»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.
00:03 FROGGS left 00:08 gdey left 00:12 Psyche^ joined 00:29 virtualsue left 00:35 colomon joined 00:46 denis_boyun left
skids r: my @f = (1,2,3); @f.say; @f[1] = 6; my ($a, $b, $c); :($a,$b,$c) := @f[0],@f[1],@f[2]; $b.say; $b = 5; # Should work? Any workarounds? 00:54
camelia rakudo 789db4: OUTPUT«1 2 3␤6␤Cannot assign to a readonly variable or a value␤ in block at /tmp/3zuvfNAA14:1␤␤»
skids oh, is rw works. 00:55
[Coke] eeeeks, as he is now committed to Frankfurt 01:04
01:05 census left 01:06 FROGGS joined
skids r,std: my $b = 4; my ($a) := $b; $a = 1 01:09
hmm.
r: my $b = 4; my ($a) := $b; $a = 1
camelia rakudo 789db4: OUTPUT«Not enough positional parameters passed; got 0 but expected 1␤ in block at /tmp/zVdg0kzRvU:1␤␤»
skids std: my $b = 4; my ($a) := $b; $a = 1
camelia std 06636e4: OUTPUT«ok 00:00 44m␤»
01:12 benabik joined 01:14 jnap joined
colomon [Coke]++ lizmat++ 01:18
[Coke] is there a gist/wiki some where for arrival/departure times? 01:20
01:22 FROGGS left 01:25 _jaldhar is now known as jaldhar 01:28 stevan_ left 01:29 stevan_ joined 01:47 jnap left
dalek rl6-roast-data: 5ccd09f | coke++ | perl6_pass_rates.csv:
Eliminate all comment lines.

Add date, sha1, and comment columns.
Temporarily lose percentages.		 01:54
[Coke] sweet. now github.com/coke/perl6-roast-data/b..._rates.csv is searchable and sortable. 01:55
(so you can easily, say, show all the "rakudo" results.)
02:01 Mouq left 02:11 Mouq joined 02:36 odoacre_ joined 02:40 Mouq left 03:01 hypolin joined, hypolin left
lue Is there any particular reason use of multiple double angles isn't mentioned in S26/"Formatting Codes" ? (like C««@a».lines»») 03:15
03:17 preflex_ joined, ChanServ sets mode: +v preflex_ 03:18 preflex left, preflex_ is now known as preflex 03:21 FROGGS joined
dalek rl6-roast-data: b8e6e8e | coke++ | perl6_pass_rates.csv:
fix some date mixups
 03:21
rl6-roast-data: f5cf8e5 | coke++ | perl6_pass_rates.csv:
add "percentage" column

Remove extraneous quotes
diakopter [Coke]: you've seen this p6 impl? hf.64p.org/list/perl6/pvip 03:22
dalek rl6-roast-data: db36980 | coke++ | perl6_pass_rates.csv:
title new column
 03:23
[Coke] diakopter: nope.
if we can build it on the new roaster, and there's a desire, I can add it to the pile. 03:25
bonsaikitten diakopter: does it have any proper homepage? 03:30
dalek rl6-roast-data: 73f72d7 | coke++ | cull:
make new records match new format
 03:34
diakopter bonsaikitten: github.com/tokuhirom/pvip 03:35
bonsaikitten right
diakopter also see github.com/tokuhirom/Perl6-PVIP
bonsaikitten sounds like I have something to play with 03:36
[Coke] github++
bonsaikitten :)
[Coke] ok. github.com/coke/perl6-roast-data/b..._rates.csv is updated to move all the data into a valid CSV format, now including the %ages. enjoy, feel free to suggest improvements. 03:38
also no more icky US dates. 03:39
dalek rl6-roast-data: 7eb58b0 | coke++ | / (8 files):
rename "rakudo" to "rakudo.parrot"
 03:43
[Coke] and now "rakudo" isn't special, it's rakudo.parrot and rakudo.jvm in the reports. 03:44
dalek rl6-roast-data: 86b4833 | coke++ | README (2 files):
Change readme to .md
 03:53
rl6-roast-data: 56f0ed3 | coke++ | README.md:
add helper notes for rakudo.jvm/niecza
04:06 odoacre_ left, odoacre_ joined 04:29 FROGGS left 04:33 mtk left 04:36 colomon left, FROGGS joined
[Coke] .tell diakopter pvip isn't quite running perl6, from what I can tell. "./pvip -e "say 3"", e.g. outputs: 04:41
yoleaux [Coke]: I'll pass your message to diakopter.
[Coke] .tell diakopter (statements (funcall (ident "say") (args (int 3))))
yoleaux [Coke]: I'll pass your message to diakopter.
[Coke] .seen tokuhirom 04:43
yoleaux I saw tokuhirom 21 Jun 2013 08:45Z in #perl6: <tokuhirom> ?
tokuhirom_ ? 04:44
[Coke] .to tokuhirom how do you run a perl6 program in pvip?
yoleaux [Coke]: I'll pass your message to tokuhirom.
[Coke] ah, hi.
tokuhirom_ hi
It's just a parser and it prints AST in s-expression form.
04:44 craw joined
[Coke] ah, ok. and there's nothing yet to execute the s-expressions? 04:45
04:45 craw left
tokuhirom_ yes,... for now. 04:45
[Coke] ok. when they can, let me know, and I can add pvip to github.com/coke/perl6-roast-data
tokuhirom_ cool 04:46
(and i'm writing transpiler between perl6 to perl5 for next yapc-nasia
using pvip.
[Coke] .to diakopter ah, it's a parser for now - I'll add it to the daily runs when it also executes.
yoleaux [Coke]: I'll pass your message to diakopter.
04:46 mtk joined
[Coke] tokuhirom_: cool! 04:47
tokuhirom_ It passed only 776 test cases XD 04:48
TimToady wonders how many tests viv -5 would pass... 04:53
diakopter :)
yoleaux 04:41Z <[Coke]> diakopter: pvip isn't quite running perl6, from what I can tell. "./pvip -e "say 3"", e.g. outputs:
04:41Z <[Coke]> diakopter: (statements (funcall (ident "say") (args (int 3))))
04:47Z <[Coke]> diakopter: ah, it's a parser for now - I'll add it to the daily runs when it also executes.
diakopter TimToady: transpoiler? 04:54
tokuhirom_: I'm interested in writing a transpiler the *other* direction... 04:57
*interested in (someone else) writing...
tokuhirom_ My idea is, writing transpiler with some xs stuff like autobox... 04:58
tokuhirom_ afk 04:59
diakopter tokuhirom_: I need p5->p6
afk tew &
05:06 SamuraiJack_ joined
tokuhirom_ ah, right. 05:07
[Coke] zzz
05:33 abnorman left 06:10 zby_home joined 06:11 zby_home left 06:19 sqirrel joined 06:21 dayangkun_ joined, dayangkun left 06:26 atroxaper joined 06:32 dayangkun_ left 06:34 domidumont joined 06:36 domidumont left 06:49 kaleem joined 06:53 BenGoldberg left 06:57 domidumont joined 07:00 zakharyas joined 07:01 domidumont left 07:02 domidumont joined 07:07 fhelmberger joined 07:09 berekuk joined 07:12 sqirrel left 07:17 berekuk left 07:18 berekuk joined
lizmat good *, #perl6! 07:29
07:32 brrt joined 07:34 mstwu001 joined 07:40 pdcawley joined 07:43 daxim joined 07:52 berekuk left 07:58 mstwu001 left
FROGGS lizmat: hi 08:01
lizmat FROGGS o/ 08:02
08:03 dmol joined, wtw joined, kaleem left 08:04 rindolf joined 08:06 domidumont left, kaleem joined
hoelzro ahoy lizmat 08:12
08:15 fglock joined
fglock hi 08:15
moritz \o FROGGS, lizmat, hoelzro, fglock, * 08:21
FROGGS hi hoelzro, hi fglock 08:22
o/
:o)
08:28 salv0 joined 08:33 berekuk joined
hoelzro holy crap 08:34
I just started up an interactive perl6 with 2013.08
it got faaaaaast
congrats!
and good work =)
08:35 berekuk left
FROGGS hoelzro: and we expect that it won't get slower in future :o) 08:36
hoelzro \o/
08:39 pmurias joined
pmurias fglock: hi 08:39
08:42 dakkar joined
fglock hi pmurias 08:47
09:11 sqirrel joined
timotimo lue: the multi-angle-brackets is one of the first things the section on formatting codes mentions 09:15
"All Pod formatting codes consist of a single capital letter followed immediately by a set of angle brackets. The brackets contain the text or data to which the formatting code applies. You can use a set of single angles (<...>), a set of double angles («...»), or multiple single-angles (<<<...>>>).
daxim ‹›〈〉 09:18
09:23 berekuk joined 09:30 berekuk left
arnsholt I really like the French name for double angles. Guillemets. ^_^ 09:36
yoleaux 26 Aug 2013 02:36Z <diakopter> arnsholt: dyncall is now compiling in the dyncall2 branch.. want to stub in the dyncall reprs and ops?
