»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'p6: say 3;' or rakudo:, std:, or /msg camelia p6: ... | irclog: irc.perl6.org | UTF-8 is our friend!
Set by masak on 12 May 2015.
00:02 vendethiel joined 00:19 anaeem1 left 00:26 vendethiel left 00:33 gcole left 00:34 kanl left 00:39 laouji joined 00:40 telex left 00:42 telex joined 00:43 laouji left
dalek kudo-star-daily: 91d64f4 | coke++ | log/ (8 files):
today (automated commit)
 00:45
01:04 Ben_Goldberg joined 01:06 BenGoldberg_ joined 01:07 aborazmeh joined, aborazmeh left, aborazmeh joined
[Coke] in backlog, but years ain't dates. they're years. 01:07
01:07 davido___ left 01:09 davido___ joined 01:10 Ben_Goldberg left 01:13 khw joined 01:14 rmgk is now known as Guest20935, rmgk_ joined, Guest20935 left, rmgk_ is now known as rmgk 01:16 vike1 left 01:23 vendethiel joined, colomon left 01:24 vike joined 01:25 gfldex left 01:29 leont left 01:33 Peter_R left 01:34 colomon joined 01:35 Peter_R joined 01:40 yeahnoob joined 01:45 vendethiel left 01:49 dayangkun joined 01:50 dayangkun left 01:54 roguelazer left 01:57 dayangkun joined 01:58 araujo_ joined 02:00 dayangkun left, vendethiel joined, dayangkun joined 02:01 araujo left 02:02 dayangkun left 02:04 dayangkun joined 02:07 TEttinger left 02:12 araujo_ left, araujo__ joined, araujo__ left, roguelazer joined 02:13 araujo__ joined 02:14 colomon left 02:17 noganex_ joined 02:18 atroxaper joined 02:19 colomon joined 02:20 noganex left 02:22 vendethiel left 02:23 atroxaper left 02:24 raiph joined, vendethiel joined 02:32 BenGoldberg joined 02:34 BenGoldberg_ left 02:40 geekosaur left 02:41 geekosaur joined 02:47 colomon left 02:48 tinyblak_ joined 02:50 vendethiel left 02:51 TEttinger joined 02:52 tinyblak left 03:02 kanl joined
kanl is "is copy" only applicable to scalars? or does it depend on .clone to copy-by-value? 03:04
03:06 colomon joined
kanl sub foo( @f is copy ) { @f.shift } my @f = 1 .. 3; foo( @f ); @f.say; 03:06
m: sub foo( @f is copy ) { @f.shift } my @f = 1 .. 3; foo( @f ); @f.say;
camelia rakudo-moar 377672: OUTPUT«5===SORRY!5=== Error while compiling /tmp/2jkzKeW2pJ␤Strange text after block (missing semicolon or comma?)␤at /tmp/2jkzKeW2pJ:1␤------> 3sub foo( @f is copy ) { @f.shift }7⏏5 my @f = 1 .. 3; foo( @f ); @f.say;␤ expecting any of:␤ …»
kanl sub foo( @f is copy ) { @f.shift }; my @f = 1 .. 3; foo( @f ); @f.say; 03:07
ok. 03:08
03:10 vendethiel joined
BenGoldberg m: (-> @f is copy { @f.shift })( 1 .. 3 ).say; 03:11
camelia rakudo-moar 377672: OUTPUT«1␤»
BenGoldberg m: (-> @f is copy { @f.shift })( 'a' .. 'c' ).say;
camelia rakudo-moar 377672: OUTPUT«a␤»
03:12 roguelazer left
BenGoldberg m: (-> @f { @f.shift })( 'a' .. 'c' ).say; 03:12
camelia rakudo-moar 377672: OUTPUT«Method 'shift' not found for invocant of class 'Range'␤ in block <unit> at /tmp/wObdPAtp49:1␤␤»
03:16 khw left
kanl so apparently 'is copy' does not depend on .clone.. 03:16
03:17 TimToady left 03:18 BenGoldberg left 03:19 kaare_ joined, TimToady joined 03:25 geekosaur left 03:26 geekosaur joined 03:30 nys left 03:34 vendethiel left 03:40 pierrot joined 03:42 vendethiel joined 04:16 laouji joined 04:21 laouji left 04:28 vendethiel left 04:34 diana_olhovik_ joined 04:37 vendethiel joined 04:47 [Sno] left 04:53 skids left 05:07 atroxaper joined 05:11 atroxaper left 05:19 aborazmeh left 05:20 vendethiel left 05:21 atroxaper joined 05:23 diana_olhovik_ left 05:27 vendethiel joined 05:46 araujo__ left 05:47 araujo__ joined 05:49 vendethiel left 05:50 araujo__ left, araujo joined 05:51 vendethiel joined 06:00 diana_olhovik_ joined, brrt joined 06:05 [Sno] joined 06:14 rurban joined 06:15 vendethiel left 06:16 diana_olhovik_ left 06:17 WABM9Q9 joined 06:18 brrt left, salva joined 06:20 anaeem1_ joined 06:21 zacts joined 06:26 kurahaupo left, kurahaupo joined, FROGGS joined 06:33 vendethiel joined
jdv79 why is the CUR part so complex? 06:35
masak morning, #perl6 06:36
jdv79: "the CUR part"? 06:37
jdv79 the finding and loading of comp units 06:38
(CompUnitRepo)
masak kanl: no, `is copy` does not clone the underlying object. it makes a copy of the reference. since a new reference to the same object will behave just like the old reference, I consider `is copy` on reference types to be pretty useless.
06:38 kanl left, btyler joined 06:39 kanl joined
masak jdv79: I think because it is trying to allow (a) Unicode module names that are not necessarily supported by the underlying file system, and (b) auths and versions. 06:39
I could be wrong though, because I've only read the spec, not been involved in actual CompUnitRepo work.
FROGGS masak is right 06:40
masak as far as I know, lizmat++ and FROGGS++ are driving in that.
FROGGS another aspect that makes it more complex is that distributions are also about media files, configuration files and other stuff
that also needs to be handled 06:41
jdv79 just feels a bit overdone/big - idk
oha masak, both is copy and is ro are misleading on that regard.
FROGGS jdv79: Perl 6's goal is to be prepared to evolve, and the CUR concept is very much sane IMO
masak: I'd vote for 'is copy' meaning 'is clone' always fwiw 06:42
06:43 brrt joined
jdv79 is S22 the only formal words on it? 06:43
FROGGS also S11 06:44
kanl cloning could be expensive though. one can still do sub foo( $foo is copy ) { my $bar = $foo.clone; ... } if he/she wants a proper by-copy, right? 06:45
FROGGS kanl: why do put an 'is copy' trait there at all?
kanl to want a 'local' effect, i guess? 06:46
if people don't want 'is copy' changed, maybe a 'is clone' can be added? 06:47
oha kanl, can i ask why people don't want it changed? (I honestly don't get the "local" effet) 06:48
jdv79 "sane" is quite subjective. RMI is same to some and its still generally a bad idea.
*sane
kanl i'm only guessing :)
jdv79 similarly, non-local CURs seems a bit out there
06:48 gfldex joined
kanl oha: 'local' as in p5. 06:49
jdv79 i kinda like my code in hand so to speak and not from the cloud
kanl { local $/ } etc.
oha kanl, yeah... but i still don't see how it would apply here. (you still have temp and let) 06:50
kanl true..
oha so in a way, i can see why someone "might" want "$arg is temp" 06:51
kanl m: class A { has $!a; submethod BUILD( Int :@a ){ $!a = [+] @a } }; my $foo = A.new( a => [ 1 .. 10 ] ); 06:52
camelia rakudo-moar 377672: OUTPUT«Type check failed in binding @a; expected 'Positional[Int]' but got 'Array'␤ in submethod BUILD at /tmp/xB6Df2y1sM:1␤ in block <unit> at /tmp/xB6Df2y1sM:1␤␤»
nine jdv79: but there are use cases where it would come in handy: metacpan.org/pod/Object::Remote
kanl isn't Array Positional?
dalek rl6-most-wanted: ad24f13 | rurban++ | most-wanted/modules.md:
Update modules.md

add Unicode::LineBreak east asian languages need access to the linebreak properties See stackoverflow.com/questions/3125050...nt-columns		 06:53
rl6-most-wanted: d3adbf4 | FROGGS++ | most-wanted/modules.md:
Merge pull request #4 from rurban/patch-1

Update modules.md
06:54 brrt left 06:55 raiph left
kanl so how so i properly do type constraints on List/Array? 06:56
should ^^
06:58 domidumont joined 07:01 rindolf joined
moritz well, you can do it with Int @a, but then one has to pass in an Array[Int], not an Array 07:03
kanl but BUILD wants named parameters no? 07:04
moritz www.nntp.perl.org/group/perl.perl6....g2155.html
kanl: so?
