»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'perl6: say 3;' or rakudo:, niecza:, std:, or /msg camelia perl6: ... | irclog: irc.perl6.org | UTF-8 is our friend!
Set by sorear on 25 June 2013.
colomon is the rakudo build failing right now?> 00:09
(on parrot)
colomon rakudobrew builds moar for me just fine, but parrot looks like it goes into an infinite recursion or something… errors stretching all the way through my scrollback buffer 00:23
benabik heh 00:26
Yes.
An optimization broke the Parrot build. I posted the breaking commit back in the clogs.
1775172: “Various return typecheck/decont optimizations.” 00:27
colomon benabik++
colomon huh. was there no response at all to your report? 00:32
benabik No. Odd hour of the night, and I’m pretty busy at the moment so haven’t followed up. 00:33
colomon moar and JVM both build fine. 00:34
if I'd been paying attention to me smoke test results, I'd have noticed the bug days ago … it hasn't built a new Rakudo successfully since that patch. 00:36
skids r: role A [ :$a = 1, :$b = $a ] { say $a ~ $b }; role B does A[ :a(2) ] { }; class C does B[ ] { }; class D does A[ :a(3) ] { }; 01:41
camelia rakudo-moar fd5f8e: OUTPUT«22␤32␤»
camelia ..rakudo-parrot fd5f8e: OUTPUT«22␤33␤» 01:41
..rakudo-jvm fd5f8e: OUTPUT«Unhandled exception: Failed to load config.properties␤ in (gen/jvm/stage2/NQPHLL.nqp:8)␤ in (gen/jvm/stage2/NQPHLL.nqp:8)␤ in (gen/jvm/stage2/NQPHLL.nqp)␤ in (src/vm/jvm/ModuleLoader.nqp:78)␤ in load_module (src/vm/jvm/ModuleLoader.nqp:60…»
grondilu wrote rosettacode.org/wiki/Window_creation/X11#Perl_6 02:12
colomon wow, installing DateTime::TimeZone on JVM incredibly slow? 02:56
dalek osystem: 22ad485 | duff++ | META.list:
Add Git::Wrapper
 06:06
sergot morning o/ 06:09
masak morning 06:19
thou m: "ab" ~~ m/(.)(.)/; say $/.keys.perl; 06:21
camelia rakudo-moar fd5f8e: OUTPUT«(0, 1, "").list␤»
thou Patch to fix that, is it good? ^ pastee.org/99sdf 06:22
masak m: "ab" ~~ m/(.)(.)/; say $/<>
camelia rakudo-moar fd5f8e: OUTPUT«｢ab｣␤ 0 => ｢a｣␤ 1 => ｢b｣␤␤»
thou m: "ab" ~~ m/(.)(.)/; say $/{''}.perl 06:23
masak thou: I guess it's good if it passes the spectests :)
camelia rakudo-moar fd5f8e: OUTPUT«[]␤»
masak oh, right.
zen slices.
a more interesting question is "what is that empty string key doing there in the first place, so that we have to delete it?"
thou I may have misread things, but it seems to be that %caps{''} := $count is used to track some state during sub capnames() 06:25
... I guess at the very least P5Regex/Actions shold be modified, too .... I'll run tests, heading to bed now. :-) 06:29
masak thou++ 06:30
dalek ecs: 37196f6 | (Tim Smith)++ | S05-regex.pod:
Fix example: $/.keys list keys start from 0, not 1
 06:48
sergot m: say [] ~~ Array; 07:04
camelia rakudo-moar fd5f8e: OUTPUT«True␤»
sergot m: say <1 2> ~~ Array;
camelia rakudo-moar fd5f8e: OUTPUT«False␤»
sergot m: say <> ~~ Array;
camelia rakudo-moar fd5f8e: OUTPUT«===SORRY!=== Error while compiling /tmp/IaGJXVhR_H␤Unsupported use of <>; in Perl 6 please use lines() to read input, ('') to represent a null string or () to represent an empty list␤at /tmp/IaGJXVhR_H:1␤------> say <⏏[…»
sergot m: say <1 2>.WHAT 07:05
camelia rakudo-moar fd5f8e: OUTPUT«(Parcel)␤»
sergot m: say [] ~~ Parcel;
camelia rakudo-moar fd5f8e: OUTPUT«False␤»
sergot m: say [].WHAT
camelia rakudo-moar fd5f8e: OUTPUT«(Array)␤»
sergot m: say (1,2).WHAT 07:06
camelia rakudo-moar fd5f8e: OUTPUT«(Parcel)␤»
masak ooh -- the above gives me an idea for a useful library function: 07:20
"most relevant common ancestor"
you pass in Array and Parcel, and you get out... whatever most significant type those two have in common.
m: say Array.^mro
camelia rakudo-moar fd5f8e: OUTPUT«(Array) (List) (Iterable) (Cool) (Any) (Mu)␤» 07:21
masak m: say Parcel.^mro
camelia rakudo-moar fd5f8e: OUTPUT«(Parcel) (Cool) (Any) (Mu)␤»
masak hm. appears to be Cool. in this case.
why aren't Parcels Iterable?
moritz why is Iterable a class?
masak :) 07:22
is there a way to ask a class "what roles did you (and your ancestors) mix in?" 07:23
m: say Array.^roles
camelia rakudo-moar fd5f8e: OUTPUT«(Positional)␤»
masak oh, gotcha.
m: say Parcel.^roles 07:24
camelia rakudo-moar fd5f8e: OUTPUT«(Positional)␤»
masak well, It Can Be Done.
lizmat good *, #perl6! 07:25
.tell Mouq should List.rotor(1,0) not be just { self } ? 07:26
yoleaux lizmat: I'll pass your message to Mouq.
lizmat .tell Mouq and could you also transmogrify the S17-supply/rotor.t tests to handle List.rotor ?
yoleaux lizmat: I'll pass your message to Mouq.
masak mornin', lizmat 07:28
lizmat masak o/
moritz \o masak, lizmat, *
masak ahoy, moritz
lizmat moritz o/
sergot: are we still on for a pre PLPW perl 6 hackathon? 07:29
lizmat is wondering whether this boilerplate "my $cmp = $by.arity == 2 ?? $by !! { $by($^a) cmp $by($^b) }" shouldn't be a function 07:43
something like: 07:44
sub BY2 (&by = &infix:<cmp>) { $by.arity == 2 ?? $by !! { $by($^a) cmp $by($^b) } 07:45
and sigs like min could then be:
lizmat hmmm... not in the sig :-( 07:46
grrr
arnsholt o/ 07:59
FROGGS o/ 08:03
masak lizmat: sounds like a step in the right direction. 08:06
lizmat: if I may bikeshed the name, I'd have it as just 'BY', though. 08:07
lizmat :-)
masak (unless there's a BY1 or BY3 hiding in there somewhere) 08:08
lizmat hehe... will look
dalek ecs: 2ac69b9 | (Elizabeth Mattijsen)++ | S17-concurrency.pod:
Spec Supply.(min|max)
 08:15
kudo/nom: 3613faa | (Elizabeth Mattijsen)++ | / (2 files):
Implement Supply.(min|max) and test them
 08:27
timotimo .o(and it should be &by every time, or $by every time, rather than a mixture of both ;) ) 08:51
sergot lizmat: what do you mean? :) 08:53
lizmat timotimo: yes, but the idea was clear, eh? 08:59
sergot: I was under the impression we would have a Perl6 hackathon before the PLPW in Poznan
I didn't hear anything about it anymore, so I started wondering :-) 09:00
fwiw, woolfy and me will be in Poznan from Wed evening
woolfy lizmat means we will drive by car on wednesday, May 14, to Poznan. 09:01
sergot lizmat: sounds great! 09:08
I'm in ! :)
lizmat well, we do need a place, right ?
sergot That's right. I can check if there's a place in company I work in. 09:10
lizmat: how many participants do you expect? 09:11
lizmat good question: I understand jnthn and FROGGS could make it there on the Friday 09:12
sergot I have never heard about this I think, so there are some questions.
lizmat tadzik hopefully :-)
sure.. :) 09:13
sergot How would it look like? How long would it last? and so on.. :)
FROGGS I'll be there on friday afternoon/evening, yes 09:14
lizmat well, Fri evening would be social time, right?
sergot Right.
