»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.
colomon o/ 00:03
diakopter eagerly awaits Pm's appearance tomorrow 01:02
cognominal linuxfr published my news :) linuxfr.org/news/calendriers-de-lav...-et-perl-6 01:04
diakopter "About Perl 6, given the limited common with Perl 5 in terms of language, I wonder why it is still called Perl. I feel that there is about as much in common between Perl 5 and Perl 6 between the C and assembler (keeping in mind that we can carry the assembler in C very easily)." 01:06
the last comment, translated
sorear does the p6 faq have an entry for that yet? 01:08
cognominal I made a mistake about BigInt being a special type. I forgot that had become the regular Int.
diakopter is having fun using Eclipse for textbook-exercises in Java. what a great IDE. 01:09
sorear diakopter: I can't tell if that was sarcasm 01:26
colomon sorear: just added the ability to increase the number of iterations for a window 01:27
(not yet pushed)
diakopter sorear: not sarcasm 01:35
djanatyn moritz: I think I'm going to need an extension >_> 02:03
colomon sorear: just pushed a patch which allows you to write out any fractal you're viewing as a PPM file. 03:29
sorear nifty 03:32
moritz good morning 05:02
djanatyn: deadline extended by 24h
djanatyn moritz: :D 05:03
djanatyn thank you. I've been playing around with roast, and Tene told me a bit about &[+] 05:03
moritz djanatyn: I'd recommened to use the example code from the ticket, and replace the call to say() with a call to a test function 05:04
japhb_ On my local system: 06:36
> ':16«f.8*:4[2]**:2<1*2**2>»/-:60[1,0]'.Numeric.perl
-62/15
\o/
moritz \o/ 06:42
jnthn *yawn* 07:46
morning, #perl6
sorear hi jnthn
jnthn japhb_: Nice .Numeric work! :)
o/ sorear
commute, bbs 08:00
tadzik gmorning 08:18
moritz \o 08:20
sorear hello
moritz jnthn: t/spec/S32-array/splice.rakudo segfaults here 08:21
jnthn: gist.github.com/1432838 08:22
japhb jnthn, thank you. :-) 08:30
g'night all! 08:33
moritz japhb: have you pushed that to a branch_ 08:34
s/_/?/
japhb: if not, please do so very soon
... and good night to japhb :-)
japhb Sure, give me a sec.
OK, should be there. 08:35
str-numeric
rebased on to nom a few hours ago.
dalek kudo/str-numeric: 499ce20 | (Geoffrey Broadwell)++ | src/core/Str.pm:
WIP snapshot of Str.Numeric rewrite
kudo/str-numeric: 7e86fa8 | (Geoffrey Broadwell)++ | src/core/Str.pm:
Remove no-longer-needed workarounds
kudo/str-numeric: d3645a8 | (Geoffrey Broadwell)++ | src/core/Str.pm:
Refactor Str.Numeric parse failures and pos updates so that failure messages can show exact position of error in source string; wrap a few long lines in same code sections
kudo/str-numeric: 6d35d8c | (Geoffrey Broadwell)++ | src/core/Str.pm:
Support :16<f.8*2**3> and :16«f.8*:4[2]**:2<1*2**2>» style radix/multiplier/exponent style numbers in Str.Numeric; add missing ASCII value comment
japhb It should *not* be merged until Failure handling is resolved. 08:36
moritz understood
japhb OK, *now* sleep & 08:37
sorear sleep
moritz nomo: my class X::Foo { ... }; my class X::Foo { }; 08:44
nom: my class X::Foo { ... }; my class X::Foo { };
p6eval nom 48b979: OUTPUT«===SORRY!===␤The following packages were stubbed but not defined:␤ X::Foo␤␤»
moritz works with s:g/X/Y/
sorear nom: say X; 08:50
p6eval nom 48b979: OUTPUT«X()␤»
sorear nom: my package X { }; class X::Foo { ... }; class X::Foo { } 08:51
p6eval nom 48b979: ( no output )
moritz nom: class X::Foo { ... }; class X::Foo { };
p6eval nom 48b979: OUTPUT«===SORRY!===␤The following packages were stubbed but not defined:␤ X::Foo␤␤»
moritz nom: ny package X { }; my class X::Foo { ... }; my class X::Foo { };
p6eval nom 48b979: OUTPUT«===SORRY!===␤CHECK FAILED:␤Undefined routine '&ny' called (line 1)␤» 08:52
moritz nom: my package X { }; my class X::Foo { ... }; my class X::Foo { };
p6eval nom 48b979: ( no output )
moritz (FWIW, the 'my' sometimes made a difference in the past)
jnthn moritz: Delightful. A crash *inside* GC. :/ 09:09
tadzik yo dawg 09:10
jnthn moritz: I thought we patched that declaration thing last week...