09:42 berekuk joined
grondilu I think "guillemet" is rather the name for double quotes (") 09:46
grondilu is french fwiw 09:48
arnsholt It is? I've always heard it used about the double-angle things
grondilu that's a common mistake, iirc
daxim it means something different in english (and german, and ………) 09:49
arnsholt Oh, neat. French Wikipedia calls both double quotes and double angles guillemets
grondilu I'm totally wrong. Sorry. fr.wikipedia.org/wiki/Guillemet 09:50
arnsholt " is guillemet anglais, and « is guillemet français
grondilu indeed, I was the one making the common mistake :)
arnsholt Dactylographie! ^_^
arnsholt learnt a new word
hoelzro lizmat: what's the schedule for the upcoming workshop?
hoelzro guesses fingerpainting
arnsholt hoelzro: Typing, actually =) 09:51
hoelzro close enough =)
timotimo fingertyping or typingpainting?
10:11 berekuk left 10:14 sqirrel left 10:15 colomon joined 10:18 berekuk joined 10:22 berekuk left
lizmat hoelzro: you mean more schedule than there was in the mail on perl6-compiler or in my blog post ? 10:23
hoelzro lizmat: I mean what time does it start/end
just so I have an idea when I should fly into/out of Frankfurt
lizmat ah, good question
well, I would think it best if you arrive Friday late afternoon, early evening 10:24
so we can start around 9am
accommodation has been booked for 3 nights
FROGGS uhh, it is still not september yet :o( 10:25
10:25 berekuk joined
dalek kudo/nom: 90cfaa0 | (Elizabeth Mattijsen)++ | src/core/ (2 files):
Make sure Bag|Set.delete don't exist (Any.delete vivified them)
 10:25
timotimo oh, i had booked the accomodation for saturday and sunday, so it'll be a good thing if i can get into the hotel lizmat++ booked 10:26
colomon lizmat++
10:26 berekuk left
hoelzro lizmat: alright, sounds good 10:27
should I expect to leave Monday morning, then?
lizmat accommodation has been booked for checkout on Mon, so you have that option 10:28
hoelzro what time to you think we'll be done on Sunday? 10:29
tadzik I leave on monday too
lizmat good question: someone commented on the program that it was a good program with only one problem in it
the problem could be fixed by s/Day/Week/
so, I think we will finish rather later than earlier 10:30
FROGGS what problem? 10:31
hoelzro ok, I figured as much
just trying to get my flight details figured out =)
10:32 berekuk joined 10:33 markov joined
diakopter hunh; nodejs co-opted /usr/bin/node on ubuntu 10:37
10:39 brrt left
lizmat r: my %h; say \%h.WHAT; say \(%h).WHAT # expecting (Capture) (Capture) 10:40
camelia rakudo 789db4: OUTPUT«().hash␤(Capture)␤»
colomon ().hash ??
timotimo weird.
colomon oh!
tadzik precedence?
colomon that's what I'm thinking
timotimo so it's (\%h).WHAT? 10:41
colomon r: my %h; say (\%h).WHAT
camelia rakudo 789db4: OUTPUT«(Capture)␤»
timotimo r: my %h; say \(%h.WHAT)
colomon yeah, precedence
camelia rakudo 789db4: OUTPUT«(Hash)␤»
lizmat guess I need to fix some tests then 10:42
moritz fwiw I don't seem the point of having prefix \ anymore 10:43
most of what prefix \ was meant to do is now item context
and I haven't had a single use case for prefix \ (except in signatures, but that's something else) 10:44
lizmat moritz: this is about S03:195, right ? 10:47
synopsebot Link: perlcabal.org/syn/S03.html#line_195
diakopter r: say \\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\9
camelia rakudo 789db4: OUTPUT«9␤»
lizmat is just trying to make sense of t/spec/S02-types/autovivification.t 10:48
moritz lizmat: yes 10:49
lizmat: and maybe it's in fact useful, and I just haven't had that use case myself yet
10:50 sqirrel joined
moritz lizmat: you mean the "capturing autovivifies" test? 10:50
I'd nuke that
lizmat yes
ok, consider it nuked 10:51
moritz the spec is pretty clear that autovivication should only happen on write accesses
and creating a capture isn't a write access on the original container
lizmat well, then we need to test it *doesn't* auto-vivify
moritz +1
lizmat++ # going through the tests 10:53
pmurias diakopter: re /usr/bin/node that's were node.js is supposed to be? 11:08
diakopter not always, I thought
pmurias not on debian 11:10
but that's silly
moritz isn't that what $PATH is for?
pmurias no 11:11
;)
11:12 benabik left 11:14 brrt joined, brrt left 11:16 pernatiy left 11:17 rindolf left 11:18 hummeleBop left 11:19 hummeleBop joined
colomon module smoke test still broken…. 11:22
FROGGS so, you cant load modules at all, right?
or was it something panda-related? 11:23
colomon I think the problem is that Panda can't load modules.
it's dying during pandabrew
dalek ast: 478c26a | (Elizabeth Mattijsen)++ | S02-types/autovivification.t:
Add some more auto-vivification tests
colomon ===SORRY!=== 11:24
Could not find Shell::Command in any of: /home/smoker/rakudo/install/lib/parrot/5.5.0-devel/languages/perl6/site/lib, /home/smoker/rakudo/install/lib/parrot/5.5.0-devel/languages/perl6/vendor/lib, /home/smoker/rakudo/install/lib/parrot/5.5.0-devel/languages/perl6/lib, /home/smoker/.perl6/2013.08-15-g789db4f/lib
11:27 atroxaper left
colomon yeah, that's what happens when you try to bootstrap panda 11:29
has the ENV stuff been cleaned up, or is that still the likely culprit 11:30
lizmat I see one ENV test still failing on parakudo 11:32
11:34 sqirrel left
FROGGS env is not working right yet AFAIK 11:35
11:39 lue left 11:41 berekuk left 11:42 berekuk joined
grondilu was trying to write a grammar for Lambda-calculus, and encountered a behavior he can't explain: gist.github.com/grondilu/6352516 11:45
shouldn't rule { <foo> | <bar> } always be the same as rule { <bar> | <foo> }? 11:46
GlitchMr grondilu: I don't think so. What if <foo> is ('a'), and <bar> is ('a'). 11:48
11:49 mtj_ joined
grondilu well, I mean at least they should either both match, or both fail. Not one matching and not the other one. 11:49
'|' is not supposed to have a sequential semantics, is it? 11:50
FROGGS as I understand it, if on branch of an alternation started to match successfully, it stays in that branch 11:51
and I think you can let it try again using backtracking
grondilu: || is sequential
GlitchMr On sidenote, I tried your code using Peg.js, and I also get fail when I swap application and variable.
gist.github.com/GlitchMr/6352572 11:52
pegjs.majda.cz/online
FROGGS s/on branch/one branch/
11:53 lue joined
GlitchMr It fails with: "Line 1, column 5: Expected [0-9] but " " found." 11:53
dalek rlito: fd628c9 | (Flavio S. Glock)++ | / (2 files):
Perlito5 - perl5 - use simple eval()
 11:55
GlitchMr Considering Perl 6 rules are PEG grammars (at least I think so), it fails. 11:56
moritz p6 rules are partially modeled after PEG
but they aren't the same
FROGGS grondilu: can you try that please? gist.github.com/FROGGS/6352618 11:58
moritz [Coke], FROGGS, tadzik, pmurias: the failure of panda to bootstrap seems to be caused by the broken mechanism to pass %*ENV to subprocesses 12:00
FROGGS okay
moritz panda modifies %*ENV<PERL6LIB> but the subshell it starts doesn't see it
grondilu r: gist.github.com/FROGGS/6352618
camelia rakudo 90cfaa: OUTPUT«#<failed match>␤»
GlitchMr if (result0 === null) {
result0 = parse_var();
if (result0 === null) {
result0 = parse_appl();
}
}
I assume something like that happens. 12:01
moritz grondilu: note that 'application' is a rule, not a token, so its whitespace terminates LTM 12:03
grondilu yeah, I should have defined it as a token, I guess. But it still fails when variable precedes application in the alternation of expression. 12:05
FROGGS now we'd need Grammar::Debugger and a working panda :/ 12:06
moritz just use the last star release, it has both.