FROGGS I hope so :o) 09:15
lizmat I was thinking from about 10am forward until social time on Fri ?
there would be at least me :-)
FROGGS hmmm
sergot: if you can make up a nice place, I try to be there at lunchtime 09:16
sergot I must talk with guys here to find out if we can do this in the venue. :)
So I'll know in about 2 hrs 09:17
lizmat ok, otherwise: it seems woolfy and me have an apartment in Poznan for the duration
so if small enough, we could have it there as well
sergot Give me a sec. ;) 09:19
lizmat sure, we still have more than a week :-) 09:22
grondilu r: subset foo of int; sub f(foo $x) { say $x }; f my int $ = 43; 09:27
camelia rakudo-{parrot,moar} fd5f8e: OUTPUT«Constraint type check failed for parameter '$x'␤ in sub f at /tmp/tmpfile:1␤ in block at /tmp/tmpfile:1␤␤»
..rakudo-jvm fd5f8e: OUTPUT«Unhandled exception: Failed to load config.properties␤ in (gen/jvm/stage2/NQPHLL.nqp:8)␤ in (gen/jvm/stage2/NQPHLL.nqp:8)␤ in (gen/jvm/stage2/NQPHLL.nqp)␤ in (src/vm/jvm/ModuleLoader.nqp:78)␤ in load_module (src/vm/jvm/ModuleLoader.nqp:60…»
grondilu r: subset foo of Int; sub f(foo $x) { say $x }; f my Int $ = 43;
camelia rakudo-jvm fd5f8e: OUTPUT«Unhandled exception: Failed to load config.properties␤ in (gen/jvm/stage2/NQPHLL.nqp:8)␤ in (gen/jvm/stage2/NQPHLL.nqp:8)␤ in (gen/jvm/stage2/NQPHLL.nqp)␤ in (src/vm/jvm/ModuleLoader.nqp:78)␤ in load_module (src/vm/jvm/ModuleLoader.nqp:60…»
..rakudo-{parrot,moar} fd5f8e: OUTPUT«43␤»
lizmat just had an idea: 09:28
what if camelia would post a gist if the output is not the same for all backaned, and post a link to that gist on the channel? 09:30
*backends
FROGGS +1 09:33
at least when it hits the '…'
lizmat yup, good point
timotimo the gist could then also contain the command in a second gist, and a link back to the clogs perhaps? 09:42
FROGGS +10 09:43
:D
and then it will to decide if it was in line with the spec and create a ticket in such a case :P
timotimo hah 09:45
maybe we'd want to create our own little site for that whole thing? 09:46
it could also incorporate these little hashtags raiph suggested
FROGGS but that sounds bigger, and will not be there soonish - at least that is my expectation 09:47
baby steps, aka using gist now would help here I think
lizmat I was just thinking of ways to make #perl6 more palatable to more people :-)
timotimo right. 09:53
Woodi hallo :)
timotimo lizmat: i heard from multiple people by now that the amount of text is a bit overwhelming to many
Woodi camelia could paste into #perl6.+ maybe ?
timotimo and i admit that sometimes when i backlog, i skip over large portions of chatlog if there's mostly evaluations :P
raydiak .tell tadzik tried through the window again on a recent moar, and it runs playably on my craptop; might have tried jvm last time...figured you'd want to know its not as grim as it seemed 09:56
yoleaux raydiak: I'll pass your message to tadzik.
timotimo moarvm doesn't yet have a spesh function for the CStruct repr, so maybe at some point it could optimize accesses to CStruct repr'd classes; jnthn said it's "complicated", so i'm probably not the one to do it any time soon :P 10:00
FROGGS timotimo: but it sounds very valuable 10:02
because, what is a fast C-lib when you have a bottleneck right before it?
timotimo well, the question is what exactly is to be gained 10:03
FROGGS and often you can't abstract accesses to structures away
timotimo my brain is kinda tired still
at least the repr op get_attribute can take a hint parameter 10:04
hm. the try_get_slot does a linear search through the mapping
er, no 10:05
FROGGS we need simple but realistic use cases
(and the profiler)
timotimo it linear-scans to find the right class
yes, i suppose we do.
the p6bench suite doesn't have any benchmarks incorporating NativeCall.
FROGGS I was hoping that you would step up and work on the profiler :o)
that's why I highlighted you the other day 10:06
timotimo aye, i saw it and thought about it
but it seems kinda scary
timotimo i think i need a nap. y'all still have time to do stuff that i can include in the weekly :) 10:07
FROGGS ohh, the weekly, yes 10:08
timotimo: about labels: done for perl6-m/nqp-m, done for perl6-p/nqp-p, in the works for nqp-j (and afterwards perl6-j) 10:09
that is all I have :/
Woodi tadzik: is it possible to add bullet sprite and wait for collision event(s) somehow ? 10:13
FROGGS If I had more time I'd port Alien::Box2D and Box2D to Perl 6 :/ 10:23
Woodi m: my %d = < 1 a 2 b 3 c 4 d 22 p 31 z >; my $ds = 2; my $de = 4; say %d{ %d.keys.grep( { $ds <= $_ <= $de } ) }; 10:25
camelia rakudo-moar fd5f8e: OUTPUT«b c d␤»
Woodi is maybe some shortcut for that ?
sergot lizmat: ok, we have a room (for ~ 10 people) near the venue 10:26
lizmat sergot: excellent news!
sergot it's a room just near plpw's room :)
FROGGS cool!
Woodi sergot: Allegro is organizing that event ? 10:27
sergot Woodi: Poznan Perl Mongers is the main organizator, they are just our sponsor and a partner.
Woodi vnice anyway :)
sergot But, we can say they organize it too :) 10:28
dalek ecs: 082012d | (Elizabeth Mattijsen)++ | S17-concurrency.pod:
Spec Supply.elems
ast: b9490bf | (Elizabeth Mattijsen)++ | S17-supply/elems.t:
Add tests for Supply.elems
FROGGS Woodi: 10:30
m: my %d = < 1 a 2 b 3 c 4 d 22 p 31 z >; my $ds = 2; my $de = 4; say %d{ $ds..$de };
camelia rakudo-moar fd5f8e: OUTPUT«b c d␤»
Woodi hmm, so I builded some additional layer ? 10:31
FROGGS yeah, kind of
Woodi thou: FROGGS++ :) 10:32
FROGGS :o)
Woodi grr # layers
dalek kudo/nom: aa668e8 | (Elizabeth Mattijsen)++ | docs/ChangeLog:
Mention Supply.(min|max|elems)
 10:34
kudo/nom: 87a4e64 | (Elizabeth Mattijsen)++ | / (2 files):
Implement Supply.elems, make sure we test it
Woodi hmm, it adds Any for non-existing keys 10:36
lizmat Woodi: try adding :v 10:38
that should filter out the non-existing keys
Woodi works ! thanx :) 10:39
lizmat S02:2468 10:40
synopsebot Link: perlcabal.org/syn/S02.html#line_2468
Woodi difference with orginal wordy try is one space so say somehow sees skipped element ? 10:42
lizmat Woodi: I'm not sure I follow you? 10:47
want to know the keys that are *not* in there ? 10:48
woolfy sergot++ for organising workshop & for hackathon-room \o/ 10:50
Woodi lizmat: example: pastebin.com/Vmb16sm1 10:51
a bit strange 'say' sees something in place of filtered elements 10:52
bat let's see if it's a problem or not
lizmat and with :v ? 10:53
I'm not sure I understand what the problem is
lizmat admires the design on this cover: akamaicovers.oreilly.com/images/978...37/lrg.jpg 10:55
Woodi line with space in front is with :v. I imagine %h{ some list/range } :v returns list of keys so I wonder why say sees invisible elements
s/keys/values/ 10:56
lizmat aha, so you're saying :v is returning an extra element ?
Woodi tested only with 'say' so far :)
checking
lizmat could you check .elems on the result and tell whether that is ok?