moritz jnthn: it seems that the presence of other X:: packages make a difference 09:11
nom: my class Y::A { ... }; my class Y::A { }
p6eval nom 48b979: ( no output )
moritz without them, it works
jnthn Oddness. That'd mean it was installing the stub in one place, but then failing to find it when it goes looking for an existing declaration. 09:12
moritz: I suspect the splice one is one of the romaing GC segfaults. I'm pretty sure it didn't happen locally for me. :( 09:16
moritz jnthn: if it helps, I could give you ssh access to a linux box where it's reproducable 09:17
jnthn moritz: May help. I will see if running under the debugger catches it, though will have to be this evening once I'm home. 09:20
moritz jnthn: I'm building as a separate user now, to see if it can reproduce the segfault 09:21
moritz jnthn: hm, can't easily reproduce :/ 09:34
moritz updated perl6advent.wordpress.com/2011/12/0...ndar-2011/ with links to the posts 09:40
jnthn moritz: yeah, that's always the issue with GC things :( 09:44
perl6advent.wordpress.com/2011/12/0...-operator/ seems to have linebreaks in odd places 09:45
moritz nom: say so 'ab' ~~ /(.)+/; say $/[0].WHAT 09:46
p6eval nom 48b979: OUTPUT«Bool::True␤Parcel()␤»
moritz nom: say so 'ab' ~~ /(.)+/; say $/[0].elems
p6eval nom 48b979: OUTPUT«Bool::True␤2␤»
moritz nom: say so 'ab' ~~ /(.)+/; say $/[0][0].Str
p6eval nom 48b979: OUTPUT«Bool::True␤a␤»
moritz nom: say so 'ab' ~~ /(.)+/; say $/.list.elems 09:47
p6eval nom 48b979: OUTPUT«Bool::True␤1␤»
moritz nom: say so 'ab' ~~ /(.)+/; say $/.list[0].elems
p6eval nom 48b979: OUTPUT«Bool::True␤2␤»
moritz fixes the line breaks
even the HTML view of the posts isn't the actual HTML that appears on the page 09:49
nom: my @a = @( 'a' ~~ /(.)+/ ); say @a.elems; say @a[0].elems 09:50
p6eval nom 48b979: OUTPUT«1␤0␤»
moritz nom: my @a = @( 'a' ~~ /(.)+/ ); say @a[0].WHAT 09:50
p6eval nom 48b979: OUTPUT«Match()␤»
moritz hm, shouldn't that be a Parcel?
jnthn nom: my @a = @( 'a' ~~ /(.)+/ ); say @a.perl 09:53
p6eval nom 48b979: OUTPUT«Array.new(Match.new(orig => "a", from => 0, to => 1, ast => Any, list => ().list, hash => EnumMap.new()))␤»
jnthn nom: say @( 'a' ~~ /(.)+/ ).perl;
p6eval nom 48b979: OUTPUT«((Match.new(orig => "a", from => 0, to => 1, ast => Any, list => ().list, hash => EnumMap.new()),),).list␤»
jnthn moritz: I guess the parcel flattens?