FROGGS true 12:10
(that was an you-we fwiw :o)
12:11 skids left 12:13 pernatiy joined 12:17 berekuk left 12:20 dmol left 12:22 dmol joined, berekuk joined 12:24 berekuk left 12:25 berekuk joined
pmurias fglock: what's the current plan with Perlito? 12:26
fglock: I have noticed there's a new XS backend
12:30 Woodi joined
dalek rlito: b38b950 | (Flavio S. Glock)++ | / (2 files):
Perlito5 - perl5 - use less code
 12:31
fglock pmurias: I'm looking at ways to add a vm 12:32
just experimenting some ideas
pmurias add a vm? 12:33
fglock compiling to low-level code - machine code, C, or jvm 12:34
12:34 kbaker joined 12:35 domidumont joined
FROGGS in a year I might suggest moarvm bytecode 12:38
but I dunno what to suggest now
12:39 jnap joined
dalek kudo/nom: 8a6a949 | (Elizabeth Mattijsen)++ | src/core/Any.pm:
Make %h<a><b> := "foo" work; alas, %h<a><b> := my $b doesn't yet
 12:40
kudo/nom: 1145e1c | (Elizabeth Mattijsen)++ | src/core/Any.pm:
Fix error message fail
12:40 sqirrel joined
lizmat fglock: MoarVM *is* intended to have all the features needed to run Perl 6 12:41
OTOH, it is also based on NQP, which Perlito of course is not
so there will be a huge API mismatch, I would think
JimmyZ lizmat: based on NQP? 12:42
I don't think so 12:43
12:43 berekuk left
moritz it's made for nqp and rakudo, but not based on it 12:43
FROGGS it is better to be based on nqp when you want to target moarvm, I guess that is what lizmat++ meant
12:44 berekuk joined
JimmyZ If you have an ast, then you don't need NQP at all 12:44
FROGGS fglock: maybe I can get an Perl5 module working that accepts your bytecode so I can run it under v5 on rakudo/nqp :P 12:45
pmurias I doubt creating QAST in perl5 code would be a very big problem 12:48
JimmyZ I'm not saying QAST :) 12:49
lizmat is glad to be proven wrong
12:51 TimK1 joined
pmurias but targeting QAST would be much more similiar to targeting a high level language then a lowlevel VM 12:52
FROGGS lizmat: you could dome to that workshop *g* 12:53
come* 12:54
lizmat you know, I just might :-)
12:54 rindolf joined
JimmyZ Targeting QAST could get a compiling time pain 12:55
;)
FROGGS that is the current problem, right? 12:56
TEST=42 perl6 -e 'say %*ENV<TEST>; %*ENV<TEST> = 1; say %*ENV<TEST>; shell(q{perl6 -e "say %*ENV<TEST>"})' # 42 1 42
JimmyZ that's why niecza was borned
pmurias FROGGS: yes, I should get back to working on the windows version of shell3 12:59
FROGGS: how much work is it to get nqp to compile on windows? 13:01
FROGGS pmurias: that was linux fwiw 13:04
pmurias: it is quite a bit of work to everything in place (activeperl, msvc express, git) and then it compiles painfully slow 13:06
but I must say that my experience is with winxp only, and there it took like 5 hours to compile parrot, nqp and rakudo
it might (should) be better on windows 7
pmurias doesn't seem like much fun
FROGGS true, so my offer still stands 13:07
and btw, I'm going to debug it myself now, maybe this is quicker since I an test uncommitted stuff
13:08 jaldhar left
FROGGS ohh, my linux box does not have the shell3 branch... 13:08
FROGGS rebuilds now
pmurias FROGGS: re debug, I have figured out what's the problem
FROGGS cool 13:09
do tell
pmurias the problem with the shell3 branch on windows is that I'm passing the env in a wrong format 13:11
the winapi expects a big buffer with null ended strings ending with a null, rather then a null ending array of pointers to null ending strings 13:12
FROGGS ahh, okay
dalek kudo-js: 16d3c4f | (Pawel Murias)++ | / (2 files):
Add missing file.
kudo-js: 18f03c2 | (Pawel Murias)++ | / (4 files):
Pass a processed version of nqp/t/qregex/rx_basic.
lizmat r: sub f(Int $x) {}; say f(Nil) # would expect (Int) rather than a bomb, so is this a rakudobug ?
camelia rakudo 90cfaa: OUTPUT«===SORRY!===␤CHECK FAILED:␤Calling 'f' will never work with argument types (Nil) (lines 1, 1)␤ Expected: :(Int $x)␤»
lizmat r: sub f(Int $x) {}; my $a is default(Nil); say f($a) # like this 13:13
camelia rakudo 90cfaa: OUTPUT«Nominal type check failed for parameter '$x'; expected Int but got Any instead␤ in sub f at /tmp/ttbGCeDhvf:1␤ in block at /tmp/ttbGCeDhvf:1␤␤»
lizmat huh?
FROGGS O.o
lizmat r: sub f(Int $x) {}; my $a is default(Nil); say $a; say f($a) # like this 13:14
FROGGS r: sub f(Int $x) {}; my $a is default(Nil); say $a
camelia rakudo 90cfaa: OUTPUT«(Any)␤Nominal type check failed for parameter '$x'; expected Int but got Any instead␤ in sub f at /tmp/HsCskdrdSC:1␤ in block at /tmp/HsCskdrdSC:1␤␤»
rakudo 90cfaa: OUTPUT«(Any)␤»
13:14 BenGoldberg joined
BenGoldberg Good morning 13:14
FROGGS lizmat: it looks like you can't have Nil as default
[Coke] lizmat++
13:14 ajr joined
FROGGS hi BenGoldberg 13:14
lizmat r: say $/.VAR.default 13:15
camelia rakudo 1145e1: OUTPUT«Nil␤»
lizmat yes you can
13:15 ajr is now known as Guest88001
[Coke] r: $/.VAR.default = 3; 13:15
camelia rakudo 1145e1: OUTPUT«Cannot modify an immutable value␤ in block at /tmp/BXyXqSLNFz:1␤␤»
13:15 PacoAir joined
lizmat r: my $a is default(Nil); say $a; $a=Nil; say $a # something wrong with initialization 13:16
camelia rakudo 1145e1: OUTPUT«(Any)␤Nil␤»
FROGGS r: my $a is default(Nil); say $a.VAR.default
camelia rakudo 1145e1: OUTPUT«Nil␤»
lizmat r: say Nil ~~ Scalar 13:17
camelia rakudo 1145e1: OUTPUT«False␤»
lizmat the problem
13:19 Guest88001 is now known as ajr_
moritz why is that (a|the) problem? 13:21
r: say 42 ~~ Scalar 13:22
camelia rakudo 1145e1: OUTPUT«False␤»
13:22 iSlug joined
moritz does that mean 42 isn't a valid default value either? 13:22
lizmat in src/core/Variable: $var = $default if nqp::istype($what, Scalar)
grrr… thinko 13:23
moritz hopes the $what there is the scalar container, not the default 13:24
lizmat the $what is the type of container 13:25
FROGGS pmurias: is that the expected behaviour on linux? TEST=42 perl6 -e 'say %*ENV<TEST>; %*ENV<TEST> = 1; say %*ENV<TEST>; shell(q{perl6 -e "say %*ENV<TEST>"})' # 42 1 42 13:28
I'd expect 42 1 1
[Coke] lizmat: is there a #channel for the workshop?
lizmat eh, no
do you think we need one ?