you can pass an Array[Int] to a named parameter too
kanl m: class A { has $!a; submethod BUILD( @a ){ $!a = [+] @a } }; my $foo = A.new( [ 1 .. 10 ] ); 07:05
camelia rakudo-moar 377672: OUTPUT«Default constructor for 'A' only takes named arguments␤ in block <unit> at /tmp/736BpXDVWg:1␤␤»
kanl Int or not, i'm not getting pass on BUILD :/ 07:06
moritz if you want to pass a positional param to .new, you have to define your own .new
kanl ah, ic. thanks! 07:07
moritz m: class A { has $!a; method new($a) { self.bless(a => $a) }; submethod BUILD( @a ){ $!a = [+] @a } }; my $foo = A.new( [ 1 .. 10 ] );
camelia rakudo-moar 377672: OUTPUT«Too few positionals passed; expected 2 arguments but got 1␤ in submethod BUILD at /tmp/vqwLfLmZcV:1␤ in method new at /tmp/vqwLfLmZcV:1␤ in block <unit> at /tmp/vqwLfLmZcV:1␤␤»
moritz m: class A { has $!a; method new($a) { self.bless(a => $a) }; submethod BUILD( :@a ){ $!a = [+] @a } }; my $foo = A.new( [ 1 .. 10 ] );
camelia ( no output )
moritz m: class A { has $!a handles 'gist'; method new($a) { self.bless(a => $a) }; submethod BUILD( :@a ){ $!a = [+] @a } }; my $foo = A.new( [ 1 .. 10 ] ); say $foo; say $foo.^name; 07:08
camelia rakudo-moar 377672: OUTPUT«55␤A␤»
moritz kanl: but think twice before you do it; it makes subclassing much harder
07:08 TimToady left 07:09 RabidGravy joined
moritz kanl: see also doc.perl6.org/language/objects#Obje...nstruction 07:10
07:11 rindolf left
kanl m: class A { has $!a; submethod BUILD( Array[Int] :@a ){ $!a = [+] @a } }; my $foo = A.new( a => [ 1 .. 10 ] ); 07:12
camelia rakudo-moar 377672: OUTPUT«Type check failed in binding @a; expected 'Positional[Array[Int]]' but got 'Array'␤ in submethod BUILD at /tmp/hlzGMjdpGu:1␤ in block <unit> at /tmp/hlzGMjdpGu:1␤␤»
kanl ok. i'll just forgo the type contraint for now. 07:14
07:15 darutoko joined
moritz m: class A { has $!a; submethod BUILD( Array[Int] :@a ){ $!a = [+] @a } }; my $foo = A.new( a => (my Int @ = 1 .. 10 ) ); 07:15
camelia rakudo-moar 377672: OUTPUT«Type check failed in binding @a; expected 'Positional[Array[Int]]' but got 'Array[Int]'␤ in submethod BUILD at /tmp/t4BfSi0eNx:1␤ in block <unit> at /tmp/t4BfSi0eNx:1␤␤»
moritz m: class A { has $!a; submethod BUILD( Int :@a ){ $!a = [+] @a } }; my $foo = A.new( a => (my Int @ = 1 .. 10 ) );
camelia ( no output )
moritz that's what I wrote in www.nntp.perl.org/group/perl.perl6....g2155.html 07:16
07:17 vendethiel left
kanl moritz: i c now. good rant :) 07:18
my $a = bag( -Inf .. Inf ); my $b = bag( 1 .. 3 ); my $c = bag(); say $a(+)$b(+)$c; ## this won't work 07:21
1. Inf is not properly guarded
oha but wasn't kanl example about requiring proper types, not coercing it?
kanl 2. the (whatever) operator requires leading whitespaces.
3. can't mix the operators without grouping ()
07:22 vendethiel joined
kanl m: my $a = bag( 1 .. 4 ); my $b = bag( 1 .. 3 ); say $a(+)$b; 07:23
camelia rakudo-moar 377672: OUTPUT«5===SORRY!5=== Error while compiling /tmp/hlD7eRLLzW␤Prefix + requires an argument, but no valid term found␤at /tmp/hlD7eRLLzW:1␤------> 3 .. 4 ); my $b = bag( 1 .. 3 ); say $a(+7⏏5)$b;␤ expecting any of:␤ prefix␤»
kanl m: my $a = bag( 1 .. 4 ); my $b = bag( 1 .. 3 ); my $c = bag(); say $a (+) $b (-) $c;
camelia rakudo-moar 377672: OUTPUT«5===SORRY!5=== Error while compiling /tmp/YpB9DyyE64␤Operators '(+)' and '(-)' are non-associative and require parentheses␤at /tmp/YpB9DyyE64:1␤------> 3( 1 .. 3 ); my $c = bag(); say $a (+) $b7⏏5 (-) $c;␤ expecting any of:␤ i…»
moritz oha: kanl passed [1, 2, 3] to a 'Int @a' parameter, and was surprised it didn't work
oha: so that's exactly the case of expecting coercion where none is
07:27 anaeem1__ joined, mr-foobar joined 07:28 anaeem1_ left
jdv79 yeah ok. remote and local are distinct enough to warrant a hard distinction 07:29
imho
i guess we'll see. 07:30
if non-local CUR is just a population method for local CUR then its less insane to me 07:33
oha moritz, I wonder if things would become easier if this would be possible: f (Array[Int] $a) { ... }; f([1,2,3]);
actually, more like 'my Array[Int] $a = [1,2,3];' 07:35
moritz oha: they would, at least on the surface; but that would also totally mess up multi dispatch
oha moritz, are you impling that [1,2,3] should become Array[Int], but I'm saying that it might get "narrow typed" if no other candidates are present. yet i see your point 07:36
moritz oha: I'm not implying that [1, 2, 3] should become Array[Int] 07:38
oha: more that an 'Int @a' param should behave more like 'Int() @a' (coercive) 07:39
oha moritz, so couldn't it work like for f(Int) vs f(1)?
moritz no
nine jdv79: CUR is not just about remotes. It also allows for simpler implementation of loading from archive files (think .jar) or even generating modules on-the-fly.
moritz or, even more to the point, like 'Array[Int]() \a'
07:40 abraxxa joined
jdv79 my perhaps naive point is that including remotes seems like a bad idea 07:40
not that CUR is a bad concept in general
07:41 amurf left
jdv79 also the global db seems an even worse part 07:41
lizmat jdv79: which global db ?
jdv79 isn't there some sort of repo db that's maintained 07:42
i don't know if it has a name
if it does i donn't know it
nine jdv79: alas, none of these are really new possibilities. You can already do this in Perl 5 by pushing a CODEREF to @INC 07:43
lizmat jdv79: check out S22:114
07:44 vendethiel left
jdv79 so what happens when this recomendation mgr yields a candidate that is non-local? 07:52
07:53 laouji joined
RabidGravy marnin! 07:58
07:58 laouji left
jdv79 late aft here 07:58
is 1700 still considered aft of is that eve?
*or
08:08 cognominal joined 08:14 dakkar joined
RabidGravy jdv79, depends on whether I was at work or not really ;-) 08:21
jdv79 what if you are not currently employed? 08:23
dalek kudo/nom: eaaaff2 | lizmat++ | src/core/List.pm:
Some more List.splice tweaking
 08:27
ast: 4405592 | lizmat++ | S32-array/splice.t:
Work in progress on splice tests
 08:28
lizmat afk for a few hours&
08:36 vendethiel joined 08:37 g4 joined, g4 left, g4 joined 08:38 kurahaupo left 08:47 sivoais left 08:49 sivoais joined 08:51 Woodi left, espadrine joined 08:52 mr-foobar left 08:54 mr-foobar joined 08:58 vendethiel left 09:03 Woodi joined 09:06 atroxape_ joined, atroxaper left, vendethiel joined 09:07 cognominal left 09:09 sunnavy left 09:10 sunnavy joined 09:13 bin_005 joined 09:24 ab6tract joined
ab6tract o/ #perl6 09:26
.tell [Coke] the conversation about years clearly ended with everyone agreeing that it should be a separate class 09:27
yoleaux ab6tract: I'll pass your message to [Coke].
09:29 amurf joined 09:34 amurf left 09:37 yeahnoob left 09:38 telex left 09:40 telex joined
ab6tract but I think you all are not doing much date related things to think that subset Year of Int is a reasonably solution, or that Years shouldn't smartwatch Dateish 09:44
*smartmatch 09:45
09:46 rindolf joined
FROGGS I do quite some date related things... I calculate fees for the employees of the company I am working for... 09:53
that is, if someone works at night for example, (s)he gets extra fees, depending on the contract atc
etc*
09:57 daxim joined
ab6tract FROGGS: sorry, the "you all" was a little overbroad there.. that really just meant leont and [Coke] 09:59
FROGGS ab5tract: np 10:03
10:03 kanl left, kanl joined
FROGGS what I'd like to see is a port of DateTime::Span(Set) etc 10:03
10:09 laouji joined 10:11 TEttinger left
lizmat Camelia growing up: toons.gotblah.com/archive/wulffmorg...9-0301.jpg :-) 10:17
masak FROGGS: what's a Span? is it something like "one year"? 10:21
10:23 rindolf left 10:29 rurban left 10:33 laouji left 10:34 laouji joined 10:44 aborazmeh joined, aborazmeh left, aborazmeh joined 10:46 dayangkun left 10:49 TimToady joined 10:54 laouji left, laouji joined
FROGGS masak: aye 10:56
bbl 10:57
10:57 virtualsue joined 11:03 FROGGS left
ab6tract FROGGS: sounds a bit like TimToady's suggestion that what I'm talking about is DateInterval 11:03
11:10 user3 joined
user3 Rakudo 2015.03 on Linux and Windows 11:14
The synopsis for command line option -e says "Arguments following ... -e are always parsed as a list of values, even if they look like valid options." 11:15
Then why does this work perl6 "-e say 10-5" but not this perl6 -e say 10-5
moritz user3: because the second case executes the program with the text 'say' and the command-line argument '10-5' 11:16
a better wording would be "arguments following -e ... are always passed to the Perl 6 program being executed"
11:19 rurban joined
jnthn afternoon, #perl6 11:23
yoleaux 7 Jul 2015 05:35Z <lizmat> jnthn: Could you have a look as to why f5ebbf480c5936 is causing S17-supply hangs
7 Jul 2015 06:04Z <lizmat> jnthn: empirically, cbff68a fixes the hanging S17-supply tests
lizmat jnthn o/ 11:24
how is the red tape?
jnthn lizmat: We're not doing that red tape this time...due to...uh...other red tape... 11:25
11:25 rurban1 joined
lizmat argh 11:25
:-(
hope you're enjoying Kyiv and each other nonetheless
jnthn (Nothing out of the ordinary. Just things taking more time than we'd imagined or lacking the fast-paths we'd imagined.) 11:26
Each othr certainly. And Kyiv, yes when the temperature is agreeable to venturing outside... :)
jnthn has been a bit ill also :/
lizmat hope you're feeling better now 11:27
jnthn In the "improved" sense, at least. 11:28
I've no idea why the .assuming work might cause hangs, unless something somewhere is using .assuming
11:28 user3 left
jnthn Anyways, right now I'm feeling well enough to try doing something useful :) 11:29
And $wife++ just brought a cup of tea to help :)
I've not kept up with backlog...is there anything of note anybody is blocking on? 11:30
lizmat jnthn: that's the thing, nothing is using .assuming in core, afaik 11:32
jnthn: I think timotimo and brrt could use some input on #moarvm
11:32 bin_005 left
jnthn .tell hoelzro re RT #125574, obtaining a type object means committing to a REPR, meaning that I can't actually fix that...I should, however, make it tell you it's not going to work rather than silently fail. I'll do that. 11:33
yoleaux jnthn: I'll pass your message to hoelzro.
lizmat other than that, it has been pretty quiet (for what I've seen of the backlog)
jnthn OK, thanks :)
lizmat ah, I have one question though
is there a specific reason why splice(@a,...) delegates to @a.splice(...) rather than the reverse ? 11:34
jnthn Because we long had a policy of sub delegates to method.
e.g. prefix:<?> delegates to .Bool
lizmat and why was that policy? 11:35
I'm thinking splice() specifically here
from a performance optimizing point of view, wouldn't it be better to use splice() and have it all the logic
and @a.splice delegate to the sub ?
jnthn Because objects are more natural behavior extenion points.
