => "one", 2 => "two"; say %foo{1} 21:02
p6eval rakudo 3670b7: OUTPUT«one␤»
timotimo oooooooooh, duh!
21:02 mj41 joined
Radvendii nom: say ?('wK' !~~ 'b') 21:02
p6eval nom 3670b7: OUTPUT«Bool::True␤»
Radvendii nom: say ?('wK' !~~ 'w') 21:03
p6eval nom 3670b7: OUTPUT«Bool::True␤»
Radvendii nom: say ?('wK' !~~ /w/)
p6eval nom 3670b7: OUTPUT«Bool::False␤»
[Coke] nom: say ?('wK' !~~ /'w'/)
p6eval nom 3670b7: OUTPUT«Bool::False␤»
Radvendii <w b>.grep: -> $x {'wK' !~~ /$x/} 21:04
nom: say <w b>.grep: -> $x {'wK' !~~ /$x/}
p6eval nom 3670b7: OUTPUT«b␤»
Radvendii nom: sub a($l) { <w b>.grep: -> $x {$l !~~ /$x/} }; a('wK'); 21:05
p6eval nom 3670b7: ( no output )
Radvendii nom: sub a($l) { <w b>.grep: -> $x {$l !~~ /$x/} }; say a('wK');
p6eval nom 3670b7: OUTPUT«b␤»
Radvendii nom: sub a($l, $c = (<w b>.grep: -> $x {$l !~~ /$x/}) { say $c }; a('wK'); 21:06
p6eval nom 3670b7: OUTPUT«===SORRY!===␤Malformed block at line 1, near "($l, $c = "␤»
Radvendii nom: sub a($l, $c = (<w b>.grep: -> $x {$l !~~ /$x/})) { say $c }; a('wK');
p6eval nom 3670b7: OUTPUT«b␤»
Radvendii ugh. lazy evaluating is so annoying sometimes! 21:08
21:09 molaf left
Radvendii is there any way to 'require' evaluation, without printing out the results? 21:09
jnthn Radvendii: eager
Radvendii can i just say eager $a;? 21:10
jnthn Think so
Radvendii nom: my $a=1; my @a = $a xx *; eager @a; $a=2; say @a[0]; 21:11
p6eval nom 3670b7: OUTPUT«2␤»
Radvendii nope
nom: my $a=1; my @a = $a xx *; @a; $a=2; say @a[0]; 21:12
p6eval nom 3670b7: OUTPUT«2␤»
Radvendii nom: my $a=0; my @a = $a xx *; while $a++ < 5 {say @a[$a]} 21:14
p6eval nom 3670b7: OUTPUT«1␤2␤3␤4␤5␤»
Radvendii tee hee
PerlJam nom: my $a=1; my @a = eager $a xx *; $a=2; say @a[0];
p6eval nom 3670b7: OUTPUT«2␤»
Radvendii nom: my $a=1; eager my @a = $a xx *; $a=2; say @a[0]; 21:15
p6eval nom 3670b7: OUTPUT«2␤»
[Coke] I think when you say lazy you're not meaning what perl 6 lazy means.
I think you're confused about container semantics.
Radvendii maybe?
nom: sub a($b is copy,$c=$b+1) {$b++; say $c}; a(1) 21:17
p6eval nom 3670b7: OUTPUT«2␤»
Radvendii nom: sub a($b is copy,$c=$b+1) {$b=3; say $c}; a(1) 21:18
p6eval nom 3670b7: OUTPUT«2␤»
Radvendii that's not happening for me...
well, not in that exact form, but close enough
nom: sub a($b is copy,$c=(<w b>.grep: -> $x { $l !~~ /$x/})) {$b='b'; say $c}; a('w') 21:19
p6eval nom 3670b7: OUTPUT«===SORRY!===␤Variable $l is not predeclared at line 1, near " !~~ /$x/}"␤»
Radvendii nom: sub a($b is copy,$c=(<w b>.grep: -> $x { $b !~~ /$x/})) {$b='b'; say $c}; a('w') 21:20
p6eval nom 3670b7: OUTPUT«w␤»
Radvendii see?
nom: sub a($b is copy,$c=(<w b>.grep: -> $x { $b !~~ /$x/})) {say $c;$b='b';}; a('w')
p6eval nom 3670b7: OUTPUT«b␤»
Radvendii it's the .grep that's lazy 21:21
moritz that's because grep is lazy
right
nom: sub a($b is copy, @c = (<w b>.grep({ !/$x/})) { $b = 'b'; say @c }; a('w') 21:22
p6eval nom 3670b7: OUTPUT«===SORRY!===␤Variable $x is not predeclared at line 1, near "/})) { $b "␤»
moritz nom: sub a($b is copy, @c = (<w b>.grep(-> $x { !/$x/})) { $b = 'b'; say @c }; a('w')
p6eval nom 3670b7: OUTPUT«===SORRY!===␤Malformed block at line 1, near "($b is cop"␤»
moritz too tired to code 21:23
Radvendii ;) go to bed 21:24
moritz I will. 21:25
Radvendii why does "say" return 1 or 0 and not what it said... 21:29
moritz why should it?
