»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'perl6: say 3;' or rakudo:, niecza:, std:, or /msg p6eval perl6: ... | irclog: irc.perl6.org/ | UTF-8 is our friend!
Set by sorear on 4 February 2011.
00:04 bonsaikitten joined 00:11 Targen left 00:12 b1rkh0ff left 00:21 Chillance left, slayer101_ left 00:32 benabik left 00:33 benabik joined 00:59 bowtie left 01:09 Targen joined 01:12 anuby joined 01:13 hypolin joined 01:33 bluescreen10 joined 01:46 skids joined 02:11 thou left 02:25 xinming left 02:28 xinming joined 02:32 slayer101_ joined 02:38 slayer101_ left 02:39 slayer101_ joined 02:59 slayer101_ left 03:20 sevin joined 03:36 alester left 03:50 dayangkun joined 04:05 stevan_ joined 04:21 sevin left 04:28 preflex left, preflex_ joined, preflex_ is now known as preflex 04:43 thou joined 04:49 thou left 05:01 geekosaur left 05:02 geekosaur joined 05:07 bluescreen10 left 05:09 eternaleye_ joined, thou joined 05:12 kaleem joined 05:26 telex left 05:28 telex joined 05:34 sevin joined 05:38 anuby left 05:39 kaleem left 05:41 arlinius left 05:43 anuby joined 05:45 anuby left 05:46 anuby joined 06:04 SamuraiJack joined 06:11 kaleem joined 06:21 PacoAir joined 06:29 PacoAir left 06:43 thou left 06:47 slayer101_ joined 06:48 slayer101_ left 06:57 For-Odin joined 07:04 For-Odin left, For-Odin joined 07:07 FROGGS[mobile] joined
FROGGS[mobile] gmorning 07:08
tadzik hello mobile FROGGS
07:30 kaleem left 07:41 sevin left 07:45 am0c joined 07:48 FROGGS joined 07:50 FROGGS[mobile] left
dalek ecs: eb2a85b | (Konrad Borowski)++ | S02-bits.pod:
?ROUTINE has & sigil, not $ sigil
07:51
07:52 thou joined, b1rkh0ff joined 07:57 PacoAir joined 08:01 thou left, PacoAir left 08:04 kaleem joined 08:27 SamuraiJack left 08:29 SamuraiJack joined
cedrvint hello #perl6 08:32
rn: say("test" ~~ /. ** 0/)
p6eval niecza v24-17-gd343a2a: OUTPUT«#<match from(0) to(0) text() pos([].list) named({}.hash)>␤»
..rakudo c3f565: OUTPUT«「test」␤␤»
cedrvint is this^ a bug in Rakudo?
sorear I would say so 08:34
cedrvint submits a ticket 08:35
moritz nqp: my $x := ('0' ~~ /./); say('yes') if $x; 08:40
p6eval nqp: OUTPUT«yes␤»
08:46 hoelzro|away is now known as hoelzro
moritz it seems to generate the right QAST::Regex node 08:47
so the bug is likely in the codegen
diakopter my code probably 08:48
I did the quant iirc
moritz nqp: /a ** 5..3/ 08:49
p6eval nqp: ( no output )
moritz std: /a ** 5..3/
p6eval std 7deb9d7: OUTPUT«===SORRY!===␤Empty range at /tmp/t202LjNh1y line 1:␤------> /a ** 5..3⏏/␤Parse failed␤FAILED 00:00 41m␤»
moritz std: /a ** 0/
p6eval std 7deb9d7: OUTPUT«ok 00:00 42m␤»
bbkr r: "a" ~~ m/{ say &?ROUTINE.WHAT } a/; 08:50
p6eval rakudo c3f565: OUTPUT«===SORRY!===␤Undeclared name:␤ &?ROUTINE used at line 1␤␤»
diakopter er, I did the quant for NFA
not sure which is DTWT
moritz: would 0 repetitions mean *doesn't match*? 08:51
bbkr r: sub foo { say &?ROUTINE.name }; foo()
p6eval rakudo c3f565: OUTPUT«foo␤»
moritz diakopter: no, it means 'match zero characters, successfull' 08:52
diakopter k
I was mostly joking :P
(I thought it extremely unlikely the answer was yes) 08:53
bbkr is above "m/{ say &?ROUTINE.WHAT } a/" a bug? previously &?ROUTINE was available in regexp blocks and returning Nil 08:54
moritz r: say Regex ~~ Routine 08:55
p6eval rakudo c3f565: OUTPUT«True␤»
moritz should refer to the regex itself
diakopter r: say("test" ~~ /{ say 5 } ** 0/)
p6eval rakudo c3f565: OUTPUT«(timeout)5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤5␤
bbkr reports 08:56
diakopter so it's not NFA
moritz diakopter: of course not if you add {}
diakopter: {} terminates LTM 08:57
diakopter that's my point; I was trying to determine whether the problem was with NFA quant.
