»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.
diakopter Guest84531: you around? 00:34
whiteknight: you around? 00:35
Guest84531 yessir 00:36
diakopter I tried out your patches for nqp&rakudo for PackfileView 00:36
wknight8111 they need work, I'm re-doing them 00:37
diakopter oh
wknight8111 what did you see with them? 00:40
diakopter I manually applied each change so as not to get the whitespace & stage0 changes; it passed all tests in nqp, then tried to build rakudo with that nqp (also applying the changes to rakudo), and it failed many spectests 00:41
see github.com/diakopter/rakudo/commit...fe18b5f5c9 and github.com/perl6/nqp/commit/d3a9cc...1ad7bb8d93 00:43
afk& 00:46
wknight8111 okay, thaks 00:47
thanks 00:48
skids r: (1, (4,4), h => (3,3)).perl.say; 01:22
p6eval rakudo aeca2f: OUTPUT«(1, (4, 4), "h" => $(3, 3))␤»
skids Why the $ nowadays? 01:23
sorear the right side of => has been forced to be an item 01:24
skids thanks. 01:25
skids sorear: may want to fix the url in the niecza release announcements to include the 'v' in the newer download url. 01:52
sorear skids: hmm? github changed something? 01:56
ah, I see 01:57
adu diakopter? 04:48
phenny adu: 13 May 17:28Z <moritz> tell adu that .join is a method on lists, not strings, so it won't appear in S32::Str
adu o right 04:49
diakopter adu 05:21
adu hi
i was going to tell you about the man/boy test 05:22
diakopter oh?
adu droscheme maxes out at about 15
adu I just learned something very interesting from rosettacode.org 05:25
Grammar.parse takes an action argument! 05:26
I wish I had known that earlier
moritz adu: we have the man-or-boy test in the official Perl 6 test suite, integration/man-or-boy.t 05:29
adu moritz: I'll check it out 05:30
TimToady it's also on rosettacode.org 05:55
Timbus a 07:11
sorear b 07:14
tadzik c 07:15
r: say DateTime.new('2012-05-14T12:53:20').posix 07:16
p6eval rakudo aeca2f: OUTPUT«1337000000␤»
Timbus heh
mikec_ hehe 07:18
jnthn oh shark, #perl6 07:31
moritz shark it's jnthn! 07:32
tadzik shark shark
sergot hi o/ 07:43
sorear o/ 07:44
tadzik \o 07:47
moritz sayspy.blogspot.de/2012/05/my-very-...-dart.html 08:32
quite a refereshingly different perspective from the usual "OMG Dart sucks" rants on the internez 08:33
dalek ast: 11ffc02 | moritz++ | S16-filehandles/chmod.t:
remove a test that is just duplicated inside an eval
 08:57
kudo/nom: a144e29 | moritz++ | / (3 files):
implement chmod as a function

because a test file wants it that way :-). Run that test file		 08:59
jnthn will be curious to run that one on Windows this evening :)
moritz it checks $*OS and does a skip_all; exit 09:00
jnthn ah :)
tadzik moritz: what's that change in 09:02
tools/update_passing_test_data.pl
?
moritz tadzik: that was accidental :( 09:03
dalek ast: 470da6d | moritz++ | S14-roles/super.t:
remove test for SUPER::, which has been unspecced
 09:04
tadzik yeah, I asked Justin Case 09:05
moritz tadzik: yes, thanks 09:06
tadzik: though I consider leaving it in
basically it reduced the output
which was fine when we had many test files which we didn't run
dalek ast: 21f9cd3 | moritz++ | S06-signature/multidimensional.t:
remove a wrong test, update test plan
 09:09
kudo/nom: 4ea6d60 | moritz++ | t/spectest.data:
run another test file
dalek ast: 59ce8bb | moritz++ | S12-methods/lvalue.t:
fix lvalue method tests
 09:13
moritz colomon: that commit might lead to some passing TODOs in niecza (because the tests were simply wrong before) 09:14
dalek kudo/nom: 56e86dd | moritz++ | t/spectest.data:
run method lvalue tests
 09:16
moritz p6: eval 'END { say "its over" }'; say 'end of main line'; 09:18
p6eval pugs, rakudo a144e2, niecza v17-4-ge5b869e: OUTPUT«end of main line␤its over␤»
dalek ast: 4ba3eed | moritz++ | S04-exceptions/die-and-end.t:
remove a wrong, short test file that is not easy to fix

in particular it assumes that END blocks inside an eval run when the eval is returned, but they are truly run as late as possible. Probably needs testing with Test::Util		 09:19
moritz r: say zip((1, 2); (8, 9)) 09:50
p6eval rakudo a144e2: OUTPUT«1 8 2 9␤»
moritz r: say [1, 2, 3] Z <a b> 10:09
p6eval rakudo a144e2: OUTPUT«1 a 2 b␤»
moritz I'm pretty sure that's wrong
dalek ast: a42f0a6 | moritz++ | S32-container/roundrobin.t:
redo roundrobin tests; delete tests for unspecced features
 10:12
colomon moritz: worrong how? 10:12
wrong how?
moritz colomon: simply confused variables/values 10:13
copy/paste errors
colomon moritz: no, I meant the 1 a 2 b. (But thanks for the heads up on the other thing!) 10:13
moritz [1, 2, 3] is a single item 10:14
zip shouldn't flatten it
dalek kudo/nom: abb8f3a | moritz++ | / (2 files):
basic implementation of &roundrobin; tests
 10:19
ast: 31af067 | moritz++ | S (2 files):
fix some uses of eval

those were still assuming that eval catches exceptions		 10:30
moritz there might be some niecza unfudges possible after this commit too
not_gerd r: say +<foo bar>.grep({$_ ~~ "foo"}); say +<foo bar>.grep(* ~~ "foo") 10:40
phenny not_gerd: 12 May 15:20Z <[Coke]> ask not_gerd if there's a ticket for the intval/opcode_t bug in parrot
p6eval rakudo 56e86d: OUTPUT«1␤0␤» 10:40
not_gerd bug?
jnthn r: say (* ~~ "foo").WHAT 10:42
p6eval rakudo 56e86d: OUTPUT«Bool()␤»
not_gerd phenny: tell [Coke] the intval/opcode_t but is filed as github.com/parrot/parrot/issues/705
phenny not_gerd: I'll pass that on when [Coke] is around.
jnthn Seems ~~ doesn't auto-curry... 10:43
jnthn Wonder if it's spec'd as not doing so or if this is a bug. 10:43
tadzik it's certainly confusing 10:44
not_gerd any comments on gist.github.com/2693261 10:48
it's a version of the regex-dna benchmark that works on current Rakudo by not using .subst()
colomon moritz: you were right about niecza passing todos in lvalue.t. :) 10:54
dalek pan style="color: #395be5">perl6-examples: 220cec7 | (Gerhard R)++ | shootout/regex-dna.p6-v2.pl:
[shootout/regex-dna] add version that works on Rakudo

the .subst() call is what causes the excessive memory consumption of the original version		 11:00
bbkr does not understand this "but" magic in this regexp 11:03
jnthn bbkr: It mixes in a method Str that returns the string after the but. 11:05
bbkr thanks, looks like I have to re-read spec about "but" operator 11:07
rakudo: my $x = 1 but "foo"; say +$x; 11:10
p6eval rakudo 56e86d: OUTPUT«foo␤»
Timbus butt operator
⍵
bbkr hm, why it invoked mixed-in Str method in numeric context?