21:30 ggoebel left
Radvendii because it's always more useful to get back what it said than it's success 21:30
masak I disagree. 21:32
moritz well, it's certainly p5ish to return the success status
masak though I never use the return value of &say in any form whatsoever.
yeah, was just going to say that :)
moritz I'd expect it die violently if it fails 21:33
benabik What is said may also be larger than you really want to keep a string around for.
masak IIRC, Rakudo simply returns True always.
moritz benabik: that would have been my next point to make
benabik moritz: GMTA
21:33 mj41 left
masak benabik: well, a sink context optimizer would take care of that... 21:33
moritz benabik: fsvo'G' 21:34
:-)
21:34 alester left
masak doing it with the other letters are worse insults :P 21:34
moritz std: #`
p6eval std 47c76b3: OUTPUT«===SORRY!===␤Opening bracket is required for #` comment at /tmp/WjGbkadWON line 1:␤------> #`⏏<EOL>␤Parse failed␤FAILED 00:01 96m␤»
21:35 ggoebel joined 21:37 molaf joined
Radvendii that's just my point, always returning true isn't helpful, and when i want to throw a say in to see what a function is returning, the return then stops returning what i want... 21:39
prammer sounds like you want something like Dwarn (p5) metacpan.org/module/Devel::Dwarn 21:43
21:44 MayDaniel left
Radvendii yes, unfortunately that's only for perl _5_ . also, it still doesn't make sense to me that say would return a boolean at all. it's just not helpful. or intuitive. 21:45
nom: say print "a"
p6eval nom 3670b7: OUTPUT«aBool::True␤»
bbkr1 rakudo: class A { has Int $.id; }; class B is A { submethod BUILD ( :$!id = 10 ) { } }; B.new' 21:51
p6eval rakudo 3670b7: OUTPUT«===SORRY!===␤Confused at line 1, near "B.new'"␤»
bbkr1 rakudo: class A { has Int $.id; }; class B is A { submethod BUILD ( :$!id = 10 ) { } }; B.new
p6eval rakudo 3670b7: OUTPUT«Null PMC access in get_pmc_keyed_str()␤ in submethod BUILD at /tmp/NGnaIv23u2:1␤ in method BUILDALL at src/gen/CORE.setting:583␤ in method bless at src/gen/CORE.setting:573␤ in method new at src/gen/CORE.setting:558␤ in block <anon> at /tmp/NGnaIv23u2:1␤ in <…
bbkr1 known bug? it only happens on inherited attributes.
another variation 21:52
rakudo: class A { has Int $.id; }; class B is A { submethod BUILD ( $!id? ) { } }; B.new
p6eval rakudo 3670b7: OUTPUT«Null PMC access in get_pmc_keyed_str()␤ in submethod BUILD at /tmp/qLZGBlS5WY:1␤ in method BUILDALL at src/gen/CORE.setting:583␤ in method bless at src/gen/CORE.setting:573␤ in method new at src/gen/CORE.setting:558␤ in block <anon> at /tmp/qLZGBlS5WY:1␤ in <…
jnthn bbkr1: That attribute ain't visible there. 21:54
Why it doesn't explode is a bit mysterious though...
masak nom: class A { has Int $.id; }; class B is A { submethod BUILD ( $!id ) { } }; B.new
p6eval nom 3670b7: OUTPUT«Not enough positional parameters passed; got 1 but expected 2␤ in submethod BUILD at /tmp/8FMWmEVmBd:1␤ in method BUILDALL at src/gen/CORE.setting:583␤ in method bless at src/gen/CORE.setting:573␤ in method new at src/gen/CORE.setting:558␤ in block <anon> at /tmp…
masak aha!
jnthn Of course, we can catch that one at compile time.
masak yeah, but what about the error above? 21:55
jnthn masak: The error above is "right" though we should really catch the other one at compile time.
masak: I mean, we'd not make it to runtime there.
Radvendii nom: say 1 ^...^ 4 21:56
p6eval nom 3670b7: OUTPUT«===SORRY!===␤Confused at line 1, near "say 1 ^..."␤»
Radvendii ..?