*extension
e.g. it's more natural to subclass and override, or implemnet a new type, than it is to export a multi candidate to change behavior 11:36
RabidGravy m: use Test; my $b = Buf.new; isa-ok($b, Buf); # er what's all this about?
camelia rakudo-moar eaaaff: OUTPUT«not ok 1 - The object is-a 'Buf'␤␤# Failed test 'The object is-a 'Buf''␤# at /tmp/0xDoqNyPhy line 1␤# Actual type: Buf␤»
jnthn m: say Buf.HOW.^name
camelia rakudo-moar eaaaff: OUTPUT«Perl6::Metamodel::ParametricRoleGroupHOW␤»
jnthn RabidGravy: Buf ain't a class.
RabidGravy ah 11:37
jnthn lizmat: So really it's a case of what we expect people to implement if they want to create their own custom Stringy thing, or Listy thing, for example 11:38
lizmat ok, so we see subs like splice() more as a delegator then as a processor
RabidGravy is($b.^name, Buf.^name) then ;-) 11:39
jnthn RabidGravy: Smart-match is the natural way to write it 11:40
RabidGravy so it is, cheers 11:44
11:46 breinbaa1 joined 11:50 atroxape_ left
dalek kudo/nom: ed8270a | jnthn++ | src/core/Supply (2 files):
Move 'is repr' to the predeclaration.
 11:52
11:54 anaeem1__ left 11:55 leont joined
lizmat jnthn: aaah, a bug fixed :-) 11:55
an unfound one, right ?
jnthn Right, found by the stricter checking I just put in :) 11:56
kanl :w 11:58
oops 11:59
12:00 ab6tract left 12:02 breinbaa2 joined
kanl is there a way to prevent something like any( 1..Inf ) from eagerly running OOM, other than being very careful? 12:03
jnthn Not really; a Junction has to have all the values 12:05
And flattens its args
Maybe GLR will have it not doing so, I don't know. 12:06
oha using a quantum computer would do the trick, tho :)
12:06 laouji left
kanl jnthn: thanks 12:07
dalek kudo/nom: a51ac08 | jnthn++ | src/ (2 files):
Catch too-late application of REPR trait.

Fixes RT #125574.		 12:08
ast: 87560fd | jnthn++ | S32-exceptions/misc.t:
Test for RT #125574.
masak oha: no, I don't think that's what quantum computers do. 12:11
(I know you were joking, but... it's not true) :P
oha :)
kanl you'd still need infinitely many q-bits :)
masak yeah 12:12
oha even for really small values of Inf? :)
kanl Inf++ 12:13
timotimo hm, synopsebot isn't here? 12:14
lizmat timotimo: indeed
dalek kudo/nom: 341f6e5 | jnthn++ | src/Perl6/World.nqp:
Look harder for existing *%_ or | pre-autogen.

Fixes RT #125513.		 12:16
ast: 1cda041 | jnthn++ | S12-class/interface-consistency.t:
Test for RT #125513.
 12:17
kanl what's the rationale for bag/set to have custom made arithmetic operators, instead of overloading the Numeric ones? 12:19
jnthn kanl: Example of an arithmetic operation on a set/bag? 12:20
kanl sorry, i mean set operations. 12:21
(+) (-) (|) etc.
jnthn Well, that's the answer though. :) In Perl we avoid overloading *semantics* onto the same symbol.
kanl they seem to me unwieldy and introduce parsing complexity
jnthn There's two sane ways to make a language with operators. Etiher you statically know your types and overload operator meanings, or you have dynamic types and operators always have consistent semantics. 12:22
That's why in Perl you never have to guess if + means numeric addition or concatenation, or whether == with do numeric or string equality. 12:23
This is just another application of that overarching design rule.
12:24 muraiki joined 12:25 brrt joined
brrt \o 12:26
itz_ does any here have access to update modules running under rakudobrew user on hack.p6c.org? 12:27
kanl jnthn: i see your point, thank you. i suppose i can always overload +/- etc. if i really wanted to :)
JimmyZ itz_: I can't switch to rakudobrew user ... 12:28
12:31 virtualsue left
JimmyZ itz_: github.com/perl6/infrastructure-do...6c.org.pod 12:31
jnthn m: use Test; my $m = "AAA" ~~ /$<letter>=(A)**3/; is +$<letter>, 3; 12:33
camelia rakudo-moar ed8270: OUTPUT«ok 1 - ␤»
jnthn m: use Test; my $m = "AAA" ~~ /$<letter>=(A)**{3}/; is +$<letter>, 3;
camelia rakudo-moar ed8270: OUTPUT«not ok 1 - ␤␤# Failed test at /tmp/iH6vbs_Lcf line 1␤# expected: '3'␤# got: (Failure)␤»
jnthn hm, bet I know why :)
timotimo jnthn: did you notice the thing where substr on a string built with the x operator can give bogus results? 12:34
masak by the way, speaking of `is copy`, I agree with the people who say that it's currently useless for reference types, and who feel inclined to give it cloning semantics... but I would like the reaction of TimToady, pmichaud_, jnthn, moritz (or an appropriate subset) before going ahead with that.
timotimo i think it's about concatenating a character at the end of a single character x'd multiple times and then trying to get one of the x'd characters together with the added character results in the x'd character twice
itz_ timotimo: can you update Text::VimColour on hack.p6c.org rakudobrew please? :)
timotimo let's see 12:35
… 12:36
how do i sudo
well, Vim::TextColour is already in MODULES, i wonder why it doesn't install
JimmyZ sudo su rakudobrew - 12:37
timotimo JimmyZ: that's what i ended up doing
itz_ it is installed . I need a bumped version
timotimo ah!^
m) 12:38
i can't install it if i try to have Vim::TextColour
itz_ sorry I don't understand
timotimo ==> Successfully installed Text::VimColour
i creatively typo'd the module name 12:39
to such an extent that panda didn't know how to suggest the right one :)
12:39 mr-foobar left
itz_ ah 12:39
kanl m: say say any( 1..3 ) + any( 2..5 );
camelia rakudo-moar ed8270: OUTPUT«any(any(3, 4, 5, 6), any(4, 5, 6, 7), any(5, 6, 7, 8))␤True␤»
kanl m: say say any( 1..3 ) ~ any( 2..5 );
camelia rakudo-moar ed8270: OUTPUT«any(1, 2, 3)any(2, 3, 4, 5)␤True␤»
itz_ maybe perl6 should allow several versions of English spelling in module names :)
timotimo: ty 12:40
hoelzro morning #perl6
yoleaux 11:33Z <jnthn> hoelzro: re RT #125574, obtaining a type object means committing to a REPR, meaning that I can't actually fix that...I should, however, make it tell you it's not going to work rather than silently fail. I'll do that.
timotimo maybe there shoud be a unicode codepoint for ou that casefolds into o? :)
hoelzro jnthn: thanks for having a look! that solution is fine with me 12:41
dalek pan style="color: #395be5">perl6-examples: 8432216 | (Steve Mynott)++ | / (2 files):
enable Text::VimColour
 12:42
12:42 xinming_ joined
kanl why not any(any(12, 13, 14, 15) ... in the case of any(...) ~ any(...)? 12:43
12:44 rindolf joined 12:46 xinming left
jnthn timotimo: Yes, I'm aware of that issue 12:46
timotimo: though had forgotten about it until you just reminded me... 12:47
12:47 mr-foobar joined 12:49 rurban1 left
lizmat m: my @a; say @a.splice(1,1).perl; say @a.perl # seems splice has some unwanted vivification behaviour 12:49
camelia rakudo-moar ed8270: OUTPUT«[]<>␤[Any]<>␤»
12:50 atroxaper joined 12:55 atroxaper left 13:00 g4 left
timotimo i don't know why it just popped into my mind 13:04
13:10 VinceDee joined
timotimo loadbytecode.c has another 13:10
jnthn Maybe sudden optimism that I'm smart enough to debug it today... :)
timotimo oops, ww
13:10 mr-foobar left 13:11 dakkar left, mr-foobar joined
timotimo to be honest, i'd not point you at any more bugs myself until the code-gen thing for local/localref is done, because i'm interested in that :P 13:11
13:11 dakkar joined
timotimo hope i'm not nagging too much 13:12
lizmat is starting to question non-infinite ranges not autovivifying slices more and more
jnthn did after a few attempts manage to fix le regex engine bug with dynamic quantifiers and captures (subject to spectesting...)
13:13 asdfqerzcvasdfqw joined
asdfqerzcvasdfqw hello 13:13
timotimo hi there
13:14 asdfqerzcvasdfqw left 13:17 mr-foobar left 13:18 g4 joined, g4 left, g4 joined 13:21 vendethiel left
lizmat m: my @a = ^2; say @a.splice(1,3,3,4).perl; say @a.perl # this feels incorrect 13:24
camelia rakudo-moar 341f6e: OUTPUT«[1]<>␤[0, 3, 4]<>␤»
lizmat sort of expected [0,Any,Any,3,4]
afk&
looks like P5 has a different problem there: 13:25
$ perl -E 'my @a = 0,1; say splice(@a,1,3,3,4); say @a'
034
aka, the 1 is not returned
dalek p: cb78f9e | jnthn++ | src/QRegex/P6Regex/Actions.nqp:
Account properly for dynquant in capture handling.