or my understanding of "but" operator is incorrect 11:11
not_gerd bbkr: + probably returned the original value as it still is numeric
Timbus r: my $x = 1; $x does role { method Str {"foo"} }; say +$x; say "$x"; 11:12
p6eval rakudo 56e86d: OUTPUT«foo␤foo␤»
Timbus pf
bbkr looks like bug to me 11:13
jnthn What bit of spec is in not in line with? 11:14
Timbus my $x = 1 but "foo"; say $x.Int;
uh
r: my $x = 1 but "foo"; say $x.Int;
p6eval rakudo abb8f3: OUTPUT«foo␤»
not_gerd r: my $x = 1 but "foo"; say $x * 1;
p6eval rakudo abb8f3: OUTPUT«1␤»
jnthn Right, Int.Int is identify
not_gerd careful with optimizing the *1 away 11:15
not_gerd r: say +((42 but "foo") but role { method gist { self + 0 } }) 11:19
p6eval rakudo abb8f3: OUTPUT«42␤»
colomon r: my $x = 1 but "foo"; say +$x 11:21
p6eval rakudo abb8f3: OUTPUT«foo␤»
bbkr ok, now I understand this operator :) 11:22
thanks
dalek ast: 9c53615 | moritz++ | S02-lexical-conventions/comments.t:
rakudo unfudge
 11:31
not_gerd r: +Int.^methods.grep(*.name eq "Str") 11:33
p6eval rakudo abb8f3: OUTPUT«(signal SEGV)»
moritz SEGV is the new NPMCA :/ 11:34
r: say Int.^methods.WHAT
p6eval rakudo abb8f3: OUTPUT«Parcel()␤»
moritz r: say Int.^methods[0].WHAT
p6eval rakudo abb8f3: OUTPUT«Method()␤»
moritz r: say Int.^methods.map(*.WHAT).gist 11:35
p6eval rakudo abb8f3: OUTPUT«Method 'count' not found for invocant of class 'Whatever'␤ in method reify at src/gen/CORE.setting:4818␤ in method reify at src/gen/CORE.setting:4743␤ in method reify at src/gen/CORE.setting:4743␤ in method gimme at src/gen/CORE.setting:5130␤ in method eager a…
moritz r: say Int.^methods.map({ say .name}).eager
p6eval rakudo abb8f3: OUTPUT«Method 'gist' not found for invocant of class 'String'␤ in sub say at src/gen/CORE.setting:6630␤ in block <anon> at /tmp/bDS8a1geuo:1␤␤»
moritz ok, at some point it returns a parrot String instead of a Str
jnthn Method calls should auto-coerce return values though 11:36
I wonder if that's being missed out for the .foo form.
r: say Int.^methods.map({ say $_.name}).eager 11:37
p6eval rakudo abb8f3: OUTPUT«Int␤Num␤Rat␤FatRat␤abs␤Bridge␤chr␤succ␤pred␤sqrt␤base␤floor␤round␤ceiling␤sign␤conj␤rand␤sin␤asin␤cos␤acos␤tan␤atan␤atan2␤sec␤asec␤cosec␤acosec␤cotan␤acotan␤sinh␤asinh␤cosh␤acosh␤tanh␤atanh␤sech␤asech␤cosech␤acosech␤cotanh␤acotanh␤unpolar␤cis␤Complex␤log␤exp␤trunca…
jnthn Yes.
not_gerd r: say Int.^methods.grep({ $_.name eq "Str"}).[0].(42) 11:40
p6eval rakudo abb8f3: OUTPUT«42␤»
not_gerd what's the sane way to do that?
moritz 42.Str :-) 11:41
or
r: say Int.^can('Str')[0](42)
p6eval rakudo abb8f3: OUTPUT«42␤»
moritz r: say Int.^find_method('Str')(42) # more efficient and less specced 11:42
p6eval rakudo abb8f3: OUTPUT«42␤»
not_gerd r: my $foo = (42 but "foo") but role { method gist { Int.^find_method("Str").(self) } }; .say for $foo, +$foo, ~$foo 11:45
p6eval rakudo abb8f3: OUTPUT«42␤42␤foo␤»
not_gerd bbkr, Timbus: ^^
Timbus hah 11:47
jnthn .^find_method will end up spec when I get around to spec'ing MOP bits, I suspect. 11:48
not_gerd r: my $foo = (42 but "foo") but role { method Numeric { self + 0 } }; .say for $foo, +$foo, ~$foo 11:50
p6eval rakudo abb8f3: OUTPUT«foo␤42␤foo␤»
bbkr not_gerd: thanks for nice example 12:18
masak <Timbus> man domt tell me .* was re-specced 12:26
<Timbus> because that would pretty much ruin all of the things i did in perl 6 D:
(fortunately .* wasn't re-spec'd)
Timbus peers in
excellent
masak protos were :)
but this brings up a risk that no-one ever brings up, I think: 12:27
as the totality of the spec congeals and solidifies.
we're ever more restricted in what changes we *can* bring in without losing downstream happy-appy peeps.
it's like we're building up a baggage of ecosystem, and we have to take care of it by not doing too much of the upheaval thing. 12:28
Timbus well thats fine. i just dont want useful things to keep vanishing. 12:29
:I
masak that might be my single biggest critique of the OKness proposal -- it's not that it doesn't have good intentions, or even will lead to improvements in how the language works.
but to me it feels "too late" in some sense to introduce such a big change into the spec.
mikec_ i think as a user you accept that to some degree 12:31
jnthn masak: Yes, but by re-spec'ing protos the upshot is that the behavior of .* will be different on multis.
masak jnthn: right but it's still not like Timbus' nightmare that it disappeared altogether. the code can probably be rewritten to still use .* in a way that works. 12:31
I'm sure TimToady++ has considered the delay that such a change would introduce to the whole universe of compilers and modules and apps. he seems to be more willing to take such a cost than I am. 12:32
Timbus well it would be nice to have modules dispatching to events via some kind of.. existing dispatch mechanism. is all.. the larger scope and implications of this okness thing. waaay over my head 12:34
masak jnthn: wouldn't an appropriate proto in the base class make Timbus' code runnable again? 12:35
r: class A { multi foo(Int $b) { say $b } }; A.*foo("OH HAI") 12:36
p6eval rakudo abb8f3: ( no output )
masak hm, why does this survive but Timbus' case die?
Timbus r: class A { multi foo(Int $b) { say $b } }; A.new.*foo("OH HAI") 12:37
p6eval rakudo abb8f3: ( no output )
Timbus shrugs
flussence r: class A { multi method foo(Int $b) { say $b } }; A.new.*foo("OH HAI") 12:38
p6eval rakudo abb8f3: OUTPUT«Cannot call 'foo'; none of these signatures match:␤:(A , Int $b, Mu *%_)␤␤ in method foo at src/gen/CORE.setting:310␤ in method dispatch:<.*> at src/gen/CORE.setting:839␤ in block <anon> at /tmp/yYQz6hx3Rv:1␤␤»
Timbus oh
method
jnthn masak: 'cus you missed the method :) 12:39
masak: (appropriate proto) huh? protos don't disambiguate any more...
uh, serve as fallbacks I mean...
[Coke] drinks coffee. 12:40
phenny [Coke]: 10:42Z <not_gerd> tell [Coke] the intval/opcode_t but is filed as github.com/parrot/parrot/issues/705
masak whaddya mean "missed the method"? wasn't Timbus' case also just a failed signature binding? 12:40
jnthn masak: Timbus one also just had "multi", not "multi method" 12:41
flussence++ got it right :)
masak oh!
silly masak, missed typing "method". 12:42
jnthn: and you're saying adding a proto method won't fix that? :(
masak ok, let me broaden the question then: how should the base class look for Timbus-style hooks in deriving classes to be possible? without getting those failed signature binding errors. 12:43
[backlog] hm, should .trans have a case-insensitivity flag? or will that create more trouble than it soaks up? 12:44
jnthn masak: I don't get what you mean about "adding a proto method". There already *is* a proto method - the generated one.