[Coke] nom: sub llog(*@a) { .say for @a ; return @a}; foo($a,$b) { return llog $a+$b}; foo(3,4);
p6eval nom 3670b7: OUTPUT«===SORRY!===␤Variable $a is not predeclared at line 1, near ",$b) { ret"␤»
Radvendii shouldn't that produce 2,3?
colomon ^,,,^ doesn't exist
^...^, I mean
[Coke] nom: sub llog(*@a) { .say for @a ; return @a}; sub foo($a,$b) { return llog $a+$b}; foo(3,4);
p6eval nom 3670b7: OUTPUT«7␤»
bbkr1 reports LTA message in this BUILD example 21:57
[Coke] nom: sub llog(*@a) { .say for @a ; return @a}; sub foo($a,$b) { return llog $a+$b}; say "foo(3,4) is" + foo(3,4);
p6eval nom 3670b7: OUTPUT«7␤1␤»
[Coke] nom: sub llog(*@a) { .say for @a ; return @a}; sub foo($a,$b) { return llog $a+$b}; say "foo(3,4) is" ~ foo(3,4);
p6eval nom 3670b7: OUTPUT«7␤foo(3,4) is7␤»
Radvendii i thought ^ at one end excluded the endpoint there
[Coke] Radvendii: there you go.
jnthn bbkr1: Thanks. I'm done with $dayjob tasks, had a stroll and have a few tuits, so may see if I can improve that now. But do file the ticket, so we can get test coverage. ;) 21:58
colomon Radvendii: that's true for Range, not true for sequence
Radvendii but isn't ... also used if you dont know which side is larger?
nom: say 5..3
p6eval nom 3670b7: OUTPUT«5..3␤»
Radvendii nom: say (5..3).list 21:59
p6eval nom 3670b7: OUTPUT«␤»
Radvendii nom: say (5...3).list
p6eval nom 3670b7: OUTPUT«5 4 3␤»
PerlJam Radvendii: note that ... and ^... and ...^ and ^...^ would all be separate operators if they all existed.
Radvendii PerlJam: why? there's no way to make that a meta-op?
PerlJam perhaps with macros 22:00
22:00 bluescreen10 left
colomon The idea of using ... if you don't know which side is larger is very dangerous in general 22:00
.. and ... are very different 22:01
Radvendii how so... 22:02
[Coke] grok ..
is the grokbot still about? 22:03
masak nope.
[Coke] .u sad
phenny U+0635 ARABIC LETTER SAD (ص)
22:05 chee joined
colomon Radvendii: .. creates a Range object, which (if numeric) represents a portion of the real line. ... creates a sequence of numbers. 22:06
for instance:
nom: say 4..5.5;
p6eval nom 3670b7: OUTPUT«4..11/2␤»
colomon nom: say 4...5.5 22:07
p6eval nom 3670b7: OUTPUT«(timeout)»
colomon timeout because 4...5.5 is the infinite sequence 4, 5, 6, 7...
another example:
nom: say 4.1 ~~ 4..5.5 22:08
p6eval nom 3670b7: OUTPUT«Bool::True␤»
colomon nom: say 4.1 ~~ 4...6
p6eval nom 3670b7: OUTPUT«(timeout)»
colomon errr... not sure why that one blows up
Radvendii ah...
:P 22:09
colomon if you want a Range from $a to $b, but you don't know which is bigger, you can say
nom: my $a = rand; my $b = rand; say $a ~ " " ~ $b; say ($a, $b).minmax.perl
p6eval nom 3670b7: OUTPUT«0.0948716237480802 0.257618284409062␤0.0948716237480802e0..0.257618284409062e0␤»
colomon nom: my $a = rand; my $b = rand; say $a ~ " " ~ $b; say ($a, $b).minmax.perl
p6eval nom 3670b7: OUTPUT«0.357001702374877 0.587949935670139␤0.357001702374877e0..0.587949935670139e0␤»
Radvendii that seems like a bit much for such an operation... 22:11
colomon ?
($a, $b).minmax is too much work for you?
Radvendii oh 22:12
i just realized that the rest of it wasn't part of the actual operation
colomon you can also say it $a minmax $b
Radvendii :P
PerlJam minmax++
Radvendii and minmax does the same thing as .. except orders them first? 22:13
wait...
benabik nom: say 2..1; say 2 minmax 1
p6eval nom 3670b7: OUTPUT«2..1␤1..2␤»
Radvendii nom: say 5 ^minmax^ 1
p6eval nom 3670b7: OUTPUT«one(5, Inf..-Inf, 1)␤» 22:14
colomon no, minmax orders them and then makes a Range (ie ..)
cognominal nom: say Signature.^methods[0].WHAT'
p6eval nom 3670b7: OUTPUT«===SORRY!===␤Confused at line 1, near "say Signat"␤»
Radvendii colomon: that's what i said...