We sometimes ended up with a clearly quantified (by **) capture only capturing once and the whole match.		 13:26
lizmat oops, differetnt P5 issue altogether
afk for a while
dalek kudo/nom: 6a000f8 | jnthn++ | tools/build/NQP_REVISION:
Bump NQP_REVISION for regex dynquant/capture fix.
 13:27
13:27 skids joined
dalek ast: 302b5c2 | jnthn++ | S05-metasyntax/repeat.t:
Test for RT #125521.
 13:27
kudo/nom: 261accb | lizmat++ | src/core/List.pm:
Handle splice with offset past end
 13:28
cdc Hello #perl6! 13:29
timotimo BBL
13:29 vendethiel joined
cdc I'd like to format -1 as an uint8 hexadecimal number, something like "-1.fmt('%x')" but that returns "0xff". 13:30
please, could you tell me how I should proceed?
jnthn hoelzro: re 125531, do you have any concrete ideas on what you'd like to happen? 13:31
oha cdc: (1 % 256).fmt() ?
jnthn (I agree it's LTA, not immediately got a feeling for right way forward)
oha cdc, (-1 % 256) actually 13:32
cdc m: (-1 % 256).fmt('%x').say
camelia rakudo-moar 341f6e: OUTPUT«ff␤»
moritz m: say (-1).fmt('%x')
camelia rakudo-moar 341f6e: OUTPUT«-1␤»
cdc oha: Thanks!
moritz: do you think it's a bug? 13:33
moritz cdc: I don't think so; %x just says hex, not unsigned hex, does it?
I mean, how would even know if it should say ff or ffff or fffffffff?
cdc m: (-10).fmt('%x') 13:34
camelia ( no output )
cdc m: (-10).fmt('%x').say
camelia rakudo-moar 341f6e: OUTPUT«-a␤»
cdc moritz: I don't know :) 13:35
dalek ast: 60ffb35 | lizmat++ | S32-array/splice.t:
Some more splice tests, still work in progress
 13:39
lizmat really afk for a while now
13:39 cognominal joined
dalek ast: 6d6c4fb | jnthn++ | S12-introspection/meta-class.t:
Add tests covering RT #80694.
 13:46
13:51 vendethiel left 13:52 Begi joined
Begi Hey ! I began with Perl6. I've code a little GuessNumber. Please, can you tell me what can I improve ? Thanks a lot ! pastebin.com/STCzpv7Z 13:53
tadzik looks good, though you may want to use loop {} instaed of while 1 {} :) 13:54
also $attempts++ is probably more suited over += 1 here 13:55
PerlJam or maybe even for 1..* -> $attempts { ... }
DrForr You can also remove the () around your if() statements as well, and $attempts++ # guh, like tadzik said.
13:58 vendethiel joined
PerlJam Begi: I guess you've never run this program? 13:58
13:58 zakharyas joined
DrForr Because of the $attemps tyops? :) 13:58
PerlJam yeah
Begi Yes , I hesitated between loop and while. I'll watch that. Thanks again 13:59
In fact, I've code in French. When I post it, I've change the variable names
PerlJam Begi: you won't hurt our feelings if your variables have french names :) 14:00
14:00 notjack joined
Begi Oh yes, I've forget a t =) 14:00
Just a question, do you have another exercise to do ? 14:01
PerlJam Begi: hangman? 14:02
Begi: or take your existing code and make it more DRY
Begi Hangman, good idea ! Make it more DRY, how can I do it ? 14:03
14:04 FROGGS joined
PerlJam Begi: you can get rid of the repeated $attempts += 1 by using a for loop that keeps track of the attempts as I said, or keep the loop the same and add a NEXT phaser that handles incrementing $attempts and outputting their number. 14:04
cdc Begi: loop (my $i = 0;; $i++) { state $r = round 100.rand; given prompt('? ') { when * < $r { say "+, $i" }; when * > $r { say "-, $i" }; when * { say "=, $i"; last } } } 14:05
jnthn m: given 2 { when * > 1 { when 2 { say 'here' }; say 'hmm'; } say 'oh'; }
camelia rakudo-moar 261acc: OUTPUT«5===SORRY!5=== Error while compiling /tmp/CHg8nqRyuw␤Strange text after block (missing semicolon or comma?)␤at /tmp/CHg8nqRyuw:1␤------> 3 1 { when 2 { say 'here' }; say 'hmm'; }7⏏5 say 'oh'; }␤»
cdc Begi: a shorter version :)
jnthn m: given 2 { when * > 1 { when 2 { say 'here' }; say 'hmm'; }; say 'oh'; } 14:06
camelia rakudo-moar 261acc: OUTPUT«here␤»
PerlJam Begi: or use some of the "tricks" that cdc mentions :)
jnthn m: for 2 { when * > 1 { when 2 { say 'here' }; say 'hmm'; }; say 'oh'; }
camelia rakudo-moar 261acc: OUTPUT«here␤»
jnthn m: { $_ = 2 { when * > 1 { when 2 { say 'here' }; say 'hmm'; }; say 'oh'; } }
camelia rakudo-moar 261acc: OUTPUT«5===SORRY!5=== Error while compiling /tmp/AZdgLDSCUM␤Unexpected block in infix position (missing statement control word before the expression?)␤at /tmp/AZdgLDSCUM:1␤------> 3{ $_ = 27⏏5 { when * > 1 { when 2 { say 'here' }; s␤ expectin…»
jnthn m: { $_ = 2; { when * > 1 { when 2 { say 'here' }; say 'hmm'; }; say 'oh'; } }
camelia rakudo-moar 261acc: OUTPUT«here␤»
jnthn m: { $_ = 2; { when * > 1 { when 2 { say 'here' }; say 'hmm'; }; say 'oh'; } }; say 'yay';
camelia rakudo-moar 261acc: OUTPUT«here␤yay␤»
ggoebel made a small donation to perl6 development fund...
PerlJam ggoebel++
jnthn ggoebel++
m: sub foo { $_ = 2; { when * > 1 { when 2 { say 'here' }; say 'hmm'; }; say 'oh'; } }; foo; say 'here' 14:07
camelia rakudo-moar 261acc: OUTPUT«here␤here␤»
ggoebel its the least I can do
14:07 anaeem1_ joined
jnthn Of the camelia evals I just did with when, does anybody have any objections to the semantics? 14:07
It turns out some of the details of the semantics in S04 are out of line with what we've long implemented, and what is being usefully used in the wild. 14:08
This relates to RT #71368, which a while back I tried to resolve by implementing the suggested semantics in the ticket and S04 14:09
It turns out that this didn't play well with various useful code/patterns in the wild.
So, I'm going to get S04 in line with what is actually implemented, since that seems to be working out well enough. 14:10
ggoebel might be able to give more later. It would be appreciated if there were some on-going updates... progress toward monthly and annual goals and the like breaking out one time and reoccuring donation levels. It would be nice to hear how much needs to be raised by July 31 to keep jnthn++ on 50% time next month.
PerlJam jnthn: Seems fine to me. 14:11
I was worried about one of them, but it turned out I'd misread the position of a closing }
jnthn ggoebel: Anything specific you're looking for on upates? I try to blog every 1-2 weeks with technical reports and recently did a summary of what got done in the first 3 months at a high level, at 6guts.wordpress.com/ 14:12
ggoebel: I also have a detailed time sheet that I'm fine with being public.
ggoebel jnthn: I'm content with your updates. I'm more concerned with updates from the perl6 foundation on your funding status and whether we're moving toward something that is sustainable. -And doesn't have you guessing from month to month whether or not the next month will be funded. 14:13
hoelzro jnthn: I think it would be nice if it said "Cannot assign to a readonly variable $/"
perhaps including the fact that ~~ is responsbiel
*responsible
ggoebel when I made a donation there was no where to leave a comment for me to express why I was giving or what I wished the funds to support 14:14
different people are motivated by different things
14:14 notjack left 14:15 anaeem1_ left
dalek ast: 5aa33ab | jnthn++ | S04-statements/when.t:
Toss tests for semantics we've decided against.

We won't try to statically work out whether a "when"/"default" is a valid thing, since it pretty much always can be, and folks have come to make good use of those semantics.		 14:16
vendethiel jnthn: what's the usecase for non-`given`'d `default`? 14:18
jnthn vendethiel: Documentary, just like given'd default...
14:18 anaeem1 joined
jnthn vendethiel: Note that, short of introducing a lexical scope, the behavior of the code in default is as if the default block wasn't there at all. 14:19
Well, and any code after it is dead
m: given 1 { when * > 2 { say 'big number' }; default { say 'here' } } 14:20
camelia rakudo-moar 261acc: OUTPUT«here␤»
jnthn m: given 1 { when * > 2 { say 'big number' }; say 'here' }
camelia rakudo-moar 261acc: OUTPUT«here␤»
vendethiel m: for ^5 { when * > 10 { say 'big number' }; default { say 'here' }; say 'ded'; }
camelia rakudo-moar 261acc: OUTPUT«here␤here␤here␤here␤here␤»
14:21 anaeem1 left
ggoebel I'd like to give a regular monthly or annual donation. But my motivation is to support developers... not the perl foundation. 14:21
skids Re: RT#125537 it looks like I overgolfed that so the fix for that did not fix what I had hoped. Should this work?: 14:22
m: :(::T $a, T $b).params[1].type.WHICH.say
camelia rakudo-moar 261acc: OUTPUT«Method 'WHICH' not found for invocant of class 'T'␤ in block <unit> at /tmp/8B9EXitm_e:1␤␤»
PerlJam ggoebel: you make me wonder now how much of the donated funds end up in the hands of developers vs how much end up paying for the overhead of TPF. 14:23
jnthn skids: Not really...it's senseless to do anything on a type var 14:25
14:25 rindolf left
jnthn skids: We could improve the error 14:25
Somehow... :)
skids :-) So what I really need is to be able to compare .type to Mu. Which can't use ===. 14:26
moritz Mu ~~ .type ? 14:28
14:28 AlexDaniel joined
jnthn You'll probably need to handle type vars explicitly 14:28
14:28 VinceDee left
cdc I tried to reduce Begi's program again, but I hit something strange: 14:29
my $r = 100.rand.round; while prompt('? ') Rcmp $r { say "$^x ($r)" }; say 'OK'
skids No it's more like $type1 === $type2 and either could be Mu, Nil, or a capturing type.
cdc "$^x" is always "More"
jnthn cdc: Try R<=>
moritz right, prompt returns a Str
jnthn Yes, that's what's up. 14:30
cdc jnthn: Thanks! It works :)
my bad
jnthn <=> forces numeric comparsion
cdc Begi: my $r = 100.rand.round; while prompt('? ') R<=> $r { say $^x }; say 'OK' 14:31
ggoebel PerlJam: I assume it is all operating in good faith. non-profit organization need funds too. And there is a lot of information on their website if you go digging for it. However, I was surprised that no where in the process of giving a donation did anyone think to ask why I was doing so, nor offer to bring me into the communications fold via an email list, etc.
