»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.
00:07 Guillaume2 left 00:12 Psyche^_ joined 00:15 Psyche^ left 00:17 kbaker joined 00:18 colomon joined, Guillaume2 joined 00:21 jnap joined
flussence n: multi sub foo (Int $f, $g) { 1; }; multi sub foo ($f, Str $g) { 1; }; foo(4,"d"); 00:32
camelia niecza v24-90-gd827fa8: OUTPUT«Potential difficulties:␤ $f is declared but not used at /tmp/mfBQYP9ww9 line 1:␤------> multi sub foo (Int ⏏$f, $g) { 1; }; multi sub foo ($f, Str $␤ $g is declared but not used at /tmp/mfBQYP9ww9 line 1:␤------> multi sub foo (In…
flussence argh
n: multi sub foo (Int $f, $g) { $f,$g,1; }; multi sub foo ($f, Str $g) { $g,$f,1; }; foo(4,"d");
camelia niecza v24-90-gd827fa8: OUTPUT«Unhandled exception: Ambiguous call to &foo; these signatures all match:␤ Int, Any␤ Any, Str␤ at /tmp/Vw525aLVLM line 1 (mainline @ 5) ␤ at /home/p6eval/niecza/lib/CORE.setting line 4579 (ANON @ 3) ␤ at /home/p6eval/niecza/lib/CORE.setting line 45…
00:33 Guillaume2 left
flussence which one's supposed to win? I'd guess specific-first but I'm not familiar with that bit of spec 00:34
skids first declared I think.
Though whether the type counts as a "constraint" vs a where clause is a question.
00:42 Guillaume2 joined 00:53 hypolin joined 00:59 panchiniak left, jnap left 01:02 shinobicl_ joined 01:10 colomon left 01:11 shinobicl_ left
TimToady neither is supposed to win in this case, since there are no constraints other than nominal typing 01:12
01:15 benabik joined 01:16 FROGGS_ joined 01:17 jnap joined
flussence yeah, my assumption kinda falls apart if those are both using named params... 01:18
01:18 jnap left 01:20 FROGGS left 01:22 dayangkun joined 01:24 BenGoldberg joined 01:30 colomon joined 01:31 Guillaume2 left
flussence script one-liner that causes a bizarro error message: «sub MAIN($foo, $bar) { constant baz = $foo + $bar; }» 01:49
Nominal type check failed for parameter '$a'; expected Any but got Mu instead 01:50
TimToady well, it's a bit understandable, insofar as you can't create a compile-time constant with values that aren't available till run time. 01:58
but the error is LTA 01:59
JimmyZ std: sub MAIN($foo, $bar) { constant baz = $foo + $bar; } 02:00
camelia std 06636e4: OUTPUT«ok 00:00 46m␤»
TimToady maybe there's some way to mark $foo and $bar to indicate there's no possible way for them to have a value yet
the syntax is fine; it's the semantics that are impossible
JimmyZ n: sub MAIN($foo, $bar) { constant baz = $foo + $bar; }
camelia niecza v24-90-gd827fa8: OUTPUT«Unhandled exception: System.NullReferenceException: Object reference not set to an instance of an object␤ at Builtins.plus (Niecza.Constants c, Niecza.Variable a1, Niecza.Variable a2) [0x00000] in <filename unknown>:0 ␤ at Anon.1.C0baz init (Niecza.Frame…
TimToady constant is just like BEGIN 02:01
if you want a symbol there, you'd have to say my \baz = $foo + $bar 02:02
that would run at, er, run time
constant and BEGIN should probably trap undefined errors and complain "Value of $foo is not known at compile time" 02:04
[Coke] r: say now - constant now
camelia rakudo 26a36b: OUTPUT«===SORRY!=== Error while compiling /tmp/JPvNj2md_s␤Missing initializer on constant declaration␤at /tmp/JPvNj2md_s:1␤------> say now - constant now⏏<EOL>␤ expecting any of:␤ postfix␤ infix stopper␤ infix o…
TimToady r: say now - constant started = now;
camelia rakudo 26a36b: OUTPUT«0.044363␤»
TimToady you can't redeclare now like that 02:05
nor can you declare a constant without initializing it
[Coke] yah, I didn't expect it to work.
geekosaur r: say now - constant $ = now
camelia rakudo 26a36b: OUTPUT«0.04833237␤»
geekosaur ...should that have worked?
TimToady sure, you just declared an anonymous constant 02:06
benabik r: say now - BEGIN now
TimToady just as you do with BEGIN
camelia rakudo 26a36b: OUTPUT«0.04930257␤»
geekosaur I was wondering if a sigil could be used there these days
02:06 BenGoldberg left 02:07 colomon left
JimmyZ Perl6 allows you do good things with leaving a anonymous name 02:07
geekosaur not quite what I was asking. more like, wondering if (a) something like `constant $foo = ...` is still legal given the sigil-less form; (b) if an anonymous scalar would be too close to that form 02:08
somehow
TimToady well, rakudo is still a bit buggy in the constants, since it basically ignores the sigil 02:10
r: constant %foo = a => 1, b => 2; say %foo.WHAT
camelia rakudo 26a36b: OUTPUT«(Parcel)␤»
TimToady n: constant %foo = a => 1, b => 2; say %foo.WHAT 02:11
camelia niecza v24-90-gd827fa8: OUTPUT«(Hash)␤»
TimToady niecza++ does it better
02:19 BenGoldberg joined 02:35 Colby` joined, Pleiades` left, Colby` is now known as Pleiades` 02:45 Pleiades` left 02:46 fridim_ joined 02:52 Pleiades` joined 02:53 BenGoldberg left 02:55 bbkr left 02:56 bbkr joined, jaldhar joined 02:57 colomon joined
clkao w^W 03:14
03:16 preflex left, preflex_ joined, ChanServ sets mode: +v preflex_, preflex_ is now known as preflex 03:28 kbaker left 03:57 Sean__ left 04:02 Pleiades` left 04:05 colomon left 04:08 Pleiades` joined, jaldhar left 04:11 jaldhar joined 04:28 Pleiades` left 04:33 Pleiades` joined 04:46 daniel-s joined 04:48 hummeleBop joined 04:57 dayangkun left 05:08 SamuraiJack_ joined 05:09 prammer left 05:15 rjbs joined 05:23 virtualsue joined 05:33 atroxaper joined 05:41 teslos joined 06:05 abnorman left 06:14 Celelibi left 06:18 Celelibi joined 06:30 abnorman joined 06:34 krunen left 06:44 Pleiades` left 06:47 Pleiades` joined 06:49 berekuk joined 06:51 FROGGS_ left 06:52 berekuk left 07:01 zakharyas joined 07:04 Pleiades` left 07:09 FROGGS joined 07:14 FROGGS left 07:15 kaleem joined 07:21 virtualsue left 07:23 berekuk joined 07:30 domidumont joined 07:35 rindolf joined 07:40 pdcawley joined 07:46 dmol joined 07:52 sqirrel joined 07:57 denis_boyun joined 07:58 berekuk left 08:09 berekuk joined, dmol left 08:12 denis_boyun__ joined 08:13 denis_boyun left 08:16 FROGGS joined 08:20 pecastro joined 08:22 pecastro_ left 08:24 denis_boyun__ left 08:26 berekuk left 08:36 daxim joined 08:40 dakkar joined 08:47 SamuraiJack__ joined 08:49 SamuraiJack_ left 08:57 atroxaper left 08:58 shinobicl_ joined, shinobicl_ left 09:03 berekuk joined 09:09 dmol joined
lizmat drops a pin 09:13
09:14 sqirrel left
moritz catches it before it falls 09:16
tadzik gosh, that scared me
lizmat :-) 09:17
good *, #perl6!