Woodi m: my %d = < 1 a 2 b 4 d 22 p 31 z >; my $ds = 2; my $de = 4; say %d{ $ds..$de } :v; 11:01
camelia rakudo-moar fd5f8e: OUTPUT«b d␤»
Woodi probably filtered elements are returned as undefined 11:02
masak is it still a fair description to say that type inference is "planned for Perl 6"? en.wikipedia.org/wiki/Type_inference 11:04
to what extent does Rakudo already do type inference? 11:05
moritz afaict it does type propagation, and inspecting return types of subroutines that it can resolve at compile time 11:07
masak oki 11:07
moritz but in Perl 6, type inference is strictly an optimization
not a language feature
masak right.
moritz because you can always just leave out the type completely 11:08
masak so maybe the sentence should simply be removed from that article.
moritz +1 11:10
Woodi my %d = < 1 a 2 b 4 d 22 p 31 z >; my $ds = 2; my $de = 4; for %d{ $ds..$de } :v -> $a { say $a }; 11:11
m: my %d = < 1 a 2 b 4 d 22 p 31 z >; my $ds = 2; my $de = 4; for %d{ $ds..$de } :v -> $a { say $a };
camelia rakudo-moar fd5f8e: OUTPUT«b␤d␤»
lizmat m: my %h = (a => 1); say (%h{<a b c>}:v).elems 11:14
camelia rakudo-moar fd5f8e: OUTPUT«1␤»
lizmat Woodi: not sure what you're seeong
m: my %h = (a => 1); for %h{<a b c>}:v { say 1 } 11:15
camelia rakudo-moar fd5f8e: OUTPUT«1␤»
Woodi lizmat: I expected "b d", not "b d" in example above. 11:16
lizmat "b d" being a newline between the b and the d ? 11:17
but you're doing 2 say's
timotimo does the :v apply if you have a space in front, too?
lizmat yes, afaik 11:18
timotimo m: my %d = < 1 a 2 b 4 d 22 p 31 z >; my $ds = 2; my $de = 4; say (%d{ $ds..$de } :v).perl;
camelia rakudo-moar fd5f8e: OUTPUT«("b", (), "d")␤»
timotimo ah. empty list
lizmat mmmm
intriguing
Woodi m: my %d = < 1 a 2 b 4 d 22 p 31 z >; my $ds = 2; my $de = 4; say %d{ $ds..$de } :v; my @a = %d{ $ds..$de } :v; say @a; 11:19
camelia rakudo-moar fd5f8e: OUTPUT«b d␤b d␤»
lizmat not sure if that's a bug 11:20
TimToady ^^^
sergot m: my %h = <1 a 2 b>; say (%h{5}).perl;
camelia rakudo-moar fd5f8e: OUTPUT«Any␤»
sergot m: my %h = <1 a 2 b>; say (%h{1..5}).perl;
camelia rakudo-moar fd5f8e: OUTPUT«("a", "b", Any, Any, Any)␤»
sergot m: my %h = <1 a 2 b>; say (%h{1..5} :v).perl; 11:21
camelia rakudo-moar fd5f8e: OUTPUT«("a", "b", (), (), ())␤»
moritz I guess it should return an explicit Nil in there
and the Nil shouldn't flatten out to ()
lizmat hmmm
moritz known not-up-to-date-spec-ness
lizmat I seem to recall that () is the only way, as Nil currently doesn't flatten to nothing?
moritz m: say (Nil, Nil).flat.perl 11:22
camelia rakudo-moar fd5f8e: OUTPUT«().list␤»
moritz it flattens to nothing.
lizmat hmmm... ok
lizmat goes mucking about in src/core/Any
$ perl6 -e 'my %h = (a => 1); say (%h{<a b c>} :v).perl' 11:28
(1,)
that looks better
Woodi on moar: (1, (), ()) 11:29
lizmat yes, s/()/Nil/ causes some breakage in spectest 11:33
investigating
moritz lizmat: maybe wait for clarification from TimToady++
lizmat S02:491 11:35
synopsebot Link: perlcabal.org/syn/S02.html#line_491
lizmat so s/()/Nil is not the answer 11:36
actually, thinking about it, one could argue that "say" is wrong 11:39
or basic interpolation
Woodi "Nil value represents the absence of a value where there should be one, so it does not disappear in list context,* but relies on something downstream to catch it* or blow up" - 'for' and array assigment catch it, 'say' not
lizmat as .elems and for do the right thing
I wonder whether the problem isn't really in "nqp::p6argvmarray()" 11:41
Woodi can :v just skip over nonexistent values, pls ? ooO(hmm, but how to skip above nothing...) 11:42
lizmat :v does that 11:44
it's just that if you "say" the result, you still see an artefact of that skipping 11:45
my %h = (a => 1); say (%h{<a b c>} :v).elems # only 1 element, not 3!
m: my %h = (a => 1); say (%h{<a b c>} :v).elems # only 1 element, not 3! 11:46
camelia rakudo-moar fd5f8e: OUTPUT«1␤»
Woodi for what such Nil-the-non-exsisted could be usefull ?
lizmat wel, e.g. with first(), to indicate nothing was found
Woodi but if nothing is found then result container should be empty :) 11:47
moritz Woodi: depends on the method
Woodi: grep returns lists; it makes sense to return an empty list there
Woodi: first returns an item; it doesn't make as much sense there
Woodi can we introduce some out-of-band communication for 'return' ? something like ANSI control sequences... ;) 11:51
would be great in C :) 11:52
sergot m: say (Any, (), Any, ()).elems; 11:56
camelia rakudo-moar fd5f8e: OUTPUT«2␤»
timotimo will the rakudo star release for this month make it out the door in the next ~hour? 12:08
lizmat afaik jnthn is teaching 12:10
so ETA would be tonight at the earliest, I would think
Woodi hmm, extending Bool class could be a pain - it have 0s and 1s hardcoded all over it ;) 12:11
timotimo OK
moritz you're not supposed to mokey-patch Bool :-) 12:15
FROGGS IIRC someone just needs to upload the tarball and spread the world 12:16
dalek ecs: e5079f8 | (Elizabeth Mattijsen)++ | S17-concurrency.pod:
Spec Supply.minmax
 12:23
dalek kudo-star-daily: b83ae29 | coke++ | log/ (4 files):
today (automated commit)
 12:34
rl6-roast-data: 97c8c48 | coke++ | / (6 files):
today (automated commit)
sergot $ perl6 -e 'say ("a" x 1073741824).perl' 12:37
Segmentation fault (core dumped)
is it known?
FROGGS not sure
lizmat do you have 10G of RAM ? 12:40
ah, 1G
but still
masak failure mode could be better, even if Rakudo ends up not being able to do it. 12:41
sergot: feel free to submit a rakudobug.
sergot Mem: 3858
[Coke] (parrot broken) yes, still broken
sergot $ perl6 -e 'say ("a" x 1073741825).perl'
repeat count > 1073741824 arbitrarily unsupported...
This is perl6 version 2014.04-169-gbeba99f built on MoarVM version 2014.04-49-gaa7bbef
masak heh. 12:42
timotimo sergot: we're currently forcing every string to use 32bit for every character
on moarvm
sergot oh
masak .oO( where's Rakudo for your billion-string repetition use case? ) :P
masak .oO( clearly not ready for release! ) 12:43
[Coke] in the daily runs, 18b7e2e was the first failing commit.
72f5b96 was the last one that worked before that. 12:44
timotimo lizmat: see #moarvm; should .roll() return an item instead of a list? 12:47
sergot m: my @a = (1); my @b = (2); say @a + @b; say @b + @a 12:53
camelia rakudo-moar fd5f8e: OUTPUT«2␤2␤»
sergot Why 2 twice? ;)
masak sergot: because addition is on numbers.
sergot: not on lists.
timotimo p6weekly.wordpress.com/?p=155&s...789c53b3a9 - i hope i'm missing a few things :)
masak sergot: (we talked about this) :)
masak sergot: so infix:<+> numifies first (list length; two 1s), and then adds the numbers. 12:54
sergot masak: I know, but I'm just curious what does this "2" mean.
Ok, it's clear now
FROGGS m: my @a = (1); my @b = (2); say @a (+) @b; say @b (+) @a
masak that the combined length of those lists is 2.
camelia rakudo-moar fd5f8e: OUTPUT«bag(1, 2)␤bag(2, 1)␤»
masak no matter which order you add them.
sergot masak: right, I should have known that. 12:55
m: say (1).WHAT; 12:56
camelia rakudo-moar fd5f8e: OUTPUT«(Int)␤»
masak it's because you put them into arrays that they act like that. 12:57
m: my @a = 42; say +@a
camelia rakudo-moar fd5f8e: OUTPUT«1␤» 12:58
sergot thanks masak++, it's clear now :) 13:00
masak tadzik++ # rakudobrew, which I somehow wasn't really aware of until now 13:03
timotimo did my sharedraft link get noticed? :P 13:09
masak notices 13:10
lizmat timotimo: s/especially// ?
masak timotimo: inconsistent 'jnthn'/'Jnthn' at the start of list items. 13:11
timotimo i suppose that's fair :)
fixed both
i would like additions to the list, mostly :P
masak timotimo: the blog post you're referring to is from Feb 2013... 13:12
timotimo oh, the holy grail or utopia one?
lizmat masak: but we only found out about it this week, no?
masak I think I've seen it flash by before, actually.
timotimo wow, i didn't even know
masak timotimo: I'm not 100% we should mention Parrot failing in a weekly report. maybe after it's been fixed?