moritz hm, right 09:55
dalek kudo/nom: e9733c5 | moritz++ | src/core/Any.pm:
implement item()
 10:26
ast: 8d0856d | moritz++ | S32-array/splice.t:
unfudge an item() test
jnthn wonders what shape the star repo is in with respect to nom 10:28
dalek ast: 445d836 | moritz++ | S0 (2 files):
rakudo unfudges
 12:33
ast: 97c4f95 | moritz++ | S03-smartmatch/any-hash (2 files):
rename a smartmatch file, and clean in up
kudo/nom: d0c1523 | moritz++ | t/spectest.data:
track test file rename
 12:35
dalek kudo/nom: 568aed7 | moritz++ | lib/Test.pm:
[Test] fix typo and error diagnostics
 13:22
felher happily reads through the perl6 advent calendar :) 13:33
dalek kudo/nom: cd3a5d7 | coke++ | t/spectest.data:
track failure modes
 13:37
muixirt perl6: for 1..20 { .say if $_ == 9 ff $_ == 13; } 13:40
p6eval niecza v12-3-g193ca57: OUTPUT«9␤10␤11␤12␤13␤»
..pugs b927740: OUTPUT«*** No such subroutine: "&infix:ff"␤ at /tmp/WJIEomF78Z line 1, column 21-42␤»
..rakudo d0c152: OUTPUT«===SORRY!===␤Flip flop operators are not yet implemented at line 1, near " $_ == 13"␤»
dalek ast: ae9d8ff | moritz++ | S (2 files):
fudge and correct tests
 13:44
moritz nom: class C1 { }; class C2 is C1 { }; sub f(C1 @x) { }; my C2 @y; f @y
p6eval nom d0c152: OUTPUT«===SORRY!===␤CHECK FAILED:␤Calling 'f' will never work with argument types (C2) (line 1)␤ Expected: :(Positional @x)␤»
moritz jnthn: bug, or NYI? 13:45
jnthn moritz: Definite bug, I know pretty much exactly what it is.
moritz jnthn: tested in S14-roles/parameter-subtyping.t 13:46
jnthn moritz: There's a far simpler case of it too
jnthn nom: my Int @x; my Int @y; @x := @y; 13:46
p6eval nom d0c152: OUTPUT«Type check failed in binding␤ in block <anon> at /tmp/3zslD9JDm7:1␤ in <anon> at /tmp/3zslD9JDm7:1␤»
jnthn nom: my @x; @x := 42; # also related
p6eval nom d0c152: ( no output )
dalek kudo/nom: 15ba074 | moritz++ | t/spectest.data:
run another test file
 13:48
colomon moritz++ 14:05
aloha Karma for moritz is now 2842
jnthn wonders why aloha found that number more noteworthy than, say, any other one, and decided to tell us about it :) 14:10
moritz jnthn++ 14:17
aloha Karma for jnthn is now 2464
moritz oh NOE!
[Coke] aloha, source?
aloha, master?
it's github.com/bacek/aloha, btw. looking now. 14:18
[Coke] the source hasn't been updated since February 16, 2011. 14:20
moritz maybe it uses some modules that have been updated
[Coke] search.cpan.org/~diz/Bot-BasicBot-P...e/Karma.pm 14:22
Responds with the new karma for <thing> unless karma_change_response is set to a false value. 14:23
looks fixable, anyway. ;)
moritz [Coke]++
[Coke] explain [Coke
aloha positive: nothing; negative: nothing; overall: 0.
[Coke] explain [Coke]
aloha positive: nothing; negative: nothing; overall: 0.
[Coke] explain moritz
aloha positive: | S02-types/hash.t: (dalek), | src/metamodel/how/NQPMu.pm: (dalek), | LHF.markdown: (dalek); negative: git commit -a without checking git diff first (moritz_), we have your karma. fix bug to get it back !:) (masak); overall: 2842.
[Coke] seen diz? 14:24
aloha diz was last seen in 15313 days 14 hours ago .
[Coke] nom: say 15313 / 365.24
p6eval nom 15ba07: OUTPUT«41.925856970759␤»
jnthn ...is IRC evem 42 years old? :) 14:25
[Coke] seen thingthatdoesntexit?
aloha thingthatdoesntexit was last seen in 15313 days 14 hours ago .
[Coke] Looks like a bug, going back to epoch.
(which is slightly older than I am, so I spot it often. ;)
[Coke] That's probably search.cpan.org/~diz/Bot-BasicBot-P...le/Seen.pm 14:26
[Coke] iwbni perl's advent calendar shared a feed every year. 14:42
(perl 5, that is)
[Coke] aloha: ask rjbs if we can get a perl5 advent calendar RSS feed that will show all years, not just the current year. 15:12
phenny: ask rjbs if we can get a perl5 advent calendar RSS feed that will show all years, not just the current year.
phenny [Coke]: I'll pass that on when rjbs is around.