[Coke] Iunno. 13:29
FROGGS: your expectation matches mine, and the test suite (which is still failing that test, IIRC)
FROGGS [Coke]: k, thanks
13:31 berekuk left 13:32 lowpro30 joined 13:33 sqirrel left
TimK1 p6: say (sub (--> Array[Int]) {return my Int @a = 3, 4, 5;})(); 13:38
camelia rakudo 1145e1: OUTPUT«Type check failed for return value; expected 'Array[Int]' but got 'Array[Int]'␤ in sub at /tmp/UyRhPeA_1Y:1␤ in block at /tmp/UyRhPeA_1Y:1␤␤»
..niecza v24-88-g1f87209: OUTPUT«Potential difficulties:␤ @a is declared but not used at /tmp/RtuJskSR3v line 1:␤------> ay (sub (--> Array[Int]) {return my Int ⏏@a = 3, 4, 5;})();␤␤3 4 5␤»
lizmat known probkem
dalek kudo/nom: 386a090 | (Elizabeth Mattijsen)++ | src/core/Variable.pm:
It's always in the timing of things: fix "is default(Nil)"
 13:39
TimK1 Yeah, thought so. 13:40
tadzik I read that as "krown problem". I'm now trying to think of a Norway joke
"Did you know Norway doesn't use Euro?" "Yeah, it's a krown problem" 13:41
13:41 lue left
FROGGS :o) 13:41
13:43 skids joined
lizmat refrains from running up the Preikestolen and jumping off 13:43
13:43 btyler joined
arnsholt Hee hee! \D 13:44
13:45 SmokeMac_ joined 13:47 SmokeMachine left, berekuk joined
lizmat TimToady: S32/Containers:1036 mentions methods "from" and "to" for Range 13:48
there are some tests for it, indicating they're really the same as min/max
13:48 kbaker left
lizmat but they're currently not implemented 13:49
Q: should they be implemented, or should "from" and "to" be removd from the spec
I would suggest eradicating from the spec, as they don't have any additional value to "min" and "max" 13:50
pmurias FROGGS: are you working on the shell3 branch? (I'm planning to work on it now) 13:51
13:51 kaleem left, kbaker joined
FROGGS pmurias: just started to print a bit debug output... 13:52
(on linux)
13:52 colomon left, lue joined 13:53 lowpro30_ joined 13:56 lowpro30 left
pmurias FROGGS: on linux when using the shell3 branch you would need to remove the #ifdef'ed for parrot code in run 13:57
dalek rl6-roast-data: e264650 | coke++ | README.md:
more info
 14:00
[Coke] pmurias: ah, the child env stuff is in a branch? 14:01
I thought I saw commits go by earlier, that would explain why it didn't work in mainline, though. 14:02
pmurias [Coke]: the child env stuff only works on linux atm, and it needs some changes to rakudo 14:03
lizmat r: my num $a = 1; $a = Nil # should this fail, as some tests seem to indicate, or just set to 0 like it does now;
camelia rakudo 1145e1: ( no output )
lizmat r: my num $a = 1; say $a; $a = Nil; say $a # should this fail, as some tests seem to indicate, or just set to 0 like it does now; 14:04
camelia rakudo 1145e1: OUTPUT«1␤0␤»
[Coke] it has been days since we had a clean run for any implementation, and that was pugs.
github.com/coke/perl6-roast-data/b...es.csv#L31
pmurias: whatever changes were made in nom are breaking the test that was working: 14:05
github.com/coke/perl6-roast-data/b....out#L3950
er, the change in nom there was the update to the nqp revision.
geekosaur lizmat, it was recently re-specced to return variables to their defaults. tests are probably lagging 14:06
[Coke] anyone working on these failures:
S32-io/io-spec-unix.t aborted 1 test(s)
S32-io/io-spec-win.t aborted 1 test(s)
FROGGS pmurias: k
lizmat [Coke]: both are clean for me on parakudo on OSX 14:07
dalek ast: e4b9e51 | (Elizabeth Mattijsen)++ | S02-types/ (6 files):
Some more fudges / unfudges / skip -> todo changes
 14:08
14:14 benabik joined
lizmat geekosaur: figured as much 14:16
skids std: multi sub fn(buf8|blob8 $g) { 1; }; 14:18
camelia std 06636e4: OUTPUT«===SORRY!===␤Unable to parse signature at /tmp/Oa546ipsVM line 1:␤------> multi sub fn⏏(buf8|blob8 $g) { 1; };␤Couldn't find final ')'; gave up at /tmp/Oa546ipsVM line 1:␤------> multi sub fn(buf8|blob8 ⏏$g) { 1; };…
skids std: multi sub fn((buf8|blob8) $g) { 1; };
camelia std 06636e4: OUTPUT«===SORRY!===␤Unable to parse signature at /tmp/QuGDQr1P9u line 1:␤------> multi sub fn⏏((buf8|blob8) $g) { 1; };␤Couldn't find final ')'; gave up at /tmp/QuGDQr1P9u line 1:␤------> multi sub fn((buf8|blob8) ⏏$g) { 1…
14:25 berekuk left
FROGGS when I use binding it works? TEST=42 perl6 -e 'say %*ENV<TEST>; %*ENV<TEST> := 1; say %*ENV<TEST>; shell(q{perl6 -e "say %*ENV<TEST>"})' # 42 1 1 14:26
is that how it is meant to be?
14:28 rindolf left 14:31 sqirrel joined
FROGGS no, that is totally weird 14:33
when I do the thing above using the unpatched rakudo, I get 42 42 42 14:34
r: say %*ENV<TERM>
camelia rakudo 386a09: OUTPUT«screen␤»
pmurias FROGGS: on the shell3 brnach with patched rakudo?
FROGGS on shell3 with unpatched rakudo
r: say %*ENV<TERM>; %*ENV<TERM> = 1; say %*ENV<TERM>
camelia rakudo 386a09: OUTPUT«screen␤1␤» 14:35
FROGGS r: say %*ENV<TERM>; %*ENV<TERM> := 1; say %*ENV<TERM>
camelia rakudo 386a09: OUTPUT«screen␤screen␤»
FROGGS see
that is what I call "weird"
pmurias it's not supposed to do anything sane currently 14:36
FROGGS star: say %*ENV<TERM>; %*ENV<TERM> = 1; say %*ENV<TERM>
camelia star 2013.02: OUTPUT«screen␤1␤»
FROGGS star: say %*ENV<TERM>; %*ENV<TERM> := 1; say %*ENV<TERM> 14:37
camelia star 2013.02: OUTPUT«screen␤screen␤»
pmurias so it's a waste of time examining an unpatched rakudo which expects a crazy parrot pmc from getenvhash and gets a hash
FROGGS well, I just wondered what programs like panda and our testsuite depend on 14:38
14:38 bluescreen10 joined 14:40 colomon joined
FROGGS @all: should assignment and binding to %*ENV<something> should do different things? 14:40
lizmat / moritz / masak / jnthn ? 14:41
moritz either one should die, or both should do the same
lizmat FROGGS: I'm not sure anymore, TimToady seemed to be in two minds about that 14:42
FROGGS moritz: I am thinking of %*ENV of just a regular hash, with no sideeffects (accept that it is being passed to subprocesses via shell())
14:43 berekuk joined
FROGGS okay, so I think I know now what I want 14:43
lizmat and %?ENV would be the environment under which compiled 14:44
and PROCESS::<%ENV> would be the immutable ENV of the process ?
14:44 berekuk left
dalek p/shell3: 257f172 | (Pawel Murias)++ | src/vm/parrot/ops/nqp.ops:
Attempt at corretly passing the env variables on windows.
 14:45
pmurias FROGGS: ^^ could you try this commit?
FROGGS sure
timotimo oh, interesting idea to have the environment available separately that was in effect when the module was compiled
FROGGS yeah, I like %?ENV, but I no opinion on PROCESS::<%ENV> 14:47
timotimo is the COMPILING pseudopackage for that? i suppose that's for while it's being compiled rather than being kept around after compilation has finished
14:51 gdey joined
pmurias FROGGS: try = run the nopaste 14:53
14:53 markov left
FROGGS will do when it finished compiling 14:54
which takes about eight minutes (for rakudo)
lizmat is there any reason to keep PIR code around, if there is an NQP equivalent that also works under Parrot ?
TimToady lizmat: from/to are fossils from when we allowed :by(-1), and the from/to could be reversed from min/max, so from/to should be removed
dalek p/shell3: dd3c146 | (Pawel Murias)++ | src/vm/parrot/ops/nqp.ops:
Use real \0
lizmat TimToady: reading the chainsaw
readying
:-)
14:55 kaare_ joined
pmurias FROGGS: the new commit is better 14:55
FROGGS: don't build rakudo, just build nqp
FROGGS: the %ENV won't work for rakudo unless the #ifdef'ed code is removed 14:56
[Coke] lizmat: oh, I get many more failures on OSX - those fail for me on feather.
14:56 markov joined
[Coke] er... that's not feather, that's diakopter's test machine (still linux) 14:56
dalek ecs: bdec1f7 | (Elizabeth Mattijsen)++ | S32-setting-library/Containers.pod:
Bring the spec of Range in sync with the implementation
pmurias lizmat: re PIR code, where is it?
lizmat src/core/Range.pm 14:57
spectesting its removal now
colomon shouldn't we decide whether we're going to use - or _ to separate words in core method names, and stick with it?