tadzik good star :) 09:20
moritz speaking of star, I should make the newest one available on camelia 09:22
09:23 abnorman left
moritz star: say 42 09:23
camelia star 2013.02: OUTPUT«42␤»
lizmat *: say 42
:-(
09:28 berekuk left 09:29 berekuk joined, Rotwang1 joined 09:34 berekuk left 09:35 wtw left
lizmat TimToady: wondering about compile time constants, shouldn't they be per compilation unit ? 09:35
things like %?ENV, %?DISTRO 09:36
09:37 jesther joined 09:39 berekuk joined 09:40 dayangkun joined 09:41 dayangkun left, jesther is now known as jestherd 09:43 atroxaper joined 09:44 wtw joined 09:51 grondilu joined
grondilu rn: say my uint $ = -1 09:52
camelia rakudo 26a36b: OUTPUT«(timeout)»
..niecza v24-90-gd827fa8: OUTPUT«===SORRY!===␤␤Malformed my at /tmp/sOXENM2ldm line 1:␤------> say my⏏ uint $ = -1␤␤Parse failed␤␤»
grondilu rn: say my uint $ = 10
camelia rakudo 26a36b: OUTPUT«(timeout)» 09:53
..niecza v24-90-gd827fa8: OUTPUT«===SORRY!===␤␤Malformed my at /tmp/GHq3gY2liF line 1:␤------> say my⏏ uint $ = 10␤␤Parse failed␤␤»
09:56 hypolin left 10:11 denis_boyun joined 10:15 rindolf left 10:17 colomon joined 10:23 pmurias joined
pmurias FROGGS: ping 10:23
FROGGS: we do we still have code to treat the hash returned by nqp::getenvhash as some crazy parrot pmc? 10:25
FROGGS pmurias: can you rephrase that? 10:27
pmurias when creating %ENV we have a #ifdef parrot 10:28
why hasn't it been deleted? ;)
FROGGS pmurias: the ifdef about the iterating?
it is still there because I didnt remove (and tested) it 10:29
pmurias hmm I still get the error about incorrect nqp --version 10:30
10:31 xenoterracide left
FROGGS pmurias: do a git pull && make clean && perl Configure.pl && make install 10:31
(in the correct nqp directory, if you have more than one)
what does nqp --version say?
pmurias I don't have the git revision 10:33
This is nqp version 2013.08 built on parrot 5.5.0 revision RELEASE_5_5_0
10:34 berekuk left
FROGGS but you need HEAD 10:35
what does `git describe` say?
pmurias 2013.08-38-gaa41dd3 10:36
FROGGS okay
pmurias: is that nqp on parrot?
pmurias yes 10:39
FROGGS then remove the content of your install dir, cd into parrot, make install, and then configure nqp again and make install
after that nqp --version should be right 10:40
pmurias it's a problem with tools/build/gen-version.pl 10:43
ahh I figured it out 10:46
10:46 pernatiy joined
pmurias tools/build/gen-version.pl doesn't work in a submodule 10:46
10:48 Rotwang1 left, berekuk joined 10:49 atroxaper left
dalek p: 3156a67 | (Pawel Murias)++ | tools/build/gen-version.pl:
Make tools/build/gen-version.pl work correctly in git submodules.
10:52
10:53 pecastro left
pmurias FROGGS: you have a rakudo commit bit? 10:55
10:56 pmurias left
FROGGS .tell pmurias yes, I have 10:56
yoleaux FROGGS: I'll pass your message to pmurias.
10:58 pecastro joined 11:06 marasgp joined
marasgp moritz: do you mind helping me with an ilbot installation? 11:06
11:07 marasgp left, marasgp joined 11:16 berekuk left 11:17 wtw left 11:19 xinming left 11:20 berekuk joined, xinming joined 11:25 wtw joined, wtw left, wtw joined
lizmat r: say 42 11:26
camelia unresponsive?
camelia rakudo 26a36b: OUTPUT«(timeout)»
11:26 sqirrel joined
lizmat I guess feather is suffering under load 11:27
11:29 jestherd is now known as jesther
jesther Not true. 11:29
lizmat r: say True
camelia rakudo 26a36b: OUTPUT«(timeout)» 11:30
jesther OUTPUT-0001.
26b35d. 11:31
It begins momentarily whithin a few months of release, the old fashioned way. 11:32
moritz marasgp: I can try 11:40
marasgp moritz: sorry I posted in the wrong channel. can we do it in #ilbot? 11:41
moritz marasgp: sure
lizmat colomon: finally getting to the end of the Set/Bag integration from niecza 11:47
r: 1 ≼ 2 # what is ≼ ?
camelia rakudo 26a36b: OUTPUT«===SORRY!=== Error while compiling /tmp/PIcX3ILsPG␤Two terms in a row␤at /tmp/PIcX3ILsPG:1␤------> 1 ⏏≼ 2 # what is ≼ ?␤ expecting any of:␤ postfix␤ infix stopper␤ infix or meta-infix␤ statement en…
tadzik r: 1 love 2 # what is love? 11:48
camelia rakudo 26a36b: OUTPUT«===SORRY!=== Error while compiling /tmp/p09Ewv4EOx␤Two terms in a row␤at /tmp/p09Ewv4EOx:1␤------> 1 ⏏love 2 # what is love?␤ expecting any of:␤ postfix␤ infix stopper␤ infix or meta-infix␤ stateme…
lizmat I'm not seeing a texas equivalent of ≼
it's not (<=), as that is linked to ⊆ 11:50
tadzik it's probably bigger :>
lizmat bigger than all of us, surely
nwc10 except Alaska 11:53
lizmat problem is that I can't leave a ≼ in the core setting, or the parse time for rakudo will explode :-( 11:55
11:55 chenryn joined
colomon lizmat++ 11:55
lizmat colomon: so what *is* ≼ ?
colomon lizmat: I never figured out a texas version of ≼
lizmat aha, that explains 11:56
colomon it's bag subset
lizmat problem is, I need it for the ACCEPTS :-(
11:56 berekuk left
lizmat guess I'll have to think of a texas equivalent myself ? 11:56
11:57 grondilu left
colomon n: my $a = (a => 10, b => 20).bag; my $b = (a => 2, b => 3); say $b ⊆ $a; say $b ≼ $a 11:57
camelia niecza v24-90-gd827fa8: OUTPUT«Unhandled exception: Unable to resolve method bag in type Parcel␤ at /tmp/m8sqyQRPVd 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…
colomon :(
lizmat .Bag ?
colomon n: my $a = (a => 10, b => 20).Bag; my $b = (a => 2, b => 3); say $b ⊆ $a; say $b ≼ $a
camelia niecza v24-90-gd827fa8: OUTPUT«True␤True␤»
colomon lizmat++
doh!
11:57 berekuk joined
colomon n: my $a = (a => 10, b => 20).Bag; my $b = (a => 2, b => 3); say $a ⊆ $b; say $a ≼ $b 11:58
camelia niecza v24-90-gd827fa8: OUTPUT«True␤False␤»
colomon there you go
11:58 jesther left
colomon ⊆ converts both sides to sets, so counts are lost 11:58
≼ converts both sides to bags, and counts are important
we can help you come up with a texas version 11:59
benabik Why does parse time explode with Unicode on Rakudo, and is there a planned fix?