timotimo mhh 13:13
lizmat masak: which might mean we never mention Parrot again ? 13:14
timotimo psst
lizmat :-) 13:15
timotimo parrot's not dead yet
timotimo hoelzro: o/ 13:15
FROGGS I am looking at that parrot issue right now
timotimo sounds good
masak FROGGS++ 13:18
rurban which parrot issue? 13:23
masak rakudo doesn't build, apparently. 13:24
see justrakudoit.wordpress.com/2014/05/...akudobrew/
FROGGS rurban: not parrot's fault 13:25
rurban interesting fail 13:26
timotimo rurban: do you have something for me to report in the weekly? about parrot? 13:27
colomon m: my $mix = (a => 3, b =>2).Mix; say $mix.roll.perl 13:28
camelia rakudo-moar fd5f8e: OUTPUT«("a",).list␤»
colomon m: say (1..6).roll.perl
camelia rakudo-moar fd5f8e: OUTPUT«6␤»
rurban should have been caught by my rakudo buildbot also, but wasn't
colomon m: say (1..6).roll(2).perl
camelia rakudo-moar fd5f8e: OUTPUT«(1, 2).list␤»
rurban timotimo: just that GSOC started with optimizing the handling of GC write barriers
colomon is happy to declare that a bug
rurban we expect only minor performance improvements (one less nested function call in every method or vtable call) 13:29
colomon rurban: the rakudo build fail does not appear to be parrot's fault, it just is limited to the parrot version of rakudo 13:30
masak colomon: that does seem pretty inconsistent, yes. either it should be lists/parcels all the time, or items for all .roll calls. 13:31
dalek kudo/nom: f579a0a | (Elizabeth Mattijsen)++ | src/core/Mixy.pm:
Fix listiness of Mixy.roll, as reported by colomon++
rurban colomon: yes, I saw. I still have a rakudo buildbot which should have caught it, but it was dead.
masak lizmat++
timotimo rurban: that's part of the optimization for method signatures gsoc project?
rurban yes
timotimo i can mention that, thanks!
colomon lizmat++ \o/
rurban thank you :)
masak m: say (1..6).roll(1).perl 13:32
camelia rakudo-moar fd5f8e: OUTPUT«1␤»
lizmat m: (a=>10,b=>20).Mix.roll.perl.say
camelia rakudo-moar fd5f8e: OUTPUT«("b",).list␤»
lizmat masak ^^^
only for Mixes
colomon lizmat++ # Mixes are perfect for what I'm doing 13:34
lizmat colomon: yw :-) 13:35
masak lizmat: yes, I was just testing the list case when 1 was explicitly passed in. 13:37
[Coke] FROGGS++ 13:40
dalek kudo/nom: b5ee59f | (Elizabeth Mattijsen)++ | / (3 files):
Implement Supply.minmax, test and announce it
 13:41
colomon cannot begin to say how awesome it is to find a bug in your code, realize your code is correct and the compiler is wrong, and have the compiler fixed within the hour. :) 13:42
timotimo tadzik: got anything i can already link to for your game stuff? 13:44
timotimo colomon: i'd prefer less compiler bugs, tbh :) 13:45
well ... post publish'd 13:46
colomon timotimo: sure, less compiler bugs is a great goal. 13:47
FROGGS timotimo++
colomon timotimo++ 13:48
dalek kudo/nom: 049c431 | (Elizabeth Mattijsen)++ | src/core/List.pm:
De-TAB
 13:49
lizmat timotimo++ :-)
lizmat ETA on NLPW videos: not this week :-( 13:56
[Coke] hey, lizmat. 13:57
lizmat [Coke] /o 13:58
masak ooh, there'll be NLPW videos? that's great news! 13:59
timotimo not this week either? :( 14:00
[Coke] /home/coke/sandbox/perl6-roast-data/rakudo.jvm/t/spec/fudge: No such test file 't/spec/S17-supply/max.t' 14:01
/home/coke/sandbox/perl6-roast-data/rakudo.jvm/t/spec/fudge: No such test file 't/spec/S17-supply/min.t'
lizmat the video guy is on holiday :-(
timotimo ah, ok 14:01
lizmat [Coke] have you pulled roast ? 14:01
FROGGS lizmat: mix.t/min.t missing here: github.com/perl6/roast/tree/master/S17-supply 14:03
err, max.t/min.t
lizmat oops 14:04
dalek ast: 20fd106 | (Elizabeth Mattijsen)++ | S17-supply/m (3 files):
Add tests for Supply.(min|max|minmax)
 14:05
lizmat [Coke] did you pull roast ? :-)
oO( pompedipompedipom ) 14:06
FROGGS *g* 14:06
lizmat whistles as if nothing happened
FROGGS lizmat: hmm, weird, as if it was there already :P
lizmat for some definition of there, that would be true :-)
dalek kudo/nom: cba80b6 | (Elizabeth Mattijsen)++ | src/core/ (3 files):
More de-TABbing
 14:11
lizmat cycling& 14:12
btyler m: (1, 2, 3)>>++; #LTA error compared to r-j 14:30
camelia rakudo-moar fd5f8e: OUTPUT«No such method 'STORE' for invocant of type 'Int'␤ in sub postfix:<++> at src/gen/m-CORE.setting:4556␤ in sub flatmap at src/gen/m-CORE.setting:17410␤ in sub METAOP_HYPER_POSTFIX at src/gen/m-CORE.setting:17313␤ in block at /tmp/FBEwBDG4Hg:1…»
btyler j: (1, 2, 3)>>++; 14:31
camelia rakudo-jvm fd5f8e: OUTPUT«Unhandled exception: Failed to load config.properties␤ in (gen/jvm/stage2/NQPHLL.nqp:8)␤ in (gen/jvm/stage2/NQPHLL.nqp:8)␤ in (gen/jvm/stage2/NQPHLL.nqp)␤ in (src/vm/jvm/ModuleLoader.nqp:78)␤ in load_module (src/vm/jvm/ModuleLoader.nqp:60…»
btyler whoops
well, r-j gave me "can't modify immutable value", which is rather clearer than the r-m message
masak btyler: agreed. maybe submit a rakudobug? 14:32
hoelzro morning #perl6 14:36
moritz masak: there's already a bug report, and failing tests, for that error message 14:40
btyler oh, I should have read the test file further, sorry for the noise 14:41
masak oh 14:55
dalek ecs: 59480b5 | (Timo Paulssen)++ | S17-concurrency.pod:
fix formatting of a text paragraph
 14:59
hoelzro o/ timotimo
I started up a repo on GH for keeping syntax highlighting/editor support in one place; anyone want a commit bit? 15:00
timotimo saw it :)
colomon hoelzro++
lolibloggedagain: justrakudoit.wordpress.com/2014/05/...-reminder/ 15:04
timotimo colomon: isn't a mix with "higher status levels" giving you those tunes more often? 15:08
oh!
there was code that was scrolled off to the side
colomon timotimo: no, the mix is weighted via 1 / 2 ** $status
yeah, wordpress is awkward that way
timotimo 'tis
colomon just realized he still remembers how to play both parts of the Flying Reel, but doesn't remember how many times each is repeated. :\ 15:11
tadzik timotimo: not really, it's not quite ready 15:19
yoleaux 09:56Z <raydiak> tadzik: tried through the window again on a recent moar, and it runs playably on my craptop; might have tried jvm last time...figured you'd want to know its not as grim as it seemed
tadzik raydiak: yay! \o/ 15:19
timotimo: I wanted this one to have menus and all that, but it seems that bugs will stop me :/
dalek kudo/nom: fb27b44 | (Tobias Leich)++ | src/Perl6/Optimizer.nqp:
unbust parrot build

There are two problems popping up as shown in the comment. Sadly I am unable to provide a fix so I just unbreak the build by commenting out the triggering code.		 15:20
tadzik but I at least want to add explosions, I found some nice sprites :o
timotimo: you could link to the prereleased code though
timotimo huh, menus? 15:22
tadzik: github.com/tadzik/rakudobrew/pull/7 pls2mrg? 15:23
tadzik timotimo: merg'd 15:24
timotimo: yeah, like "Play, Instructions, Exit"
and it goes to the menu when you win/lose, instead of closing the game
timotimo mhm
sdl2 has font rendering for us now? :)
timotimo or do we need sdl_ttf for that? 15:24
tadzik if not, then sdl2_ttf surely does
I was thinking of just cheating and using pngs... :P 15:25
tadzik then I don't have to include a font and worry about its license 15:25
timotimo hah
aye.