FROGGS pmurias: I removed the #ifdef fwiw, but testing nqp only is better of course 14:58
lizmat [Coke]: must have done at least 5 spectest today already, no unexpected failures 14:59
*spectests
dalek ecs: 850de3f | (Konrad Borowski)++ | S32-setting-library/Containers.pod:
bounds returns Parcel, not List
 15:00
TimToady PROCESS::<%ENV> is immutable; if an interpreter started under, say, apache, wants to modify %ENV globally, it should copy it into GLOBAL so that it is not shared by other interpreters, or it should do 'temp %*ENV;' around any scope it wishes to have a mutable environment
a temp at the top of the mainline code would be roughly equivalent to copying into GLOBAL 15:01
lizmat GlitchMr: I was just changing the implementation of Range.bounds to be a List :-(
pmurias FROGGS: pastie.org/8270977 - the debugging pastie
GlitchMr sorry
I'm going to revert it.
TimToady I suppose if we're tracking CWD separately, we have to merge that into ENV at the last moment before spawning 15:02
lizmat GlitchMr: I'm not saying that is needed
I'm just wondering what would be the best: Range.bounds to return a Parcel, or a List
15:02 TimK1 left
GlitchMr Well, I think Parcel makes sense. It's just pair of values. 15:03
(but not actual pair)
lizmat yeah, but .bounds.key wouldn't make sense, .bounds[0], would, no?
15:03 berekuk joined
grondilu rn: say 42.REPR 15:04
camelia niecza v24-88-g1f87209: OUTPUT«Unhandled exception: Unable to resolve method REPR in type Int␤ at /tmp/mCnNDx5cl1 line 1 (mainline @ 3) ␤ at /home/p6eval/niecza/lib/CORE.setting line 4579 (ANON @ 3) ␤ at /home/p6eval/niecza/lib/CORE.setting line 4580 (module-CORE @ 576) ␤ at /home/p…
..rakudo 386a09: OUTPUT«P6opaque␤»
grondilu what's REPR exactly? Is it rakudo specific?
GlitchMr Well, for me, Parcel makes sense. 15:05
It doesn't make sense, to push element to bounds, for example.
15:05 hummeleBop left
lizmat r: say (1,2).Parcel[0] 15:06
camelia rakudo 386a09: OUTPUT«1␤»
GlitchMr List, I think, is for lists - and in lists, the ordering doesn't matter much.
order*
FROGGS pmurias: gist.github.com/FROGGS/393b2db32957ba103f97
lizmat fair enough, Parcel it is then
GlitchMr But with .bounds, order doesn't matter.
does*
lizmat assuming TimToady doesn't mind
TimToady order matters in lists too
GlitchMr Yes, but if I sort the list, nothing bad should happen. 15:07
TimToady in this case either parcel or list is fine; parcel is probably less overhead
lizmat parcel it is then
GlitchMr++ for changing the spec
GlitchMr It makes sense to sort list. It doesn't make sense to sort a tuple. 15:08
TimToady the bounds should always come out min,max, and should not be sorted
since the semantics change if min > max
say (1..0).bounds
nr: say (1..0).bounds 15:09
camelia rakudo 386a09, niecza v24-88-g1f87209: OUTPUT«1 0␤»
TimToady 1..0 is very different from 0..1
GlitchMr Well, for me 1..0 looks like empty list, but I assume there are uses for such ranges.
dalek kudo/nom: 956aa09 | (Elizabeth Mattijsen)++ | src/core/Range.pm:
Eradicate .PIR code for range reification

There is no point keeping it around if we have an NQP version around that also works on Parrot. And which in the future will be better optimisable.
TimToady I just don't understand why we're talking about sorting, but then, I haven't had any coffee yet... 15:10
GlitchMr My argument is that if you randomly change positions in Parcel (containing phone number parts, for instance), something bad would happen. 15:11
colomon TimToady: Is that why I'm so frequently confused here? No coffee at all?
GlitchMr: please do not randomly change positions in my Lists…. 15:12
TimToady yes, parcels are more like records that way, and lists tend to be homogenous, but those are both oversimplifications, I suspect
as soon as the rest of a parcel is bound to a slurpy, it's essentially a list instead 15:13
FROGGS pmurias: I can confirm that your test script works well on linux and windows 15:14
TimToady but of course positionals care greatly about, er, position
15:14 btyler left
fglock re targetting QAST - I think it would be easier to start targetting Perl6 (translate perl5 to perl6) 15:14
when it gets running in Perl6, QAST can be generated using native libraries 15:15
15:15 kaleem joined
fglock same thing with JVM, targetting a JVM language first (maybe java, but javascript would do) 15:16
dalek ast: 2713e65 | (Elizabeth Mattijsen)++ | S02-types/range.t:
Remove tests for Range.from and Range.to
 15:17
lizmat cycling and dinner&
fglock FROGGS: what is the state of v5? 15:18
FROGGS fglock: 2500 passing tests
(of >40000)
github.com/rakudo-p5/v5/blob/master/STATUS.md
it is hard to describe what is missing and what is there 15:19
__DATA__ pseudo filehandles are missing, typeglobs are missing
diakopter ;) No such symbol '&DynaLoader::boot_DynaLoader'
FROGGS but you can use/require modules and get their symbols exported for example
P5 regexes needs love, and many built-ins are NYI 15:20
diakopter: :o)
fglock: indirect is sort of there, but only to a small degree 15:21
indirect object syntax*
15:24 btyler joined
skids So buf* are now immutable and blob* are immutable versions of buf* ?... ??... ??? 15:27
FROGGS r: temp %*ENV; say "alive"
camelia rakudo 386a09: ( no output )
benabik Yes. Although when you put it that way I would somewhat expect the names to be reversed. </bikeshedding>
FROGGS -.-
benabik FROGGS: Impressive. 15:28
dalek p/shell3: 432b962 | (Pawel Murias)++ | t/nqp/78-shell.t:
Test that passing env variable to child processes works correctly.
FROGGS benabik: I had another word in mind :o)
pmurias FROGGS: I added a proper test for the env variables 15:29
benabik Temp is horribly borken, I guess.
FROGGS pmurias: I'll test it on both machines...
benabik Hm. Temp in mainline is broken? 15:30
r: { temp %*ENV; say 1; }; say 2
camelia rakudo 386a09: OUTPUT«1␤2␤»
FROGGS pmurias++ # passes on both
benabik r: temp %*ENV; say 1; say 2
camelia rakudo 386a09: ( no output )
FROGGS ohh
benabik ohh? 15:32
FROGGS nr: my $a = 42; { temp $a = 22; say $a }; say $a
camelia rakudo 386a09, niecza v24-88-g1f87209: OUTPUT«22␤42␤»
FROGGS nr: my $a = 42; temp $a = 22; say $a; say $a
camelia niecza v24-88-g1f87209: OUTPUT«22␤22␤»
..rakudo 386a09: ( no output )
FROGGS okay, has nothing todo with dynamic vars
benabik Yeah. 15:33
FROGGS ohh meant "that is surprising"
benabik Drat, hoped it was "ohh, that's the problem"
FROGGS no, sadly not 15:34
15:34 bluescreen10 left
fglock FROGGS: thanks for the link, I'm taking a look at the code :) 15:35
15:35 fglock left
FROGGS have fun 15:38
TimToady buf is mutable, blob is immutable 15:39
diakopter <- is a buf today 15:41
tadzik .tr pl en bufon 15:43
yoleaux tadzik: Sorry, that command (.tr) crashed.
tadzik yolo
FROGGS .tr tr tr tr 15:44
yoleaux tr tr tr (en → en)
dalek p/shell3: d4257cd | (Pawel Murias)++ | src/vm/parrot/ops/nqp.ops:
Avoid leaking memory with shell3 on windows.
pmurias FROGGS: we now need to update rakudo
FROGGS w/e
tadzik .tr pl en bufon
yoleaux tadzik: Sorry, that command (.tr) crashed.
FROGGS pmurias: yeah
tadzik your mom crashed
FROGGS *g*
diakopter tadzik: buffoon? 15:45
tadzik yeah, something like this
dalek p: 3adbc0a | (Pawel Murias)++ | src/vm/parrot/ops/nqp.ops:
Make nqp::shell respect the 3 argument on linux.
 15:48
p: 6cbe2ef | (Pawel Murias)++ | src/vm/parrot/ops/nqp.ops:
Pass the env variables on windows.
p: 257f172 | (Pawel Murias)++ | src/vm/parrot/ops/nqp.ops:
Attempt at corretly passing the env variables on windows.
p: dd3c146 | (Pawel Murias)++ | src/vm/parrot/ops/nqp.ops:
Use real \0
nqp: 432b962 | (Pawel Murias)++ | t/nqp/78-shell.t:
nqp: Test that passing env variable to child processes works correctly. 15:49
pmurias who has a rakudo commit bit?