11:59 berekuk left
lizmat there is no planned fix as yet, hoping for pmichaud to get some tuits to look at this / give pointers to others to fix this 11:59
it is mightily annoying
daxim define texas 12:00
lizmat pure ascii
colomon daxim: pure ascii version of a unicode operator
lizmat (<+)
12:00 jferrero left 12:01 jferrero joined
lizmat the plus indicating the importance of the numbers ? 12:01
12:01 sqirrel left
FROGGS lizmat: I thought jnthn said that you can have a infix:«\xsomething» in the setting, without loosing speed 12:01
lizmat doesn't work 12:02
I've put in the code in comments, so that when it starts to work, we can just uncomment
colomon (+<+) indicating numbers on both sides?
lizmat FROGG: #only sub infix:<"\x222A">(\|$p) is equiv(&infix:<|>) { infix:<(|)>(|$p) } # NO UNICODE OPS YET 12:03
colomon actually (+<=+) I guess, by analogy with (<)
lizmat wow, that's a long one
colomon lizmat: long is okay, IMO
lizmat but we don't have (=<=) 12:04
so why (+<+) ?
colomon right, but we do have (<=)
lizmat: I was just thinking of emphasizing the numbers on both sides
but I'm just throwing out ideas here
lizmat but (<=) is quivalent to ⊆
12:05 locsmif joined, wtw left
lizmat feels more natural to do (<+) for ≼ 12:05
I'll go for (<+) for symmetry with (<=) for now
colomon lizmat: you maybe can't see it, but ≼ has the = sign in there
lizmat I know, I can see 12:06
colomon it's the bag version of ⊆
lizmat yes, and bags have numbers not booleans
colomon the point is that there is a distinction between (<) and (<=), and ≼ is closer to (<=) 12:07
12:08 teslos left
lizmat ⊂ = (<), ⊆ = (<=) and ≼ would be (<+) 12:09
maybe ⊆ should become (=<=) and ≼ (+<+) ? 12:10
or (<-) and (<=) ?
FROGGS r: sub infix:«\x227c»(\a, \b) { 1 + 2 }; say 1 ≼ 2; 12:11
camelia rakudo 26a36b: OUTPUT«3␤»
lizmat huh?
lizmat tries again
12:13 wtw joined, iSlug left 12:14 teslos joined
FROGGS dunno about slowdown though 12:16
12:16 teslos left
lizmat don't know yet either 12:16
FROGGS arglbargl! Unhandled exception: java.lang.OutOfMemoryError: Java heap space
I can't run the spectest suite on windows :/ 12:17
it starts to fail all test after the 20nth or so
it seems it is unable to consume more than 295MB 12:18
r: sub infix:<\x227c>(\a, \b) { 1 + 2 }; say 1 ≼ 2; # < > dont interpolate
camelia rakudo 26a36b: OUTPUT«===SORRY!=== Error while compiling /tmp/cWDNJEwXYY␤Two terms in a row␤at /tmp/cWDNJEwXYY:1␤------> infix:<\x227c>(\a, \b) { 1 + 2 }; say 1 ⏏≼ 2; # < > dont interpolate␤ expecting any of:␤ postfix␤ infix stopper␤…
12:19 Pleiades` joined
FROGGS r: sub infix:«"\x227c"»(\a, \b) { 1 + 2 }; say 1 ≼ 2; # < > dont interpolate # dont add extra quotes 12:19
camelia rakudo 26a36b: OUTPUT«3␤»
FROGGS ohh, that actually work
s
timotimo hehe
FROGGS r: sub infix:«"(\x227c)"»(\a, \b) { 1 + 2 }; say 1 ≼ 2;
camelia rakudo 26a36b: OUTPUT«===SORRY!=== Error while compiling /tmp/91Wc70PwYY␤Two terms in a row␤at /tmp/91Wc70PwYY:1␤------> x:«"(\x227c)"»(\a, \b) { 1 + 2 }; say 1 ⏏≼ 2;␤ expecting any of:␤ postfix␤ infix stopper␤ infix or meta-i…
FROGGS w/e
k, trying again with Xmx512m set 12:21
12:24 skids left 12:29 SamuraiJack__ left, spider-mario joined 12:33 prammer joined, pmurias joined
JimmyZ FROGGS: ping 12:34
pmurias FROGGS: I'll try to remove the old env leftovers
yoleaux 10:56Z <FROGGS> pmurias: yes, I have
JimmyZ github.com/perl6/nqp/pull/116 Could someone apply it? 12:35
FROGGS JimmyZ: pong 12:36
moritz does it
dalek p: 432f444 | jimmy++ | Configure.pl:
fixed potential build failure on mingw32
p: 51c1fb6 | moritz++ | Configure.pl:
Merge pull request #116 from zhuomingliang/patch-1

fixed potential build failure on mingw32
FROGGS pmurias: cool :o)
JimmyZ thanks
FROGGS k
JimmyZ FROGGS: nvm
FROGGS :o)
Juerd aaa 12:37
-
Oops, ignore that please :) 12:38
JimmyZ another one: github.com/perl6/nqp/pull/117 12:39
actually from original dyncall 12:40
12:43 denis_boyun left
dalek p: 7ecee93 | jimmy++ | 3rdparty/dyncall/Makefile.embedded:
Updated dyncall to support mingw32
12:49
p: daae9b5 | moritz++ | 3rdparty/dyncall/Makefile.embedded:
Merge pull request #117 from zhuomingliang/patch-2

Updated dyncall to support mingw32
lizmat colomon: it looks like "is iffy" is a nieczaism 12:50
colomon lizmat: I think it may be in std?
moritz maybe ported from STD to niecza, and later removed from STD
lizmat definitely not in the spec 12:51
moritz but it is in std.pm6 12:52
r: 1 !+ 2
camelia rakudo 26a36b: OUTPUT«===SORRY!=== Error while compiling /tmp/L6YVJfuKwj␤Cannot negate + because it is not iffy enough␤at /tmp/L6YVJfuKwj:1␤------> 1 !+⏏ 2␤»
moritz std: 1 !+ 2
camelia std 06636e4: OUTPUT«===SORRY!===␤Cannot negate + because additive operators are not iffy enough at /tmp/TThxAIHveq line 1:␤------> 1 !+⏏ 2␤Parse failed␤FAILED 00:00 42m␤»
moritz lizmat: oh, you meant the actual trait, no the concept
lizmat yes, "is iffy"
moritz yes, that might be an nieczaism, but a helpful one IMHO 12:53
because otherwise you can't make user-declared operators iffy
lizmat well, the spec doesn't have the string "iffy" in it
maybe something needs to be specced then :-)
12:54 marasgp left
pmurias how do I run a single test file? 12:54
(without installing rakudo)
12:55 jnap joined
lizmat make t/spec/dir/name.t 12:55
12:58 dayangkun joined, kbaker joined
moritz \o/ new HPMoR chapter 13:00
13:00 ajr joined, ajr is now known as Guest3044 13:01 Guest3044 is now known as ajr_
PerlJam ... there goes another few minutes of lost productivity this morning 13:08
13:10 PacoAir joined
pmurias FROGGS: how should I supply a patch for you to apply? 13:17
FROGGS pmurias: a pull request would be nice 13:18
moritz star: say 42 13:19
camelia star 2013.08: OUTPUT«42␤»
moritz updated to current R*
lizmat *: say 42
:-(
pmurias hates the github ui changes making things confusing 13:20
moritz lizmat: you can patch that in evalbot if you want :-)
lizmat I might 13:21
meanwhile, I'm looking for a stray unicode char that's causing the parse to explode :-(
(as in taking very long)
13:22 donaldh joined
moritz lizmat: is that committed and pushed? 13:23
pmichaud ack '[\x80-\xff]'
lizmat no, only in my local repo
moritz pmichaud: it's not that simple, since Unicode chars in the Latin-1 range are OK 13:24
and ack doesn't decode as UTF-8
FROGGS but it is a start
pmichaud ack '[\xc3-\xff]' # or something like that
but I don't think there are that many latin-1 chars in the setting anyway 13:25
pmichaud checks
oh, except for the french angles :-/
moritz 106
lizmat found them, I think
13:26 bluescreen10 joined
pmurias the github rakudo-js has been accidently deleted and recreated, hopefully it dosn't inconvinience anyone 13:26
13:27 dmol left
JimmyZ Error occurred during initialization of VM 13:29
Could not reserve enough space for object heap
:(
pmichaud for future reference: ack '[\xc4-\xff]' will find lines with chars with codepoint > 0xff
(in a utf8 encoded file) 13:30
pmurias FROGGS: I have made a pull request
lizmat pmichaud++
13:31 PacoAir left
FROGGS pmurias: for rakudo? I can't see it 13:32
13:32 PacoAir joined
JimmyZ change -Xmx1600m to -Xmx1200m works 13:32
13:33 iSlug joined
lizmat r: constant term:<<"\x2205">>= "foo"; say ∅ # rakudobug ? 13:33
camelia rakudo 26a36b: OUTPUT«===SORRY!=== Error while compiling /tmp/U8gVPcX4lq␤invoke() not implemented in class 'QAST::Want'␤at /tmp/U8gVPcX4lq:1␤------> ␤»
lizmat or just NYI with LTA error message
pmichaud nyi, i suspect.