Ulti if I do "somepath" .= IO and the somepath doesn't exist what is that failure because it doesn't appear to be an exception I can catch 14:32
m: my $file = "oogabooga"; try { $file .= IO; CATCH { say "getting here" } } 14:33
camelia ( no output )
14:34 anaeem1_ joined 14:36 colomon left 14:37 colomon joined
dalek ast: c43e6a7 | jnthn++ | S04-statements/when.t:
More tests for when/default.

To exercise semantics update coming in S04 design doc.		 14:38
14:39 atroxaper joined
ugexe IO doesnt give you a file handle 14:41
14:43 atroxaper left
ugexe m: my $file = $?FILE.IO; say "{$file.e} | {$file.d} {$file.f}"; say $file.slurp; 14:44
camelia rakudo-moar 261acc: OUTPUT«True | False True␤my $file = $?FILE.IO; say "{$file.e} | {$file.d} {$file.f}"; say $file.slurp;␤»
14:46 rindolf joined
cdc is "default" same as "when *"? 14:46
Ulti ugexe huh? I'm saying it complains when the file doesn't exist but doesn't give me a chance to do anything about that 14:47
you can call .lines which opens that path
or slurp
but I dont even get that far because I want to test if the file exists but it errors on that 14:48
ugexe thats what the $file.e shows you
Ulti I get that I can do this some other way I was just wondering what exactly is happening since its not a normal exception
hoelzro jnthn: thanks for that is repr fix! would be difficult or encourage bad code to allow is repr on the definition only if the is repr matches? ex. class F is repr(A) { ... } class F is repr(A) {} 14:52
jnthn hoelzro: Feels like a DRY violation...
cdc: Yes 14:53
dalek osystem: 51b15e8 | (Matt Oates)++ | META.list:
Added BioInfo

Finally got annoyed at not being able to install wherever I want to release :3		 14:58
ecs: 5f132ab | jnthn++ | S04-control.pod:
Revise and simplify when/default semantics.

This matches behavior seen in implementation and relied upon in the wild. There are various examples of existing code making use of the fact every block has an implicit C<$_>, which it sets so it can use C<when> and C<default>. This means trying to identify a topicalizer in a clever way, or even complain in the absence of one, breaks a common and useful pattern (yes, I tried it!) Added some notes on the semantics of nested when/default, also bringing the design doc in line with existing and tested behavior.
ugexe Util: ah. if you do "oogabooga" .= IO i think you actually will catch it. if you do $file = "oogabooga"; $file .= IO you get uncatchable 14:59
15:00 xinming_ left
ugexe CATCH { default { say "caught" }}; say "start"; "xxx" .= IO 15:00
m: CATCH { default { say "caught" }}; say "start"; "xxx" .= IO
camelia rakudo-moar 261acc: OUTPUT«start␤caught␤»
15:00 xinming joined
ugexe m: CATCH { default { say "caught" }}; say "start"; my $y = "xxx"; $y .= IO 15:00
camelia rakudo-moar 261acc: OUTPUT«start␤»
jnthn Phew, down below 1030 RTs again. 15:01
Util Ulti: ^^^ (7 lines up from ugexe)
hoelzro jnthn: alright, I'll just change my code then 15:02
ugexe oh oops 15:03
Ulti heh yeah 15:06
well at least its not just me, yeah I get something you can catch direct from the string literal like that and even just doing .IO.e on the variable its specifically the .= IO thats odd 15:07
ugexe well IO is just an empty role mostly
Ulti yeah but $thing .= IO should be the same as just doing $thing = $thing.IO right? and I can catch that 15:08
I think... *double checks*
ugexe $thing.IO doesnt call IO.new($thing) i dont think
15:09 zakharyas left
ugexe if th ecompiler knows its a string it looks like it knows what to do 15:10
m: CATCH { default { say "caught" }}; say "start"; my Str $y = "xxx"; $y .= IO
camelia rakudo-moar 261acc: OUTPUT«start␤caught␤»
dalek ast: 2098c0c | jnthn++ | S04-blocks-and-statements/pointy.t:
Correct and unfudge test.

  Routine.name is not fully qualified.
15:10 domidumont left
Ulti hmm either way what I was doing was kind of gross just wondered about why the error was slippery 15:11
ugexe yeah the code is not right. but its not doing the right not right thing 15:12
15:14 brrt left
jnthn m: sub foo($a) { sub bar($n) { say $a; $n == 1 ?? 1 !! $n * &?ROUTINE($n - 1) } }; my $r1 = foo('a'); my $r2 = foo('b'); $r2(4); 15:14
camelia rakudo-moar 261acc: OUTPUT«b␤b␤b␤b␤»
jnthn m: sub foo($a) { sub bar($n) { say $a; $n == 1 ?? 1 !! $n * &?ROUTINE($n - 1) } }; my $r1 = foo('a'); my $r2 = foo('b'); $r1(4);
camelia rakudo-moar 261acc: OUTPUT«a␤b␤b␤b␤»
jnthn bingo
15:18 nys joined 15:22 vendethiel left 15:24 FROGGS left
itz_ w00t! examples.perl6.org is colour syntaxed now .. and not a line of python insight :) 15:28
jnthn itz_++ 15:29
hoelzro itz_: you didn't use pygments? how'd you highlight it? 15:32
15:33 vendethiel joined
itz_ basically shelling out to a vim line from a perl6 wrapper and using that. perl 6 syntax support is in vim 7.4 15:35
15:35 aborazmeh left
Ulti itz_ is it using the (relatively) recently updated Perl 6 vim highlighting? 15:35
itz_ Ulti: not yet I was just wondering about importing those files from github.com/vim-perl/vim-perl 15:36
or using git to pull them in via a submodule (?) or something
Ulti yeah they are way better now 15:37
vim tended to get quite confused frequently before the updates in my experience
itz_ yes I'm using the newer files myself and they are better (although slower which I guess is hard to avoid) 15:38
Ulti yeah looking at them I get why :) they are HUGE compared to other language syntax files 15:39
really we want vim shelling out to perl6 to do the highlighting with the rakudo parser :P 15:40
RabidGravy has anyone had any further thoughts about the rt.perl.org/Ticket/Display.html?id=125408 ? really getting in my way now
15:40 Begi left
itz_ yes long term I'm sure using Perl6's own grammar to highlight is the way to go 15:41
15:45 petercommand left, petercommand joined, anaeem1_ left 15:46 anaeem1 joined 15:47 virtualsue joined 15:49 g4 left, zakharyas joined 15:50 anaeem1 left 15:55 FROGGS joined, vendethiel left 15:57 nowan left 15:59 vendethiel joined 16:00 nowan joined
dalek kudo/nom: 3db9326 | jnthn++ | src/Perl6/Actions.nqp:
Correct closure semantics of &?ROUTINE.
 16:04
kudo/nom: 3bdd0af | jnthn++ | src/Perl6/Actions.nqp:
Implement &?BLOCK.

We go with a similar approach to $?FILE/$?LINE, which lets us cheaply get the closure semantics correct (cheaply as in "only pay if the symbol is used").
ast: 4331787 | jnthn++ | S02-magicals/sub.t:
Test closure semantics of &?ROUTINE.
ast: aaf36de | jnthn++ | S02-magicals/block.t:
Unfudge &?BLOCK test; add closure semantics test.
jnthn Time for break, dinner, etc. 16:06
[Coke] waves from the AWS summit, where there is no metion of p5, let alone p6. 16:08
yoleaux 09:27Z <ab6tract> [Coke]: the conversation about years clearly ended with everyone agreeing that it should be a separate class
16:08 brrt joined
[Coke] . tell ab6tract - yes, I see that, it was clear. I obviously responded before backscroll completed, but does that mean I shouldn't have voiced my opinion, even if I had waited until I caught up? 16:09
.tell ab6tract - yes, I see that, it was clear. I obviously responded before backscroll completed, but does that mean I shouldn't have voiced my opinion, even if I had waited until I caught up?
yoleaux [Coke]: I'll pass your message to ab6tract.
16:16 synbot6 joined
[Coke] 1028 tickets. wonder if we can get that under 1K today. 16:18
I will have a few uninterrupted hours on the train later today, will see what I can do.