dalek kudo/nom: d9a22f4 | moritz++ | src/core/List.pm:
coerce the RHS to Int in infix:<xx>

fixes things like $val xx @array		 15:18
ast: d9bb314 | moritz++ | S03-operators/repeat.t:
rakudo unfudge
moritz enjoyed watching vimeo.com/9270320 15:21
moritz the information density isn't all that high, but it's... enjoyable nonethelss 15:24
dalek kudo/nom: b8ee2ed | coke++ | t/spectest.data:
track failure modes
 15:31
moritz nom: sub f(:$a!, $b) { } 15:43
p6eval nom d9a22f: OUTPUT«===SORRY!===␤Cannot put required parameter after variadic parameters at line 1, near ") { }"␤»
moritz :$a! isn't optional, so that error message makes little sense 15:44
std: sub f(:$a!, $b) { }
p6eval std be1f10e: OUTPUT«===SORRY!===␤Cannot put required parameter after variadic parameters at /tmp/xI4t8RynP1 line 1:␤------> sub f(:$a!, $b⏏) { }␤Potential difficulties:␤ $b is declared but not used at /tmp/xI4t8RynP1 line 1:␤------> sub f(:$a!,…
jnthn hm, fail 15:45
Should probably say "after named parameters" or so...
At least Rakudo is consistent with STD here ;)
tadzik 6 failures in t/spec/S05-mass/properties-general.rakudo, known? 15:46
dalek ast: 3f59192 | moritz++ | S12-methods/what.t:
correct tests that called positional arguments by name
[Coke] tadzik: I'll kick off a spectest here. 15:53
jnthn I have vague memories of hearing about differeing pass/results on those tests. 15:54
Maybe different ICU version or something...
tadzik moritz: commented github.com/perl6/roast/commit/3f591925fe 15:55
moritz (properties-general) fails here too
I suspect different ICU version too
tadzik: there's no mention of Multi or MultiSub in the specs 15:56
jnthn It doesn't make any sense since the multi spec changed 15:57
tadzik so what does this RT expect?
oh, it's outdated I guess
jnthn decommute ^ 15:58
&
moritz tadzik: it is
tadzik okay
moritz the ticket just wanted it to not Null PMC Access 16:00
tadzik okay 16:01
dalek ast: 3d1207b | moritz++ | S02-names-vars/names.t:
unfudge passing tests
 16:02
dalek kudo/nom: bef928a | moritz++ | src/Perl6/Grammar.pm:
fix parsing of "enum" as identifier
 16:02
lichtkind i think i have a advent topic 16:06
colomon \o/
lichtkind the new hash methods
or is that taken?
lichtkind i checked old calendars 16:07
by taken i mean in your plan
tadzik new hash methods?
lichtkind y
moritz what new hash methods? 16:08
lichtkind allright last year they were new
moritz what hash methods?
classify? 16:09
... and categorize
lichtkind yes
moritz lichtkind: then please add your name and topic to the schedule file in the mu repo 16:10
lichtkind moritz: i better write first the text
moritz nom: $_ = 3; for 1 -> $a { $_ = 5 }; .say 16:11
p6eval nom b8ee2e: OUTPUT«5␤»
moritz nom: say &[<=>].WHAT 16:15
p6eval nom b8ee2e: OUTPUT«Sub()␤»
moritz nom: say &[<=>].('1', '4'( 16:16
p6eval nom b8ee2e: OUTPUT«===SORRY!===␤Unable to parse postcircumfix:sym<( )>, couldn't find final ')' at line 1␤»
moritz nom: say &[<=>].('1', '4')
p6eval nom b8ee2e: OUTPUT«-1␤»
TimToady nom: say '1'.&[<=>]('4') 16:17
p6eval nom b8ee2e: OUTPUT«-1␤»
moritz nom: say sort &[<=>], <1 4 3 6 2> 16:18
p6eval nom b8ee2e: OUTPUT«No applicable candidates found to dispatch to for 'infix:<<=>>'. Available candidates are:␤:(int $a, int $b)␤:(num $a, num $b)␤:(Int $a, Int $b)␤:(Num $a, Num $b)␤:(Instant $a, Instant $b)␤:(Date $a, Date $b)␤:(Real $a, Real $b)␤:(Any $a, Any $b)␤␤ in sub infix:<<=>>…
moritz doesn't understand why
nom: say &[<=>].arity 16:19
p6eval nom b8ee2e: OUTPUT«1␤»
moritz nom: say &[<=>].count
p6eval nom b8ee2e: OUTPUT«1␤»
moritz ah 16:20
so it tries to invoke it with one argument only
jnthn: see above
moritz goes through the nom regressions in roast
TimToady though &[] is supposed to force 2-arg
moritz I know 16:21
TimToady maybe because chaining?
er, wait, no
moritz and that implies that the tests that do ok &infix:<==> === &[==] are wrong 16:22
TimToady nom: say &[cmp].arity
p6eval nom b8ee2e: OUTPUT«2␤»
moritz TimToady: no, default values for reduction
(I guess)
because if &[==] forces arity 2, and &infix:<==> does not, they can't be the same 16:23
TimToady nom: say &[leg].arity
p6eval nom b8ee2e: OUTPUT«1␤»
TimToady leg is also 1
but not cmp?!?