lizmat n: constant term:<<"\x2205">>= "foo"; say ∅ # rakudobug ?
camelia niecza v24-90-gd827fa8: OUTPUT«set()␤» 13:34
lizmat NYI then :-)
huh?
nope, it was supposed to say "foo", not an empty set
pmichaud I wonder if ∅ is predefined in niecza
lizmat it is
13:34 kaleem left
moritz defining terms is also NYI in rakudo 13:35
I've tried twice to implement it, and failed both times
pmurias FROGGS: github.com/pmurias/rakudo/commit/2...f0d498fc86
13:36 dmol joined
pmichaud I'm afk again 13:36
pmurias FROGGS: could I just mail you a patch?
13:36 btyler joined
FROGGS pmurias: sure 13:37
pmurias: hold on
github.com/rakudo/rakudo/pull/203
dalek kudo/nom: 23d310a | (Pawel Murias)++ | src/core/terms.pm:
Unify the %*ENV creating across the jvm and parrot backends.
13:38
kudo/nom: 59b2dfb | (Tobias Leich)++ | src/core/terms.pm:
Merge pull request #203 from pmurias/nom

Unify the %*ENV creating across the jvm and parrot backends.
pmurias FROGGS: thanks
FROGGS yw
bbl 13:41
13:41 FROGGS left 13:45 rindolf joined 13:46 bluescreen10 left, FROGGS[mobile] joined 13:52 dayangkun left
[Coke] yawns 13:52
moritz fwiw panda bootstrap seems to work with this latest patch 13:53
lizmat morning [Coke] !
colomon what moritz said, I'm running another smoke test now. 13:54
FROGGS[mobile] does somebody know how to spectest rakudo@jvm without running into memory limits?
13:55 fridim_ left
moritz 'make spectest' does that for me :-) 13:55
FROGGS[mobile] on windows
moritz oh.
colomon FROGGS[mobile]: if it's java running out of memory rather than windows, you can probably set a flag to fix that somewhere. 13:56
FROGGS[mobile] I tried Xmx1024m without luck 13:57
13:57 btyler left
FROGGS[mobile] but it is possible that I did it wrong 13:57
[Coke] needs コーヒー 13:58
14:00 thou joined 14:02 skids joined
jnthn is home 14:05
yoleaux 01:38Z <benabik> jnthn: To fix 3rdparty/dyncall try `git submodule sync && git submodule update`
nwc10 and still in one piece?
lizmat maybe jnthn has been a bit osloed ? 14:06
nwc10 norwai!
(not sure who first used that one, but I got in first with it today)
jnthn Yes, in one piece 14:09
No transportation accidents in this trip. :)
...thankfully, that hasn't happened for well over a year by now. :) 14:11
btw, for those of you curious about that podcast that interviewed me about architecture stuff last week, it's up at t.co/DnVBweWXmA 14:13
FROGGS[mobile]: I've not had problems spectesting Rakudo JVM on Windows, fwiw 14:14
(didn't do it for a few days, mind...)
timotimo yays 14:15
jnthn (It's nothing Perl 6 related; more about what I do at $dayjob) 14:16
timotimo yup 14:17
14:19 TimK1 joined
FROGGS[mobile] I'm on windows xp 14:20
jnthn XP?!
FROGGS[mobile] and I think I set the memory option fot jre instead of jdk 14:21
yeah, "good" old xp
colomon still uses XP in the windows VM on his MBP
colomon hates every post-XP Windows he has tried.
colomon switched to OS X because of this, actually. 14:22
FROGGS[mobile] interesting, I hate osx
lizmat r: only sub infix:<(.)>(\|$p) {"foo"} # what's wrong with this ? 14:23
camelia rakudo 59b2df: OUTPUT«===SORRY!=== Error while compiling /tmp/eFXGyJf3Mq␤Missing block␤at /tmp/eFXGyJf3Mq:1␤------> only sub infix:<(.)>(\⏏|$p) {"foo"} # what's wrong with this ?␤ expecting any of:␤ new term to be defined␤ constrain…
lizmat is this a nieczaism? 14:24
FROGGS[mobile] Is that signature valid?
lizmat copy/pasted from niecza 14:25
I guess not
jnthn std: only sub infix:<(.)>(\|$p) {"foo"}
camelia std 06636e4: OUTPUT«===SORRY!===␤Unable to parse signature at /tmp/6et9jNqcTg line 1:␤------> only sub infix:<(.)>⏏(\|$p) {"foo"}␤Couldn't find final ')'; gave up at /tmp/6et9jNqcTg line 1:␤------> only sub infix:<(.)>(\⏏|$p) {"foo"}[…
jnthn That looks like syntax we had once upon a time...
I forget what it was for :)
lizmat I suspect \|$p is now just |p 14:26
jnthn It may just be spelled \p now
oh, or that
I forget :)
It's hard enough to remember current spec :P
lizmat my $base_set = $p ?? $p[0].Set !! set(); # seeing that, I assume \p would be better 14:27
nwc10 "this week" isn't what it used to be
14:27 denis_boyun joined
lizmat tries to compile settings again 14:30
jnthn just got himself up to date builds of stuff 14:31
14:33 iSlug left
lizmat ok, am about to give up: any suggestions from this massive stacktrace ? gist.github.com/lizmat/6366683 14:34
ah, found it
jnthn Suggest trying wiht --ll-exception if you didn't already
lizmat this is while doing the make 14:35
apparently it cannot handle "!! die "Routine given to equiv does not appear to be an operator";"
jnthn Sure...just ram it into the Makefile :)
lizmat right
guess I need to check all of the is equivs 14:36
jnthn Apparenlty they ain't all equivalent :P 14:37
donaldh is anyone working on sockets on nqp jvm ? 14:41
jnthn donaldh: tadzik has a branch with a good bit of progress on them, not sure of current status/activity 14:42
donaldh ah
jnthn tadzik: ping :)
donaldh I started hacking but then thought I should check first.