PerlJam colomon: 4.1 ~~ 4...6 blows up because it means (4.1 ~~ 4) ... 6, or False ... 6
cognominal nom: say Signature.^methods[0].WHAT
p6eval nom 3670b7: OUTPUT«Can only use get_what on a SixModelObject␤ in block <anon> at /tmp/S_ZupZj2K4:1␤ in <anon> at /tmp/S_ZupZj2K4:1␤»
cognominal this probably explains rt.perl.org/rt3//Public/Bug/Displa...?id=108968
colomon PerlJam++: ah! thanks!
cognominal nom: say Int.^methods[0].WHAT
p6eval nom 3670b7: OUTPUT«Method()␤» 22:15
Radvendii PerlJam: shouldn't False evaluate to 0 in numerical context?
colomon who said ... was numerical context?
Radvendii oh right... 22:16
22:16 kaare_ left
PerlJam Radvendii: ... is deep and powerful magic :) 22:16
Radvendii nom: say 5 ^minmax 1
p6eval nom 3670b7: OUTPUT«one(5, 1..1)␤»
Radvendii cocks an eyebrow
oh
colomon nom: say (True, * ^^ True ... *)[10] 22:17
p6eval nom 3670b7: OUTPUT«===SORRY!===␤Unable to parse postcircumfix:sym<( )>, couldn't find final ')' at line 2␤»
Radvendii i see what it's doing
colomon you can't use ^ with minmax like you can with Range
std: say 5 ^minmax 1
Radvendii colomon: so then it doesn't solve my original problem. and why not?
masak colomon: you have control chars in your programs again.
p6eval std 47c76b3: OUTPUT«ok 00:01 109m␤»
22:18 ksi left
colomon nom: say (True, * ^^ True ... *)[10] 22:18
p6eval nom 3670b7: OUTPUT«Bool::False␤»
colomon doh!
nom: say (True, !* ... *)[10]
p6eval nom 3670b7: OUTPUT«Bool::True␤»
colomon nom: say !True
p6eval nom 3670b7: OUTPUT«Bool::False␤»
colomon hmmm, I don't remember False stopping sequences 22:19
Radvendii: I don
t
Radvendii: I don't think anyone ever thought of adding ^minmax 22:20
PerlJam I'm not sure what ^minmax would even mean
Radvendii PerlJam: minmax, except not including the operand on that side?
colomon so 10 ^minmax 1 would be 1 ..^ 10 22:21
Radvendii yeah.
or it could be that 10 ^minmax 1 would be 1 ^.. 10
and by the way, i do want ...
[Coke] if part of the point of minmax is that you might not know the order ahead of time, that seems iffy.
Radvendii because i am doing something like: 22:22
colomon I'd think the first explanation makes a lot more sense
Radvendii nom: ((10 ... 1) «,» 3).map -> $i, $j { say $i, ", ", $j } 22:23
p6eval nom 3670b7: OUTPUT«===SORRY!===␤Confused at line 1, near "((10 ... 1"␤»
Radvendii ...? 22:24
nom: (10 ... 1 «,» 3).map -> $i, $j { say $i, ", ", $j } 22:25
p6eval nom 3670b7: OUTPUT«===SORRY!===␤Confused at line 1, near "(10 ... 1 "␤»
Radvendii nom: (10 ... 1 »,» 3).map -> $i, $j { say $i, ", ", $j }
p6eval nom 3670b7: OUTPUT«===SORRY!===␤Confused at line 1, near "(10 ... 1 "␤»
PerlJam Radvendii: so ... what would @a ^minmax @b mean? 22:27
Basically the ^ there would just exclude the end point on that side? 22:28
Radvendii i thought it would make more sense if it excluded it on the lesser side 22:29
cognominal jnthn, in BOOTSTRAP.pm I think the added method should be put in a container so as not to later trigger a bug. What it the nqp:: sub for that?
Radvendii because you wouldn't use minmax if you knew which side that was
PerlJam so if @a minmax @b == $a..$b, then @a ^minmax @b == $a^..@b (and presumably @a minmax^ @b == $a..^$b) ?
cognominal nom: say Proxy.^methods[0].WHAT 22:30
p6eval nom 3670b7: OUTPUT«Can only use get_what on a SixModelObject␤ in block <anon> at /tmp/5EqRkdDmEM:1␤ in <anon> at /tmp/5EqRkdDmEM:1␤»
jnthn cognominal: Well, probably need to re-arrange things a little in there to make sure Method etc exists.