tadzik but yeah, brokens 15:26
it's probably not brokens on parrot, but parrot inbearably slow
timotimo sad to hear it
tadzik so I may just ship it later today with no menus 15:27
but I want to add them explosions :o
I could workaround the bug, it'd just be a little bit annoying
timotimo if you make sure it's compatible with rakudo star moar ... :P
tadzik heh 15:28
masak FROGGS++
colomon Done, parrot-HEAD built \o/
colomon FROGGS++ 15:29
jnthn back 15:41
jnthn Hm, I think that patch shouldn't have commented it out, but instead check if we're on Parrot with an "if" and just done tried the opt. 15:42
jnthn Is there anybody about with a rakudo.org upload key? My laptop does not, apparently, have one. :/ 16:00
If so, jnthn.net/tmp/rakudo-star-2014.04.tar.gz is the thing to upload
[Coke] checks. 16:02
FROGGS jnthn: Perl6::Optimizer is not under fudge's control 16:03
so I was unable to #?if that out
[Coke] jnthn: uploaded, I think. 16:05
FROGGS timotimo / tadzik: about font rendering... at some point you might want to support pango
[Coke] jnthn++
FROGGS yes, it is available at rakudo.org/downloads/star/rakudo-st....04.tar.gz 16:06
tadzik [Coke]++ 16:09
[Coke] it was pretty much the least I could do! ;)
tadzik FROGGS: probbly, yeah 16:10
FROGGS damn, we need cpantesters so much! 16:13
TimToady :v is not supposed to return anything for missing elements; returning either () or Nil would be wrong 16:21
PerlJam m: say $*OUT: "hi"; # Is this supposed to work or am I just mis-remembering something basic? 16:32
camelia rakudo-moar fd5f8e: OUTPUT«===SORRY!=== Error while compiling /tmp/bVZcaSG_7L␤Two terms in a row␤at /tmp/bVZcaSG_7L:1␤------> say $*OUT⏏: "hi"; # Is this supposed to work or a␤ expecting any of:␤ postfix␤ infix sto…» 16:32
benabik m: print $*OUT: ‘hi’ 16:33
camelia rakudo-moar fd5f8e: OUTPUT«===SORRY!=== Error while compiling /tmp/3_3FPsAtkK␤Two terms in a row␤at /tmp/3_3FPsAtkK:1␤------> print $*OUT⏏: ‘hi’␤ expecting any of:␤ postfix␤ infix stopper␤ infix or meta…»
TimToady rakudo doesn't parse that yet 16:33
benabik Erm, no, that’s basically the same thing you did. Whoops. 16:34
benabik std: say $*OUT: ‘hi’ 16:34
camelia std ec3d853: OUTPUT«===SORRY!===␤Confused at /tmp/U5EBrXIZGm line 1:␤------> say $*OUT: ⏏‘hi’␤ expecting any of:␤ feed_separator␤ infix or meta-infix␤ infixed function␤ statement modifier loop␤Parse failed␤FAILED 00:01 …»
PerlJam Then I guess I was misremembering that rakudo did parse it.
TimToady um, those aren't ASCII quotes
std: say $*OUT: 'hi' 16:35
camelia std ec3d853: OUTPUT«ok 00:01 125m␤»
benabik Is OS X mucking with my quotes now?
TimToady something is
benabik std: 'a' # fixed? 16:36
camelia std ec3d853: OUTPUT«ok 00:01 121m␤»
TimToady we may have to break down and allow fancy quotes if this becomes more of a problem
and things like … too 16:37
PerlJam I have no desire for fancy quotes, but I've occasionally wondered why … didn't work. 16:37
TimToady blames dwimmy telephones for changing people's expectations
BenGoldberg m: say q􏿽x91test􏿽x92 16:39
camelia rakudo-moar fd5f8e: OUTPUT«test␤»
BenGoldberg m: $*OUT.say 'hi'; 16:40
TimToady it wouldn't be hard to support them, it's just a bit more bloat
camelia rakudo-moar fb27b4: OUTPUT«===SORRY!=== Error while compiling /tmp/T7w4Xi_D7K␤Two terms in a row␤at /tmp/T7w4Xi_D7K:1␤------> $*OUT.say ⏏'hi';␤ expecting any of:␤ method arguments␤ postfix␤ infix stopper␤…»
PerlJam btw, is there a way to make !!! a little more dwimmy? (I always tend to use it like so: has $.foo = !!! 'foo is required'; # And I'd like to not have to type that message all the time. 16:41
BenGoldberg: you need the : or parens
BenGoldberg m: $*OUT.say: 'hi';
camelia rakudo-moar fb27b4: OUTPUT«hi␤»
vendethiel thinks that the object target specifier use of : is confusing
PerlJam vendethiel: confusing how?
vendethiel $*OUT.say: <= the : denotes the call 16:42
say $*OUT: the : denotes the receiver
TimToady std: has $.foo = !!!;
camelia std ec3d853: OUTPUT«Potential difficulties:␤ 'has' declaration outside of class at /tmp/7Z1WF23XpE line 1:␤------> has⏏ $.foo = !!!;␤ok 00:01 123m␤»
TimToady std is fine with leaving out the message
PerlJam yeah, but I want which attribute is required to still be output. fail, fail with warning, and die aren't quite dwimmy enough without some other introspective information. (I find anyway) 16:45
PerlJam I guess that's more of a compiler issue than a spec issue though 16:46
TimToady it would certainly be possible to make the message dwimmier, it's just a smop
PerlJam I guess I'll have to look at it when I have some time and see if I can't make it better. 16:47
FROGGS PerlJam: FYI, it is not a typed exception yet 16:51
so that would be the first step, throw a X::Stub or similar 16:52
PerlJam aye
FROGGS currently it is just: rakudo/src/Perl6/Actions.nqp:4182: $past.push($*W.add_string_constant('Stub code executed')); 16:53
TimToady that's not the fundamental problem; regardless of type, the exception needs to know what the user was intending to stub, so in this case it needs to know that we're in the initializer of a declarator 16:58
and that's more of a lexical issue than a dynamic issue 16:59
Martin90 Hello, what are the changes in perl 6 in comparsion to perl 5 ? 17:00
PerlJam Martin90: lots of things
Martin90: why do you ask?
Martin90 I am just curious of news in perl 6 17:01
jnthn FROGGS: Yeah, but we can check at runtime... nqp::getcomp('perl6').backend.name
Martin90 is there any sable version ?
stable*
hoelzro Martin90: rakudo is fairly stable
PerlJam Martin90: what do *you* mean by stable?
hoelzro and it's just getting better
Martin90 I've heard about it, what it means "rakudo" ?
hoelzro rakudo is one implementation of Perl 6 17:02
PerlJam Martin90: Rakudo is a Perl 6 compiler
BenGoldberg perl6 is definited by the specification.
hoelzro originally targetting Parrot, but it now targets Parrot, the JVM, and MoarVM
MoarVM is the new Perl 6 awesome sauce
Martin90 I've heard that perl 6 is able to produce exacutable code, true ?
hoelzro better described as a VM that's designed specifically for Perl 6 (more accurately, the Six Model Object system) 17:03
BenGoldberg There are a couple different implementations of that specification. There's the Rakudo implementation (with three different backends), and the Niecza implementation (which targets .net/mono)
PerlJam Martin90: Depends on what you mean by 'executable code' :-) But, the answer is probably "no" for what you mean. 17:04
Martin90 like .exe ?
on windows
tadzik that even perl 5 can do, if you put the perl compiler in that exe
jnthn [Coke]: Thanks for uploading!
tadzik you can do that with every language if you are clever enough 17:05
Martin90 true ;)
tadzik so, did you mean: produce machine code?
Martin90 but the key factor is "speed" of it
BenGoldberg ponders a befunge executable ;)
Martin90 yes machine code
PerlJam jnthn: Are you putting the release announcement on rakudo.org? (I was going to point Martin90 at rakudo.org and noticed it's still announcing March's release)
tadzik no, it cannot produce machine code 17:06
Martin90 Where can I find full list of changes from perl 5 to 6 ?
tadzik and producing machine code wouldn't really make it fast by itself
TimToady we should have a JIT by end of summer though
hoelzro I'd say it's unlikely for a Perl 6 implementation to produce machine code in the future, but we may get some very nice JITing in the future
BenGoldberg Well, if you use Rakudo with the jvm backend, and your computer uses java bytecode as it's native machine code... ;)
hoelzro Martin90: it's probably best to just read about Perl 6 itself; there are so many changes that a diff might not help
tadzik dynamic languages tend to fare better when they're dynamically optimized, rather than being compiled upfront
Martin90 Where can I find full list of changes from perl 5 to 6 ?
PerlJam Martin90: start reading at perl6.org. Particularly the synopses 17:07
tadzik or the "Perl 5 to 6" articles
PerlJam Martin90: You could say that everything changed but the spirit
tadzik they might be a bit dated, but they're a nice introduction
Martin90 perlJam what could you pelase provide linke to this synopis 17:08
?