surely that was clear by "so it's not NFA"
moritz no, wasn't clear to me 08:58
diakopter 'course, the problem could also be in the NFA quant
moritz because I thought you meant it's clear that /{ say 5 } ** 0/, but generalized it to ** 0 in all cases
* that /{ say 5 } ** 0/ was not NFA, 08:59
09:05 bbkr__ joined, kaleem left
diakopter I checked NFA.nqp; if min/max are zero, it hits line 255 (emits a no-op) 09:06
DTRT++ 09:07
09:07 kaleem joined, sqirrel joined
moritz diakopter++ 09:07
# also starts with a 'd' :-)
jnthn morning o/ 09:24
arnsholt chromatic is one hell of a backseat driver 09:25
'lo jnthn \o
diakopter is a trunk driver
nwc10 that's so svn
diakopter jnthn: that's boot to you 09:26
jnthn Why boot me? I just arrived :P 09:27
09:30 xinming left 09:31 bowtie joined
moritz we should rather boost jnthn++ :-) 09:32
diakopter STL is fine thank you
nwc10 no-one mentioned this: blogs.perl.org/users/hercynium/2013...-well.html
There, I fixed that for you :-)
09:32 kaleem left
moritz I just read it :-) 09:33
and it's interesting, but kinda missing a point
diakopter pray for the unsung genius hero(es) chromatic mentions
arnsholt I just read the chromatic post linked from that post (thus my backseat driver comment)
moritz afaict scala was designed to be run by the JVM 09:34
09:34 xinming joined
moritz whereas neither Perl 5 nor Perl 6 are 09:34
so it's not obvious that just because scala on the JVM is faster for a particular application than native Perl 5 is, Perl-[56] on JVM will be fast
nwc10 IIRC I've read people frustrated that Scala doesn't have some things, or compramises on some things, simple because it has to, to work on the JVM 09:35
(ie, they knew why it was done the way it was)
diakopter I need 20,000 threads _now_ ;) 09:36
moritz it's probably really easy to get a 90% subset implementation of Perl 5 to be fast on the JVM
s/really/relatively/
diakopter smop == wink and a node
er, nod
moritz but the last 10% (magic, lvalue substr(), whatever), they might slow things down considerably 09:37
diakopter tieing..
nwc10 that would be consistent with what fglock said about implementing loops in Perlito. They were much faster. Much much faster. (Forget the number). But then adding continue and next, and they were slower than Perl 5
moritz and if you want to use 20 modules, even if all of them are pure perl, chances are that some of them sorely need some of the missing 10% features
arnsholt moritz: Which I think moe's more or less explicitly acknowledged, by saying right away that backwards compatibility is going to be broken 09:39
nwc10 thinks all of this. Glad I don't need to type anything
but this doesn't make Moe wrong. :-) 09:40
it makes Moe interesting.
checking the assumptions. And trying different trade offs
moritz arnsholt: right. The hard part will be figuring out which compromises to make, so that Moe feels both perlish and fast
arnsholt Yeah, that's gonna be really, really hard 09:41
09:41 isBEKaml joined
isBEKaml OHHAI, #perl6! 09:41
moritz thinks it's a good excuse/reason to start to learn scala
diakopter the >100ms startup time won't be funnest
isBEKaml moritz: scala's always a good reason, even if it's just to learn FP on the jvm with static types. :) 09:42
09:42 Psyche^ joined 09:43 Patterner left, Psyche^ is now known as Patterner 09:44 sqirrel left
moritz isBEKaml: yes, but I usually fail at learning languages unless I have a project where I actually use it alongside 09:44
isBEKaml moritz: I have yet to find a suitable project to hack along using Scala. :-) I'm a big phel that way.
09:57 For-Odin left, dakkar joined 10:04 hypolin left
jnthn has a fix for the quant bug 10:06
cedrvint jnthn++ 10:07
10:08 wk_ left
moritz has a small patch that checks for empty range at parse time 10:09
std: /a**5..3/
p6eval std 7deb9d7: OUTPUT«===SORRY!===␤Empty range at /tmp/AVQa_aK8K2 line 1:␤------> /a**5..3⏏/␤Parse failed␤FAILED 00:00 41m␤»
moritz like this
10:10 SmokeMachine joined
cedrvint std: /. ** -1/ 10:10
p6eval std 7deb9d7: OUTPUT«===SORRY!===␤Unrecognized regex metacharacter * (must be quoted to match literally) at /tmp/qdfIkwa49r line 1:␤------> /. *⏏* -1/␤Unable to parse regex; couldn't find final '/' at /tmp/qdfIkwa49r line 1:␤------> /. ** ⏏…
cedrvint std: /.**-1/
p6eval std 7deb9d7: OUTPUT«===SORRY!===␤Unrecognized regex metacharacter * (must be quoted to match literally) at /tmp/6abJBmo_EH line 1:␤------> /.*⏏*-1/␤Unable to parse regex; couldn't find final '/' at /tmp/6abJBmo_EH line 1:␤------> /.**⏏…
cedrvint std: /.**1/ 10:11
p6eval std 7deb9d7: OUTPUT«ok 00:00 42m␤»
jnthn r: say 'aa' ~~ / a ** 0/
p6eval rakudo c3f565: OUTPUT«「aa」␤␤»
jnthn r: say 'aa' ~~ / a ** 0..0/
p6eval rakudo c3f565: OUTPUT«「aa」␤␤»
jnthn Yeah, fixed those two
jnthn spectests 10:12
moritz guesses his and jnthn's patches are orthogonal, but waits to see
cedrvint hum, according to S05 negative ranges are allowed
moritz according to STD not 10:13
I'm going with STD here :-)
jnthn moritz: oh, sorry...I didn't know you were patching too
moritz jnthn: no problem, go ahead
cedrvint n: 'test' ~~ /. ** -1/
p6eval niecza v24-17-gd343a2a: OUTPUT«===SORRY!===␤␤Unrecognized regex metacharacter * (must be quoted to match literally) at /tmp/_jC1K05nry line 1:␤------> 'test' ~~ /. *⏏* -1/␤␤Unable to parse regex; couldn't find final '/' at /tmp/_jC1K05nry line 1:␤---…
10:13 anuby left
jnthn moritz: Hm, my fix breaks some of the :P5 tests... 10:18
r: say "abbbbc" ~~ rx:P5/ab{0,}bc/ 10:19
p6eval rakudo c3f565: OUTPUT«「abbbbc」␤␤»
jnthn ah...looks like P5Regex actions bug. 10:21
10:23 chromis joined 10:27 Chillance joined 10:36 cognominal joined, dayangkun left 10:39 isBEKaml left
dalek p: d8431d2 | jnthn++ | t/qregex/rx_quantifiers:
Test for ** 0 and ** 0..0 quantifiers.