Radvendii yeah. i think so
jnthn cognominal: And they probably want some kinda signature object piecing together too
cognominal all the classes that uses the HOW.add_method have the problem.
jnthn cognominal: Yeah.
cognominal: The problem is that they're adding a Parrot sub. 22:31
(more)
Radvendii nom: say (1..4).map(&say)
p6eval nom 3670b7: OUTPUT«1␤2␤3␤4␤True True True True␤»
Radvendii :P
jnthn That could be resolved by wrapping them all in Perl 6 Methods. Alternatively, NQP may be about to get code objects too.
PerlJam Radvendii: I'm dubious of the utility, but otherwise it seems okay 22:32
cognominal ok
Radvendii because right now i need a ^minmax^ operator
well... what i really need is ^...^
PerlJam Radvendii: that's another kettle of fish though
Radvendii why?
nom: say 1...^10 22:33
p6eval nom 3670b7: OUTPUT«1 2 3 4 5 6 7 8 9␤»
Radvendii nom: say 1^...20
p6eval nom 3670b7: OUTPUT«1^..1/5␤»
Radvendii nom: say 1^...10
p6eval nom 3670b7: OUTPUT«1^..1/10␤»
Radvendii nom: say 1(^...)10
p6eval nom 3670b7: OUTPUT«===SORRY!===␤Confused at line 1, near "say 1(^..."␤»
Radvendii nom: say 1 ^... 10 22:34
p6eval nom 3670b7: OUTPUT«===SORRY!===␤Confused at line 1, near "say 1 ^..."␤»
Radvendii what is going on?
PerlJam ^... is probably not implemented?
Radvendii right
oh... 22:35
that's because...
nope
can't think of a reason
colomon because it would mean "drop the first element of your sequence"
Radvendii yeah...
PerlJam btw, I think there's either a spec bug or a rakudo bug. Compare the text at S03:1721 with the output of my @a = 1..^10; my @b = 5..7; @a minmax @b 22:36
nom: my @a = 1..^10; my @b = 5..7; @a minmax @b
p6eval nom 3670b7: ( no output )
PerlJam nom: my @a = 1..^10; my @b = 5..7; say @a minmax @b
p6eval nom 3670b7: OUTPUT«1..9␤»
Radvendii that makes sense at least...
PerlJam the spec says "...except that if the corresponding C<excludes> flag is set in Range, the excludes flag is also set in the returned C<Range>." 22:37
which means it should be 1..^10
colomon nom: my @a = 1..^10; say @a.WHAT
p6eval nom 3670b7: OUTPUT«Array()␤»
Radvendii PerlJam: ah.
colomon it's not a Range
PerlJam oh.
There's a PerlJam bug then :)
colomon nom: say (1..^10) minmax (5..7) 22:38
p6eval nom 3670b7: OUTPUT«1..9␤»
colomon that, on the other hand, looks more like a bug
Radvendii :P
colomon hmmm...
nom: (1..^10, 5..7).minmax.say
Radvendii (1..^10).WHAT
p6eval nom 3670b7: OUTPUT«1..9␤»
Radvendii nom: say (1..^10).WHAT
PerlJam yeah, there you go
p6eval nom 3670b7: OUTPUT«Range()␤»
Radvendii good catch 22:39
colomon nom: my $a = 1..^10; my $b = 5..7; say ($a, $b).minmax
p6eval nom 3670b7: OUTPUT«1..^10..5..7␤»
colomon that's a bug for sure
PerlJam heh!
indeed
Radvendii haha
colomon niecza: my $a = 1..^10; my $b = 5..7; say ($a, $b).minmax
p6eval niecza v13-352-g6694950: OUTPUT«5..7..1..^10␤»
colomon blast
Radvendii ouch
colomon well, spectests to be added, bugs to be fixed...
ng: my $a = 1..^10; my $b = 5..7; say ($a, $b).minmax 22:40
Radvendii colomon: i still dont get why ^... wouldn't make sense
colomon b: my $a = 1..^10; my $b = 5..7; say ($a, $b).minmax
p6eval b 1b7dd1: OUTPUT«123456789␤»
Radvendii std: my $a = 1..^10; my $b = 5..7; say ($a, $b).minmax
p6eval std 47c76b3: OUTPUT«ok 00:01 113m␤»
Radvendii i take it std is not a perl6 implementation... 22:41
colomon could have sworn he had minmax working properly at one point back in the days of b. :( 22:42
PerlJam Radvendii: not one that has a runtime, no.
Radvendii PerlJam: understood
PerlJam Radvendii: std is basically a perl 6 syntax checker
Radvendii PerlJam: got it. and std is for standard?