[Coke] gist.github.com/1434231 # for tadzik - i miss being able to just paste this into a bot on the web. 16:42
tadzik [Coke]: I didn't get that failures, and I have the same TODO passing 16:43
moritz I can declare a my constant i = 1i; if I put it into SAFE.setting
[Coke] tadzik: just unfudged that todo passing.
tadzik ok cool
moritz but it loops if I put it into src/core/ somwhere
[Coke] tadzik: you have a 64 bit box? 16:44
moritz (compilation loops, that is)
tadzik [Coke]: yes
dalek ast: 29ee946 | coke++ | S03-operators/range.t:
unfudge for rakudo/nom
jnthn moritz: compilation loops probably means exception handling fail :/ 16:45
[Coke] tadzik: that'll do it.
jnthn Or at least, that's what I've tended to encounter.
jnthn &[+] is kinda cheating at the moment - I didn't make it filter to arity 2. Not sure what they best way to do that is...it'd kinda suck to re-compute the arity-2 list every time, but I'm not sure where to stash it either. 16:45
PerlJam jnthn: got a second? Could you explain to me the process by which method lookup happens? I think the first time the method is called, we get find_method, and the result of that is put in the cache and then executed. Future method calls read from the cache. Is that correct? 16:47
jnthn PerlJam: no 16:48
PerlJam: At compose time, we pre-emptively populate the method cache.
PerlJam ah.
jnthn PerlJam: NQP doesn't have to support any kind of dispatch failover, so it goes a step further and marks the cache as authoritative, meaning that normal dispatches never have to call .^find_method
PerlJam Hmm 16:50
jnthn Well, I should say NQP's meta-objects
We have an interest in being able to say "no such method" quite fast in NQP. 16:51
PerlJam jnthn: I was looking at Grammar::Debugger ... correct me if I'm wrong, but doesn't it create a new routine on each method call?
jnthn: I was thinking the way to optimize that was to call .wrap() on the method once and put that in the cache, but I guess I'd have to introspect the method names and call .wrap on them at compose time to populate the cache appropriately. 16:52
jnthn PerlJam: If you wrap it then you'll damage to original. 16:53
*the
PerlJam: The answer is to cache the computed closures. 16:54
PerlJam: It'd make things faster, though I'm not sure how much.
PerlJam okay, thanks for the input (I just think about these things in the interstices when I'm procrastinating at work :) 16:56
tadzik bug of the day: Foo *a = malloc(...); memset(&a, ...); 17:04
jnthn oops :) 17:05
tadzik (wtf, why did this malloc return 0)
mux yes, why? that's actually an interesting question
tadzik memset expects a pointer to the memory location
mux oh scratch that, I didn't see the erroneous &
tadzik me neither
mux I thought you were talking about something else entirely :D
tadzik me too 17:06
:)
flussence I vaguely remember reading somewhere that 0 is a valid return value for malloc...
mux flussence: it is
the thing is it rarely happens in practice
tadzik it is, but it would be... extremely suprising :) 17:06
mux (by 0 I meant NULL here)
flussence or was it that it's a valid input value...
tadzik I started to think "so malloc got broken and the system still runs? wow"
ashleydev on the iphone that wouldn't be as rare 17:25
but if you wait a while it'll clear out some memory space and it'll probably work the 2nd time 17:26
sorear good * #perl6 17:36
ashleydev * 17:38
colomon o/ 17:42
arnsholt flussence: malloc will return NULL if it failed to allocate the requested amount of memory 17:43
colomon woah. one of the political blogs out there has just done a post inspired by the Flip-Flop operator post. 17:48
moritz colomon: URL? 17:50
colomon theothermccain.com
there's a trackback, that's how I learned about it.
theothermccain.com/2011/12/05/softw...flip-flop/
arnsholt colomon: That was one of the less coherent posts I've read in a while =) 17:54
colomon I wouldn't be surprised if it doubles our page views just by accident, though. :)
arnsholt Indeed =) 17:55
Speaking of page views, has the advent calendar been submitted to HN/Slashdot?