(maybe the last thing that is required for panda ) 14:43
jnthn donaldh: I doubt that tadzik will object to you contributing to the branch(es) 14:44
donaldh Let me scrape around for the branch and take a look
hah, the same branch name as I had chosen. who'd a thunk that jvm-sockets would be a popular name 14:45
lizmat could it be that "is equiv" is simply not (well) implemented yet in rakudo ? 14:46
tadzik jnthn, donaldh, pong :) 14:47
donaldh tadzik: enquiring about sockets on jvm ^^
tadzik jvm-sockets is present in both nqp and rakudo
lizmat jnthn: does this look sane?
multi trait_mod:<is>(Routine $r, :&equiv!) {
nqp::can(&equiv, 'prec')
?? trait_mod:<is>($r, :prec(&equiv.prec))
!! die "Routine given to equiv does not appear to be an operator";
}
tadzik in rakudo it implements IO::Socket::INET that works
doesn't pass the tests, because we have no open(..., "p")
lizmat jnthn: it looks like any "is equiv" dies in the settings 14:48
14:48 FROGGS joined
donaldh open (…, "p") ? 14:48
lizmat and the only "is equiv" I see in the settings, are the once I just added frmom niecza
tadzik donaldh: pipes
io-socket-inet.t relies on that 14:49
donaldh surely that's unix, no?
tadzik I'm not sure
oh, sure
it's using some different hack on windows
jnthn lizmat: is equiv? I think it does, which iirc is why there's a big chunk of stuff at the end of the setting.
lizmat: Which sets all the precedence traits up. 14:50
lizmat aha! ok 14:51
jnthn lizmat: See end of src/core/operators.pm
lizmat found it, yes
jnthn 'tis there
lizmat I guess I need to add them there, rather than at definition time
jnthn I musta decided it was better to avoid the issue than take it on. That probably means it's hairy, or I was just lazy :)
donaldh tadzik, I started hacking sockets but using NIO. I'll maybe try and meld my SyncSocket into the rest of your work. 14:52
FROGGS r: my $a; temp $a; say "jnthn, have you seen that?" 14:53
camelia rakudo 59b2df: ( no output )
tadzik donaldh: my sockets are only synchronous 14:54
jnthn NIO can do dync too :)
uh, sync 14:55
tadzik then it's probably saner for us to use it for everything
donaldh tadzik: also. My impl shared 90% with FileHandle.
tadzik sounds good :) 14:56
feel free to replace mine with yours
donaldh tadzik: looks like you have done most of the hard work in Ops.java, Compiler.nqp and most importantly, rakudo. 14:57
FROGGS do I have to set the memory limit for C:\Programme\Java\jdk1.7.0_11\bin\javaw.exe, or the javac.exe? 14:58
or for the jre only?
jnthn FROGGS: It tends to just be set when invoking the jvm (with java) as done in perl6[.bat] 14:59
FROGGS ahh
hmm
jnthn FROGGS: Note that the spectests use an evilserver
um
evalserver
donaldh lol 15:00
jnthn So maybe its memory use needs a tweak
heh, what a thinko :)
FROGGS MUHAHAHAHAA
donaldh jnthn: you were right the first time :D
FROGGS it was -Xms100m
15:00 ajr_ left 15:01 ajr joined, ajr is now known as ajr_
FROGGS cool, it is working now I think 15:04
15:04 kaare_ joined, ajr_ left 15:05 jaldhar left 15:06 jaldhar joined 15:10 ajr joined, ajr is now known as Guest93807 15:11 Guest93807 is now known as ajr_
colomon evilserver! 15:13
evilserver++
15:15 btyler joined
jnthn bbi20 15:17
15:17 dalek left, dalek joined, ChanServ sets mode: +v dalek
FROGGS T-2min 15:35
15:40 prevost joined
tadzik donaldh: good :) 15:41
15:42 berekuk joined, pdcawley left
TimToady FROGGS: you're assuming he meant minutes :) 15:42
FROGGS TimToady: and I'm assuming decimals :o) 15:43
15:44 prevost left 15:45 pdcawley joined 15:46 prevost joined
TimToady lizmat: $?foo is lexically scoped, but most the the predefined ones can be put in the setting, though some of the values we might want to generate lazily; we could populate %?ENV on first mention, for instance, since most programs won't use it 15:47
lizmat TimToady: I was thinking about the case of the core setting vs something like lib/Test.pbc
TimToady what's the problem? 15:48
lizmat one could envision that .pbc files were compiled on another time / place than the setting
15:48 zakharyas left
jnthn There was a queue... :P 15:48
lizmat so that the %?ENV that lib/Test.pbc had seen, would be different from the %?ENV that the settings saw when they were compiled 15:49
jnthn lizmat: Does %?ENV exist? 15:50
TimToady well, and what's wrong with that?
jnthn: it was speculated that it exists yesterday
lizmat at the moment we only have one %?ENV, no?
or is that already "local" to the compunit anyway ?
15:51 kaleem joined
jnthn At the moment Rakudo has no %?ENV afaik :) 15:51
There's %*ENV which is typically found in PROCESS
FROGGS r: say %?ENV
camelia rakudo 59b2df: OUTPUT«===SORRY!=== Error while compiling /tmp/D9s2jQYDbu␤Variable '%?ENV' is not declared␤at /tmp/D9s2jQYDbu:1␤------> say %?ENV⏏<EOL>␤ expecting any of:␤ postfix␤»
jnthn For any of the $?FOO, though, the assumption is that it's about what was there at compile time
lizmat well, %?ENV is just an example, same goes for any $?foo variable
15:52 kaleem left
jnthn lizmat: Sure, but that's the way $?FOO variables work. That's the point of them. :) 15:52
lizmat ok, then I will shut up :-)
jnthn $?OS # what OS was running when we compiled
$*OS # what OS are we running on
I'd guess the $?OS like ones live in UNIT... 15:53
TimToady funny that you picked $*OS...
lizmat blames being exposed to Sets/Bags too long
jnthn TimToady: oh, what did I forget about it? )) 15:54
TimToady that the spec considers it an underspecified them, like 'length'
my rakudo runs on a VM supplied by a DISTRO built around some KERNEL 15:55
which of those is really the OS?
FROGGS DISTRO can be like debian, ubuntu, centos then? 15:56
TimToady that's the notion
FROGGS is microsoft a distributor or windows? 15:57
of*
tadzik I think vista is a distro of windows
FROGGS yeah, and all have the NT kernel I suppose
(except the one that don't)
like windows ce on powerpc or whatever will be there in five years 15:58
TimToady well, any system is going to map poorly to reality sometimes, but one level of OS maps particularly poorly
FROGGS is there a map of examples for DISTRO/KERNEL?