PerlJam aye
dalek kudo/nom: 12f3505 | jnthn++ | src/Perl6/Metamodel/MethodContainer.pm:
Parentheticize a remark in an error.
 22:44
kudo/nom: f9d635c | jnthn++ | src/Perl6/Actions.pm:
Catch undeclared attributes mentions in signatures.
Radvendii still, what's wrong with ^...?
colomon Radvendii: because $a ^... $b would be exactly equivalent to $a+1 ... $b -- just less clear 22:46
Radvendii nope
5 ^... 1 would not be the same as 6 ... 1
PerlJam colomon: would it? $a would just be omitted wouldn't it?
5 ^... 1 --> 4 3 2 1, 'a'^...'d' --> 'b', 'c', 'd' 22:47
jnthn phenny: tell moritz Things that use $*W.throw(...) (e.g. sub foo($x! = 42) { }) gives an error in -e or when run in a file, but the error is never shown in the REPL - it seems to swallow them. 22:48
colomon sure, if it's a decreasing sequence it's $a-1..$b
phenny jnthn: I'll pass that on when moritz is around.
PerlJam Radvendii: the problem for me is when it's like 1, 3, 5 ^... 13 Is tha the same as 3, 5, 7, 9, 11, 13 ? 22:49
Radvendii yes, but the whole _point_ is that i dont know whether it's bigger or not
if i did, i would use ^..
22:49 phenny left
benabik Radvendii: Use minmax to sort out the size, _then_ use ^.. 22:49
Radvendii PerlJam: yes, why wouldn't it be?
benabik: i could, but that doesn't sound a little silly to you? 22:50
PerlJam just seems weird is all
Radvendii PerlJam: why? it's non-inclusive
22:50 phenny joined
benabik "Non-inclusive" and "series" don't go together in my head 22:51
Radvendii: Given the potential confusions with what you're excluding with ^minmax, it doesn't seem silly at all.
Radvendii benabik: why not? how else would i do $a, $b ^... $c without knowing what $a and $b are?
22:52 overrosy left
benabik Non-inclusive series... "I need a set of numbers defined by these numbers that excludes one of them" 22:52
Radvendii i could do ($a, $b ... $c)[1..*-1]
rather ($a, $b ... $c)[1..*] 22:53
22:53 benabik left
Radvendii benabik: that's true, if you know what all of those numbers are. 22:53
22:53 overrosy joined
PerlJam nom: 22:54
nom: 1,3,5 ...^ 13,16,27
colomon In general, if you don't know what the numbers $a, $b. and $c are, then $a, $b ... $c will almost certainly be an infinite sequence
p6eval nom 3670b7: ( no output )
PerlJam blah
nom: say 1,3,5 ...^ 13,16,27 22:55
p6eval nom 3670b7: OUTPUT«1 3 5 7 9 11 16 27␤»
PerlJam I can't concile that with the proposed ^...
colomon Radvendii: I think this is when sorear++ would quite sensibly say, can you please tell us what you are actually trying to do, rather than the way you think it ought to be done?
Radvendii colomon: uless you know they are all integers
PerlJam Radvendii: eh?
Radvendii that was at his previous point 22:56
colomon Radvendii: not even then, unless you know that $c % ($b - $a) == 0
errr, ($c - $a) % ($b - $a) == 0, I mean
Radvendii what i want to do is $a ... $b except not include $a or $b. without knowing which is bigger 22:57
PerlJam Radvendii: you forgot "... in one operation" because you can certainly do that in multiple steps :) 22:58
(that ... was just an ellipsis) 22:59
Radvendii yes. but that seems like the kind of thing that should be doable, considering that ... doesn't care about order and ^..^ exists
doesn't it just seem like the task involves putting those two together?
colomon and you're sure $a and $b are integers?
Radvendii yes 23:00
they are coordinates
so yeah
(on a grid)
or, more appropriately, they are indicies
it's just as much of a problem to say @a[$b] 23:01
because $b might be 1.5
hm...
colomon @a[1.5] is perfectly legal
Radvendii nom: my @a = <a b c>; say @a[1.5]
p6eval nom 3670b7: OUTPUT«b␤»
Radvendii wth?
dalek p/bs: 36df449 | jnthn++ | t/serialization/01-basic.t:
Tests for P6num and P6str serialization.
p/bs: fd071cc | jnthn++ | src/6model/reprs/P6 (2 files):
Implement serialize and deserialize for P6num and P6str representations.
Radvendii that's bizzare
it just truncates 23:02
but reguardless, yes. i knwo they are integers.
masak that's not bizarre.
that's Perl.
colomon it floors, actually, I believe
masak right.