10:43
p: 82793b5 | jnthn++ | src/QAST/Compiler.nqp:
Fix compilation of ** 0 and ** 0..0 quantifiers.
p: 5401c4e | jnthn++ | src/QRegex/P5Regex/Actions.nqp:
Fix {1,} compilation in P5Regex.
kudo/nom: efac0f1 | jnthn++ | tools/build/NQP_REVISION:
Bump NQP_REVISION to get ** fix.
10:44
p-jvm-prep: 79ffc9d | jnthn++ | src/org/perl6/nqp/sixmodel/REPRRegistry.java:
Make sure REPR's id and name are set.
10:47
p-jvm-prep: e81fd35 | jnthn++ | src/org/perl6/nqp/sixmodel/ (2 files):
Add Uninstantiable REPR.
p-jvm-prep: 29f8ce7 | jnthn++ | t/qast_6model.t:
Test for reprname and newtype.
p-jvm-prep: 1d7c623 | jnthn++ | / (2 files):
Implement reprname and newtype.
10:55 kaleem joined 11:03 wk joined 11:04 For-Odin joined 11:05 For-Odin left, Su-Shee_ joined 11:06 Su-Shee left, Su-Shee_ is now known as Su-Shee
dalek p-jvm-prep: 49fe3ee | jnthn++ | t/qast_6model.t:
Fix typo'd test plan.
11:07
p-jvm-prep: 8e58be5 | jnthn++ | / (3 files):
Implement and test nqp::substr.
11:22 cognominal left 11:23 vikram joined 11:24 vikram left 11:31 cognominal joined 11:42 cognominal left 11:43 FROGGS left 11:44 cognominal joined 12:10 sftp joined
masak good definitely afternoon already, #perl6 12:12
12:12 MikeFair__ joined 12:13 MikeFair left
cedrvint o/ masak 12:14
masak timotimo: re rt.perl.org/rt3//Public/Bug/Displa...l?id=77272 -- that error message is *horrible*. 12:18
timotimo: so, while I'm happy that it now has line and file in it, I'd like for it not to spew 28 lines of useless inventory of trait_mod:<is> candidates, none of which will help the user realize what went wrong. (typo in class name) 12:19
is it possible to hijack the useless error message and replace it with a useful one in this case?
dalek p-jvm-prep: 773a84a | jnthn++ | / (2 files):
Set up some of the nqp-specific ops.
12:20
p-jvm-prep: d5691cb | jnthn++ | t/qast_core.t:
Tests for while and for with control exceptions.
p-jvm-prep: 89aaf7a | jnthn++ | lib/QAST/JASTCompiler.nqp:
Implement control op for next/redo/last.
p-jvm-prep: 6c5ecc9 | jnthn++ | src/org/perl6/nqp/runtime/Ops.java:
Fix control exceptions crossing block boundaries.
masak I mean, we know that what comes after 'class A is' must be either something kinda special, like 'rw', or an already-declared class/role name.
if it isn't, we should be able to hijack.
jnthn Apart from the "something kinda special" is not special at all.
It's a perfectly ordinary multi-dispatch to trait_mod:<is> 12:21
arnsholt If any trait_mod:<is> fails I think we want line and file, really
jnthn Yes, you get line and file now :)
The problem is how to try and do better than listing the candidates.
masak right.
arnsholt r: class A is NoSuchType {}
p6eval rakudo efac0f: OUTPUT«===SORRY!===␤Cannot call 'trait_mod:<is>'; none of these signatures match:␤:(Mu:U $child, Mu:U $parent)␤:(Attribute:D $attr, :rw(:$rw)!)␤:(Attribute:D $attr, :readonly(:$readonly)!)␤:(Attribute:D $attr, :box_target(:$box_target)!)␤:(Routine:D $r, …
12:22 sqirrel joined
jnthn r: multi trait_mod:<is>(Mu:U $type, *%fail) { die "No such parent" }; class A is NoSuchType {} 12:22
p6eval rakudo efac0f: OUTPUT«===SORRY!===␤No such parent␤at /tmp/UyilwwusGR:1␤------> ␤»
jnthn r: multi trait_mod:<is>(Mu:U $type, *%fail) { die "No such parent" }; class A is Int {}
p6eval rakudo efac0f: ( no output ) 12:23
jnthn r: multi trait_mod:<is>(Mu:U $type, *%fail) { die "No such parent" }; class A is rw {}
p6eval rakudo efac0f: ( no output )
jnthn So that looks promising *but*...