16:21 vendethiel left
[Coke] rt.perl.org/Ticket/Display.html?id=77170 - how clever can we be here? Is it worth it? 16:22
16:22 uncleyear left, uncleyear joined 16:28 abraxxa left 16:34 brrt left, vendethiel joined 16:37 spider-mario joined 16:42 rurban left 16:45 amurf joined 16:46 molaf joined 16:47 synbot6 left
[Coke] I'm seeing failures in S32-exceptions/misc.t 16:47
16:50 dakkar left, amurf left 16:54 diana_olhovik joined 16:55 brrt joined 16:57 vendethiel left 17:02 tinyblak_ left 17:07 zakharyas left 17:11 Hor|zon left 17:12 uncleyear left, uncleyear joined
b2gills m: -> $a { say &?BLOCK }($) # gist doesn't match definition 17:15
camelia rakudo-moar 3bdd0a: OUTPUT«-> ($a) { #`(Block|48836336) ... }␤»
17:18 vendethiel joined 17:29 domidumont joined 17:40 vendethiel left 17:42 telex left 17:44 espadrine left, telex joined
FROGGS jnthn++ # &?BLOCK 17:45
17:45 vendethiel joined 17:46 molaf_ joined 17:48 breinbaa2 left 17:49 molaf left 17:52 rurban joined
dalek rl6-roast-data: 70b00e6 | coke++ | / (9 files):
today (automated commit)
 17:59
[Coke] Maybe it's mac only. :P 18:01
itz_ twitter.com/bryanl/status/61920409...16/photo/1
[Coke] m: use Test; throws_like 'sub foo;', X::SemicolonForm::Invalid, what => 'sub'; 18:03
camelia rakudo-moar 3bdd0a: OUTPUT«Could not find symbol '&Invalid'␤ in block <unit> at /tmp/aEzbf6OspH:1␤␤»
18:04 baest_ joined
[Coke] huh. why doesn't that fail in the daily run, I wonder. but it's failing on OSX, failing here. 18:04
18:06 baest left 18:07 atroxaper joined, Hor|zon joined 18:08 roguelazer joined 18:12 atroxaper left, bin_005 joined, Hor|zon left
bartolin [Coke]: that was changed to 'X::UnitScope::Invalid' with github.com/perl6/roast/commit/1c4b217809 18:14
[Coke] ... arglebargle. 18:15
I would have figured that out in a few hours that I needed to update roast. Thanks! ;) 18:16
bartolin yw :-)
18:18 domidumont left
itz_ I'm still getting t/spec/S32-io/IO-Socket-INET.t fail on OSX 18:20
brrt anything changed recently that'd make s05-metasyntax/repeat.t die 18:21
dalek ast: cf0d490 | coke++ | S32-exceptions/misc.t:
Add test for RT #114274
 18:23
itz_ which is golfed for perl6 -e 'qx{PERL6_BINARY=/Users/steve/.rakudobrew/moar-nom/perl6-m sh -x t/spec/S32-io/IO-Socket-INET.sh 7 1024}'
Unhandled exception: Unable to execute...
[Coke] 1027 tickets! 18:24
r: say 1.0000000000000000000000000000001 18:25
camelia rakudo-{moar,jvm} 3bdd0a: OUTPUT«1.000000000000000000000000000000100␤»
18:25 rurban left, muraiki left 18:26 muraiki joined 18:29 domidumont joined 18:31 [Sno] left 18:34 Ven joined, amurf joined
[Coke] r: my $a = :b; say $a.key =:= $a.key 18:37
camelia rakudo-jvm 3bdd0a: OUTPUT«False␤»
..rakudo-moar 3bdd0a: OUTPUT«True␤»
[Coke] m: sub a(int $i) is cached { return unless $i; a($i-1) }; say a(4); 18:38
camelia rakudo-moar 3bdd0a: OUTPUT«Cannot find method 'perl': no method cache and no .^find_method␤ in any enter at src/gen/m-Metamodel.nqp:3560␤ in sub a at /tmp/wZutjuAe0s:1␤ in any call_with_capture at src/gen/m-Metamodel.nqp:3488␤ in any enter at src/gen/m-Metamodel.nqp:3560…»
[Coke] m: sub a(int $i) { return unless $i; a($i-1) }; say a(4); 18:39
camelia rakudo-moar 3bdd0a: OUTPUT«Nil␤»
18:39 amurf left
colomon jnthn++ 18:43
18:47 konkeydong joined 18:50 Hor|zon joined
masak FROGGS: sorry; conversation scattered across the day... 18:52
FROGGS masak: we have a conversation? :S
masak FROGGS: I didn't do something like DateTime::Span or Interval or whatever, because '1 year' does not have a single defined length.
FROGGS was very busy today at $work :o(
masak FROGGS: instead, it depends where you put either endpoint. 18:53
FROGGS masak: true
masak FROGGS: also, things like 'one month after Jan 31'. the day you end up on might even depend on whether it's a leap year or not.
brrt masak: fwiw, take a look at the current way i use macros :-) 18:54
github.com/MoarVM/MoarVM/blob/even...t/exprlist
masak FROGGS: instead, I ended up mapping those needs to the .earlier and .later methods, where you can only add/subtract one time unit at a time. many problems thus go away.
FROGGS I use usually do: DateTime::Span->new(from => $today-at-midnight, $today-at-midnight->add($one-day))
masak looks
FROGGS: oh, so a Span is always rooted in two DateTime endpoints? 18:55
FROGGS then my span is from midnight to midnight, and I can use set-like ops to get intersections etc
brrt i've chosen to go for consistency between definition and use; adding a 'keyword' as the first list item also resolved parsing special-casing
FROGGS masak: you have to apply a span to an endpoint to work with it, yes
brrt because the macro name, with a ^, isn't resolved as a macro if it isn't the first item in the list
masak brrt: nice.
brrt: yeah, looks clean and usable. brrt++ 18:56
brrt thanks for the advice again :-) masak++
masak brrt: glad I could apply my (meagre) knowledge of macros to something useful :P
18:57 pyrimidine joined
brrt now i'm really off for today :-) 18:57
FROGGS masak: my use case is that I have a "calender" (a spanset) of spans defining the possible fees one can get, and another calender/spanset describing when that someone actually were at work
18:57 brrt left
FROGGS the intersection is the fees (s)he gets 18:57
(it is not that simple in fact, but the underlying principle works like that) 18:58
masak FROGGS: what does a Span give you that a range of DateTime doesn't?
18:58 konkeydong left
FROGGS well, I usually prepare spans to apply them later to different endpoints... 18:59
but my intend is to do math with datetime ranges only
ohh way, I think it is stilled called span when a span has concrete endpoints 19:00
masak: in fact I dont need a Span that has loose endpoints 19:02
masak all the better. 19:03
then it looks like what you have is two DateTimes that you want to manipulate in concert.
FROGGS what I want: my $span = DateTime::Span.new( $dt1, $dt2 ); [...]; if $span1 (&) $span2 -> $intersection { ... }
masak ooh, yes.
that would come out nicely.
(|) would only be closed if the spans intersect, though. 19:04
FROGGS a spanset is about a bunch of spans btw 19:08
like today from 0-6 o'clock and 20 to midnight 19:09
and when I work from 5 to 14, and 22 to midnight, I get night-fees for three hours 19:10
so you can just calc the intersections of both sets, and all is fine
and it also just works out on the days when we switch to daylight savings time for example 19:11
(which was a nice bug report I got once from our HR department.... "Ohh noes, we paid for 25 hours on-call fee that day!" - "That's right :o)") 19:12
19:13 anaeem1 joined
dalek ast: 0aa3fde | usev6++ | S02-literals/numeric.t:
Unfudge some tests for RT #124559
 19:17
19:26 domidumont left
moritz FROGGS: funny, we had an actual bug in our time accounting regarding DST 19:37
FROGGS: when entering a time that ended at midnight, it would miscalculate the date, becaue it assumed each day had 24 hours
FROGGS moritz: well, I had (of course) bugs in that area in the past... also involving endless loops when we added 24 hours and expecting to get to the next day 19:38
though that was when doing most stuff in php... it got better when we switched to Perl :o)
Ven oh, we have &?BLOCK now? amazing :D. jnthn++
skids b2gills: I'll look into that block .signature problem probly next week. 19:41
And yes jnthn++. Now I have to go figure out all the places where I worked around not having it :-) 19:42
Ven m: { my $a = &?BLOCK; }
camelia ( no output )
19:43 FROGGS left 19:44 virtualsue left 19:46 virtualsue joined 19:47 molaf_ left 19:49 [Sno] joined
skids m: say (for 1..2 -> $a { &?BLOCK.last(3); })# uhp, well I guess we need .last now :-) 19:50
camelia rakudo-moar 3bdd0a: OUTPUT«Method 'last' not found for invocant of class 'Block'␤ in block <unit> at /tmp/aIzmoZwI1r:1␤␤»
19:53 notjack joined
Ven skids: and leave :P 19:54
19:54 muraiki left
dalek ast: f6a7dfb | usev6++ | S06-advanced/wrap.t:
Change test from 'dies-ok' to 'throws-like'
 19:57
notjack m: 3.WHAT.say; 19:58
camelia rakudo-moar 3bdd0a: OUTPUT«(Int)␤»
notjack m: Int.WHAT.say;
camelia rakudo-moar 3bdd0a: OUTPUT«(Int)␤»
dalek ast: 865092a | usev6++ | S32-str/substr.t:
Don't fudge whole block for JVM (RT #124692)
 20:08
masak notjack: the .WHAT of a type object is the type object itself. 20:10
20:12 darutoko left 20:15 notjack left 20:18 Alina-malina left 20:19 Alina-malina joined 20:23 amurf joined 20:25 pecastro left 20:28 amurf left 20:30 zakharyas joined 20:32 diana_olhovik left 20:34 diana_olhovik joined 20:36 ab6tract joined 20:37 itz__ joined 20:40 davido___ left, itz_ left, davido___ joined 20:41 rurban joined
dalek kudo-star-daily: 74c1fe2 | coke++ | log/ (9 files):
today (automated commit)
 20:53
20:53 zakharyas left 20:59 rurban left 21:03 ab6tract left 21:04 beastd joined 21:05 skids left 21:09 diana_olhovik left
jnthn The report on last week's work: 6guts.wordpress.com/2015/07/09/thi...enty-more/ 21:13
'night, #perl6
masak jnthn++ # blug 21:14
'night, jnthn
lizmat gnight jnthn 21:15
21:16 rurban joined, rurban left
cognominal jnthn++ 21:26
21:27 TEttinger joined, pyrimidine left
dalek kudo/nom: 01a60df | lizmat++ | src/core/List.pm:
Handle offset past end as failure (as per spec)
 21:28
21:34 kaare_ left 21:36 ab6tract joined, pyrimidine joined
ab6tract masak, FROGGS: I enjoyed reading your DateTime::Span discussion in the backlog 21:37
yoleaux 16:09Z <[Coke]> ab6tract: - yes, I see that, it was clear. I obviously responded before backscroll completed, but does that mean I shouldn't have voiced my opinion, even if I had waited until I caught up?
masak how long was the interval you enjoyed it? :P
PerlJam
.oO( many many many light-meters )
 21:38
ab6tract [Coke]: no, of course not
masak: at least a parsec ;)
ab6tract hopes the Han Solo reference carries through 21:39
masak indeed.