If you want any dispatch to work out, just write in the base class a multi foo(|$) { } 12:45
masak ah. yes.
jnthn That'll sort looser than anything else.
masak \o/
jnthn You could probably even write a trait so you can do proto method foo(|$) is failless { * } 12:46
(and it just adds the candidate I suggested above) 12:47
masak r: sub foo(Bool $b = sub { False }) {}; say "alive"
p6eval rakudo abb8f3: OUTPUT«alive␤»
masak jnthn: would it be possible to get a compiler warning (or even an error) for the case of non-smartmatching defaults? 12:48
jnthn r: sub foo(Bool $b = sub { False }) {}; foo(); say "alive"
p6eval rakudo abb8f3: OUTPUT«Nominal type check failed for parameter '$b'; expected Bool but got Sub instead␤ in sub foo at /tmp/y1PGydjjlZ:1␤ in block <anon> at /tmp/y1PGydjjlZ:1␤␤»
jnthn phew
masak :)
jnthn masak: If you mean "non-typechecking"?
Then probably, yes.
If we know the type of the RHS, which we do in this case. 12:49
Guess an error is fine.
masak aye. 12:49
jnthn I mean, if it's gonna fail anyway... :)
masak not necessarily.
not if you always call it with the argument provided.
but it's a ticking bomb, for sue.
sure*
masak .oO( and some angry end user might sue, for sure ) 12:50
jnthn: will you be helped by me submitting a TODO RT ticket 'bout it?
should I make a spec change saying that we should explicitly check for this?
s/we/compilers/
jnthn masak: We don't tend to spec such things as "must". But a TODO RT is fine. 12:52
Or just implement it...it's only gonna be 3 or so lines of code. :)
masak submits TODO rakudobug, and might come around to implementing it leytur 12:53
jnthn (in parameter action method...you already have the nominal_type to hand, and can see if the default has a known compile time value, which we already consider anyway)
masak jnthn: that's cool. means I can probably do it all by myself ;) 12:54
pmurias fglock: hi 12:56
fglock: i want to hook in an optimalisation pass/backend to Perlito5 as an external process what's the best way to do that? 12:57
dalek blets: a9ae891 | (Herbert Breunung)++ | docs/appendix-g-glossary.txt:
polish entry for sigil
 13:05
masak .oO( many Polish entries ) 13:06
pmurias shouldn't the tablets be linked to the spec? 13:08
masak pmurias: by "be linked to", do you mean "link to"? 13:29
pmurias masak: yes 13:37
masak sure, why not? 13:41
pmurias i'm not volunteering for that just suggesting ;) 13:49
fglock pmurias: hi 13:53
I think you could use the -Cast-perl5 output
alternately you need to plug it in at the compiler itself, in 2 places: perlito5.pl, and eval() 13:55
pmurias fglock: what's i'm currently doing is adding an -Cast-hs option 13:59
i want to write the optimizer in haskell so i can't use the hoopl library
fglock you will probably want to read the AST back to emit code, it should be easy to add a new switch that does that 14:00
nice
or just run haskell!
p5 -> AST -> haskell -> bootstrap :) 14:01
pmurias how will BEGIN block be handled in Perlito5? 14:03
fglock it is executed by eval(), and the result is serialized - the details are a bit more complicated, see README-perlito5-js#Compile-time / Run-time interleaving (TODO) 14:05
fglock the -Cast-* output happens after all BEGIN and "use", it should be ready to run (but there is a long way ahead implementing all these) 14:11
pmurias BEGIN {$foo = create_a_complicated_object()} is most concerning thing 14:13
fglock: we will serialize the whole environment? (like kp6 used to do?) 14:14
fglock yes 14:16
and if you need eval(), we throw in a compiler too 14:17
fglock (perlito is a kp6 without the slow parts) 14:20
[Coke] # 05/13/2012 - rakudo++ (22080); niecza (92.24%); pugs (35.26%) 14:40
masak p6: say (2, 3, 4) leg (2..4)
p6eval rakudo abb8f3, niecza v17-4-ge5b869e: OUTPUT«Same␤»
..pugs: OUTPUT«0␤»
masak p6: say (2, 4) leg (2..4)
p6eval pugs: OUTPUT«1␤»
..rakudo abb8f3, niecza v17-4-ge5b869e: OUTPUT«Decrease␤»
dalek ast: 03de2f7 | (Kyle Hasselbacher)++ | S32-num/base.t:
[base.t] Test for RT 11290
 14:57
[Coke] hurm. I can't build rakudo with --gen-parrot behind work proxy. 15:04
moritz :(
[Coke]: does 'git pull' work for both ssh and git:// URLs?
[Coke] [email@hidden.address] works. git:// does not. 15:05
the former goes through the http proxy correctly.
masak [Coke]: I've had the same at times. 15:06
I just do the pull manually, and re-run the Configure.pl --gen-parrot
if it finds a new enough nqp and parrot, it doesn't try to git-pull it.
tadzik I can't use git protocol on my university either 15:07
moritz [Coke]: if you clone parrot once with the right URL yourself, rakudo will re-use the right URL
[Coke] I'd rather make the proxy work. ah well.
dalek ast: 8ba936c | (Kyle Hasselbacher)++ | S32-str/split-simple.t:
[split-simple.t] Test for RT 112868
 15:32
[Coke] if we switched to using git submodules, would git clean -xdf in rakudo not wipe out t/spec, nqp/, and parrot/ ? 15:34
masak does anyone actually use git submodules? 15:36
(honest question, not trying to be sarcastic)
r: class A { has $!x; submethod BUILD { my $!x = 42 }; method foo { say $!x } }; A.new.foo 15:37
p6eval rakudo abb8f3: OUTPUT«===SORRY!===␤Cannot use ! twigil on my variable␤at /tmp/8Zmo0LUNcp:1␤»
masak :)
masak b: class A { has $!x; submethod BUILD { my $!x = 42 }; method foo { say $!x } }; A.new.foo 15:38
p6eval b 922500: OUTPUT«42␤»
masak right.
[Coke] masak: I don't. 15:39
moritz there's a newish tool, 'git subtree', which supports a slightly different working style and which has been praised by many people 15:40
I haven't used it myself, can't comment on how good or how fitting it is
tadzik masak: my coworker was using them some time ago, he was mostly cursing and getting rid of them
JimmyZ git subtree is a good tool 15:43
moritz but would it be a good fit for rakudo to include parrot and nqp as subtrees? 15:44
masak I've also heard about 'git subtree'. maybe worth checking out. but I'm not actually sure there's a problem there that particularly needs fixing.
I mean, company proxies will exist even if we employ 'git subtree'. 15:45
moritz troo
masak and otherwise things seem to be working pretty well.
moritz because we've written our own tools, which we still need to maintain 15:45
masak yeah.
JimmyZ moritz: masak see github.com/zhuomingliang/skytools, there is a lib dir , but I don't how it works 15:46
[Coke] rakudo seems happy enough building with perl 5.16 15:48
(OS X 10.7.3)
moritz [Coke]++ for testing
moritz I'm not surprised, our configure system doesn't do anything "fancy" 15:48
jnthn decommute, at last & 15:56
bbkr std: /[]/ 15:57
p6eval std 8632387: OUTPUT«===SORRY!===␤Null pattern not allowed at /tmp/MMjcI7Gz6a line 1:␤------> /[⏏]/␤ expecting regex atom␤Parse failed␤FAILED 00:00 41m␤»
[Coke] building rakudo is probably the most complicated thing I use perl5 for in a week. 16:00
*sadface*
goraki [Coke]: Do you not have problems building from nom on Lion?
[Coke] I'm using 10.7.3. I'm sorry, I have no idea if that's lion. 16:01
goraki os x, right?
[Coke] 11:48 < [Coke]> (OS X 10.7.3) 16:02
bbkr 10.7.x = Lion, 10.8.x = Mountain Lion
goraki yeah, that. (-: it's Lion, also the same subversion I'm running - which is why I asked.
[Coke] goraki: works fine here, using rakudo nom, i686-apple-darwin11-llvm-gcc-4.2 (GCC) 4.2.1 16:03
you having a build issue?
goraki yeah.
during the building of parrot it seems.