r: multi trait_mod:<is>(Mu:U $type, *%fail) { die "No such parent" }; multi trait_mod:<is>(Mu:U $type, :omgz!) { say "ok" }; class A is omgz {} 12:24
p6eval rakudo efac0f: OUTPUT«===SORRY!===␤Malformed parameter␤at /tmp/IS77AjXuud:1␤------> t" }; multi trait_mod:<is>(Mu:U $type, :⏏omgz!) { say "ok" }; class A is omgz {}␤ expecting any of:␤ pair value␤ formal parameter␤ constraint␤…
jnthn r: multi trait_mod:<is>(Mu:U $type, *%fail) { die "No such parent" }; multi trait_mod:<is>(Mu:U $type, :$omgz!) { say "ok" }; class A is omgz {}
p6eval rakudo efac0f: OUTPUT«ok␤»
jnthn Ah, OK, it gets it right.
So maybe it can work.
12:24 SamuraiJack left
jnthn But I don't wanna do something that could have ecosystem fallout the day before release :) 12:25
arnsholt How hard would it be to make it something like "Cannot call trait_mod at $file:$line"?
moritz we could also define a proto that does a re-dispatch, and catches a dispatch failure
though we don't have typed exceptions for those yet 12:26
so, probably not
jnthn arnsholt: The file/line is output already now
The problem is that a common mistake triggers an accurate, but unhelpful error.
A fail to apply a trait mod really *is* a multi-dispatch failure. I don't think we should be hiding that in the general case. 12:27
12:28 sqirrel left
arnsholt There is that. But moving the line and file info to the top instead of after the pseudo-stacktrace would help a bit at least 12:28
jnthn I don't see how putting the file/line in a different place than the user is used to seeing it helps anything. 12:29
The pointy thing should be better placed, though (actually point to the trait name)
masak ooh, ++isBEKaml is doin' this month's release. 12:30
any volunteers for March? :)
moritz is the eternal fallback
masak I can be a fallback too, provided nothing messes up my March schedule. 12:31
jnthn End of March is ungood for me. 12:32
masak oh, right.
no, I was meaning more "any of the lurkers volunteering for March?" :)
jnthn (So would be good for somebody to take the Star release also then)
I'm happy to do this month's Star release if nobody else wishes to.
12:47 FROGGS[mobile] joined 12:52 grondilu joined
grondilu You know what would be nice? A *persistent* command-line history in the REPL. 12:53
moritz grondilu: wouldn't be too hard to do
there's a history thingy that comes with readline
grondilu I actually have no idea about how the REPL works. Is it in perl6 or is the perl6 interpreter encapsulated in Perl5 or something? 12:56
arnsholt Mostly inherited from NQP IIRC
grondilu so the readline API calls are in NQP?
12:57 FROGGS joined
moritz well, NQP calls $stdin.readline_interactive 12:58
which in turn is a parrot method
from src/pmc/filehandle.pmc
grondilu ok 13:00
moritz it seems it already has a call to add_history 13:01
so I guess it's just a matter of setting a history file first, and later closing it
or...something
grondilu yes, a convention for the file would have to be chosen.
but since we already have a ~/.perl6/ directory, it's not absurd to use a ~/.perl6/history file. 13:02
moritz correct 13:03
13:07 araujo joined 13:10 jaldhar left 13:12 isBEKaml joined
isBEKaml good evening, #perl6! 13:13
moritz \o isBEKaml 13:14
jnthn hi isBEKaml 13:15
isBEKaml hello, jnthn, moritz 13:16
13:16 cognominal left 13:17 jaldhar joined
FROGGS hint: to add a deadlock to rakudo just use perl 5's ternary syntax 13:17
isBEKaml Wait, I'm doing this month's release? (I totally never got how my name came up :)
jnthn r: my $a = 1 ? 2 : 3; 13:19
p6eval rakudo efac0f: OUTPUT«===SORRY!===␤Unsupported use of ?: for the conditional operator; in Perl 6 please use ??!!␤at /tmp/ud0vQFsD8h:1␤------> my $a = 1 ?⏏ 2 : 3;␤»
FROGGS jnthn: I added it in the internals...
isBEKaml r: my $a = 1 ?? 2 !! 3; #lemme see...
p6eval rakudo efac0f: ( no output )
FROGGS stage parse got stuck
isBEKaml r: my $a = 1 ?? 2 !! 3; say $a; #lemme see... 13:20
p6eval rakudo efac0f: OUTPUT«2␤»
isBEKaml ...
jnthn FROGGS: Where exactly? 13:21
FROGGS in Match.pm method ACCEPTS
Match.ACCEPTS( $thing ) is the same like $think ~~ Match, right? 13:22
13:22 cognominal joined
FROGGS or is there a difference between $think ~~ Match and Match ~~ $thing? 13:22
13:22 jaldhar left
jnthn It's not symetric 13:22
But if you want to write a type check in the setting then the optimal way is nqp::istype($thing, Match) 13:23
masak isBEKaml: irclog.perlgeek.de/perl6/2012-11-11#i_6143331
isBEKaml Anyway, since I'll be doing a release in a couple of hours now, will it be okay if I do it through a series of pull requests with a final upload to rakudo.org?