moritz arnsholt: I'm not aware of a /. submission
arnsholt: but it would be great to have one 17:56
[Coke] colomon++ 17:58
aloha Karma for colomon is now 183
colomon [Coke]: er, what have I done now? ;) 17:59
arnsholt moritz: This is the third year of Perl 6 advent, no? 18:00
colomon arnsholt: yes
[Coke] moritz: and still no real production release? ;) 18:05
colomon: whoops! I thought you wrote the blog post. anyway, thanks for noting the linkback. 18:08
arnsholt There we go. Submitted to slashdot as well: slashdot.org/submission/1870962/per...t-calendar
[Coke] PerlJam++ # No, I still don't think I'll ever use the flipflop operator.
aloha Karma for perljam is now 55
colomon I've never used the flipflop, but after reading this it does feel to me like there is some cool application for it, hiding off in the shadows somewhere... 18:13
sorear flipflop's reason for existing is, like so many other Perl feaures, s2p 18:14
sorear '/foo/,/bar/d' becomes 'print unless /foo/ ff /bar/' 18:15
PerlJam I used to used flip-flop all the time in Perl 5 (not so much lately) but I still had a hard time explaining it in words.
moritz PerlJam: the weird thing is that the flip-flop operator carries state 18:39
PerlJam: unlike any/most other operators
colomon moritz: yeah, I was wondering how it was implemented.... 18:41
[Coke] my $foo = 42; ($foo, "stuff")[0] = 24; say $foo; 19:09
perl6: my $foo = 42; ($foo, "stuff")[0] = 24; say $foo; 19:10
p6eval pugs b927740: OUTPUT«*** Can't modify constant item: VRef <Scalar:0x7f41cc366fa1>␤ at /tmp/g6RkzZzOIu line 1, column 15-38␤»
..rakudo bef928: OUTPUT«Cannot assign to a readonly variable or a value␤ in block <anon> at /tmp/eROaY_J02E:1␤ in <anon> at /tmp/eROaY_J02E:1␤»
..niecza v12-3-g193ca57: OUTPUT«24␤»
[Coke] std: my $foo = 42; ($foo, "stuff")[0] = 24; say $foo;
p6eval std be1f10e: OUTPUT«ok 00:01 124m␤»
[Coke] b: my $foo = 42; ($foo, "stuff")[0] = 24; say $foo;
p6eval b 1b7dd1: OUTPUT«24␤»
[Coke] are b and niecza right?
moritz I think so 19:11
dalek ast: 88b5007 | coke++ | S02-types/lists.t:
Convert regression to RT
 19:18
raiph greetings, flying camels 19:55
tadzik hello raiph
colomon nom: say 'This "is" "really" a test'.comb(/ <-["]>/) 20:03
p6eval nom bef928: OUTPUT«T h i s i s r e a l l y a t e s t␤» 20:04
colomon oooo, right in one! :)
felher If i have a grammar, consisting only of regexes (which i think do backtrack) and i have a regex of structure 'regex A { \h* [<B>] \h* }' and change it to 'regex A { \h* [<A>|<B>] \h* }', how can the first parse something that the second can not? I'm only adding alternatives, am i not? 21:08
[Coke] could have negative assertions in <A>, no? 21:09
arnsholt Or could be some kind of LTMy bug
jnthn We don't do LTM on alternations yet in nom, fwiw. 21:10
arnsholt Then probably not an LTMy bug =) 21:10
felher What workes is paste.pocoo.org/show/516795/
what doesn't is: paste.pocoo.org/show/516793/
diakopter are you using nom? 21:11
felher diakopter: yes 21:11
diakopter probably it doesn't work because backtracking into subrules doesn't work 21:12
felher diakopter: Yeah, that would explain it. Thnx :) 21:14
mberends \o/ 21:59
mberends iz riting teh next p6advent blogpost :0 22:00
colomon mberends++ 22:01
colomon is cooking noms and swearing at LilyPond
jnthn is feeling way too tired after somehow only managing 4 hours of sleep last night 22:02
colomon sometimes thinks LilyPond was designed by the STEP specific committee... 22:08
jnthn manages to make 5000 tests fail 22:23
Fixed a longstanding RT though ;) 22:24
benabik 5000 failures to fix one bug? Nice ratio. ;-) 22:28
mux that's one badass fix
jnthn Yeah, turns out that losing the type constraints on every parameter works wonders for failing tests. 22:30
japhb_ jnthn, I think you should call that a win. ;-) 22:35
jnthn :P 22:40
Got it fixed now
jnthn Busted split...well, because it does something bad. 22:40
japhb_
.oO( Bad method! No biscuit! )
 22:42
jnthn Relied on @foo := ... not actually enforcing the sigil. 22:43
Which I've got fixed locally.