TimToady not really, it's kinda fuzzy, as our initial understanding of 'length'-like units was 16:00
16:00 donaldh left
TimToady but the three levels seems adequate for linux-y distrubtions 16:01
I suppose I'd be open to naming the kernel one OS
since that's shorter, and people have trouble remembering how to spell kernel
FROGGS hmmm, DISTRO=Ubuntu, KERNEL=3.5.0, OS=linux, ARCH=x86_64 16:02
TimToady but each of those is supposed to be an object, so current usage of $*OS would have to become $*OS.name
no
they aren't strings
though they might stringify that way 16:03
FROGGS yeah, I've seen that VM has attributes
16:03 daxim left
TimToady in particular KERNEL is linux-3.5.0 16:03
or OS
each of these levels tends to have a name/auth/ver 16:04
(in linux-land anyway)
FROGGS DISTRO.name = Ubuntu, DISTRO.ver=13.04, DISTRO.auth=canocial
that fits
TimToady even if we rename KERNEL to OS, it's probably not going to stringify as just 'linux' 16:05
and DISTRO probably gistifies to Ubuntu-13.04 16:07
since the authority is usually implicit in the name
FROGGS yeah 16:08
TimToady or a given name has a default authority, but forks can have an explicit authority
Teratogen sporks 16:09
sporking, a new Perl 6 idiom
FROGGS k, so KERNEL or OS .name would be linux, and .ver = 3.5.0
TimToady linux-FROGGS-1.0.0 your new fork of the linux kernel :)
FROGGS I think I'm making a map of the >10 OSes I have, then we might see if it fits or not
TimToady that would be good 16:10
note how your fork can reset its version number, since authority really comes before version
FROGGS nod 16:11
TimToady maybe we'll need a BROWSER/WEBSERVER level in there too :) 16:12
FROGGS it is getting complicated these days 16:13
TimToady we probably need to make sure these object dispatch nicely to a set of multi-sigs
*jects
so we can have constraint matching that can talk about 3.5+ and such 16:14
er, v3.5+
FROGGS well, you could have a `where OS.name eq '...'`
TimToady well, would be nice to do it with positionals or nameds
FROGGS or *.name or what the syntax is
that would be nice for all kinds of objects 16:15
16:15 dmol left
TimToady multi sub kernel ('linux', *, v3.5+) { do_this() } 16:16
FROGGS like I could have a multi for a thing that has length > 10, and another multi for length <= 10
TimToady er, $ rather than *
well, maybe
certainly with 'where'
don't think we quite allow 0..10 as a literal in a sig 16:17
std: sub foo(1..10) {...}
camelia std 06636e4: OUTPUT«===SORRY!===␤Unable to parse signature at /tmp/B73ARrsesh line 1:␤------> sub foo⏏(1..10) {...}␤Couldn't find final ')'; gave up at /tmp/B73ARrsesh line 1:␤------> sub foo(1⏏..10) {...}␤ expecting any of:␤ co…
TimToady std: sub foo($ where 1..10) {...}
camelia std 06636e4: OUTPUT«ok 00:00 45m␤»
FROGGS std: multi sub test ( $a where .length 1..10 ) { } 16:18
camelia std 06636e4: OUTPUT«===SORRY!===␤Unable to parse signature at /tmp/5h9mivTjK1 line 1:␤------> multi sub test ⏏( $a where .length 1..10 ) { }␤Couldn't find final ')'; gave up at /tmp/5h9mivTjK1 line 1:␤------> multi sub test ( $a where .length[3…
FROGGS err, missing an infix
bbl, dinner
TimToady multi sub kernel($k where (:name<linux>, :ver(v3.5+))) {...} # I guess this should work already 16:23
oh, only without the where 16:30
and I suppose a named interface is less fragile in a fractal landscape (see BUILD) 16:34
16:35 teslos joined, ajr_ left
lizmat ok, I'm giving up for now 16:36
the patches for Set/Bag update behaviour and associated error while compiling at: gist.github.com/lizmat/6367558
cycling and dinner& 16:37
TimToady o/
sjohnson ( `ー´) 16:39
16:41 teslos left
diakopter jnthn: u not in the other channel 16:43
16:43 Pleiades` left 16:44 Pleiades` joined
jnthn diakopter: ? 16:53
16:55 panchiniak joined 16:59 prevost left 17:00 abnorman joined
diakopter jnthn: er, nm; oops 17:01
17:04 abnorman left 17:05 dakkar left
jnthn wonders if dinner will wake him up a bit... 17:06
TimToady is dubious
sjohnson it will if there's a thumbtack hidden in it 17:11
17:11 xinming_ joined 17:14 xinming left 17:16 jnap left 17:22 jaldhar left, jaldhar joined 17:23 FROGGS[mobile] left
jnthn sjohnson: um...nej tack :P 17:23
17:24 dmol joined
sjohnson what's a nej? :) 17:24
is that the sound of "no" after you accidently bit on one? ;~;
17:26 mauke joined
moritz sjohnson: just the Swedish spelling of "no" 17:28
geekosaur "nay!" 17:29
sjohnson heh
jnthn And together, "nej tack" is Swedish for "no thanks" :)
sjohnson i'm impressed 17:31
jnthn: that worked out well :) 17:34
dalek ar: 398f993 | moritz++ | docs/announce/2013.08.md:
fix typo, (Cole Keirsey)++
17:36
17:37 pernatiy left
mauke who here might have handed out lots of karma but not under the nick "dalek"? 17:39
moritz mauke: historically? or recently? 17:40
mauke recently
moritz dunno; but many of us are generous with karma
mauke because I don't think you had 4946 karma a while ago 17:41
moritz .karma mauke
I don't even think we track karma in here at the moment
FROGGS ARGGGHHH, in the middle of S29 spectests it runs into memory limit >.<
PerlJam moritz++ ;)
moritz (though I may eventually extract it from the logs)
FROGGS after hours!
mauke preflex does :-)
preflex: karma moritz 17:43
preflex moritz: 4947
mauke irclog.perlgeek.de/perl6/search/?ni...p;q=%2B%2B doesn't work, sadly 17:44
sjohnson .karma sjohnson 17:47
@karma sjohnson 17:48
sjohnson gives up.
mauke preflex: karma sjohnson
preflex sjohnson: 4
tadzik sjohnson++ # fatfaces
sjohnson heh
flussence wonders how useful I am by some arbitrary number metric 17:49
preflex: karma flussence
preflex flussence: 58
flussence oh wow, bit higher than I thought
sjohnson mauke: preflex is written in Haskell? 17:50
mauke yes
sjohnson is the source code available for the public eye? 17:51
i'm curious what kind of technology is employed to allow it to connect to the internet
mauke it's not really public because I think it's kind of shit 17:52
sjohnson ah, so perhaps on an "ask nicely, don't criticize it" basis?
mauke mauke.hopto.org/preflex/ 17:53
you can criticize all you want :-)
17:53 ponbiki left
mauke but the magic technology is sockets 17:53
sjohnson moritz: while we're on the topic of bot stuff, is it possible to update hugme without having to ^C it / have it reconnect to freenode? 17:54
i'm kind of thinking it
might be very handy if one could change its codebase with constant join/part/quits
without*
i suppose some sort of "plugin paradigm" might allow such a thing. 17:55
flussence that's kind of exactly what an irc bouncer is for 17:56
mauke preflex can do that
sjohnson mauke: did you do it in such a way that you can "unload" and "load" its functionality sort of like a plugin?
mauke this part *is* magic, though
no, its "plugins" are statically compiled in
sjohnson i'm starting to think i need to start using your bot for my own bizarre ideas. 17:57
flussence irssi can update in-place, but I'm scared to look at how it does it
mauke that's how preflex does it, more or less
it exec()s itself, passing serialized state as command line parameters 17:58
tadzik ooh :)
flussence looks like /upgrade does indeed work (at the cost of all my scrollback)
mauke the socket logic is much simpler, though, because preflex only connects to one server at a time
17:58 SamuraiJack__ joined
mauke (I'm running multiple instances) 17:59
flussence: fails horribly with SSL connections :-)
flussence I noticed, good thing I'm too lazy to use ssl on anything I actually talk on :) 18:00
18:20 Rotwang joined, Rotwang left, Rotwang joined 18:21 denis_boyun left
jnthn om nom 18:22
dagurval preflex: karma dagurval 18:24
preflex dagurval: 18
jnthn preflex: karma jnthn 18:25
preflex jnthn: 4222
colomon preflex: karma colomon
preflex colomon: 264
dagurval o_o
jnthn moritz works harder than me here :)
FROGGS preflex: karma Jonathan Worthington
preflex Jonathan Worthington: 1
FROGGS O.o
jnthn preflex: karma jonathan 18:26
preflex jonathan: 2169
18:26 grondilu joined
FROGGS preflex: karma FROGGS 18:26
preflex FROGGS: 147
FROGGS preflex: karma Tobias Leich
preflex Tobias Leich: 138
FROGGS :o)
diakopter preflex: karma not_gerd
preflex not_gerd: 60
diakopter preflex: karma diakopter 18:27
preflex diakopter: 190
diakopter preflex: karma lwall
preflex lwall: 415
diakopter preflex: karma larry
preflex larry: 57
diakopter preflex: karma TimToady
preflex TimToady: 390
diakopter heh
preflex: karma autrijus
preflex autrijus has no karma
diakopter preflex: karma audreyt
preflex audreyt: 47
diakopter preflex: karma au
preflex au: 64
18:28 abnorman joined, araujo left, araujo joined 18:29 btyler left 18:31 ssutch joined 18:32 btyler joined, REPLeffect joined 18:35 domidumont left 18:38 berekuk left 18:39 REPLeffect left 18:49 denis_boyun joined
FROGGS TimToady: we are missing arch somehow: ethercalc.org/wrwsbhvuim 18:50
TimToady: I think it matters if you have a x86_64 or 68k 18:51
18:54 berekuk joined
moritz preflex: karma lizmat 18:54
preflex lizmat: 148
moritz preflex: karma moritz
preflex moritz: 4947
dalek p: 360e781 | (Dagur Valberg Johannsson)++ | src/vm/parrot/6model/reprs/VMArray.c:
Check for out of bounds. Fixes rakudo crash on "Buf.new.subbuf(0, 1)\;
18:57
moritz dagurval++
18:59 berekuk left
dagurval \o/ 19:00
19:01 PZt left, rindolf left
dagurval Would it make sense to make subbuf behave like substr? 19:02
moritz yes
19:02 berekuk joined
dagurval I'd like to work on that 19:02
jnthn ++dagurval 19:04
moritz also hopes that in the future the VMs will over some of memcopy-like interface which makes subbuf, concatenation and other operations a lot faster 19:08
dagurval: it would make sense to throw an X::OutOfRange exception from Perl 6 land in subbuf 19:09
flussence I'm stuck travelling offline tomorrow so I might make a HTML5::Entities module, reason being that the part of rakudo that's supposed to handle those in E<> is... really outdated.