FROGGS I want to change the return value of that operation
masak isBEKaml: that's how it happened. you obviously didn't expect us to survive this far :P
isBEKaml: it's only Wednesday. you have until tomorrow ;) 13:24
FROGGS r: say "abc" ~~ s[x] = "y"
p6eval rakudo efac0f: OUTPUT«Cannot assign to a non-container␤ in sub infix:<=> at src/gen/CORE.setting:12250␤ in block at /tmp/BlU1VY0wtQ:1␤␤»
FROGGS ohh
isBEKaml (I'm currently going through the release guide now)
FROGGS r: my $s = "abc"; say $s ~~ s[x] = "y"
p6eval rakudo efac0f: OUTPUT«True␤»
isBEKaml r: say 42 # up, are you?
p6eval rakudo efac0f: OUTPUT«42␤»
masak isBEKaml: but yes, the release process should be doable on a clone of the rakudo/rakudo repository.
FROGGS should be false, thats what I'm trying to do 13:25
masak isBEKaml: though note that the perl6/nqp repository is involved, too.
FROGGS I think I add some print()s so I get a clue what happens 13:26
13:26 cognominal left
isBEKaml man, my network strength is weak... :/ 13:28
moritz if you don't have commit access to perl6/nqp, I can fix that for you 13:29
just shout :-)
isBEKaml moritz: aye, nqp bits please! :)
moritz isBEKaml: what's your github ID? 13:30
isBEKaml moritz: svatsan
FROGGS
.oO( the moment you recognize you read to much crap... I read "bitch please" )
isBEKaml FROGGS: yeah, blame p6eval. :-)
moritz isBEKaml: done 13:31
isBEKaml masak: Well, I must have raised my hand in a moment of weakness. (I know... )
moritz: thank you.
masak isBEKaml: a weakness for which we will profusely thank you tomorrow. 13:33
isBEKaml masak: but no matter, I'm doing it anyway. I'm just as happy to be able to contribute something.
masak does anyone here know what "triangulating" means in the context of TDD?
FROGGS masak: I can just image that you have to triangulate the problem if the tests are too vague 13:34
13:35 dayangkun joined
masak does anyone here *know*...? 13:36
FROGGS ;o)
bowtie hi, all, I like the new colours in error messages, some info for you - data blocks in S26 don't seam to be supported in 2012-12 even though perl6.org/compilers/features suggest that it should now work, can you please get feature web site updated, and thanks, perl 6 is cool :)
isBEKaml masak: according to codemanship.co.uk/parlezuml/blog/?postid=987, triangulate apparently means running a series of tests to arrive a general solution to a problem.
13:37 cognominal joined
FROGGS bowtie: it might help if you would make an RT ticket with a short example code and the expected behaviour 13:38
bowtie: that note here might get lost
isBEKaml masak: it's only 16th today, eh... Here's the catch, I may be offline tomorrow or worse, without a network until saturday. :|
bowtie FROGGS, will do if you kick me in correct direction of where to do the 'rt' 13:39
masak isBEKaml: re triangulate -- thanks. I searched and didn't find anything. 13:40
isBEKaml: may I offer to switch release dates with you?
isBEKaml: I can do tomorrow, if you do March.
arnsholt tadzik: You around?
tadzik arnsholt: yeah, sort of
isBEKaml masak: March eh? no no.. I totally want to pull a joke on the Mayans!
arnsholt Regarding bowtie's issue. You happen to know where data blocks are supposed to be at ATM? 13:41
tadzik bowtie? blocks?
arnsholt Being the POD-master and all =)
tadzik oh
data blocks, right
bowtie tadzik, where do I create the 'rt' 13:42
masak isBEKaml: you can still pull the joke in March if you want. I was just thinking of getting the right release date on the release.
tadzik email [email@hidden.address]
arnsholt RT is rt.perl.org
tadzik yeah, data blocks are NYI
arnsholt (The Perl bugtracker)
tadzik probably lhf though
arnsholt Right, so it's sort of a bug in the feature matrix as well
FROGGS bowtie: here is that mail posted too, and some explanation: rakudo.org/tickets/
bowtie FROGGS, will do 13:43
arnsholt bowtie: Right, so that bit of POD isn't implemented yet (NYI == not yet implemented), and the feature matrix isn't entirely correct it would seem
isBEKaml masak: If I do a release tomorrow 0530 hrs IST, it'll be just past midnight UTC. :-)
masak oh, ok. 13:44
isBEKaml: on your head be it.
my offer for switching release dates still stands if you don't want to be awake that early :)
isBEKaml masak: So, if I do it, the date is still in. (On my own head) :P
masak: I travel tomorrow, so I guess I'll be awake. :)
masak ok. 13:46
arnsholt isBEKaml: Where in India are you, if I may ask?
isBEKaml arnsholt: Mumbai.
masak let me/us know if you need any kind of help with the release process.
arnsholt Oh, cool! I was there just before christmas
isBEKaml arnsholt: Ah, awesome. I didn't know you were here then. (I usually follow along IRC logs) Your visit was good/fruitful, I guess? (No troubles?) 13:47
arnsholt It was pretty good 13:48
isBEKaml masak: sure, I'll ask here if I need anything.