PerlJam Martin90: if you'd like to compare perl5 code against perl6 code, checkout rosettacode.org
tadzik it's on perl6.org
PerlJam Martin90: what tadzik said
(there's also a link to rosettacode on perl6.org too)
[Coke] looks at rosettacode.org/wiki/Sieve_of_Eratosthenes#Perl and thinks that might be a little too many ways to do it. 17:09
Martin90 ok thanks ;) 17:10
jnthn PerlJam: Yeah, working on posting that right now :)
Martin90 And the most important quetion
how looks the performance comparsion between perl 5 and 6 ?
PerlJam Martin90: Depends on which compiler you use and which features you exercise. 17:11
TimToady short answer: not as bad as it used to
Martin90 in favor ?
tadzik also short answer: not too good :)
PerlJam Martin90: Currently the Rakudo compiler is slower, but has the ability to get faster :) 17:12
tadzik real world example: a 60fps space invaders with 100 collision checks per frame is basically pushing the limits of rakudo's speed
Martin90 haha, nice answer ;)
what is your generaln impression of perl 6 ? 17:13
TimToady likes it
Martin90 syntax speed etc
PerlJam heh
Martin90: if you're looking for speed, you probably don't want to use any of the perl6 compilers yet.
jnthn Perl 6 always feels very refactorable to me. It's easy to evolve code from one form to another.
PerlJam Martin90: (You realize that you're talking to the people that use and hack on perl6 all the time, right?) 17:14
Martin90 refactorable ? 17:15
PerlJam, rather yes, am I wring ? 17:16
wrong*
tadzik I'd say "yes", but it might not be the answer you're looking for :)
tadzik it makes sense to ask us, but we're also slightly biased 17:16
Martin90: I think, performance or not, that it's worth to try Perl 6 out, even if you won't end up using it 17:17
Martin90 heh, what you mean by biased ?
tadzik it has some nice ideas in it, and will certainly teach you something new about programming in general
PerlJam Martin90: The best thing to do, if you want to get a feel for Perl 6, is to install a compiler and use it.
TimToady perceives that Martin90 is not a native English speaker
tadzik biased as in, we put a lot of effort into it ourselves, so our opinion may be a bit off compared to what the outsiders would say
Martin90 PerlJam right ;)
but what you feel will perl 6 push perl 5 out in next two years ? 17:18
PerlJam Martin90: Perl 5 isn't going anywhere for a long time.
vendethiel well, if you want to compare language X and Y, I don't think the best idea is too ask somebody who used both
tadzik I don't think anything will push Perl 5 out in the next 20 years :)
vendethiel it's probably better to ask people working on X and people working on Y :-)
jnthn PerlJam: rakudo.org/2014/05/05/announce-raku...e-2014-04/
BenGoldberg There is still perl4 code out there, in use today. 17:19
PerlJam jnthn++
BenGoldberg feels sorry for the poor fellows who have to maintain it ;)
PerlJam Martin90: see rakudo.org :)
Martin90 do you know who is in developers teem for perl 6 ? 17:20
tadzik we are :o
PerlJam Martin90: us!
vendethiel people in this channel :p
Martin90 Lary Wall ? ;)
BenGoldberg Surprisingly, no
tadzik he's here somewhere to
PerlJam no?
vendethiel Unsurprisingly, no :p
tadzik too
nwc10 BenGoldberg: your information is now out of date
Martin90 I ve heard Randal Shwartz is there 17:21
PerlJam I've never seen Randal here. 17:21
tadzik M<TAB> says he's not
nwc10 BenGoldberg: but only fairly recently, and not massively wrong either.
BenGoldberg Ahh
PerlJam In fact, I've seen him more on #git than #perl even.
Martin90 right 17:23
Damin conway works on perl 6 ?
any info aut there ;)
out* 17:24
PerlJam huh ... looks like merlyn was here once last year (and earlier too)
hoelzro Martin90: Damian used to be pretty involved, but I think that he has other priorities these days 17:25
suffice to say, anyone who's seriously contributing (and a few cheerleaders like myself ;) ) are here
PerlJam He's still touring and giving Perl 6 presentations
tadzik I know for a fact that he offers his guidance when you ask him how to interpret the specs he wrote :D
hoelzro heh 17:26
PerlJam Martin90: anyway, install Rakudo and play with it. You'll find there's lots of neat stuff in the language. 17:27
Martin90 PerlJam, I am going to do that ;) 17:28
hoelzro there's a Perl 6 book online
but it's something like 2 years out of date, I think?
jnthn Sent the R* announce to the mailing lists too.
tadzik it is
hoelzro someone should bring it up to speed
jnthn If anybody wants to do the Wikipedia updates, feel free...I need dinner :)
TimToady the main difference between Perl 5 and Perl 6 is that Perl 6 fixes most of the problems mentioned in the 361 RFCs that people came up with in 2000 17:29
PerlJam jnthn++ get some food! :)
hoelzro I noticed when I read it long ago that it skimped on some of the more advanced features as well
Martin90 Why is see perl 6 always connected with parrot ?
PerlJam hoelzro: well volunteered! ;)
Martin90: because parrot was originally designed as the VM that Perl 6 would use. 17:30
jnthn bbl &
Martin90 I always see perl 6 and parrot in articles
hoelzro PerlJam: I would if I had time =)
I might get a round to it one day
but if I did, I would need a proofreader with better Perl 6 fu than I
PerlJam hoelzro: woolfy and lizmat have *bunches* of round tuits if you need some :)
Martin90 I see but do we really need V for perl ?
VM*
colomon ==> Fetching IUP
==> Building IUP
No such method 'stage' for invocant of type 'X::Parameter::InvalidType' 17:31
in sub gen-result-failure at ./smoker:11
in sub MAIN at ./smoker:53
hoelzro I have a tendency to talk big about Perl 6 projects, but never get around to them =/
colomon in block at ./smoker:4
[Coke] Martin90: spoken like someone who has never hacked on the perl 5 core! ;)
colomon Both parrot and jvm smoke tests die there ATM.
TimToady Martin90: for a long time people treated them as synonymous, but parrot evolved into a VM for many languages, and Perl 6 evolved into a langauge for many VMs, so the visions diverged
[Coke] going forward, you're more like to see rakudo tied to moarvm, not parrot. (though it will continue to target parrot for the foreseable future)
Martin90 Coke, yo uare right in some way ;) Why you are poke in pler core ;) 17:32
PerlJam hoelzro: yeah, I know the feeling. There's always lots to do, it's hard to do them all. 17:33
Martin90 TimToady, hanks ;)
thanks
PerlJam hoelzro: though, last night I did something I'd been meaning to do, I got started on github.com/perlpilot/p6-Git-Wrapper
hoelzro oh, nice! 17:34
I started on the perl6-syntax repo myself =)
PerlJam it's not nice yet, but it's workable for simple needs.
hoelzro so I'm finally getting around to some Perl 6-y things
FROGGS colomon: dunno if this open PR would help: github.com/mrhdias/perl6-IUP/pulls 17:37
colomon FROGGS: I dunno, it failed gracefully up until sometime this month.... 17:38
retupmoca I was looking at making IUP use LibraryMake a couple weeks back, but I'm not sure how to install the C library on my system :/
PerlJam What's the little S11 crossed out on the panda icons on modules.perl6.org supposed to signify exactly? 17:44
benabik That they don't conform to S11? 17:45
PerlJam yeah, but what does that *mean*?
or, how is conformance determined? 17:46
retupmoca whether the META.info has a 'provides' line
PerlJam so ... S11 doesn't contain the word "provides" and S22 says that it must be META6.json and that " 17:49
"provides" is optional
FROGGS PerlJam: yes, there was a redesign in March, and we need to catch up 17:50
PerlJam oh, that would explain my confusion. I hadn't really noticed the S11 designations before, so I assumed they were part of the "redesign" 17:51
FROGGS we're are currently working on the cpan.org indexer, and when that is in place, the S11 icon thingy will change anyway
no, the icon you see was introduced in december
PerlJam panda has made me look at modules.perl6.org less than I once did :-) 17:52
panda++
masak aww, missed a Martin90! :) 18:00
others++ # good job
TimToady obviously he loves Perl if he has a .pl extension on his network address :) 18:01
FROGGS yeah, I wanted to mention the PLPW but then he vanished :/ 18:02
FROGGS because there is much Perl 6 going on 18:02
vendethiel there is :-) 18:04
TimToady some people think there's too much going on, but that's a nice problem to have 18:07
vendethiel TimToady: as a beginner/newcomer, that doesn't seem like a problem ! :)
timotimo when will we remove "non-blocking I/O" from the "not quite there yet" feature list in the rakudo star announcements? 18:20
tadzik is now officially going to yapc::eu 18:22
colomon tadzik++ 18:28
FROGGS damn, we still have that nqp-m Win32 x86 problem... "MAST::Local index out of range" 18:32
nwc10 when/how does that happen? 18:33
the error message could do with improving - I feel that it should also show the two values used in the comparison that causes the failure
and also, the two places that currently have that same message should differ 18:34
FROGGS nwc10: Building NQP ...