(only recognises xhtml1.0 names and lives in parrot-land) 19:10
moritz ++flussence
flussence and anyway, how many other programming languages can claim to have partial HTML5 support in the core? :) 19:12
mauke PHP
tadzik :)
dagurval moritz: not if we want subbuf to behave like Str.substr 19:14
r: "".substr(0, 2).say
camelia rakudo 9faa30: OUTPUT«␤»
flussence mauke: looks like <? htmlspecialchars() ?> only supports HTML5 as of PHP 5.4, so we're not that far behind 19:15
moritz r: ''.substr(-2, )
camelia rakudo 9faa30: OUTPUT«Start argument to substr out of range. Is: -2, should be in 0..Inf; use *-2 if you want to index relative to the end␤current instr.: 'throw' pc 381820 (src/gen/CORE.setting.pir:162005) (src/gen/CORE.setting:10155)␤called from Sub 'sink' pc 417689 (src/gen/CORE.sett…
moritz dagurval: well, in some cases it does throw :-)
FROGGS r: "".substr(2, 2).say # index out of bounds, not length out of bounds :o)
camelia rakudo 9faa30: OUTPUT«Start of substr out of range. Is: 2, should be in 0..0␤ in method gist at src/gen/CORE.setting:11528␤ in method gist at src/gen/CORE.setting:987␤ in sub say at src/gen/CORE.setting:12424␤ in method say at src/gen/CORE.setting:984␤ in method say at src/gen/CORE…
dagurval ah, I see the difference :) 19:17
moritz fwiw I'm not sure if the way rakudo does it with substr is actually specced 19:18
dagurval "While it is illegal for the start position to be outside of the string, it is allowed for the final position to be off the end of the string; in this case the entire rest of the string is returned, whatever is available." 19:20
Seems to be!
19:23 SamuraiJack__ left
diakopter moritz: or a mem-non-copy ;) 19:37
moritz diakopter: Buf ropes! 19:38
diakopter bufcowropes 19:39
like buffalo cow ropes
jnthn tries to decide if that'd be something you eat or not 19:41
moritz jnthn: ropes don't sound very tasty to me
jnthn But buffalo mozerella is good...
masak evening, #perl6 19:42
moritz good UGT morning, masak
masak: feel reminded of t3 reviews! 19:43
masak moritz: acknowledged! 19:45
19:45 cognominal left
jnthn masak: I really want to know how good terminator 3 is too! 19:45
masak jnthn: not nearly as good as t2. 19:46
moritz jnthn: it's, well, finished. Terminated. :-)
PerlJam masak: It's been so long since the contest that it took me almost a full minute to grok what "t3" was in this context. ;) 19:47
masak: at this pace you'll be finished just in time to start all over again
jnthn wonders if it might wrap around... 19:48
"No contest this year, still reviewing last year's" :)
19:49 pdcawley left
PerlJam moritz, masak: weren't the two of you taking a collaborative approach to reviews? Or did I just imagine that? 19:49
masak that was last year.
sorry about it taking a long time. even I am surprised.
I guess $dayjob is using me more than I've been able to internalize. 19:50
or maybe I'm just worse these days at finding tuits for p6cc reviews. :/
anyway, I strongly doubt there'll be a p6cc2013.
PerlJam masak: no need to apologize (to anyone as far as I'm concerned). 19:51
in some sense the contest itself was the meat, and the reviews are "just" gravy.
tadzik masak: I'll do it if you don't
maybe it won't be that awesome, or of that scale, or of that formula, but there'll be one
PerlJam tadzik++
19:52 cognominal joined, cognominal left 19:53 cognominal joined
flussence my netbook can't compile jakudo :( 19:54
spider-mario I thought it was called Rakujo
masak tadzik: huh. :)
flussence I guess I'll just sacrifice this sd card for swap space...
masak spider-mario: actually, it's called Jajujo. 19:55
spider-mario :D
flussence Jjjjjj?
masak JAJUJO! \o/
spider-mario flussence : you might want to consider zram
tadzik this is how we call
spider-mario it uses some RAM as compressed swap
masak which contrasts it to its other incarnations, Papupo and Mamumo.
spider-mario that’s a funny idea but it might help
PerlJam is that ja-ju-jo or JA-ju-jo or ja-JU-jo or ja-ju-JO?
masak PerlJam: it's JA-JU-JO! 19:56
:P
jnthn flussence: You might be able to build it on a more powerful machine and then copy the JARs over... :)
tadzik well, aren't the jars identical everywhere?
We could just have a buildbot :)
jnthn tadzik: aye 19:57
tadzik: Well, paths aside perhaps
masak tadzik: if you do p6cc2013, I'd be happy to assist you in various ways. 19:58
tadzik: I have three years of prior experience which might be of use... :)
flussence spider-mario: I already had zram (1.5GB total); it barely makes it through "Stage jast" then gets OOM-killed on the next part 19:59
spider-mario ah.
flussence hopefully with an extra 16GB it'll work :) 20:00
tadzik hehe
spider-mario hope it’s a fast sd card, though
flussence with the CPU in this thing I don't think I'd notice a difference
it takes *exactly* 1 hour to build parrot on it though :) 20:02
20:18 hummeleBop left 20:21 REPLeffect joined 20:24 ajr joined, xenoterracide joined, ajr is now known as Guest30237, Guest30237 is now known as ajr_ 20:33 kaare_ left 20:35 denis_boyun left 20:38 TimK1 left 20:39 denis_boyun joined 20:41 denis_boyun left
flussence it worked! and I can even run rakudo-jvm outside the directory and have it complain about none of these modules being installed :D 20:43
lizmat is back and sad that nobody had an idea about her gist gist.github.com/lizmat/6367558 20:44
jnthn
.oO( what gist? :) )
diakopter didn't see it
jnthn called from Sub 'missing' pc 12817 (src/gen/perl6-grammar.pir:4629) (src/Perl6/Grammar.nqp:236)
That is probably a big hint.