arnsholt I was at a conference at IIT-B, which was excellent (my first major conference), and then a five days in Goa for relaxation and a couple days in SoBo before going back home 13:49
isBEKaml [Release Prepping] Please update changelogs, issues or anything that goes along into this release. I'm planning to do a release around Midnight UTC.
arnsholt The traffic was insane (obviously =), but everything went smoothly for the most part
isBEKaml arnsholt: Ah... Goa's always good for relaxation.
arnsholt A couple hiccups at the conference excursion
masak isBEKaml++ 13:50
13:50 rindolf joined
arnsholt Yeah, after two months of travelling almost every weekend, sun, sand and relaxation was wonderful =) 13:50
isBEKaml arnsholt: traffic snarls, it's pretty much expected anywhere in India.
arnsholt Yeah, I've been in Chennai once before and it was sort of similar
But Mumbai is just so much bigger =)
isBEKaml (except maybe, in the rural/remote pockets but you go without water! :) 13:51
arnsholt: Chennai too? awesome.
arnsholt Yeah, my girlfriend did fieldwork for her master's thesis in Tamil Nadu, around Kanchipuram 13:52
(She does religious history)
bowtie 'rt' - 116419Perl6 features out of date 13:56
arnsholt, the issue is that the feature list implies it works, not that it dose not 13:58
arnsholt Yeah, that's true
Is there someone who's curator of the feature matrix? 13:59
tadzik everyone is
it's on github/perl6/features or so
arnsholt Right
14:01 Targen left
jnthn bowtie: Where does the feature list say data blocks work, ooc? 14:02
Oh, unless it's counted as part of Pod... 14:03
arnsholt It's specced together with POD at least
bowtie data blocks are in S26 which is the reference next to Pod in features, it's just an observation 14:06
jnthn *nod* 14:07
bbkr_ r: say 1.0000000000000000000000000000001 + 1 14:08
p6eval rakudo efac0f: OUTPUT«2␤»
bbkr_ rn: say 1.0000000000000000000000000000001 + 1
p6eval rakudo efac0f, niecza v24-17-gd343a2a: OUTPUT«2␤»
14:09 renormalist joined 14:10 PacoAir joined
bbkr_ r: 1.0000000000000000000000000000001 + 1.0 14:11
p6eval rakudo efac0f: ( no output )
bbkr_ r: say 1.0000000000000000000000000000001 + 1.0
p6eval rakudo efac0f: OUTPUT«2␤»
bbkr_ weird...
14:11 FROGGS left
bbkr_ r: say 1.0000000000000000000000000000001 14:12
p6eval rakudo efac0f: OUTPUT«1.000000000000000000000000000000100␤»
bbkr_ those trailing 0s looks like a bug
14:12 SmokeMac_ joined, bluescreen10 joined 14:13 PacoAir_ joined, PacoAir left, SmokeMachine left, PacoAir_ is now known as PacoAir
moritz r: say ( 1.0000000000000000000000000000001 + 1.0 ).WHAT 14:18
p6eval rakudo efac0f: OUTPUT«Num()␤»
moritz you are witness to Rat overflowing, and auto-casting to Num
masak it's either that or using up more processor/memory trying to juggle all the precision. 14:19
the BigRat rabbit hole runs pretty deep :) 14:20
er, FatRat
14:21 thou joined 14:24 FROGGS joined
bbkr_ what about trailing 0s in example above? bug? 14:25
[Coke] wr: say 1.0000000000000000000000000000001.WHAT
r: say 1.0000000000000000000000000000001.WHAT
p6eval rakudo efac0f: OUTPUT«Rat()␤»
jnthn Gee, I'm rocketing thorugh getting nqp-mo compiling. That's bad 'cus it means I'm less than a day of work from the point where my blocker becomes porting the bounded serialization stuff... 14:26
;)
[Coke] r: say 1.0000000000000000000000000000001.denominator
p6eval rakudo efac0f: OUTPUT«10000000000000000000000000000000␤»
jnthn r: say 1.0000000000000000000000000000001e1
p6eval rakudo efac0f: OUTPUT«10␤»
moritz bbkr_: if the spec says trailing zeros aren't allowed, then it's a bug
14:27 xinming_ joined 14:28 xinming left
bbkr_ moritz: spec does not forbid trailing zeros. it's "common sense" type of issue. 14:29
masak I don't really see what's wrong with trailing zeroes. 14:33
rn: say 1.0
p6eval rakudo efac0f, niecza v24-17-gd343a2a: OUTPUT«1␤»
masak rn: say 1e1
p6eval rakudo efac0f, niecza v24-17-gd343a2a: OUTPUT«10␤»
masak rn: say 1e0
p6eval rakudo efac0f, niecza v24-17-gd343a2a: OUTPUT«1␤»
masak rn: say sprintf '<%5.2f>', 3/2 14:34
p6eval rakudo efac0f, niecza v24-17-gd343a2a: OUTPUT«< 1.50>␤»
[Coke] r: eval eval eval eval eval eval eval eval eval eval eval "squirrel" 14:35
p6eval rakudo efac0f: OUTPUT«===SORRY!===␤Undeclared routine:␤ squirrel used at line 1␤␤»
bbkr_ masak: IMO number should be printed in simplest form possible. and it works that way up to 1.0000000000000000001 (no trailing zeros). 14:36
r: say 1.0000000000000000001
p6eval rakudo efac0f: OUTPUT«1.0000000000000000001␤»
masak I'll go so far as agree that it's a bit weird that those zeroes are there.