15:49 nebuchadnezzar left, zakharyas left
FROGGS I have 15:49
tadzik I do, for example
FROGGS off to dinner though 15:50
colomon have commit bit, just hanging around the kitchen working on dishes and playing flute
skids OK, so rakudo's new "feature" of "Buf is now a role, and Buf objects are immutable" should read "we had to make Buf-based objects immutable for a while, we'll get back to it"? 15:51
15:53 markov left 15:55 nebuchadnezzar joined, kaleem left
FROGGS pmurias: I think this does it for now: gist.github.com/FROGGS/97be80800999006a1263 15:56
15:56 SamuraiJack_ left
FROGGS pmurias: after that we can care about making %*ENV constant, and fixing the temp bug 15:56
timotimo skids: i think it means if an object just provides the Buf role, it'll be immutable, but it may offer mutability some other way? 16:00
benabik src/core/Buf.pm:200: # TODO: override at_pos so we get mutability 16:01
timotimo oh 16:02
FROGGS pmurias: hmmm, no, my patch is not ideal... 16:05
TimToady grondilu: rakudo currently does not implement the specced semantics regarding whitespace that have no token in front of them
FROGGS the rakudo spectest fails
16:08 markov joined
FROGGS it feels like it is loosing cwd 16:08
16:09 domidumont left 16:10 ajr_ left 16:11 sqirrel left
TimToady btw, in case it's not clear, the design principle behind the recent changes to ENV and CWD is that each dynamic scope should be able to pretend that it's a separate process 16:12
the limitations of JVM just pushed this over the edge of 'we need this right now'
FROGGS yeah, we are just trying to pass the current %*ENV to a shell() call 16:13
TimToady which I understand is problematic on parrot :) 16:14
TimToady wonders if there are other VMs with this "feature"
FROGGS the problem should be in rakudo now 16:15
skids r: my @to = Buf,buf8,Blob,blob8,Any,Mu; my $a = blob8.new(1); $a.WHAT.say; $a.isa($_).say for @to; $a = buf8.new(1); $a.WHAT.say; $a.isa($_).say for @to # this is not making updating Sum:: very easy :-)
camelia rakudo 956aa0: OUTPUT«(Blob)␤False␤False␤False␤False␤True␤True␤(Buf)␤False␤False␤False␤False␤True␤True␤»
benabik r: say Buf ~~ buf8.new(1) 16:18
camelia rakudo 956aa0: OUTPUT«False␤»
benabik r: say buf8.new(1) ~~ Buf
camelia rakudo 956aa0: OUTPUT«True␤»
benabik Does isa not check rules?
roles
tadzik isa is about inheritance. .does checks roles 16:20
TimToady lizmat: there's no default for native int, but I don't see why num can't just default to NaN, since that's a native value. 16:22
timotimo oh, the pure-perl checksum library, skids? 16:23
skids Well, only pur perl until I think it's time to put in the NativeCall stuff.
timotimo i'd be interested to see how the jvm performs for that
skids Right now it mainly stands as an optimization challenge :-)
I haven't had the time to invest in compiling jakudo yet to see if it even works. 16:26
16:31 donaldh joined, pernatiy_ joined
[Coke] lizmat: the only reason to keep PIR at this point is for speed on parrot, and that should be pushed to nqp, not rakudo, IMO. 16:31
16:32 pernatiy left
pmurias FROGGS: re you patch if we use %*ENV instead of CALLER::DYNAMIC:<%*ENV> do the problems persist? 16:35
FROGGS pmurias: yes
pmurias and what's the problem? 16:36
FROGGS: I'll build a rakudo and check things 16:38
FROGGS pmurias: test 14 failes like this: Unable to open filehandle from path 'getout-22432.out': No such file or directory(2) 16:39
in block at t/spec/packages//Test/Util.pm:104
if the file is there, so I think it doesnt know about its Cwd anymore
pmurias on linux? 16:42
skids r: multi sub f(blob8 $f) { 8.say }; multi sub f(blob16 $f) { 16.say }; my $a = buf16.new(1); f($a); # \o/ I can work with that.
camelia rakudo 956aa0: OUTPUT«16␤»
FROGGS pmurias: yes
donaldh need a better value for $*EXECUTABLE_NAME on JVM. It's currenty 'java' which is not useful. 16:43
I propose we change the constant to 'perl6' for now.
16:44 gdey left
colomon donaldh: I ran into this with Niecza. 16:45
as far as I know, the real general use for $*EXECUTABLE_NAME is to be able to call it again.
donaldh It is perl6 when installed. Would help unblock panda.
16:46 spider-mario joined
colomon is rakudo-jvm actually installable now? 16:46
donaldh colomon: Yes that's how panda uses it.
colomon: Yes, rakudo-jvm has an install target. Panda doesn't work yet.
benabik Perhaps it should be `java MAGIC-OPS path/to/rakudo.jar` ;-) 16:47
*OPTS
donaldh benabik: rakudo installs a shell script called perl6 that runs rakudo-jvm
colomon donaldh: my point is, $*EXECUTABLE_NAME is a clumsy way to spwan another p6 process. that blocked panda on Niecza, too
donaldh colomon:sure. a spawn self type thing would be nice. 16:48
colomon and it seems to conflict with the classic use of executable-name, seeing what *actual* name the script running started with. Or is that a different variable in p6?
donaldh not sure it exists in p6 16:49
s/exists/is implemented/
16:50 dmol left, dmol1 joined
[Coke] colomon: if we don't require an executable wrapper, how would niecza, for example, declare how to run the executable it was declared with? 16:50
16:51 xenoterracide left
[Coke] ... wouldn't it end up looking a lot like $*EXECUTABLE_??? 16:51
benabik $*INTERPRETER.spawn($script, @args) ?
TimToady Perl 5 distinguishes $0 from $^X
[Coke] I think that's probably a reasonable thing. 16:52
colomon what benabik said, more or less
16:53 fhelmberger left
colomon ah, is $0 vs $^X $*PROGRAM_NAME vs $*EXECUTABLE_NAME in p6? 16:53
TimToady nodnod 16:54
decommuting from TO & # well, packing, anyway 16:55
rjbs (One of my favorite little changes in recent p5s is that $^X is, if possible, found via the OS API for getting the full path name. Mmmm.) 16:56
16:57 xenoterracide joined 16:59 Ayiko joined, dakkar left 17:00 daxim left
TimToady afkr & 17:04
17:06 virtualsue joined 17:14 rindolf joined, frdmn joined
dalek osystem: 4cb49e9 | duff++ | META.list:
Add sake
 17:15
[Coke] is pod6 far enough along that we can convert the rest of the syn? 17:16
PerlJam [Coke]: last I checked, the semantic blocks weren't supported yet. 17:17
17:18 pernatiy_ left 17:20 pmurias left
PerlJam (that would be "I think not" in answer to your question :) 17:20
17:23 donaldh left 17:24 hummeleBop joined 17:30 pmurias joined 17:34 SamuraiJack_ joined 17:37 ajr joined, ajr is now known as Guest26117, Guest26117 is now known as ajr_ 17:38 colomon left 17:41 pdcawley left 17:43 abnorman joined 17:47 btyler left 17:48 btyler joined 17:54 abnorman left 17:58 btyler left 18:00 btyler joined 18:04 donaldh joined 18:07 benabik left
lizmat TimToady: I'm not sure making NaN the default for "num", is very DWIM. It feels more like a WAT to me, especially since int defaults to 0 18:08
18:08 kbaker left
lizmat also: I think it would be rather hard to implement, as "num" doesn't have a container to keep the default value 18:08
18:09 stevan_ left
lizmat so this would need to live in the general assign logic at vm guts level 18:09
18:10 colomon joined
lizmat jnthn, FROGGS: is there an easy way to find out if a given type object is unboxed (aka "int", "num",etc ? 18:10
18:10 denis_boyun joined
lizmat r: my num $a is default(42) # need a more awesome error message here 18:10
camelia rakudo 956aa0: OUTPUT«===SORRY!=== Error while compiling /tmp/S7emHuDExR␤Cannot look up attributes in a type object␤at /tmp/S7emHuDExR:1␤------> ␤»
FROGGS r: my num $a; say say nqp::isnum($a); my Num $b; say nqp::isnum($b) 18:11
camelia rakudo 956aa0: OUTPUT«0␤True␤0␤»
FROGGS r: my num $a; say nqp::isnum($a); my Num $b; say nqp::isnum($b)
camelia rakudo 956aa0: OUTPUT«0␤0␤»
FROGGS r: my num $a; say nqp::objprimspec($a); my Num $b; say nqp::objprimspec($b) 18:12
camelia rakudo 956aa0: OUTPUT«0␤0␤»
FROGGS hmmm
18:13 donaldh left 18:14 donaldh joined 18:17 rindolf left
FROGGS lizmat: no idea, sorry 18:20
lizmat too bad, thanks for trying!