20:45 spider-mario left
lizmat perhaps a big hint for you, but not for me 20:45
jnthn lizmat: It's some kind of syntax error, looks like
lizmat yeah, I figured as much as well, but *where* !
jnthn lizmat: Was that --ll-exception output?
lizmat no
jnthn Try that maybe, but it's failing to construct the exception object... 20:46
lizmat not sure where to put that in the makefile
jnthn on the line that compiles CORE.setting
before --setting=NULL is fine enough
20:47 denis_boyun joined
lizmat found it 20:47
masak .oO( call 'missing' maybe ) 20:49
lizmat compiling now
diakopter masak: this method
masak: this method is 20:50
lizmat no difference in output :-(
diakopter might as well not botch the joke
masak diakopter: ;)
jnthn lizmat: Well, it must be some syntax error inside an if block in your local diff...
lizmat are you sure it is in an if? not in a only sub ? 20:51
jnthn yes, two down from missing
called from Sub 'statement_control:sym<if>'
lizmat ok, looking
commenting out lines like Bag.new-from-pairs($set.map({ ; $_ => [*] @bags>>.{$_} })); 20:53
nope, that's not it 20:55
flussence is reminded why the general populace hates java... 21:04
real 0m20.792s
diakopter ?
the general populace loves java
flussence rakudo-parrot only takes 6 seconds to start up here
[Coke] startup cost is irrelevant for everything I use java for. 21:05
21:05 jnap joined
diakopter right, either very long-running things or effectively indefinitely running things 21:05
flussence I guess I need to write bigger scripts then
diakopter which is nearly every software that exists 21:06
21:07 kbaker left
[Coke] certainly won't help us in the "write a quick script to get stuff done" categorgy, though. 21:07
21:11 jferrero is now known as JoaquinFerrero, skids left 21:13 kbaker joined
lizmat apparently there is more than 1 problem: now down to Could not find sub &METAOP_NEGATE 21:14
benabik Huh. My REPL accepts both (|) and (&) from lizmat's gist and that's the only bits I could find with if blocks.
lizmat I bet you just tried that with a built rakudo 21:15
it only fails as part of bulding the core settings, afaics
benabik Ah.
21:16 wtw left
flussence okay I take that back 21:16
benabik patch does not apply?
flussence this random script I'm messing with is 23s for parrot vs 28s for jvm (1 module, precompiled on both) 21:17
which works out to about...
21:17 Guillaume2 joined
flussence r: (28 - 20)/(23 - 6) 21:17
camelia rakudo 9faa30: ( no output )
flussence r: say (28 - 20)/(23 - 6) 21:18
camelia rakudo 9faa30: OUTPUT«0.470588␤»
flussence r: say 1/((28 - 20)/(23 - 6))
camelia rakudo 9faa30: OUTPUT«2.125␤»
Guillaume2 Hello how can I access the Cocoa API from OS/X using Perl6 ?
flussence 212% faster
Guillaume2 Can I ?
benabik flussence: Using crazy math, maybe?
flussence my brane's not in numbers mode right now, so probably :) 21:19
112% faster sounds about right...
benabik A 5s difference is over 100% faster? 21:20
Guillaume2: Rakudo can talk to C APIs, but I don't think there's any way to access Cocoa without Objective C
benabik could be wrong.
Guillaume2 I see...right it's Objective C
flussence it's a 9 second difference not counting the startup times, on a short-running script 21:21
geekosaur wrappers exist for various languages but I don't think anyone has written such for perl6 yet
Guillaume2 yeah the object mapping between objective c and perl...dont know how that would work
maybe in a few years 21:22
:)
benabik Someone did it for Ruby, but I don't know how.
(Other than a lot of (ab)use of method_missing)
Guillaume2 I guess I'll write a web app 21:23
flussence for a native GUI, GTK or EFL is probably the most straightforward route right now
Guillaume2 yeah could do that too
I wanna write a Bible software with Bibles in many language, I wanna be able to cross-reference between Hebrew, Greek and let's say English 21:24
I need to be able to print Hebrew and Greek characters 21:25
benabik You could use rakudo-jvm and then use Swing or SWT, maybe?
diakopter Guillaume2: I just wanted to make sure you know there are thousands of programs out there that do that...
Guillaume2 oh yeah that would be multiplatform
yeah for $500 21:26
lizmat jnthn: does this ring a bell:
$ perl6 src/core/Set.pm
===SORRY!===
CHECK FAILED:
Calling 'infix:<(-)>' will never work with argument types (Set, Set) (lines 144, 144, 144, 144)
Expected: :(\p)
21:26 xenoterracide left
benabik hates writing GUIs, but scala's Swing wrappers made it not too bad. 21:26
Guillaume2 the SWORD project..the only open source project about this particular topic dont produce anything
21:27 pmurias left, abnorman left
lizmat only sub infix:<(-)>(\p) { 21:27
colomon: any thoughts on this only sub infix ?
Guillaume2 that's why I wanna write my own app in Perl6
21:27 abnorman joined
jnthn lizmat: Those errors are usually correct 21:28
lizmat: Is the proto for that op wrong, maybe?
lizmat: Yes, the error is correct. An infix should take two arguments! :)
rakudo++ # catching it at compile time :)
lizmat well, the niecza version says: (\|$p) 21:29
but that's not valid either
21:29 Guillaume2 left
lizmat maybe it should be a slurpy? 21:30
jnthn lizmat: What's the whole sub look like?
lizmat only sub infix:<(-)>(\p) {
return set() unless p;
if p[0] ~~ Baggy {
my @bags = p.map(*.Bag);
my $base = @bags.shift;
Bag.new-from-pairs($base.keys.map({ ; $_ => $base{$_} - [+] @bags>>.{$_} }));
}
else {
my @sets = p.map(*.Set);
my $base = @sets.shift;
Set.new: $base.keys.grep(* !(elem) @sets.any );
}
in niecza, the sig was: only sub infix:<∖>(\|$p) is equiv(&infix:<^>) { 21:31
benabik Is it |p?
\p is one argument, I think(?) |p is capture.
jnthn It may want to be |p or **@p
lizmat ok, trying to build settings again 21:33
21:35 xenoterracide joined
lizmat jnthn: why **@p and not *@p ? 21:39
jnthn lizmat: Avoid flattening 21:40
lizmat ah, yup 21:41
jnthn
.oO( No LoL )
21:42 berekuk left
masak .oO( No PrOjEcT EuLeR ) 21:42
21:43 kbaker left
masak Tene your twitter account sent me spam through DM. you might want to change your password or something. 21:43
argh, and I fail at IRC. :/
Tene: anyway, above. :/ 21:44
colomon lizmat: set difference? what's the question?
21:45 cognominal left
lizmat what your intent was with " only sub infix:<∖>(\|$p)" 21:45
|p **@p ?
\|$p isn't valid syntax in rakudo 21:46
colomon it's supposed to capture all N arguments passed to the operator. 21:48
lizmat **@p then
colomon maybe? 21:49
jnthn **@p would match other list-assoc ops, I think
colomon right, it cannot flatten tehem
benabik \|$p isn't valid syntax anymore. It's supposed to be just |p
21:50 virtualsue joined
lizmat all the changed files compile ok when run seperately (except some "The following packages were stubbed but not defined:" errors) 21:51
stll it doesn't build
I'll update the gist with the current state and call it a day 21:52
maybe a good night of sleep will bring inspiration
gist updated, comments/ideas/patches welcome! 21:54
gist.github.com/lizmat/6367558
sleep&
diakopter o/ 21:57
22:02 Rotwang left, FROGGS left 22:03 abnorman left
masak 'night, #perl6 22:04
diakopter o/
22:06 abnorman joined 22:11 FROGGS joined 22:13 Ayiko left 22:23 zwut00 joined 22:24 PacoAir left 22:25 virtualsue left 22:26 Rotwang joined 22:30 grondilu left 22:43 denis_boyun left 22:53 dmol left 22:54 cognominal joined, teslos joined 22:58 benabik left 23:03 skids joined, jnap left
dalek d: d699c5f | Mouq++ | STD.pm6:
Allow multiple unspaces in sequence
23:11
d: 83f92a5 | larry++ | STD.pm6:
Merge pull request #5 from perl6/multiunsp

Allow multiple unspaces in sequence
23:13 Rotwang left 23:14 teslos left 23:21 zwut00 left 23:22 btyler left 23:23 panchiniak left 23:25 benabik joined 23:32 fridim_ joined 23:56 benabik left 23:58 jnap joined, FROGGS left