(during perl Configure.pl --gen-parrot) 16:04
[Coke] throw the build log somewhere I can see it?
bbkr also tell us which XCode do you use 16:05
goraki XCode Version 4.3.2 (4E2002)
build log coming. 16:06
bbkr have you installed it directly from AppStore on clean system or over 4.2 standalone series? 16:07
goraki it wasn't a clean system I'm afraid. 16:08
[Coke] my gcc is from xcode 4.3 (4E109)
goraki this is an upgrade from a 10.6.x, and I've been upgrading Xcode since I got the machine (a couple of years ago). 16:09
masak niecza: say 4 ~~ Failure ?? 5 !! 42 16:12
p6eval niecza v17-4-ge5b869e: OUTPUT«===SORRY!===␤␤Expecting a term, but found either infix ?? or redundant prefix ?␤ (to suppress this message, please use space between ? ?) at /tmp/3SRuSARtcP line 1:␤------> say 4 ~~ Failure ??⏏ 5 !! 42␤␤Parse failed␤␤»…
masak sorear: why do I get this? Failure is a term, no?
std: say 4 ~~ Failure ?? 5 !! 42
p6eval std 8632387: OUTPUT«ok 00:00 41m␤»
masak submits nieczabug
bbkr goraki: I remember that after Xcode 4.3 upgrade I had to explicitly select Xcode version to use (despite the fact that I uninstalled standalone 4.2). "sudo xcode-select -switch /Applications/Xcode.app/Contents/Developer/" - as far as I remember 16:14
goraki bbkr: I've uninstalled the old version, so in theory there should be no conflict. 16:15
goraki however I'll double check. 16:15
geekosaur in theory. in practice, uninstalling doesn't remove saved state (like the xcode-select -switch value) etc. 16:16
...which will be /Developer if you have used an earlier Xcode
bbkr hence explicit selection of AppStore version in Applications/ 16:17
goraki right, going to check that now. 16:18
in the mean time, here's the end of the build log.
pastebin.com/NBnaQ24s
I can post the rest of it if you like, but it's long, obviously.
sorear masak: Failure is an unknown name, so it's treated as a postdeclared listop ;) 16:19
masak ah. 16:20
sorear masak: also, dunno if you saw, but you got the first link in announce.v17 wrong
masak :/
bbkr goraki: do you build 2012.04 release or HEAD version? 16:22
goraki bbkr: I can build 2012.04 fine. HEAD doesn't. 16:23
bbkr: that's the odd thing.
dalek ecza: fe453ec | sorear++ | docs/release.txt:
fix small glitch in release.txt skids++
masak sorear: I'm not quite synched up with the backlog yet.. 16:26
masak sorear: ah. see github.com/sorear/niecza/commit/b8...72b2ac4a31 -- this was the only place where the release guide led me wrong. I could fix it in FETCH_URL -- didn't even think about the announcement, which was already sent at that point. 16:29
I don't remember exactly what it was about the release guide that was confusing. only that it was.
jnthn home
masak jnthn! \o/ 16:30
dalek blets: 590b09c | (Herbert Breunung)++ | docs/appendix- (2 files):
getting quoting terms in better shape
masak greetings, adu. 16:31
colomon masak: no Failure in niecza
niecza: Failure 16:32
p6eval niecza v17-4-ge5b869e: OUTPUT«===SORRY!===␤␤Undeclared name:␤ 'Failure' used at line 1␤␤Unhandled exception: Check failed␤␤ at /home/p6eval/niecza/boot/lib/CORE.setting line 1364 (die @ 3) ␤ at /home/p6eval/niecza/src/STD.pm6 line 1147 (P6.comp_unit @ 33) ␤ at /home…
goraki bbkr: so far no joy switching the Xcode instance, I've done everything apart from reboot.
bbkr goraki: file from your backlog that causes trouble is in NQP, not Parrot
goraki ahhh, right, sorry. 16:33
masak colomon: yeah, sorear explained.
masak no &fail either, IIRC. 16:34
sorear o/ colomon 16:35
[ good * #perl6 ]
[Coke] looks like one of the 3rd party libraries isn't quite up to date with latest xcode. wonder if there's a later version we can merge in.
jnthn [Coke]: Which one, ooc? 16:36
bbkr goraki: unfortunately I know very little about NQP :(
goraki bbkr: okay, well don't worry about it. (-:
bbkr: before looking at it seriously I'm going to upgrade Xcode and double check that there is no legacy stuff hanging around, otherwise I could be chasing ghosts. 16:37
dalek blets: 5eed3a4 | (Herbert Breunung)++ | docs/appendix-g-glossary.txt:
prepare to wrap up argument related terms
jnthn [Coke]: ah, nm, just saw the link to the log
goraki bbkr++: thanks for your help.
jnthn Hmm. dyncall related... 16:38
geekosaur is on his second "newer xcode breaks stuff" in 12 hours... increasingly glad he stayed at 4.2.1 16:39
(mind, would be nice to be able to use it, but so much stuff still breaks...)
sorear why does xcode break stuff? 16:43
tadzik 'evening 16:44
sorear greetings tadzik.
masak tadziku! \o/ 16:45
tadzik masaku, sorearze! 16:47
geekosaur sorear, Apple moved the internals of Xcode around in the 4.3 releases; a lot of stuff that refers to said internals has not caught up yet 16:48
sorear are those declensional endings?
ingy waves
geekosaur ideally nothing would refer to them directly, in practice it's still necessary sometimes
tadzik sorear: yes
geekosaur although the changes in 4.3 should hopefully make it less necessary once stuff is rewritten to use it... 16:49
masak sorear: vocative.
tadzik if "declensional" is what I think it is :)
masak yes. 16:50
masak there's a metaphor going on there, about how nouns are being "bent" (declined) for various purposes. 16:51
PerlJam masak: so ... someone who does that can be reasonably considered a "bender"? 16:52
masak grammar. it's addictive. 16:53
dalek ast: efd9d5b | (Kyle Hasselbacher)++ | S05-mass/rx.t:
[S05-mass] Test for RT 112450
ast: cb72ab0 | (Kyle Hasselbacher)++ | S05-mass/rx.t:
[S05-mass/rx.t] Test for RT 112454

These are the only tests in this file that use $/, so maybe they should be somewhere else.
sorear masak: What is 'sorearze
sorear masak: What is 'sorearze'? w:Polish_morphology#Nouns doesn't mention -ze 16:53
masak I bet it's also the vocative. 16:54
masak the vocative doesn't seem regular, though I bet there are patterns. 16:54
haven't figgered them out yet.
it's "masaku" and "tadziku" but "sergocie" and "sorearze". 16:55
sergot :)
sorear en.wikipedia.org/wiki/Vocative#Polish 16:57
tadzik sorear: I can't really tell how this works. sorearze is like "boże" or "serze", "bóg" and "ser" being god and cheese consequently
sorear nods
tadzik I have to admit that I never learned the rules behind those things, I just can use them properly 16:58
masak decommutes 16:59
moritz moritz.faui2k3.org/tmp/changelog-lines.svg # number of changelog entries per rakudo release 17:04
for that I've lumped the 2012.04.1 and the 2012.05 releases together 17:05
gist.github.com/2695100 is what I used to extract that info 17:07
az5112 Funny, I have just checked and google translate thinks that both sorear and sergot are females :) 17:08
sorear does polish have the silly gramattical gender = social gender for humans thing?
moritz az5112: well, 'sorear' is neither a word nor a traditional name, and 'sergot' is a lst name 17:09
jnthn++ # Grammar::Trace 17:12
az5112 Yeah, but I entered "My dear Sorear/Sergot" to force the translator to assign the gender (for the adjective), and it always picks the adjective form for the female; so I guess there must be some variable controlling political correctness somewhere in the translator :)
moritz :-) 17:14
az5112 Funny... It doesn't do that for French though... 17:18
az5112 My dear sorear is Mon cher sorear 17:19
Nor for Russian 17:20
az5112 Nor for Slovak 17:21
Nor Slovenian!