PerlJam That's just Perl since it was born
jnthn
.oO( That's totally floored! )
Radvendii colomon: isn't floor and truncate the same?
masak probably rounds towards 0 :)
colomon: it's impossible to tell since negative numbers are no longer allowed. 23:03
colomon Radvendii: not for negative numbers.
Radvendii also, when would i ever want to access @a[1.5] and _not_ be confused
masak good night, #perl6
colomon can't you (at least hypothetically) create an array with negative indices?
Radvendii o/
masak colomon: yes, when S09 is implemented.
masak sleeps 23:04
Radvendii nom: my @a; @a[-4]=3; say @a;
p6eval nom f9d635: OUTPUT«Cannot assign to a non-container␤ in block <anon> at /tmp/jBVoKsOCl7:1␤ in <anon> at /tmp/jBVoKsOCl7:1␤»
colomon Radvendii: @a[@a.elems / 2] where there are three elements in the array
PerlJam Radvendii: have a look at gist.github.com/21bbf1da86d0e495408b Are those the semantics you think should exist?
Radvendii indeed 23:05
23:05 slavik left
PerlJam the asymmetry bothers me 23:07
Radvendii yeah, i would also prefer it if the 19 were removed
but
i understand why that cannot be
PerlJam I could make a case for 1,3,5 ^...^ 13,16,19 --> 1 3 7 9 11 16 19
:-)
Radvendii just think of it as (1,3,5 ^...^ 13),16,19 23:08
because that's what it is
23:08 slavik joined
PerlJam Radvendii: convince TimToady and you're done modulo the implementation :) 23:11
23:11 zamolxes left
Radvendii modulo the implementation? 23:11
oh
got it
it can't be too hard... 1,3,5 ^... 13 is exactly the same as (1,3,5 ... 13)[1..*] 23:12
jnthn ugh, tired...'night o/
Radvendii jnthn: \o
o> <--- check out his muscles 23:14
PerlJam: am i wrong? ^^ up there? 23:15
23:16 xinming left
Radvendii nom: say (1,3,5 ... 13)[1..* 23:18
p6eval nom f9d635: OUTPUT«===SORRY!===␤Unable to parse postcircumfix:sym<[ ]>, couldn't find final ']' at line 2␤»
Radvendii nom: say (1,3,5 ... 13)[1..*]
p6eval nom f9d635: OUTPUT«3 5 7 9 11 13␤»
23:18 xinming joined
Radvendii nom: say (13...4)[1..*] 23:18
p6eval nom f9d635: OUTPUT«12 11 10 9 8 7 6 5 4␤»
Radvendii yup
nom: say (13...^4)[1..*]
p6eval nom f9d635: OUTPUT«12 11 10 9 8 7 6 5␤»
Radvendii nom: say (4...^4)[1..*] 23:22
p6eval nom f9d635: OUTPUT«␤»
Radvendii nom: say (4...^4) 23:23
p6eval nom f9d635: OUTPUT«␤»
Radvendii nom: say (4...4)
p6eval nom f9d635: OUTPUT«4␤»
Radvendii nom: say (4...4)[1..*]
p6eval nom f9d635: OUTPUT«␤»
Radvendii nom: say <a b c> «,» () 23:24
p6eval nom f9d635: OUTPUT«(timeout)»
Radvendii nom: say 'a' ~~ 0 23:25
p6eval nom f9d635: OUTPUT«Bool::True␤»
Radvendii so why does it convert the 'a' into a number and not the 0 into a string? 23:26
Timbus 0.ACCEPTS('a') makes sense to me 23:28
Radvendii oh. is that what ~~ calls? 23:31
Timbus yeah. 23:32
Radvendii so what about 'a' ~~ /0/. does that do /0/.ACCEPTS('a') ?
or is that something else...
Timbus i think that's what it does, but there's been a few issues regarding that 23:33
Radvendii oh
Timbus nom: say /0/.ACCEPTS('a')
p6eval nom f9d635: OUTPUT«#<failed match>␤»
Radvendii apparently
Timbus nom: say /\w/.ACCEPTS('a') 23:34
p6eval nom f9d635: OUTPUT«=> <a>␤␤»
Timbus so yeah thats how it works
Radvendii o-oh
so it's easy to overload ~~ 23:35
Timbus yes
Radvendii clever
:P
Timbus the problem with ~~ and matches is that $_ is overridden during the match, but sometimes people use magical values in their matches and submatches
Radvendii magical values? 23:36
Timbus theres also the issue of how ~~ and substitution matches should work (i think they should not work at all)
Radvendii i think it should return the string after substitution
Timbus yes, like previous match values and the like
Radvendii ah 23:37
Timbus anything tied to $_ i guess..