FROGGS npr: say 1.0000000000000000000000000000001 14:38
p6eval rakudo efac0f: OUTPUT«1.000000000000000000000000000000100␤»
..pugs: OUTPUT«1.0000000000000000000000000000001␤»
..niecza v24-17-gd343a2a: OUTPUT«1␤»
bbkr_ this may cause problems when using in some fixed-width environment (tables, protocols using stringified value, etc) 14:41
FROGGS well, might use sprintf for that
[Coke] if you're using a fixed width, i'd argue you shoudl be using sprintf.
FROGGS npr: print 1.0000000000000000000000000000001 14:42
p6eval pugs: OUTPUT«1.0000000000000000000000000000001»
..niecza v24-17-gd343a2a: OUTPUT«1»
..rakudo efac0f: OUTPUT«1.000000000000000000000000000000100»
14:42 cognominal left 14:43 cognominal joined 14:48 FROGGS[mobile] left, FROGGS[mobile] joined
masak note though that not even sprintf guarantees a maximum width if numbers are too large. 14:50
formats are better for that.
FROGGS I usually have a hash table for that, field name => regex 14:51
(before doing formatting)
so I dont write crap to systems like SAP
15:00 cognominal left, cognominal joined
isBEKaml hmmm - I got connection reset when I git pull --rebase on nqp. :-) (after a loooong time) 15:00
fatal: read error: Connection reset by peer
fatal: protocol error: bad pack header
moritz isBEKaml: probably a network problem 15:01
just retry
15:01 bitbybit joined
masak any suggestion on how to convert a number to int in nqp? 15:02
that is, I'm after the semantic effect of rounding it down towards the closest integer. 15:03
PerlJam isn't there an nqp::floor or something?
jnthn nqp::floor_n
PerlJam yeah, that.
15:04 bitbybit left
[Coke] r: say 200000000*0.0001 # xkcd 15:07
p6eval rakudo efac0f: OUTPUT«20000␤»
masak yay. thank ye.
benabik [Coke]: Uhg, don't remind me.
15:08 FROGGS left
masak hm, but if there's only floor but not round, I have to be careful with those negative numbers... 15:09
oh, the special cases. :)
15:09 FROGGS[mobile] left 15:10 FROGGS[mobile] joined, FROGGS[mobile] left
isBEKaml network issues. BBL (release in 9-10 hrs ish) 15:12
15:14 isBEKaml left
PerlJam masak: what does "round" mean anyway? Which rounding semantics were you expecting? :-) 15:15
benabik n <= x < n.5 rounds to n. n.5 < x <= n+1 rounds to n+1. x == n.5 rounds non-deterministically to n or n+1. 15:16
15:17 SunilJoshi joined
PerlJam benabik: I just *knew* there was a reason we included a random number generator! 15:19
benabik Ooh, even better! Rounds to n|n+1 15:20
_Real_ non-determinism, not this psuedo-random stuff.
Oh, maybe that should be n&n+1
masak PerlJam: oh, indeed. I need to do more work than I initially thought. and more tests. 15:24
15:25 stopbit joined 15:28 orafu joined 15:35 isBEKaml joined 15:36 arlinius joined
isBEKaml Was there a parrot release this month? I see from git commits that the last parrot release was a month ago, not this past week or anything more recent. 15:36
15:37 hash_table joined
benabik isBEKaml: dukeleto is going to release Parrot 5.0 fairly soon. 15:37
diakopter r: say("test" ~~ /{ BEGIN die } ** 0/) 15:38
p6eval rakudo efac0f: OUTPUT«===SORRY!===␤␤»
diakopter r: say("test" ~~ /{ die } ** 0/)
p6eval rakudo efac0f: OUTPUT«「」␤␤»
isBEKaml benabik: hmm, the question now is, should we hold? Rakudo's convention is a release about 3-4 days after Parrot's.
diakopter r: say("test" ~~ /{ macro die { }; die } ** 0/) 15:40
p6eval rakudo efac0f: OUTPUT«「」␤␤»
isBEKaml benabik: I see you've asked this question on #parrot. :-)
diakopter r: say("test" ~~ /{ macro die { }; BEGIN die } ** 0/)
p6eval rakudo efac0f: OUTPUT«「」␤␤»
benabik isBEKaml: From the docs: "The Parrot project makes releases on the third Tuesday of each month." So the release is scheduled for the 18th. 15:41
masak benabik: no, the 18th is a Thursday.
benabik isBEKaml: I lurk here and in #parrot. Less asking, more osmosis.
masak er, a Friday.
benabik masak: D'oh
jnthn The third Tuesday was yesterday.
benabik My calendar app was stuck on November for some reason. 15:42
isBEKaml masak: Apologies.
isBEKaml If parrot's not released yet, I don't think I can cut a release this month. (I'll be on road and out of network most of the next couple of weeks) 15:43
masak well, nqp is still on Parrot 4.10. 15:44
I don't immediately see a reason to upgrade to 5.0.
isBEKaml not even on 4.11?
masak (we generally don't upgrade to the latest Parrot unless there's a reason to, methinks)
isBEKaml: I just checked, and it's on 4.10.