So it's either the bug or they have a female programmer on the team! :D 17:22
spider-mario it’s a shame google can’t translate from or to lojban ! 17:22
-' '
I’m too used to putting a space before “!”, “?”, “;” and “:”
we do in French
(a thin non-breakable space, more specifically) 17:23
jnthn moritz: Nice graph!
sorear it can't do Japanese either :|
dalek ast: a4148aa | (Kyle Hasselbacher)++ | S29-context/eval.t:
[eval.t] Test for RT 112472 -- bad error message for eval of bogus language
 17:24
KyleHa Hi #perl6. 17:24
sorear graph? where?
tadzik hi KyleHa
sorear Hello, KyleHa
tadzik moritz.faui2k3.org/tmp/changelog-lines.svg
sorear: ^
jnthn o/ KyleHa 17:25
KyleHa Hey, jnthn, you can help me!
sorear masak: I have a vague desire to see that integrated.
KyleHa I put a test for this bug: rt.perl.org/rt3/Ticket/Display.html?id=112472
sorear moritz: I have a vague desire to see that integrated.
KyleHa I want to know if the test is good enough to close the ticket, or if it needs to check more. 17:26
jnthn KyleHa: Looks decent to me. 17:27
KyleHa Thanks jnthn!
Is Moritz in the house? 17:28
My stomach is burning. I'll come back when I'm fed. 17:30
jnthn ooh, getting fed is a good idea.. 17:30
adu masak: greetings 17:49
TimToady intercepts masak's greetings en passant and replies o/ on behalf of mashark^Wmasak 17:50
TimToady is attempting to backlog with migraine jaggies, but at least they've moved off my fovea now 17:51
adu ooo the chess rule nobody knows about 17:52
I realize now that masak's hello was over an hour ago 17:53
TimToady: have you tried Aspirin? 17:54
TimToady I am not allowed to take aspirin; I get GI bleeding 17:56
no NSAIDs for me, except aceteminaphen
TimToady doesn't really want to end up in the hospital again 17:57
geekosaur ...which isn;t an NSAID :p
PerlJam TimToady: Every time I hear about some health issue that you have, the more I think Perl a miracle.
TimToady geekosaur: it's non-steroidal, and it's somewhat anti inflammatory :P
sorear is it 'only' an antipyretic/analgesic? 17:58
TimToady fortunately this is (I hope) just an ophthalmic migraine
geekosaur (and it doesn't do anything noticeable to me unless I take so much that I'm risking worse liver problems than the stomach issues from NSAIDs...)
TimToady so while I have the "mean feeling" of a migraine, it doesn't really hurt
geekosaur ah, bodies. more ways to be messed up than there are people... 17:59
TimToady well, eventually your doctors give you completely contradictory advice, and then you die :)
geekosaur they already have, a few times.
TimToady well, then you're gonna die :P 18:00
geekosaur 's social circle somewhat encourages talking about "the last time my doctor tried to kill me"
adu TimToady: like "get plenty of rest" while you're working 60-hour work weeks
TimToady sure, but they didn't tell you to work the 60-hour work week 18:01
I'm supposed to eat a low-fat diet because of family history, and I'm suppoed to eat a high-fat diet because of having had a partial gastrectomy 18:02
slavik1 TimToady: eat something in between?
medium fat diet?
TimToady mostly I try to eat healthy fats along with the unhealthy ones :)
slavik1 potato chips for the win :) 18:03
[Coke] TimToady: bleargh. i do wish I could just have food handed to me. I do a terrible job picking out things that I should eat.
sorear Life with organs is hard. Brain in a jar sounds like a pretty good deal to me. 18:04
TimToady except it's pretty hard to clone yourself without sacrificing brain tissue that way
plus bits of your brain can also be construed as organs 18:07
the problem isn't so much organs and organics... 18:08
ingy date +%s|cut -b1-4
TimToady s/and/as/
ingy++ is caught redhanded lurking 18:09
sorear Yeah. Down with organics.
[Coke] ... if you're caught redhanded, does that make you move more gingerly next time?
sorear caught redheaded! 18:10
TimToady oddly, pipe organs are mostly not organic
but organ pipe cactuses are, go figure...
[Coke] runs igny's command. nifty. 18:11
ingy It's 1337 week!
sorear o/ fglock 18:12
TimToady is this something intended to offset Mother's Day?
fglock o/
moritz niecza: say (4 ~~ Failure) ?? 5 !! 42 18:17
p6eval niecza v17-4-ge5b869e: OUTPUT«===SORRY!===␤␤Undeclared name:␤ 'Failure' used at line 1␤␤Unhandled exception: Check failed␤␤ at /home/p6eval/niecza/boot/lib/CORE.setting line 1364 (die @ 3) ␤ at /home/p6eval/niecza/src/STD.pm6 line 1147 (P6.comp_unit @ 33) ␤ at /home…
moritz masak: ^^ Failure is simply not declared in niecza
TimToady Failure is not an option...
adu ingy: is that unix timestamp? 18:18
spider-mario r: (True ^^ True).perl.say
p6eval rakudo abb8f3: OUTPUT«Nil␤»
masak moritz: yes, I know. sorear++ explained it.
masak and then colomon. 18:18
sorear adu: yes
masak it's quite amazing that Niecza is as developed as it is, and yet doesn't have the Failure type.
adu oo the second billenium is in 2033 18:20
masak :P 18:22
adu we should plan a huge party somewhere 18:23
moritz wow, I've got a /msg on perlmonks "Perl6 grammar capability is simply beautiful!" about that gist that parses the ChangeLog
masak I'm reading the diakopter/gfldex discussion from yesterday, and I want to jump in and make all the miscommunication go away, but I can't because it happened in the backlog, not now.
adu moritz: well, it is
masak moritz: where's the gist? /me curious 18:24
moritz in the backlog :-) gist.github.com/2695100
gfldex masak: could we ease your pain by pretending it never happened? :-> 18:24
moritz I used it to produce moritz.faui2k3.org/tmp/changelog-lines.svg
masak nah. I'm not wired that way. 18:25
masak but it would be interesting to heard what you learned from it. 18:25
jnthn hah, moritz++ also used Grammar::Tracer to debug it :D
masak gfldex: you held and defended a staunchly fixed and fiercly wrong opinion against a compiler implementor. what was the fallout of that? 18:26
moritz: thanks. I'm not done with the lackbog yet. :)
that's why I'm looking so confused and underinformed.
gfldex there was no fallout on my side, at least non that i could notice
moritz jnthn: heh, yes. My grammar worked from the very start, except that it failed at the very end of the string (because there's no newline after the last changelog entry) 18:27
jnthn moritz: oh :)
diakopter niecza: my $b = ->($a=->{}) { say $a }; $b($b)
p6eval niecza v17-4-ge5b869e: OUTPUT«Unhandled exception: Unable to resolve method Capture in class Block␤ at <unknown> line 0 (ExitRunloop @ 0) ␤ at /tmp/NBMArHlyvZ line 0 (ANON @ 1) ␤ at /tmp/NBMArHlyvZ line 1 (mainline @ 2) ␤ at /home/p6eval/niecza/lib/CORE.setting line 3842 (ANON @ 3) …
masak gfldex: hm, I just meant "fallout" as in "the world events that this resulted in". i.e. maybe you realized something.
moritz jnthn: and once I ran it with Grammar::Debugger, it took so long that I figured out the solution on my own before the run finished 18:28
masak <gfldex> $ is a marker for the compiler that tells it it's not a function
diakopter oh boy
masak gfldex: I actually see where you're coming from with this one.
gfldex: but you made a crappy job at explaining it so that diakopter understood the point of it :) 18:29
diakopter I understood the point; my point was the statement equally applies to %,@
masak granted. 18:30
gfldex's point seems to have been about sigils being noun markers.
dalek blets: 9d0bd55 | (Herbert Breunung)++ | docs/appendix- (2 files):
sharpening some bits
masak but he mucked it up by going on a tangent about holding function values at runtime. 18:31
gfldex because the person who asked the original question seam to have a fixation on types, what failed for him when he tried to understand $ 18:32
sorear masak: see also me getting frustated with the direction the convo was going and leaving 18:33
masak aye.
sorear: but you got frustrated with Araq. 18:34
I'm surprised that the discussion mentioned "Perl 6 won't make sigils optional" without once mentioning the new \ sigil and how it makes sigils optional. 18:40
TimToady "won't make optional" is code for "won't banish them" 18:41
masak I guess.