C:\Perl\bin\perl.exe Configure.pl --prefix=C:\rakudo --backends=moar --make-install
Found C:\rakudo/bin/moar.exe version 2014.04, which is new enough.
Cleaning up ...
Microsoft (R) Program Maintenance Utility, Version 10.00.30319.01
Copyright (C) Microsoft Corporation. Alle Rechte vorbehalten.
Microsoft (R) Program Maintenance Utility, Version 10.00.30319.01
Copyright (C) Microsoft Corporation. Alle Rechte vorbehalten.
C:\Perl\bin\perl.exe -MExtUtils::Command -e mkpath gen\moar\stage1\gen
C:\Perl\bin\perl.exe tools\build\gen-cat.pl moar src\how\Archetypes.nqp src\how\RoleToRoleApplier.nqp src\how\NQPConcreteRoleHOW.nqp src\how\RoleToClassApplier.n
qp src\how\NQPCurriedRoleHOW.nqp src\how\NQPParametricRoleHOW.nqp src\how\NQPClassHOW.nqp src\how\NQPNativeHOW.nqp src\how\NQPAttribute.nqp src\how\NQPModuleHOW.nqp s
rc\how\EXPORTHOW.nqp > gen\moar\stage1\nqpmo.nqp
C:\rakudo\bin\moar --libpath=src\vm\moar\stage0 src\vm\moar\stage0\nqp.moarvm --bootstrap --setting=NULL --no-regex-lib --target=mbc --output=gen\moar\stage1\nqp
dalek ecs: 2cdc917 | (Elizabeth Mattijsen)++ | S17-concurrency.pod:
Spec Supply.reduce
benabik ~.~
FROGGS mo.moarvm gen\moar\stage1\nqpmo.nqp
MAST::Local index out of range
at gen/moar/stage2/QAST.nqp:4818 (src\vm\moar\stage0/QAST.moarvm:assemble_to_file:19)
from gen/moar/stage2/NQPHLL.nqp:61 (src\vm\moar\stage0/NQPHLL.moarvm:mbc:61)
from gen/moar/stage2/NQPHLL.nqp:1327 (src\vm\moar\stage0/NQPHLL.moarvm::79)
from gen/moar/stage2/NQPHLL.nqp:1315 (src\vm\moar\stage0/NQPHLL.moarvm:compile:164)
from gen/moar/stage2/NQPHLL.nqp:1073 (src\vm\moar\stage0/NQPHLL.moarvm:eval:53)
nwc10 OK, I don't see it on x86 linux, which is curious
FROGGS from gen/moar/stage2/NQPHLL.nqp:1282 (src\vm\moar\stage0/NQPHLL.moarvm:evalfiles:105)
from gen/moar/stage2/NQPHLL.nqp:1212 (src\vm\moar\stage0/NQPHLL.moarvm:command_eval:292)
from gen/moar/stage2/NQPHLL.nqp:1161 (src\vm\moar\stage0/NQPHLL.moarvm:command_line:122)
from gen/moar/stage2/NQP.nqp:3697 (src\vm\moar\stage0\nqp.moarvm:MAIN:18)
from gen/moar/stage2/NQP.nqp:3692 (src\vm\moar\stage0\nqp.moarvm::335)
nwc10 and presuambly not on jnthn's machine, which is 64 bit?
FROGGS from <unknown>:1 (src\vm\moar\stage0\nqp.moarvm::8)
damn!
gist.github.com/FROGGS/ed7f3e4455c33669f58b 18:35
I'm sorry ó.ò
I don't see that on my 64bit Win7 either
I just built a rakudo-star msi for parrot backend and wanted to do the same for moar
vendethiel damn spam 18:37
Woodi heh, I was just compiling moar on debian and wondering... Microsoft (R)... What is going on ?? 18:37
FROGGS: np :) 18:38
nwc10 FROGGS: can you fake it by compiling the MoarVM "byte"code on a different system, and just running the tests on Win32?
FROGGS nwc10: hmmm, good question 18:39
I can also try to copy over my moar binary, and use that to build nqp 18:41
... on a x64
nwc10 FROGGS: hangon, is that the first thing that NQP tries to build? 18:49
FROGGS yes, and that is weird, see: 18:50
src\mast\compiler.c:753 MAST::Local index out of range 0 >= 1
it thinks that 0 >= 1
nwc10 you edited it to show more diagnostics? 18:51
FROGGS yes
DIE(vm, "%s:%d MAST::Local index out of range %d >= %d", __FILE__, __LINE__, l->index, ws->cur_frame->num_locals);
nwc10 did you use the correct format strings?
oh, no you didn't
FROGGS ohh?
nwc10 l->index is MVMint64
FROGGS %ld then, right?
nwc10 I've been staring at that code most of Sunday. :-)
%lld
FROGGS k 18:52
nwc10 actually, %llu
and I would add that there is quite some inconsistency about what's 64 bit and what's 32 bit
but that's way down in the list of "things to see to"
FROGGS src\mast\compiler.c:753 MAST::Local index out of range 4294967296 >= 43 18:54
:/
nwc10 Oh, I've seen *that* one
OK, not quite that one 18:55
nwc10 I mailed you 2 patches, which probably won't solve it 18:58
I'm still playing at breaking stuff
FROGGS cool, trying...
nwc10 if they do fix it, I wonder what broke it and when 18:59
[Coke] rages at this SQL that goes exponential if I try to union one.more.thing. 19:06
PerlJam [Coke]: missing index? 19:07
nwc10 "it will be fixed in the next version of MySQL" :-)
[Coke] it's Oracle. I have 3 subqueries defined as "with ..."; not sure there's a place to -put- an index. 19:08
cost (via explain plan) when it's zippy is 348; cost when it's evil is only 367. :P 19:09
PerlJam you've discovered oracles' magic constant! 19:10
FROGGS nwc10: it continues!
nwc10 FROGGS: oh, *interesting* 19:11
so, I wonder what point it broke.
nwc10 Pi takes 12 or so minutes to build MoarVM, and I don't know if it ever worked on ARM, so I don't want to try to bisect 19:12
benabik bisect does require having a known good point. :-D 19:13
PerlJam nwc10: no cross compiler? 19:13
nwc10 so, yes, my question is if anyone had tried to build NQP on MoarVM on Win32 before
PerlJam: is MoarVM cross compilable? Who knows 19:14
also no known good start point
FROGGS nwc10: that it does not work for windows x86 is a fact for long time now... (long for the lifetime of moarvm at least)
nwc10 15 isn't *insanely* slow actually. It's the non good point
FROGGS: I thought that it was just the Nativecall. It was actually much earlier? The NQP stuff?
given that it's taking 8 or so hours for stage mast 19:15
I'm prepared to wait for the Pi to finish natively
FROGGS I know that the build was already broken when JimmyZ++ actively worked on implementing the IO ops
nwc10 OK. Interesting
did I accidentally fix it for Win32 too?