21:43 atroxaper joined 21:47 atroxaper left
lizmat m: say [email@hidden.address] # why are we backslashing the array ? 21:49
camelia rakudo-moar 3bdd0a: OUTPUT«"\@a.foo"␤»
masak no good reason.
ab6tract The thing about intervals that distinguishes them from a range of dates is that it would feel natural to me to have a range of intervals. since ranges can't have ranges as endpoints, a different solution must be found.
geekosaur seems like a holdover from perl5?
masak ab6tract: I'm all for having intervals of Instant. 21:50
ab6tract: intervals of DateTime just aren't well-defined. ranges of intervals of DateTime even less so.
ab6tract: the conceptual problems keep multiplying as you build bigger and bigger abstractions on a murky foundation. 21:51
lizmat geekosaur: yeah, feels like it
masak backslashing @a there is not wrong. just unnecessary. 21:52
m: say "\@a".perl
camelia rakudo-moar 01a60d: OUTPUT«"\@a"␤»
geekosaur actually I think early on (like pugs era) we still expanded simple variables
masak m: say "@a".perl
camelia rakudo-moar 01a60d: OUTPUT«"\@a"␤»
ab6tract fair enough. the ecosystem is a fine place for something more robust
dalek ecs: 5e0b4bd | (Stéphane Payrard)++ | S03-operators.pod:
suppressed spurious double quotes
masak there isn't a "more robust". it's ill-defined even in theory. 21:53
ab6tract because my idea is exactly that the murky foundations exist and I'm tired of dealing with '2002' as an numeric value
masak ab6tract: if you want to deal with '2002', define a class Year, and give it all the DateTime semantics you please.
ab6tract when it could be an object with knowledge of 2002's specifities
masak people are too afraid sometimes to build their own abstractions.
ab6tract what exactly are you talking about? I have a) agreed that it wasn't a thing for Date, but Dateish, and b) said that I will do my own thing and put it in the ecosystem 21:55
I fail to see my unwillingness to craft my own abstraction
lizmat
.oO( maybe it's an ab6traction? )
 21:58
masak m: class Year { has $.year; has $.start-date = Date.new(:$!year, :month(1), :day(1)); has $.end-date = Date.new(:$!year, :month(12), :day(31)); method leap-year { $.start-date.is-leap-year } }; say .end-date and say .leap-year given Year.new(:year(2012))
camelia rakudo-moar 01a60d: OUTPUT«2012-12-31␤True␤»
masak ab6tract: oh, it's quite possible that we agree and that I'm just arguing our side too agressively. apologies if so. ;) 21:59
ab6tract cheers :) 22:00
is anyone going to the Swiss Perl Workshop? 22:02
masak 22:04
ab6tract cool!
masak see each other there? 22:05
ab6tract I'd imagine so ;)
RabidGravy I'll be in Ibiza I think ;-)
ab6tract RabidGravy: well, shortly after I'll be talking up Perl 6 ops in Grenada :) 22:06
vendethiel should be coming as well 22:07
masak \o/
itz__ RabidGravy: rave on? 22:08
RabidGravy well probably just drinking too much and winding up deep house DJs ;-) 22:09
ab6tract I notice jnthn has a new async talk scheduled :D
and yours looks cool too, vendethiel :D
dalek kudo/nom: 1b537b9 | lizmat++ | src/core/List.pm:
Some failure mode tweaks
 22:10
ab6tract MAIN is definitely a super-sweet feature that tends to impress folks
itz__ "doves and white gloves" 22:11
;)
dalek ast: 48a5a13 | lizmat++ | S32-array/splice.t:
Add some more sanity / failure tests
Ven ab6tract: i'm particularly found of the talk title :P
ab6tract masak: both of yours sound awesome as well. it's a p6 party! :D 22:12
22:12 spider-mario left
lizmat will be at SPW as well 22:13
but probably not in the best of shapes, as we will have come back from the YAPC::Asia only days before
cognominal m: say True.'!' # github.com/perl6/specs/blame/maste....pod#L2867
camelia rakudo-moar 01a60d: OUTPUT«5===SORRY!5=== Error while compiling /tmp/pfXTxsU6br␤Quoted method name requires parenthesized arguments. If you meant to concatenate two strings, use '~'.␤at /tmp/pfXTxsU6br:1␤------> 3say True.'!'7⏏5 # github.com/perl6/specs/blame…»
masak ab6tract: glad to hear it! I'm looking forward to it all too. :)
cognominal do I misinterpret the doc, or is this not supported? 22:14
lizmat and on that note, I wish everybody a good night's sleep (unless prohibited)
ab6tract masak: now I need to decide whether to give a dry run of my Whatever talk that's already scheduled for YAPC::EU
22:14 virtualsue left
ab6tract or use the opportunity to prepare a different one 22:15
cognominal m: my @a = < a b c>; say @a.'$'
camelia rakudo-moar 01a60d: OUTPUT«5===SORRY!5=== Error while compiling /tmp/fjeUiAZw8C␤Quoted method name requires parenthesized arguments. If you meant to concatenate two strings, use '~'.␤at /tmp/fjeUiAZw8C:1␤------> 3my @a = < a b c>; say @a.'$'7⏏5<EOL>␤»
ab6tract nice to have choices :) 22:16
masak ab6tract: i0.kym-cdn.com/photos/images/newsfe...31/0fc.gif
ab6tract I've been wanting to talk about general "programming culture" trends, the cycles of buildup and teardown (*sheds a tear for jQuery*), famous flame war quotables, etc 22:17
masak: well played sir. I can honestly type lol for that one :) 22:18
labster m: say DateTime.now < now; 22:19
camelia rakudo-moar 01a60d: OUTPUT«Cannot call Real(DateTime: ); none of these signatures match:␤ (Mu:U \v: *%_)␤ in block <unit> at /tmp/uDv0emrYoi:1␤␤»
ab6tract through the lens of perl's history
m: say so DateTime.now before now 22:20
camelia rakudo-moar 01a60d: OUTPUT«True␤»
ab6tract are there any hidden gotchas behind 'before' ? I ask because I developed a hand rolled "version" of it before chancing upon it on the periodic table today 22:23
labster wow, didn't know about before.
cognominal m: my $a; @$a = 1,2,3
camelia rakudo-moar 01a60d: OUTPUT«Cannot modify an immutable List␤ in block <unit> at /tmp/Nd84kew49J:1␤␤»
ab6tract labster: me either :)
22:23 bin_005 left
cognominal hum, many things dubious and not implemented in S03 :( 22:24
masak ab6tract: only that `before` is the least specific you can get. so prefer more specific ones (`<`, `lt`) whenever it makes sense.
cognominal m: my $a; $a[] = 1,2,3
camelia rakudo-moar 01a60d: OUTPUT«Cannot modify an immutable Any␤ in block <unit> at /tmp/gonFkl7ROv:1␤␤»
labster multi sub infix:<before>(\a, \b) { (a cmp b) < 0 } # I doub't there's anything gotcha in here.
and apparently doub't is now a contraction :) 22:25
TimToady that's dub'itable 22:26
ab6tract I thought I was pretty cool doing:
m: my $a=DateTime.now; my $cmp = $a ~~ Numeric ?? * <= * !! * lt *; $cmp($a,now)
camelia ( no output )
TimToady well, but you should only do lt if it's a string, and use 'before' for the general case... 22:27
masak 'night, #perl6
labster I do feel like one should be able to compare Instants and DateTimes with the numeric comparison operators. I was thinking about having both DateTime and Instant inherit from a role, but can't decide what to name it.
I was thinking about Momentous. 22:28
ab6tract labster: note that 'lt' works here because stringified DateTime comparison DWIMs
TimToady: yeah, but this was before I know about 'before' :)
TimToady m: say DateTime.new.earlier(:2000years)
camelia rakudo-moar 01a60d: OUTPUT«Required named parameter 'year' is missing for DateTime.new()␤ in block <unit> at /tmp/gnqlOicGzA:1␤␤»
TimToady m: say DateTime.new.earlier(:2000year)
camelia rakudo-moar 01a60d: OUTPUT«Required named parameter 'year' is missing for DateTime.new()␤ in block <unit> at /tmp/1PblTEjXwN:1␤␤»
TimToady m: say DateTime.now.earlier(:2000year) 22:29
camelia rakudo-moar 01a60d: OUTPUT«0015-07-10T00:29:05+02:00␤»
TimToady I guess... :)
was there a leap second that year? :P
ab6tract night masak!
labster m: say DateTime.now.later(:1000year) lt now; 22:31
camelia rakudo-moar 01a60d: OUTPUT«True␤»
RabidGravy .tell jnthn that it seems that the rt.perl.org/Ticket/Display.html?id=125408 has gone away at 2015.06-226-g01a60df built on MoarVM version 2015.06-88-g647df11
yoleaux RabidGravy: I'll pass your message to jnthn.
ab6tract okay, so now it don't :)
s/now/no/
m: DateTime.now.later(:1000year) before now 22:32
camelia ( no output )
ab6tract m: say DateTime.now.later(:1000year) before now
camelia rakudo-moar 01a60d: OUTPUT«True␤»
labster m: say DateTime.now.earlier(:1000year) before now
camelia rakudo-moar 01a60d: OUTPUT«True␤»
labster Time is circular
TimToady um, comparing DateTimes with Instants there
m: say DateTime.now.earlier(:2000year) before DateTime.now 22:33
camelia rakudo-moar 01a60d: OUTPUT«True␤»
labster Yeah, the stringy representations are bound to be wrong. But I think it makes sense to let < and friends work.