TimToady but I'm also a bit surprised 18:42
masak when I see someone arguing for banishing sigils in Perl 6, I flip the bozo bit. 18:42
gfldex masak: i did mention \ 18:42
lichtkind is it declaring to a .. or declaring as a .. ?
ha
masak gfldex: oh! url?
lichtkind: what's the full sentence? 18:43
lichtkind: it's like asking "is it 'passing in a ...' or 'passing over a ...'?" both are correct.
gfldex irclog.perlgeek.de/perl6/2012-05-13#i_5587077
lichtkind masak: the keyword role declares the following block to a role
gfldex and i would still like to know what parcel binding is for 18:44
masak gfldex: oh, I missed that because I was feeling sorry for diakopter at that point, who had to counter your strange arguments :)
TimToady for leaving out the sigil :)
gfldex that is just nasty!
masak don't do it, then. 18:45
TimToady in my experience, masak++ is much more interested in the truth than in being nasty
diakopter I didn't know about \ 18:46
moritz gfldex: the idea is that you sometimes just want to pass something on without any modifications. But binding something to a sigiled variable always implies context
masak TimToady: oh, I thought "nasty" was a reaction to "for leaving our the sigil", not what I said.
TimToady maybe it was ;)
gfldex could very well be
masak :P 18:47
gfldex when you use the intarweb anything is possible!
masak in this instance I'm interested in doing an autopsy on a discussion with cross purposes and good intentions.
people who are benign can still fail to reach each other and make a genuine connection on IRC.
diakopter std: sub foo(label1: say 1; say 2) {} 18:48
p6eval std 8632387: OUTPUT«Use of uninitialized value $x in pattern match (m//) at STD.pm line 66537.␤Use of uninitialized value $x in concatenation (.) or string at STD.pm line 66576.␤===SORRY!===␤In parameter declaration, typename 'label1' must be predeclared (or marked as …
masak diakopter: I never cease to admire the twistedness of your mind :)
diakopter stdbug 18:49
let
sorear masak: I can't say with certainty how much of my snapping was caused by Araq, how much was caused by gfldex, and how much was caused by RL stress
n: sub foo(label1: say 1; say 2) {}
p6eval niecza v17-4-ge5b869e: OUTPUT«===SORRY!===␤␤In parameter declaration, typename 'label1' must be predeclared (or marked as declarative with :: prefix) at /tmp/5FRaY137_j line 1:␤------> sub foo(label1⏏: say 1; say 2) {}␤␤Parse failed␤␤»
diakopter stdbuglet being the unitialized $x warning
masak right. 18:50
diakopter: but signatures are their own slang. why would 'label1:' even make sense there?
diakopter sry; it doesn't...
masak that colon is bound to end up meaning 'invocant marker'.
diakopter I'm exploring semilist
masak ok.
I'm just assuming when I see that that you know something I don't :) 18:51
moritz r: say roundrobin(<a b c>; <d e>; <f g h i>)
p6eval rakudo abb8f3: OUTPUT«a d f b e g c h i␤»
sorear std: say(label1: say 1; say 2)
p6eval std 8632387: OUTPUT«===SORRY!===␤Undeclared routine:␤ 'label1' used at line 1␤Check failed␤FAILED 00:00 42m␤»
TimToady ==SORRY!== A diakopterism has been discovered at line 42; please seek a bomb shelter at your earliest convenience.
sorear std: say[label1: say 1; say 2]
p6eval std 8632387: OUTPUT«===SORRY!===␤Unsupported use of bare 'say'; in Perl 6 please use .say if you meant $_, or use an explicit invocant or argument at /tmp/RIf5LyAuDt line 1:␤------> say⏏[label1: say 1; say 2]␤Check failed␤FAILED 00:00 41m␤»…
geekosaur needs to figure out how to get x11-xchat to display higher unicode...
sorear std: die[label1: say 1; say 2]
p6eval std 8632387: OUTPUT«ok 00:00 41m␤»
colomon TimToady++
geekosaur (I did manage to get it to do the ANSI escapes... makes std and niecza errors more comprehensible) 18:52
sorear do you have a decision procedure for diakopterisms
TimToady well, first step would be to translate "Confused" to that, but it wouldn't help in this case... 18:53
diakopter niecza: 2[; 5[5]] 18:54
p6eval niecza v17-4-ge5b869e: OUTPUT«Use of uninitialized value in numeric context␤ at /home/p6eval/niecza/lib/CORE.setting line 1226 (warn @ 3) ␤ at /home/p6eval/niecza/lib/CORE.setting line 262 (Any.Numeric @ 6) ␤ at <unknown> line 0 (ExitRunloop @ 0) ␤ at /home/p6eval/niecza/lib/CORE.se…
TimToady diakopter: but yes, an arglist is not a statement or sequence of statements
diakopter k 18:55
TimToady what would foo(1,2,3 if $x) mean anyway? 18:57
same kinda issue 18:58
masak az5112: no, gender assignment doesn't have to have human intent behind it in Google Translate. it's all statistical translation based on large corpora of text. so if there's a lot of "my dear <female name>", maybe the algorithms catch on and think that's a rule. 18:58
az5112: there are funny cases of this. en fr "New York"? translating to "Paris" at some point, for example.
phenny: en fr "New York"?
phenny masak: "New York" (en to fr, translate.google.com)
masak now it gets it right :) 18:59
phenny: en sv "mockingbird"?
phenny masak: "glassbil" (en to sv, translate.google.com)
tadzik phenny: en pl "Homer Simpson"?
phenny tadzik: "Homer Simpson" (en to pl, translate.google.com)
masak *lol*
arnsholt Glassbil?!
masak that word in Swedish means "ice cream van".
how in the...?
[Coke] hopes masak has seen translationparty.com/
tadzik . o O ( to kill an ice cream van ) 19:00
masak [Coke]: ooh. no, I hadn't. 19:00
[Coke] to kill an ice cream van -> translation party -> To kill the ice cream fans 19:01
masak :P
tadzik Lasagna... working on my laptop, eat at the same time!
eat laptop?!?!?
arnsholt masak: But yes, Google translate is all based on statistics 19:02
tadzik all? So it reads webpages in two languages and compares them? 19:03
moritz basically, yes
tadzik amazing
moritz though it's much more sophisticated, of course :-) 19:04
masak gets off this train, onto a bus, to eventually end up on another train. exciting!
[Coke] urk. why can I not select the text at the top block in perl6.org?
tadzik naturally
[Coke]: I can
moritz [Coke]: dunno. I can.
[Coke] ah. I can as well, but it is not showing as hilighted.
bad css. no donut. 19:05
cogno I am trying to backlog. What is the okness proposal?
moritz cogno: it's a proposal by TimToady to overhaul the types that can indicate success and failure 19:05
[Coke] translation party's the first sentence of perl6.org and gets: 19:06
Hi, my name is Camelia. I'm the spokesbug for Perl 5, Perl 6, sexy sister.
moritz cogno: for example False, Failure, Nil and a fals Match object can express some kind of failed test/check
tadzik :)
moritz cogno: and TimToady wants to change that... somehow. The details were a bit scary to me, but I think I agree now with the overall goals 19:07
moritz there was a gist which summarized it all.. does anybody have the URL handy? 19:08
cogno I suppose that a smart compiler will eventually able weed out the uneeded part of the overloaded value. So if it makes life easier for the programmer I am all for that kind of thins 19:11
Oops i am a loosy typist, but with the iPad it gets even worse. 19:12
TimToady gist.github.com/1828667 19:13
cogno Thx 19:15
TimToady getting a bit of "normal" migraine, so better go sleep it off... &
KyleHa Hi #perl6. 19:17
tadzik welcome back
KyleHa Thank you. 19:18
moritz \o KyleHa
KyleHa Moritz! Question for you... I see now that spectest .t files have numeric plans everywhere and 'done;' at the end. Did "plan *" and "done_testing;" have a problem? 19:19
sorear KyleHa: yes 19:20
KyleHa: we need to be able to count tests without running them
moritz KyleHa: the main "problem" was that people wanted to know how many tests we had, and had a hard time figuring it out
KyleHa Spending a few minutes running the suite was not a good solution? 19:21
tadzik "few minutes", hehe
sorear 1. the entire suite cannot be run 19:21
moritz KyleHa: so, not my decision, but I relented to the majority
sorear 2. I think you mean a few hours
moritz depens very much on the machine 19:21
jnthn spectets in 3-4 minutes :) 19:21
moritz it's more like 5-10 minutes here on 4 cores 19:22
KyleHa I think it was less than ten minutes the last time I timed it, yesterday.
sorear good for you
flussence on my netbook it's most definitely >1h
KyleHa Do we still need "done;" if there's now a numeric plan? 19:23
moritz no
just a left-over that wasn't removed
KyleHa OK.