benabik ouch.
jnthn Not sure Rakudo's ever got this right before
japhb_ jnthn, out of curiosity, what are you working on? 22:49
dalek ast: c486ab7 | jnthn++ | S14-roles/parameter-subtyping.t:
Correct a test, unfudge another.
 22:50
ast: 4f77025 | jnthn++ | S14-roles/parameterized-type.t:
Turns out we still passed this test for the wrong reasons; fudge it. :(
kudo/nom: 375bbe8 | jnthn++ | src/ (3 files):
Refactor to static type handling, so it relies on the bind constraint rather than the container descriptor. Fixes my @foo := 42; not failing, plus avoids giving the optimizer a bunch of crappy type info so it gives false positives when analyzing parameterized role arguments.
 22:51
kudo/nom: 32f00e7 | jnthn++ | src/core/Str.pm:
We now enforce the sigil when binding; fix split, which worked based on that check being missing.
kudo/nom: 40b3369 | jnthn++ | src/core/Array.pm:
First bit of enforcing typed array types.
 22:52
kudo/nom: 104beb8 | jnthn++ | t/spectest.data:
We can now run and pass all of parameter-subtyping.t.
jnthn japhb_: That lot ;)
japhb_: Another yak shave for doing arrays in NCI decently.
japhb_ Ah, gotcha!
jnthn marks a 3 year old RT as testneeded 22:55
eek, German spam in the RT queue 22:56
"pray! Entspannt, Erfrischt, Macht fit!"
phenny: "pray! Entspannt, Erfrischt, Macht fit!"?
phenny jnthn: The de to en translation failed, sorry!
jnthn aww
jnthn wanted to learn 3 new words :)
benabik translate.google.com: Pray! Relaxed, refreshed, might fit! 22:57
jnthn ah 22:58
I'm pretty sure "might fit" is a bad translation.
jnthn It's "might" as in "power", iiuc. 22:58
nom: my &foo = 42 23:00
p6eval nom bef928: OUTPUT«Type check failed in assignment to '&foo'; expected 'Callable' but got 'Int'␤ in block <anon> at /tmp/DXxkH5_axL:1␤ in <anon> at /tmp/DXxkH5_axL:1␤»
jnthn \o/ 23:01
dalek ast: 3f1e455 | jnthn++ | S16-filehandles/filetest.t:
Fudging; verification on a platform other than Windows would be helpful.
 23:24
lichtkind jnthn: correct translation is: makes agile 23:25
dalek kudo/nom: e5cf070 | jnthn++ | src/core/IO.pm:
Re-instate getc.
kudo/nom: f346f0a | jnthn++ | t/spectest.data:
Run S16-io/getc.t.
kudo/nom: 7bc7c6e | jnthn++ | src/core/IO.pm:
Add back .z file test; filetest.t now seems close to working.
kudo/nom: 74c7491 | jnthn++ | t/spectest.data:
Run filetest.t.
jnthn Clearly the spam was trying to sll us scrum training. :)
*sell
lichtkind something like that :) 23:27
diakopter I will ScrumMaster your Requestor Sprint Backlog with Artifacts to Story 23:29
lichtkind diakopter: now i have a full row :) 23:30
dalek ast: 62c60f6 | jnthn++ | S12-introspection/walk.t:
Bring walk.t in line with the spec.
 23:47
mberends perl6advent.wordpress.com/2011/12/0...on-niecza/ came out slightly early, please comment :-) 23:48
jnthn mberends++ # looks good 23:50
mberends thanks jnthn 23:51
diakopter +1
diakopter learns about the flip-flop operator for the first time 23:53
dalek kudo/nom: 6a15ccd | jnthn++ | src/core/Mu.pm:
Implement Mu.WALK.
 23:54
kudo/nom: ded9695 | jnthn++ | t/spectest.data:
Turn on S12-introspection/walk.t.

Powered by Cro and the Raku® Programming Language.

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

Thank you!