TimToady then people will think they're working with normal numbers, which is misleading 22:34
labster m: say now < now 22:35
camelia rakudo-moar 01a60d: OUTPUT«True␤» 22:36
labster so that's misleading too?
ab6tract TimToady: I assumed that DateTime.now was the equivalent of calling : DateTime.new( now )
labster: I think 'before' is good enough for me
labster ab6tract: it's very similar, but .now includes the timezone
ab6tract labster: ah, gotcha 22:37
labster TimToady: I ask about the comparison ops for Date/DateTime/Instant, because of rt.perl.org/Ticket/Display.html?id=125555 22:39
TimToady m: my $dt1 = DateTime.now; say +$dt1 22:41
camelia rakudo-moar 01a60d: OUTPUT«Cannot call Numeric(DateTime: ); none of these signatures match:␤ (Mu:U \v: *%_)␤ in block <unit> at /tmp/48RaAFhT9z:1␤␤»
TimToady < is supposed to coerce to Numeric, and I do not think we should be overloading things with < that don't do that
m: +now 22:42
camelia ( no output )
TimToady m: say +now
camelia rakudo-moar 01a60d: OUTPUT«Instant:1436481797.189694␤»
labster Well, it could by coverting to TAI.
*converting
secretly, of course 22:43
TimToady we've been trying to move away from epochs, and I'm not terribly interested in re-installing them into the languge
ab6tract m: say DateTime.now.later(:1day).Instant before now 22:44
camelia rakudo-moar 01a60d: OUTPUT«False␤»
ab6tract that's pretty clear :)
[Coke] (swiss perl workshop) maybe, depending on funding.
ab6tract [Coke]++
TimToady just as we've moved away from absolute precedence levels, and we encourage everyone to add new precedence relative to existing precedence levels 22:45
ab6tract labster: tbh, that ticket almost makes me think that < could be turned off for Date instead
TimToady yes, I think that supporting < there is something like a mistake 22:46
labster m: say DateTime.now.later(:1day).Instant after now
camelia rakudo-moar 01a60d: OUTPUT«True␤»
22:46 rindolf left
labster before and after seem about right. We'd just need to support cmp for these, which is pretty easy. 22:46
TimToady well, we still haven't made a final determination about comparisons between different types 22:47
ab6tract TimToady: that seems like something that should throw an exception 22:48
TimToady suspects that we want to differentiate cmp into strict and lax versions, with sort using the lax version
[Coke] m: A::B
camelia rakudo-moar 01a60d: OUTPUT«Could not find symbol '&B'␤ in block <unit> at /tmp/QFPMXNqMbE:1␤␤»
ab6tract $a cmp $b :lax
TimToady m: say 42 before "123"
camelia rakudo-moar 01a60d: OUTPUT«False␤»
[Coke] std: A::B
camelia std 28329a7: OUTPUT«5===SORRY!5===␤Undeclared name:␤ 'A::B' used at line 1␤Check failed␤FAILED 00:00 134m␤»
TimToady m: say "42" before 123
camelia rakudo-moar 01a60d: OUTPUT«False␤»
[Coke] TimToady: std's behavior there is desired, yes 22:49
?
TimToady currently I think it just stringifies between types, but that's probably not the best thing
[Coke]: yeah, I think the full name would probably be more awesome 22:50
ab6tract TimToady: yeah, that sounds like the kind of thing that should only happen via an explicit mechanism
[Coke] TimToady: danke. digging...
ab6tract plus, you can catch the exception, and if you have known-dirty lists with routines ready to handle turning the Milk into Cheese based on a given context, you do your business and resume the exception 22:52
so to speak 22:53
TimToady or at least provide a Cheese view of the Milk, if Milk is immutable...
ab6tract well said
though the milk must already be cheese then, to be so solid? ;) 22:54
labster Perl 6 has liquid typing. 22:55
ab6tract have a great *, #perl6. catch you on .succ
labster good night, ab6tract 22:56
TimToady .tell jnthn "to implementing through implementing"?
yoleaux TimToady: I'll pass your message to jnthn.
22:56 ab6tract left
[Coke] looks at the ll exception backtrace for that and tries to figure out what is calling the Stash AT-KEY that is failing. 22:57
22:58 beastd left, skids joined
[Coke] it's soft failing - at the point that it fails, it only has the &B and the stash doesn't seem to know it's own name. wondering if there's something in the Perl6 Action I can tweak there... 22:59
22:59 roguelazer left
TimToady is tired of typing --ll-exception 23:00
[Coke] is there anything that hides from ll-exception and can I show that? 23:01
*its own, bah.
23:06 amurf joined, anaeem1 left 23:07 anaeem1_ joined
RabidGravy nighty night people! 23:07
[Coke] ~~
labster alias 6 perl6 --ll-exception 23:10
23:12 RabidGravy left
[Coke] is there a parameter to make fail die immediately? 23:14
or get the backtrace from the fail, not from where it ends up throwing?
fail--
[6~[6~
23:15 anaeem1_ left
[Coke] gives up and moves on to a more tractable RT. :P 23:16
skids [Coke]: I've the hood open on Failure right now and that is one of the things I'm looking at 23:17
psch hi #perl6 o/ 23:28
i have 2 FAILs in S04-phasers/enter-leave.t, on jvm. i wouldn't have thought that the changes in my branch are responsible, has anything there come up? 23:29
[Coke] psch: github.com/coke/perl6-roast-data/b....out#L1846 23:32
psch [Coke]: thanks, that's reassuring. i might be able to spare a look, although i suspect the guts of phasers on jvm are a bit beyond me 23:33
m: say ((rand * 10) %% 2 ?? "standalone jar" !! "roast fixes") 23:37
camelia rakudo-moar 1b537b: OUTPUT«roast fixes␤»
psch ...because what good is a deployable format if it doesn't pass the official test suite, i guess 23:38
[Coke] if we have an enum, and support ++ -- on it, what should happen at the ends of the enum?
m: Enum A <a b c>; my A $q = a; say $q++; say $q++;say $q++;say $q++; 23:39
camelia rakudo-moar 1b537b: OUTPUT«5===SORRY!5=== Error while compiling /tmp/VqZnC0IbFY␤Two terms in a row␤at /tmp/VqZnC0IbFY:1␤------> 3Enum7⏏5 A <a b c>; my A $q = a; say $q++; say $␤ expecting any of:␤ infix␤ infix stopper␤ statement end␤…»
[Coke] m: enum A <a b c>; my A $q = a; say $q++; say $q++;say $q++;say $q++; 23:40
camelia rakudo-moar 1b537b: OUTPUT«Type check failed in assignment to '$q'; expected 'A' but got 'Int'␤ in block <unit> at /tmp/B3mNplEoBs:1␤␤»
[Coke] m: enum A <a b c>; my A $q = a; say $q++;
camelia rakudo-moar 1b537b: OUTPUT«Type check failed in assignment to '$q'; expected 'A' but got 'Int'␤ in block <unit> at /tmp/GB4_KjYciJ:1␤␤»
psch m: enum A <a b c>; my $a = b; say $a++
camelia rakudo-moar 1b537b: OUTPUT«b␤»
psch m: enum A <a b c>; my $a = c; say $a++
camelia rakudo-moar 1b537b: OUTPUT«c␤»
psch oh wait
[Coke] m: enum A <a b c>; a++.say;
camelia rakudo-moar 1b537b: OUTPUT«Cannot assign to an immutable value␤ in block <unit> at /tmp/4apOGLOpJ1:1␤␤»
psch m: enum A <a b c>; my $a = c; say ++$a
camelia rakudo-moar 1b537b: OUTPUT«3␤»
psch m: enum A <a b c>; my $a = c; $a++ andthen say $a 23:41
camelia rakudo-moar 1b537b: OUTPUT«3␤»
psch makes sense to me, coerce to Numeric and assign
[Coke] right. Trying to fix that so you get the enum back, not the int val. but if you have c, should get c back?
psch: see RT #92742
or does that throw? 23:42
psch m: enum A <a b c>; say b.succ
camelia rakudo-moar 1b537b: OUTPUT«2␤»
psch that should give c
imo
but c.succ... i'm not sure
i'm leaning towards modulo arithmetic
i.e. wrap around
dalek osystem: 3481a5c | hoelzro++ | META.list:
Add Algorithm::Elo to ecosystem
 23:43
[Coke] .ask TimToady if we can get a ruling on RT#92742 - should enum <a b c>; c++ give c, a, or die?
yoleaux [Coke]: I'll pass your message to TimToady.
[Coke] m: class A { has str $.x; method BUILD(:$!x) { } }; say A.new(:x<foo>).x' 23:44
camelia rakudo-moar 1b537b: OUTPUT«5===SORRY!5=== Error while compiling /tmp/DYlXqapa8B␤Two terms in a row␤at /tmp/DYlXqapa8B:1␤------> 3 BUILD(:$!x) { } }; say A.new(:x<foo>).x7⏏5'␤ expecting any of:␤ infix␤ infix stopper␤ postfix␤ s…»
[Coke] m: class A { has str $.x; method BUILD(:$!x) { } }; say A.new(:x<foo>).x 23:45
camelia rakudo-moar 1b537b: OUTPUT«Cannot modify an immutable str␤ in method BUILD at /tmp/fVt6fwaJ3s:1␤ in block <unit> at /tmp/fVt6fwaJ3s:1␤␤»
23:45 BenGoldberg joined
psch actually, thinking more about it throwing seems more fitting 23:46
enumeration is (somewhat) counting, and counting beyond what is countable isn't feasible
kind of the inverse of "you can't count to Inf"
23:48 tinyblak joined
psch ...we wouldn't want to pidgeon hole people into modulo arithmetic :) 23:49
23:50 nightfrog left 23:55 nightfrog joined
[Coke] .ask jnthn I bet RT #107204 is closable with tests after your nfg work. 23:55
yoleaux [Coke]: I'll pass your message to jnthn.
23:57 kurahaupo joined 23:58 Ven left

Powered by Cro and the Raku® Programming Language.

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

Thank you!