KyleHa WOW, roast/TODO hasn't changed since I did some cursory update in 2011. 19:25
moritz it's probably very out of date now
dalek ast: 5b9b4c7 | (Solomon Foster)++ | S12-methods/lvalue.t:
Unfudge tests that now work in Niecza.
 19:40
dalek blets: 612a102 | (Herbert Breunung)++ | docs/appendix-g-glossary.txt:
explaining the term role, please check it
 20:03
colomon perl6: say 'abc'.split('').perl 20:11
p6eval pugs: OUTPUT«("a", "b", "c")␤»
..niecza v17-4-ge5b869e: OUTPUT«["", "a", "b", "c", ""].list␤»
..rakudo abb8f3: OUTPUT«("a", "b", "c").list␤»
colomon spectests seem to now claim all these answers are wrong? 20:12
jnthn wat
moritz well
jnthn The pugs and Rakudo ones look reasonable.
moritz I think we have special-cased that in rakudo 20:13
KyleHa I wrote that test earlier today.
moritz so, what is expected now
KyleHa It was based on a ticket with a discussion in which TimToady had said "xyz".split('') should be ("", "x", "y", "z"). 20:14
moritz there are more leading than trailing empty strings on a string? :-)
PerlJam trailing are elided, right? 20:15
moritz no
anyway
rakudo implements split with strings differently than split with regexes
(ie uses index)
so the only possible zero-width match there is '' 20:16
so we special-case that anyway
we can make it do whatever we want, but I don't see how ('', 'x', 'y', 'z') makes any more sense than any other option 20:17
KyleHa Here's the discussion I saw: rt.perl.org/rt3/Ticket/Display.html?id=112868 20:18
moritz KyleHa: I see. I'm still not convinced. Maybe I'll be convinced tomorrow :-) 20:20
moritz -> sleep
KyleHa Rules were made to be rewritten.
PerlJam I could argue that there are an infinite number of emptry strings between each character in a string too ;)
jnthn moritz: Think I've got the roles pre-comp thing fixed locally. :) 20:21
KyleHa I hope we'd do that split lazily!
PerlJam I could also argue that split('') should elicit a warning and function as if it were split(/''/) 20:22
dalek p: 3ab8ab1 | jnthn++ | / (2 files):
Add an OwnedResizablePMCArray, along the same lines as OwnedHash.
 20:23
p: 5308279 | jnthn++ | src/6model/serialization.c:
Deserialize RPAs with their object ownership. This seems to fix the leak that saw parametric roles and pre-comp sometimes interact badly.
colomon KyleHa: just reading that link you posted again, it's not clear to me what TimToady was trying to say. 20:29
"arguably niecza is more consistent with how /''/ would split" 20:30
perl6: say 'abc'.split(/''/).perl
p6eval niecza v17-4-ge5b869e: OUTPUT«["", "a", "b", "c", ""].list␤»
..rakudo abb8f3: OUTPUT«("", "a", "b", "c", "").list␤»
..pugs: OUTPUT«Error eval perl5: "if (!$INC{'Pugs/Runtime/Match/HsBridge.pm'}) {␤ unshift @INC, '/home/p6eval/.cabal/share/Pugs-6.2.13.20120203/blib6/pugs/perl5/lib';␤ eval q[require 'Pugs/Runtime/Match/HsBridge.pm'] or die $@;␤}␤'Pugs::Runtime::Match::HsBridge'␤"␤*** Can't locate P…
PerlJam wishes there was a way to say "perl6: ..." and only have the non-error ones output 20:31
colomon PerlJam: but errors are sometimes important too... 20:32
PerlJam colomon: sure. I'm wishing for more, not less
colomon PerlJam: oh, gotcha. perl6filtered: or something
PerlJam aye
KyleHa I think TimToady was saying that split('') should do the same thing as split(/''/) does, which produces a leading empty string. 20:33
PerlJam maybe "perl6!:"
KyleHa: leading and trailing, yes?
perl6: say "abc".split(/''/); 20:34
p6eval rakudo abb8f3, niecza v17-4-ge5b869e: OUTPUT« a b c ␤»
..pugs: OUTPUT«Error eval perl5: "if (!$INC{'Pugs/Runtime/Match/HsBridge.pm'}) {␤ unshift @INC, '/home/p6eval/.cabal/share/Pugs-6.2.13.20120203/blib6/pugs/perl5/lib';␤ eval q[require 'Pugs/Runtime/Match/HsBridge.pm'] or die $@;␤}␤'Pugs::Runtime::Match::HsBridge'␤"␤*** Can't locate P…
KyleHa Oh. Duh. How did I not see the trailing empty string ten times?
PerlJam perl6: say "abc".split(/''/).perl;
p6eval niecza v17-4-ge5b869e: OUTPUT«["", "a", "b", "c", ""].list␤»
..rakudo abb8f3: OUTPUT«("", "a", "b", "c", "").list␤»
..pugs: OUTPUT«Error eval perl5: "if (!$INC{'Pugs/Runtime/Match/HsBridge.pm'}) {␤ unshift @INC, '/home/p6eval/.cabal/share/Pugs-6.2.13.20120203/blib6/pugs/perl5/lib';␤ eval q[require 'Pugs/Runtime/Match/HsBridge.pm'] or die $@;␤}␤'Pugs::Runtime::Match::HsBridge'␤"␤*** Can't locate P…
dalek kudo/nom: d1f5db2 | jnthn++ | / (2 files):
Bump NQP_REVISION to get bs fixes; a small tweak in Rakudo to account for the change too.
slavik1 doesn't that say that there is an assumed "" in the beginning of every string? 20:35
KyleHa I'll "fix" the test in a few minutes.
PerlJam slavik1: It's all somewhat arbitrary if you ask me. Why do we care about certain zero-width assertions in regex, but not others? Why are ^ and $ so important? 20:37
Why not ^^ and $$ too?
dalek kudo/nom: 570f2bb | jnthn++ | docs/ChangeLog:
Few ChangeLog updates.
PerlJam why not <ww> or <wb>?
dalek kudo/nom: 7d33ee1 | kboga++ | / (2 files):
Adds a slightly adjusted but mostly copy pasted version of ng's pack & unpack
 20:42
kudo/nom: f2e5c38 | jonathan++ | / (2 files):
Merge pull request #66 from kboga/ng-pack-unpack

Adds a slightly adjusted but mostly copy pasted version of ng's pack & unpack.
dalek ast: 1641f44 | (Kyle Hasselbacher)++ | S04-phasers/end.t:
[S04-phasers/end.t] Test for RT 112408 -- Can use END as a bareword hash key
 20:44
ast: 4e65cd6 | (Kyle Hasselbacher)++ | S12-methods/what.t:
[S12-methods/what.t] Test for RT 112364 -- Parse Foo:: correctly
ast: fead744 | (Kyle Hasselbacher)++ | S32-str/split-simple.t:
[split-simple.t] Change test for RT 112868
ast: 01fd0e8 | (Kyle Hasselbacher)++ | S12-methods/lvalue.t:
Merge branch 'master' of github.com:perl6/roast
dalek kudo/nom: eb05804 | jnthn++ | docs/ChangeLog:
Note pack and unpack in ChangeLog.