moritz wrote a blog entry about it but I can't find it
Radvendii but yeah, i hardly ever want to modify the string itself. and if i did i could use ~~= 23:38
yeah, i experienced recently when i wanted to do .grep: {$i !~~ $_} or something like that
nom: say (one(<a b c>) ~~ /a/) 23:41
p6eval nom f9d635: OUTPUT«This type cannot unbox to a native string␤ in <anon> at src/gen/Metamodel.pm:2686␤ in !cursor_init at src/stage2/QRegex.nqp:878␤ in method ACCEPTS at src/gen/CORE.setting:7559␤ in block <anon> at /tmp/Z4I_FuY25p:1␤ in <anon> at /tmp/Z4I_FuY25p:1␤»
Radvendii nom: say ((one(<a b c>) ~~ /a/).perl)
p6eval nom f9d635: OUTPUT«This type cannot unbox to a native string␤ in <anon> at src/gen/Metamodel.pm:2686␤ in !cursor_init at src/stage2/QRegex.nqp:878␤ in method ACCEPTS at src/gen/CORE.setting:7559␤ in block <anon> at /tmp/aSMmL4cfBj:1␤ in <anon> at /tmp/aSMmL4cfBj:1␤»
Radvendii nom: say ?(one(<a b c>) ~~ /a/) 23:42
p6eval nom f9d635: OUTPUT«This type cannot unbox to a native string␤ in <anon> at src/gen/Metamodel.pm:2686␤ in !cursor_init at src/stage2/QRegex.nqp:878␤ in method ACCEPTS at src/gen/CORE.setting:7559␤ in block <anon> at /tmp/4XUoXuhjMY:1␤ in <anon> at /tmp/4XUoXuhjMY:1␤»
Radvendii nom: say ?(one(<a b c>) eq 'a')
p6eval nom f9d635: OUTPUT«Bool::True␤»
Radvendii nom: say (one(<a b c>) eq 'a')
p6eval nom f9d635: OUTPUT«one(Bool::True, Bool::False, Bool::False)␤»
Radvendii is there any way to find out which of them was the 'one'? 23:43
nom: say (a ^^ 0) 23:45
p6eval nom f9d635: OUTPUT«===SORRY!===␤CHECK FAILED:␤Undefined routine '&a' called (line 1)␤»
Radvendii nom: say ('a' ^^ 0)
p6eval nom f9d635: OUTPUT«a␤»
23:46 molaf left, bbkr1 left, colomon left
Radvendii nom: 4 ...^ 4 23:47
p6eval nom f9d635: ( no output )
Radvendii nom: say 4 ...^ 4
p6eval nom f9d635: OUTPUT«␤»
Radvendii nom: say qx[pwd]
p6eval nom f9d635: OUTPUT«qx, qqx is disallowed in safe mode␤ in sub restricted at src/SAFE.setting:2␤ in sub QX at src/SAFE.setting:9␤ in block <anon> at /tmp/wiLcIEXBe3:1␤ in <anon> at /tmp/wiLcIEXBe3:1␤»
Radvendii figured :P 23:48
nom: say one(5,0) 23:49
p6eval nom f9d635: OUTPUT«one(5, 0)␤»
Radvendii nom: say +one(5,0)
p6eval nom f9d635: OUTPUT«one(5, 0)␤»
Radvendii nom: say ?one(5,0)
p6eval nom f9d635: OUTPUT«Bool::True␤»
Radvendii nom: say =one(5,0)
p6eval nom f9d635: OUTPUT«===SORRY!===␤Preceding context expects a term, but found infix = instead at line 1, near "one(5,0)"␤»
Radvendii nom: say -one(5,0)
p6eval nom f9d635: OUTPUT«one(-5, 0)␤»
Radvendii nom: say -(1,2,3)
p6eval nom f9d635: OUTPUT«-3␤» 23:50
Radvendii nom: say -« (1,2,3)
p6eval nom f9d635: OUTPUT«-1 -2 -3␤»
Radvendii Woah! 23:53
23:53 bbkr1 joined
Radvendii nom: say ?(one(0,0) !== 0) 23:53
p6eval nom f9d635: OUTPUT«Bool::True␤»
23:53 baest_ joined
Radvendii uh 23:53
what??
nom: say ?(one(0,0) != 0)
p6eval nom f9d635: OUTPUT«Bool::False␤»
Radvendii so somehow !== is different from !=
and != is more correct 23:54
23:55 baest left 23:56 molaf joined
Radvendii o-oh 23:57
is it that $a !== $b is the same as !($a == $b)
i think that's it
i dont know if that's good in this case...

Powered by Cro and the Raku® Programming Language.

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

Thank you!