18:25 donaldh left
FROGGS pmurias: still there? 18:30
lizmat I think he will at least be back in ~ hour for the #parrotsketch weekly meeting 18:31
FROGGS k
t/spec/S02-magicals/env.rakudo.parrot ........................... ok 18:32
18:34 SamuraiJack_ left 18:35 stevan_ joined
pmurias FROGGS: pong 18:36
FROGGS pmurias: seems to work now, do you mind if I merge shell3 into master? 18:37
18:37 SamuraiJack_ joined
pmurias FROGGS: it's already merged 18:37
FROGGS awesome
pmurias what did you change to make it work? 18:38
18:39 abnorman joined
FROGGS pmurias: just this: gist.github.com/FROGGS/30ed8c0f93374de5a730 18:40
flussence I've noticed S26 refers to a Pod6:: naming scheme but Rakudo and Niecza both use Pod::* for those, so it probably makes sense to drop the "6". Any objections before I change it?
FROGGS but I think that this is exactly what we want
pmurias FROGGS: you kept the #ifdef parrot? 18:41
FROGGS no
PerlJam flussence: I kinda like Pod6 to distinguish it from Pod (for perl 5)
18:42 SmokeMac_ left
FROGGS pmurias: wait, there is an ifdef before that 18:42
flussence PerlJam: yeah, it's kinda awkward because there's a few mentions of Pod:: around in S26, and it also calls itself "Pod" and p5's "POD"...
FROGGS but it uses now nqp::shell instead of spawn
PerlJam: but this is Perl 6, so our Pod is Pod. when referring to other languages we can put something in front of it... (Perl5::Pod) 18:43
lizmat r: my &a is default(-> ($a) {$a}); say a(100) # wonder whether this should work or not 18:44
camelia rakudo 956aa0: OUTPUT«Not enough positional parameters passed; got 0 but expected 1 in sub-signature␤ in block at /tmp/vE4DmkSvEO:1␤␤»
pmurias NQP revision 2013.08-14-g3c2b6e3 required (currently 2013.08).
what the heck does that mean?
PerlJam FROGGS: true. I guess I just tend to think of p5-pod as "pod" and p6-pod as "pod6"
18:44 SmokeMachine joined
flussence pmurias: your nqp's 14 commits behind what rakudo wants 18:45
FROGGS pmurias: your have to reconfigure + make install
you*
pmurias in nqp?
FROGGS pmurias: yes 18:46
and then check nqp --version
pmurias FROGGS: the #ifdef code you changed in the diff should have been deleted
dalek ast: 87f477b | (Elizabeth Mattijsen)++ | S02-names/is_default.t:
Added tests for "is default(Nil)"
FROGGS PerlJam: I'd agree if we had other places where we say: hey, this is XXX6 and not the XXX from Perl 5 18:47
pmurias lizmat: I avoid the #parrotsketch meeting as I post a prereport and I don't care about the other parroty things
lizmat this will push rakudo.jvm lower again: on the brighter side, making .VAR work on rakudo.jvm will get a lot of tests passing again
pmurias visiting a friend&
flussence also all of roast uses Pod:: too; the spec looks like the odd one out here
18:48 pmurias left
FROGGS flussence: yeah 18:48
PerlJam flussence: If you also change the mentions of POD (as in Perl 5 POD) to say something like "Perl 5 POD", that would make me happier at least :)
flussence: anyway, +1 to dropping the 6 18:49
flussence I guess I can do both things 18:50
PerlJam oh, I just acked S26 and it looks like it already has "Perl 5 POD" in most places. 18:51
(and the places it doesn't, it's clear from the context that it's still talking about p5-pod) 18:52
18:52 grondilu left
flussence yep, looks like that's already done 18:53
dalek ecs: 6af26e3 | flussence++ | S26-documentation.pod:
s/Pod6::/Pod::/g to match Rakudo and Niecza
 18:54
rl6-roast-data: ee3c1da | coke++ | / (5 files):
today (automated commit)
 18:57
18:58 kbaker joined
Util #ps in 30m 19:00
19:00 krunen joined 19:02 konundra joined
FROGGS [Coke]: rakudo should be clean tomorrow 19:05
19:07 donaldh joined
colomon \o/ 19:08
[Coke] FROGGS: .parrot? yay! 19:11
FROGGS++
now I have to refix pugs and fudge niecza.
FROGGS [Coke]: yes, I'm just testing pmurias++ and my patch again
lizmat FROGGS: should it be clean now? 19:12
ah, ok
[Coke] pugs is going to drop off that list soon, I think. :(
19:15 grondilu joined 19:17 SamuraiJack_ left
colomon FROGGS: ah, that explains why it didn't work when I tried it just now. 19:18
dalek kudo/nom: 2d28fdd | (Elizabeth Mattijsen)++ | src/core/ (3 files):
X::Comp::Trait exception becomes X::Comp::Trait::Unknown
 19:28
19:30 dwarring left
grondilu upgrades his debian-sid system today and notices that it will bring perl 5.18 19:31
FROGGS cool
moritz has been using a perlbrew'ed perl 5.18 for quite some time now 19:32
[Coke] perlbrew++
dalek kudo/nom: bc7c2eb | (Tobias Leich)++ | src/core/terms.pm:
fetch/store to %*ENV directly

  ... instead of using the thing returned by nqp::getenvhash()		 19:36
kudo/nom: 17b17a4 | (Tobias Leich)++ | / (2 files):
use nqp::shell() instead of pir::spawn

Also bump NQP_REVISION, to pull in pmurias++ patch.
lizmat spectests 19:37
FROGGS ohh yeah, a mactest would be interesting 19:38
lizmat NQP rebuilds ok 19:40
FROGGS nqp t/nqp/78-shell.t # this passes? 19:41
lizmat all ok 19:42
FROGGS yay
19:42 SamuraiJack_ joined 19:45 BenGoldberg left
lizmat rakudo builds ok 19:46
19:47 jnap left 19:51 konundra left, colomon left 19:52 colomon joined
colomon looks like the module smoker is successfully running panda again! 19:52
moritz \o/
colomon and my screen blacked out and my cpu fan suddenly went off. should I be worried?
FROGGS \o/ 19:53
these are the best news!
ohh
colomon: yes
/o\
19:53 SamuraiJack_ left
colomon feels like the fan is back on but lower? 19:53
FROGGS I'd open the case (just in case) 19:54
lizmat spectest result: t/spec/S02-magicals/env.rakudo.parrot (Wstat: 0 Tests: 16 Failed: 0)
TODO passed: 15-16
FROGGS right 19:55
that's cool
lizmat will you unfudge or shall I ?
colomon I think maybe OS X killed killed a process that was out of control....
FROGGS I'm running jvm spectests now, and compile on windows xp to test there
lizmat: would be cool if you do it
dalek ast: ab66274 | (Elizabeth Mattijsen)++ | S02-magicals/env.t:
Unfudge tests because FROGGS++ and pmurias++ fixed the issue
 19:56
FROGGS cool 19:57
FROGGS is happy
colomon FROGGS++
19:58 abnorman left 20:02 abnorman joined
donaldh seen tadzik 20:10
lizmat .seen tadzik 20:13
yoleaux I saw tadzik 16:20Z in #perl6: <tadzik> isa is about inheritance. .does checks roles
20:14 benabik joined
donaldh thx lizmat++ 20:18
lizmat yw
donaldh chdir emulation in rakudo: github.com/rakudo/rakudo/pull/202
20:24 colomon left
masak waves 20:27
lizmat masak o/
20:30 kaare_ left
sjohnson yo 20:31
20:32 colomon joined 20:35 virtualsue_ joined 20:36 virtualsue left, virtualsue_ is now known as virtualsue
lue timotimo: I know, but the spec oddly leaves out "multiple double angles", mentioning only multiple single angles 20:41
20:44 colomon left 20:47 teslos joined 20:48 grondilu left 20:49 colomon joined, stevan_ left
dalek ecza: 9d2f552 | (Tobias Leich)++ | lib/CORE.setting:
properly return exit code from shell()
 20:50
tadzik .seen tadzik
yoleaux You're right there.