 20:51
Woodi hallo everyone 20:54
slavik1 PerlJam: I would say that ^ and $ anchors are special case for regex, since that is the notation for "begins with" and "ends with" vs. contains 20:55
geekosaur there's the whole huffmanization thimng, ^ and $ are extremely common 20:57
PerlJam slavik1: so why not ^^ and $$? Those are "begins with" and "ends with" for a line. 21:06
slavik1 PerlJam: are they the same for Perl5? this might be why I am confused (I haven't played with Perl6 much since rakudo* came out). 21:07
PerlJam r: say "".^methods.map(*.name); 21:10
p6eval rakudo abb8f3: OUTPUT«BUILD Int Num chomp chop substr pred succ ll-match match-list ords lines samecase samespace trim-leading trim-trailing trim words encode capitalize trans indent WHICH Bool Str ACCEPTS Numeric gist perl comb match subst split␤»
PerlJam r: say rx/''/.^methods.map(*.name); 21:11
p6eval rakudo abb8f3: OUTPUT«Unmarshallable foreign language value passed for parameter '$x0'␤ in block <anon> at /tmp/XI_iIC5rVM:1␤␤»
PerlJam That's not quite what I expected
r: say rx//.^methods.map(*.name); 21:12
p6eval rakudo abb8f3: OUTPUT«===SORRY!===␤Unrecognized regex metacharacter (must be quoted to match literally) at line 1, near ".^methods."␤»
PerlJam I guess the rx/''/ was split paranoia on my part
r: grammar X { }; say X.^methods(:all).map(*.name); 21:13
p6eval rakudo abb8f3: OUTPUT«Method 'name' not found for invocant of class 'NQPRoutine'␤ in block <anon> at /tmp/NVxKNGVJ8g:1␤␤»
PerlJam So ... classes are fully Perl6ified, but grammars aren't I guess? They still leak their implementation. 21:16
n: say rx//.^methods.map(*.name);
p6eval niecza v17-4-ge5b869e: OUTPUT«===SORRY!===␤␤Null pattern not allowed at /tmp/aInKYKf0rS line 1:␤------> say rx/⏏/.^methods.map(*.name);␤␤Parse failed␤␤»
PerlJam n: grammar X { }; say X.^methods(:all).map(*.name);
p6eval niecza v17-4-ge5b869e: OUTPUT«Unhandled exception: Unable to resolve method methods in class ClassHOW␤ at /tmp/s_QYhQBmzt line 1 (mainline @ 4) ␤ at /home/p6eval/niecza/lib/CORE.setting line 3842 (ANON @ 3) ␤ at /home/p6eval/niecza/lib/CORE.setting line 3843 (module-CORE @ 65) ␤ at …
jnthn PerlJam: It's not grammars at issue. It's code objects generally, I think.
PerlJam ok 21:16
jnthn Welcome to the exciting world of circularities. 21:17
PerlJam I had an idea to annotate strings with which grammar rules would match where, but if I can't get a list of rules, that makes it slightly more difficult. 21:18
jnthn Why can't you get a list of rules? 21:19
PerlJam I don't know how I guess.
masak home
jnthn r: grammar Foo { token TOP { <foo> }; token foo { xxx } }; Foo.^methods(:local)>>.say
p6eval rakudo eb0580: OUTPUT«foo␤TOP␤»
jnthn shark, masak 21:20
masak orca!
jnthn whales at the pun 21:21
PerlJam this obsession with sea creatures makes little sense to me too
jnthn PerlJam: I is a little bit fishy...
er, It is... :)
diakopter I won't carp about it 21:22
masak jnthn: neither the whale nor the orca are actually... oh, fuggedaboutit...
PerlJam I guess both orca and sharks are in the same phylum, so ... maybe there's some sense to be had.
masak heh.
PerlJam: your mom is in the same phylum! 21:23
PerlJam masak: as is yours!
masak :P
PerlJam one good fish deserves another
masak srsly toothless "your mom" insults, part 18.
masak .oO( sushi: a dish best served cold ) 21:24
diakopter you can't bait me to join the puns 21:26
masak hook, line, and sinker. 21:27
PerlJam diakopter: #perl6 casts a wide net. 21:28
(though, if we were doing word association, "hook, line, and sinker." makes me immediatly think of "lock, stock, and barrel.") 21:29
though it doesn't improve my spelling. 21:30
diakopter If this were #csharp I could make a pun on sealed
masak car repair guy: "looks like you've blown a seal" -- car owner: "hey, leave my personal life out of this!" 21:31
diakopter you otter keep that one to yourself 21:32
pmichaud jnthn: ping
masak diakopter: I'll keep things above the waterline from now on. 21:33
jnthn pmichaud: pong
masak .u 乒乓 21:34
phenny U+4E52 CJK UNIFIED IDEOGRAPH-4E52 (乒)
U+4E53 CJK UNIFIED IDEOGRAPH-4E53 (乓)
pmichaud got your email -- I still plan to have mark commit done before release (which I know is Thursday)
spider-mario .u ௹ 21:35
phenny U+0BF9 TAMIL RUPEE SIGN (௹)
PerlJam masak: just tread softly.
spider-mario .u ৳
phenny U+09F3 BENGALI RUPEE SIGN (৳)
pmichaud things here have been more-than-the-usual busy-ness... but nothing too problematic
jnthn pmichaud: ooh :)
pmichaud: OK :)
diakopter masak: I keep herring the puns wrongly
masak PerlJam: and carry a big floatation device? :) 21:36
PerlJam :)
masak diakopter: are you familiar with www.songlyrics.com/kip-addotta/wet-dream-lyrics/ ? 21:37
diakopter no 21:38
masak feels relevant.
diakopter haha
masak :) 21:39
pmichaud "Wet Dream" is an awesome song :) 21:47
diakopter in a coffee shop, whose rights prevail, the person who has a medically necessary dog, or the person with dog allergies? 21:57
diakopter the manager is negotiating with them now. 21:58
diakopter that's California for you. 21:58
tadzik oh, precomp fixes 22:05
masak diakopter: seems a Solomonian problem. just cut the dog in half! 22:09
diakopter the person with the allergies won; seems the no-pets rule of the coffee shop can apply if the person is not waiting in line or for a drink. 22:11
masak 'night, #perl6 22:12
tadzik 'night
slavik1 diakopter: in other words, the dude who has not given his money away (yet) has right of way everywhere
night
masak dream of medically necessary allergic dogs waiting in line for a drink. 22:13
slavik1 ... 22:13
diakopter well, technically Starbucks reserves the right to refuse service to anyone, but it also has a corporate image/goodwill to uphold. 22:15
slavik1 as in "don't do evil"? 22:17
diakopter anyone, *with some exceptions*.
www.legalzoom.com/us-law/equal-righ...se-service 22:18
geekosaur as in "getting a reputation as being assholes to special needs customers can hurt their business" 22:21
jnthn -> sleep o/ 22:49
cognominal about the okness gist. how do one compare a gist and a fork of that gist? 23:09
[Coke] cognominal: check them both out and do a diff is the low tech way, I guess. 23:12
cognominal :( 23:17
dalek blets: 49ac38e | (Herbert Breunung)++ | docs/appendix-g-glossary.txt:
pilish role entry and add new
 23:19
geekosaur does seem like something github *ought* to provide, but I see no way to do it 23:22
(also, the idea looks quite sane to me. certainly no scarier than smartmatches already are...) 23:23
am0c cognominal, add forked gist remote, and pull to another branch. finally, diff two branches. (I guess) 23:27
cognominal thx. I hoped that github would have helped here 23:28
github has some weird blind spots. 23:29

Powered by Cro and the Raku® Programming Language.

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

Thank you!