FROGGS :o) 19:16
I'm in rakudo's state mast now
it build
nwc10 Said patches are arguably ready to go in, but I was going to lob them to the list when I'd stressed them a bit more on x86 and x86_64 linux
the Pi is in the NQP tests with unsigned chars
(ie no funky flags)
dalek kudo/nom: 9995d08 | (Elizabeth Mattijsen)++ | / (3 files):
Implement Supply.reduce, test and announce it
 19:17
ast: 6b01eb2 | (Elizabeth Mattijsen)++ | S17-supply/reduce.t:
Add tests for Supply.reduce
dalek ast: 91bbf6e | (Elizabeth Mattijsen)++ | S17-supply/reduce.t:
Make (+) do the coercing
 19:21
dalek kudo/nom: 709f630 | (Elizabeth Mattijsen)++ | src/core/Any.pm:
Remove () from :v:k:kv:p: hash/array slices
 19:42
lizmat m: ({a => 1}<a b c>:v).perl.say
camelia rakudo-moar fb27b4: OUTPUT«(1, (), ())␤»
lizmat $ perl6 -e '({a => 1}<a b c>:v).perl.say' 19:43
(1,)
sergot: ^^^ 19:44
BenGoldberg n: ({a => 1}<a b c>:v).perl.say 19:56
camelia niecza v24-109-g48a8de3: OUTPUT«Unhandled exception: Excess arguments to KERNEL Hash.postcircumfix:<{ }>, unused named v␤ at <unknown> line 0 (KERNEL Hash.postcircumfix:<{ }> @ 1) ␤ at /home/p6eval/niecza/lib/CORE.setting line 186 (postcircumfix:<{ }> @ 5) ␤ at /tmp/SGN2cpc…»
FROGGS :/ 19:57
Unhandled exception: java.lang.StackOverflowError when building Perl6::Action on Win32 x86
lizmat BenGoldberg: niecza doesn't do array/hash slice adverbs afaik 19:59
BenGoldberg realizes that. :)
cognominal__ just did a small pull request : github.com/perl6/nqp/pull/168 20:04
dalek p: 03349b2 | (Stéphane Payrard)++ | src/QRegex/Cursor.nqp:
Fixing --target=parse on nqp

Was working fine on rakudo		 20:05
p: 0d764f8 | (Tobias Leich)++ | src/QRegex/Cursor.nqp:
Merge pull request #168 from cognominal/master

Fixing --target=parse on nqp
FROGGS cognominal++
cognominal__ FROGGS++ # thanks 20:07
dalek rl6-roast-data: f894557 | coke++ | / (6 files):
today (automated commit)
 20:12
[Coke] growing number of supply failures for jvm & moar. 20:13
lizmat also on moar ?
lizmat looks
[Coke] github.com/coke/perl6-roast-data/b....out#L2306 20:14
github.com/coke/perl6-roast-data/b....out#L2288
FROGGS rakudo star release candiates for windows (as msi):
froggs.de/perl6/rakudo-star-2014.04-moar.msi (7.5MB)
froggs.de/perl6/rakudo-star-2014.04-parrot.msi (17.3MB)
please test
lizmat moar's errors seem to be watch_path frragilness 20:15
the jvm errors are interesting 20:16
FROGGS ahh, the rakudo@parrot checkout is like 10 hours old
lizmat j: class A { method a (A:D:) {...} }; A.a
camelia rakudo-jvm fb27b4: OUTPUT«(timeout)» 20:17
lizmat j: class A { method a (A:D:) {...} }; A.new.a
camelia rakudo-jvm fb27b4: OUTPUT«Unhandled exception: Stub code executed␤ in throw (gen/jvm/CORE.setting:11226)␤ in sink (gen/jvm/CORE.setting:12664)␤ in MAIN (gen/jvm/main.nqp:47)␤ in (gen/jvm/main.nqp:41)␤ in (gen/jvm/main.nqp)␤␤»
lizmat j: class A { method a (A:D:) {...} }; A.a
camelia rakudo-jvm fb27b4: OUTPUT«Unhandled exception: Stub code executed␤ in throw (gen/jvm/CORE.setting:11226)␤ in sink (gen/jvm/CORE.setting:12664)␤ in MAIN (gen/jvm/main.nqp:47)␤ in (gen/jvm/main.nqp:41)␤ in (gen/jvm/main.nqp)␤␤»
lizmat m: class A { method a (A:D:) {...} }; A.a
camelia rakudo-moar fb27b4: OUTPUT«Invocant requires an instance, but a type object was passed␤ in method a at /tmp/UtLTPYwwWR:1␤ in block at /tmp/UtLTPYwwWR:1␤␤»
lizmat jvm is wrong in this respect
ergo: D: doesn't work in signatures on jvm 20:18
I will mark those tests as "todo" for jvm
benabik D: indeed.
jnthn lizmat: Whoa...how on earth did hat not get noticed agaes ago... 20:20
lizmat I have no idea ... :-(
lizmat seems (TypeName:D:) works 20:23
but (TypeName:D $self:) does not
moritz that'd explain it
lizmat e.g. method do(Supply:D $self: &side_effect) {
jnthn wat...
lizmat eh? 20:24
did I do something wrong ?
moritz j: say 42
camelia rakudo-jvm fb27b4: OUTPUT«42␤»
moritz j: class A { method a (A:D:) { 42 }; method b(A:D $self:) { 23 } }; say A.a; say A.b 20:25
camelia rakudo-jvm fb27b4: OUTPUT«42␤23␤»
jnthn lizmat: No, the bug...
moritz seems to work in neither case
dalek ast: cd68374 | (David Warring [email@hidden.address] | integration/advent2012-day14.t:
created RT ticket for fudged test
 20:26
FROGGS j: class A { method a (A:D:) { say nqp::isconcrete(self) }; method b(A:D $self:) { say nqp::isconcrete($self) } }; A.a; A.b 20:32
camelia rakudo-jvm fb27b4: OUTPUT«0␤0␤»
dalek kudo/nom: 4cd39b9 | (Tobias Leich)++ | src/Perl6/Optimizer.nqp:
disable specific optimization only on parrot, jnthn++
 20:33
jnthn FROGGS++
dalek ast: 0128a9b | (Elizabeth Mattijsen)++ | S17-supply/ (19 files):
Fudge (A:D $self:) signature related for JVM
 20:45
lizmat *tests
PerlJam m: class C { has @.a; }; my $a = C.new( :a<a b c d> ); my $b = $a.clone; $b.a.[2] = "foo"; say $a; say $b 21:02
camelia rakudo-moar 709f63: OUTPUT«C.new(a => Array.new("a", "b", "foo", "d"))␤C.new(a => Array.new("a", "b", "foo", "d"))␤»
PerlJam How do I clone properly so that $a and $b are independent of each other? 21:03
jnthn PerlJam: Write a method clone in your class
It's just a method; if the default doesn't work for you, you can do something else. 21:04
PerlJam there's no deepclone or anything?
jnthn No. That's tricky enough it's for modules, imo.
PerlJam okie. Thanks.
jnthn I think that's been the conclusion when this has come up before, anyway, and it makes sense to me.
jnthn Note in this case you can probably get away with method clone() { nextwith(a => @a.clone) } or so 21:07
dalek kudo/nom: f374d87 | (Elizabeth Mattijsen)++ | src/core/Supply.pm:
If Supply in on() is live, the resulting is also
 21:28
benabik But what if the Supply in on() is Memorex? 21:29
lizmat only if all Supplies in on()n are Memorex, is the resulting supply a Memorex 21:30
lizmat needs eleep
so, gnight #perl6!
benabik But how do you tell the difference? ;-) 21:31
'night
woolfy nwc10++ (and many more ++) for pioneering Perl 6 on Raspberry Pi. Wow. 21:44
timotimo for when seeing one letter per second on a desktop isn't slow enough for you just yet 21:46
hoelzro so I can't believe this, but I found a task at my job where I can finish faster in Perl 6 than any other language =) 21:53
(at least if I could get this thing working right)
can anyone give me a hand on this:
perl6 -n -e 'BEGIN { my $max = 0 }; $max max= .chars.say; END { $max.say }'
I can't figure out how to declare $max once but have it referred to within the loop implied by -n 21:54
PerlJam maybe you want to use a state var instead? 21:55
hoelzro ahhhh
good call
PerlJam It would be interesting if you could do something like perl6 -e 'my $max = 0' -n -e '$max max= .chars; END { $max.say }' 21:59
hoelzro that would be nice 22:03
-e is supposed to imply 'no strict' though, rigth? 22:04
*right
PerlJam I think so. But I doubt it's implemented. 22:05
hoelzro same here 22:06
cognominal__ r: [42,666]».say 22:20
camelia rakudo-jvm 4cd39b: OUTPUT«(timeout)»
..rakudo-{parrot,moar} 4cd39b: OUTPUT«666␤42␤»
thou hoelzro: state works; so does: ls -1 | perl6 -n -e 'BEGIN { our $max = 0 }; our $max += 1; END { say "Max: $max"; };' 22:22
hmm, for that particular case, my works, too: ls -1 | perl6 -n -e 'my $max += 1; END { say "Max: $max"; };' 22:23
Not sure why... 22:24
timotimo oh, that's interesting 22:25
have it give you --target=ast and/or --target=optimize
segomos 666 22:38
thou timotimo: Was that (--target=ast) directed at me (re my $max)? 22:56
By the way, I'm unable to get j-spectest to complete. I get somewhere into S06-* and almost all tests result in "No subtests run". I have plenty of RAM available on the system, but the java proc is using 2.6G or so RAM, maybe it's got some leak and hits a resource limit? 22:59
thou Anyone else have that? I'm on OS X 10.9.2, java 1.7.0_55 64-bit 23:00
colomon Cannot add breakpoint to unknown file 'Mu.pm' 23:41
darn

Powered by Cro and the Raku® Programming Language.

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

Thank you!