tadzik oh, right :0 20:51
20:51 kbaker left
donaldh tadzik: with the chdir emulation pull request ^^ panda is close to working on jvm 20:53
dalek nda: 7ea82d9 | (Donald Hunter)++ | lib/Panda/Common.pm:
Change Common.pm to use OS instead of VM.
nda: 6ec5264 | tadzik++ | lib/Panda/Common.pm:
Merge pull request #54 from donaldh/os-var

Change Common.pm to use $*OS instead of $*VM
ast: 0a714c2 | (Tobias Leich)++ | S29-context/exit.t:
unfudged passing tests
lizmat fwiw, please note that $*OS is not specced
tadzik donaldh: thank you :)
donaldh tadzik: panda bootstrap runs to the point where it fails with a parrot specific compile command 20:54
lizmat but I guess we can fix that later after a deprecation cycle
tadzik --target=pir?
donaldh we need to figure out how to package up compilation capability in the installed rakudo.
dalek ast: 07bd0ab | (David Warring)++ | S05-capture/caps.t:
more tests on regex conjunctive op (&&)
donaldh it might just be a case of different options: --target=jar 20:55
I haven't tried yet.
tadzik maybe 20:57
FROGGS donaldh: I'm testing #202 now
donaldh tadzik: Maybe should factor out the compile command into a backend specific impl
the target name will need to be different too, i.e xxx.jar
maybe it's as simple as s/pir/jar/ 20:59
FROGGS after .pir you invoke parrot to create a .pbc 21:00
tadzik FROGGS: which gives approx. 0 performance benefit 21:01
donaldh: yeah, that sounds sane
FROGGS: (been there, done that :))
FROGGS tadzik: are you sure?
okay
21:01 colomon left
tadzik or maybe I did something horribly wrong and have actually missed some huge win :) But istr it was more trouble than it was worth 21:02
FROGGS Perl6::ModuleLoader takes pir too?
tadzik yes 21:03
FROGGS hmmm, I'll test this then (tomorrow or so)
21:05 skids left
tadzik awesome 21:05
FROGGS because this would save about a minute to compile v5, and maybe a bit more for rakudo
tadzik oh? 21:07
ah, not creating pbc
21:07 ribasushi left, ribasushi joined
tadzik well, there is no reason why it shouldn't be faster in the end. It's just that for the modules the difference seemed marginable when I measured it 21:07
21:09 spider-mario left
donaldh Yay. panda bootstrap just worked with s/pir/jar/ \o/ 21:15
FROGGS cool! 21:16
donaldh I say worked. I mean ran to completion.
install is also hard-coded to do pir 21:17
FROGGS still sounds like progress :o) 21:23
21:27 colomon joined
donaldh Yes progress. 21:29
colomon module smoke test runs again! no new failures, one new (passing) module. perlpilot++ 21:32
donaldh FROGGS++ for testing #202. I missed that earlier 21:39
FROGGS :o) 21:40
donaldh 'night #perl6
FROGGS gnight donaldh
21:43 hummeleBop left 21:45 donaldh left 21:52 lowpro30 joined, virtualsue left 21:53 jnap joined 21:54 dmol1 left 21:55 lowpro30_ left 21:58 lowpro30 left
dalek p: 6f97a10 | (Tobias Leich)++ | src/vm/jvm/QAST/Compiler.nqp:
return an int as promised by b563f22e696761287243dd054a119f6334a1234d
 22:04
22:08 colomon left
jnthn evening, #perl6 22:10
FROGGS hi jnthn 22:11
jnthn oslo++ # awesome beer destination
FROGGS *g* 22:12
dalek kudo/nom: 3bdc0c3 | (Elizabeth Mattijsen)++ | src/core/ (2 files):
Make error message on 'my int $a is default(42)' etc. more friendly
lizmat jnthn o/
lizmat calling it a night 22:13
sleep&
FROGGS gnight lizmat
jnthn o/ lizmat :)
22:14 btyler left 22:15 dmol joined 22:16 btyler joined
FROGGS jnthn: should nqp::shell() should return posix style return values? thats is, should the return value be shifted +< 8 ? 22:18
I'm asking because nqp@parrot is doing that, nqp@jvm not
huf low byte is usually the signal number that killed the process, no? (and other various things, binary not found being one i recall) 22:19
FROGGS I'm talking about that: github.com/perl6/nqp/blob/master/s...p.ops#L395 22:20
huf: that might be the case, but it will be 0 in case of nqp::shell
22:20 colomon joined
jnthn FROGGS: Not sure, but if it's any help the general Perl 6 approach to things is to not throw away information. 22:21
FROGGS well, there are tests that check for an unshifted return value
huf hm. does it not shell out? how come it will always be zero? 22:22
FROGGS for nqp::shell we are faking information that is not there
huf: it returns the return value shifted 8 to the left, so the lower eight bits are 0
see: github.com/perl6/nqp/blob/master/s...p.ops#L395 22:23
huf right, i see that. it's just puzzling
FROGGS shell() from the parrot implementation just returned state, same goes for jvm
huf OH! 22:24
this is in an #ifdef WIN32, disregard me. 22:25
FROGGS ohh wait
now I see it too *g*
then it makes sense 22:26
22:26 PacoAir left
FROGGS but then I will apply that shifting to nqp::shell@jvm too 22:26
huf and the posix code just returns status
FROGGS right
huf arent preprocessor directives fun? :) 22:27
FROGGS they are :o) 22:28
22:28 stevan_ joined 22:29 ajr_ left 22:32 denis_boyun left 22:36 nebuchadnezzar left 22:37 teslos left
FROGGS okay, spec says something about a Proc::Status methods object, wich provides the exit value and state (also the combined posix for in .status) 22:37
22:41 benabik left
dalek p: aa41dd3 | (Tobias Leich)++ | src/vm/jvm/runtime/org/perl6/nqp/runtime/Ops.java:
fake posix-like return value for nqp::shell()@jvm
 22:42
kudo/nom: 26a36bf | (Tobias Leich)++ | / (2 files):
return posix-like return value for nqp::shell

This should be returned as a Proc::Status object, according to S29:552 and S29:634.		 22:47
synopsebot Link: perlcabal.org/syn/S29.html#line_552
sjn 00:29 -!- ajr_ [~quassel@ottawa-hs-64-26-155-153.s-ip.magma.ca] has quit
oops
dalek ast: 030a4ab | (Tobias Leich)++ | S29-context/exit.t:
accept posix-status until we have Proc::Status
 22:48
jnthn sjn: I see you made it home safely :)
dalek ecza: d827fa8 | (Tobias Leich)++ | lib/CORE.setting:
revert 9d2f552d768563416a0f8c7d38f8cd3c57d9a546
 22:49
jnthn gets some rest
&
FROGGS gnight jnthn
sjn jnthn: I managed to survive the tram trip home, yes :) 22:51
somehow
<- time for some Zzz's
nn #perl6
FROGGS gnight sjn 22:53
22:53 berekuk left 22:54 thou left, jnap left 22:56 berekuk joined 22:58 colomon left, berekuk left 23:02 colomon joined 23:03 nebuchadnezzar joined 23:12 btyler left 23:14 colomon left 23:16 cbk joined 23:27 Guillaume2 joined
Guillaume2 is there a way i can interface perl6 with cocoa under mac os x ? 23:28
or should i use gtk or qt
what would be the most portable gui to use
huf html4 23:30
anything that's anything can render that 23:31
Guillaume2 yeah maybe i should stick to a web app
i wanna code some bible softwares
with hebrew and greek support
unicode i guess
huf it might not be very pleasant, but it is very portable :)
23:37 skids joined 23:39 panchiniak joined 23:41 stevan_ left, stevan_ joined 23:44 nebuchadnezzar left, nebuchadnezzar joined
skids r: multi sub foo (Int $f, $g) { 1; }; multi sub foo ($f, Str $g) { 1; }; foo(4,"d"); # Tiebreaking fail. 23:47
camelia rakudo 26a36b: OUTPUT«Ambiguous call to 'foo'; these signatures all match:␤:(Int $f, $g)␤:($f, Str $g)␤ in sub foo at /tmp/RSfqXNf2FT:1␤ in block at /tmp/RSfqXNf2FT:1␤␤» 23:48
skids (and that is a regression, I think) 23:49
23:54 dmol left, Sean__ joined 23:55 ssutch left

Powered by Cro and the Raku® Programming Language.

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

Thank you!