Rakudo, however, is on a relatively new (post-release) NQP, so you'll still need to release NQP as usual. 15:45
isBEKaml masak: Okay - If there's no problem, I'll cut a release as usual. 15:47
masak no problem as far as I can see. 15:49
isBEKaml masak: okay 15:54
timotimo r: class A is NoSuchType { } 15:58
p6eval rakudo efac0f: OUTPUT«===SORRY!===␤Cannot call 'trait_mod:<is>'; none of these signatures match:␤:(Mu:U $child, Mu:U $parent)␤:(Attribute:D $attr, :rw(:$rw)!)␤:(Attribute:D $attr, :readonly(:$readonly)!)␤:(Attribute:D $attr, :box_target(:$box_target)!)␤:(Routine:D $r, …
timotimo i get "null pmc access" locally :| 15:59
say, masak, do you think it's reasonable to look through the lexpad/world object for all objects that are similarly named (levensthein distance?)
jnthn Works locally for me...
timotimo builds a newer rakudo 16:00
masak timotimo: I think it's reasonable, as long as we're talking error reporting. might give Perl 6 a reputation of almost unbelievably good error messages. 16:01
Su-Shee can I have just unbelievably good messages? ;) and no errors? 16:02
hello everyone.
masak Su-Shee! \o/
timotimo masak: yeah, i think in that case rakudo would terminate as the next step, so it would be reasonable to do some extra work i guess.
now to implement levensthein in nqp! 16:03
i've been misspelling it *again* :~
masak oh the irony.
16:04 kaleem left 16:10 MikeFair joined 16:11 isBEKaml left, MikeFair__ left
timotimo can't seem to figure out how to get the length of a string in nqp :| 16:17
jnthn nqp: say(nqp::chars('beer')) 16:20
p6eval nqp: OUTPUT«4␤»
timotimo thanks. should have tried that, too :| 16:21
since i'm assuming nqp doesn't offer "is memo", how do i do %memo{$key} :exists in nqp? 16:24
alternatively, where do i have to look to find all subs in the nqp:: namespace?
is that nqp/src/ops/nqp.ops? 16:25
jnthn No, src/QAST/Operations.nqp
nqp::existskey(%memo, $key)
timotimo thanks :)
there seems to be no "min" or "max" functions, is that right? 16:30
dalek kudo-js: 35340a6 | (Paweł Murias)++ | run (2 files):
fix return value of to_n, pass test 33
kudo-js: c3f9577 | (Paweł Murias)++ | runtime.js:
deserialize doubles
kudo-js: 529e2af | (Paweł Murias)++ | runtime.js:
Implement nqp::existskey.
kudo-js: b73198e | (Paweł Murias)++ | / (2 files):
Update to current nqp, mention that in README.
jnthn timotimo: No, we don't have those 16:31
timotimo don't really need it for only three elements anyway :) 16:32
16:33 pmurias joined
pmurias jnthn: re spurt in t/helpers.nqp the only difference to spew is the :encoding? 16:34
jnthn nqp: spew()
p6eval nqp: OUTPUT«too few positional arguments: 0 passed, 1 (or more) expected␤current instr.: 'spew' pc 734 (src/stage2/gen/NQPCORE.setting.pir:331) (src/stage2/NQPCORE.setting:388)␤»
jnthn huh, didn't even realize we had that :)
no, spurt is the spec name for spew :)
pmurias just noticed that
jnthn: so we should rename spew into spurt in nqp? 16:35
jnthn pmurias: yeah, but maybe leave it until after the release tomorrow.
pmurias ok 16:38
16:39 hoelzro is now known as hoelzro|away
timotimo sprunge.us/BLdj - this is my first attempt, but i'm getting a parse error (couldn't parse blockoid; unable to find "}"). i'll be afk for ~30 minutes, but if someone could point out the (probably) obvious flaw, that would be nice :) 16:41
moritz r: sprunge.us/BLdj 16:42
p6eval rakudo efac0f: OUTPUT«===SORRY!===␤Two terms in a row␤at /tmp/mH49DSaAlo:1␤------> http⏏://sprunge.us/BLdj␤ expecting any of:␤ argument list␤ postfix␤ infix or meta-infix␤ infix stopper␤ statement end␤ sta…
moritz grml
timotimo wha, how did that happen?
moritz it tried to parse the URL
timotimo hah, ok
locally i get │elements() not implemented in class 'Parcel' on line 10
moritz automagic fetching only works for gist URLs, iirc
timotimo unlike for gist, i have a super handy global alias "SPR" that adds | sprunge to any command, which i find super helpful 16:43
it even copies the url into my clipboard for me
bbl now.
moritz timotimo: wait, that's NQP code, right? 16:46
and you're running it with NQP?
if yes, I don't see how you can end up with a Parcel
timotimo i was running it with nqp, but to paste that error, i ran it with rakudo instead (like you tried with evalbot) 16:48
rakudo seems to parse the code just fine without any worries about missing or unreachable "}" 16:49
jnthn Try && instead of and 16:51
moritz if you write it as nqp code, you have to run it with nqp
for example nqp::elems returns a Parrot integer
you can't simply use that in Rakudo
(needs boxing)
jnthn r: say nqp::elems(nqp::list()).WHAT 16:52
p6eval rakudo efac0f: OUTPUT«Int()␤»
jnthn Since QAST it's actually a bit smarter.
Since QAST supports per-hll configurable boxing.
But that relies on you calling something that returns a native integer (which nqp::elems does) 16:53
moritz oh wow :-)
jnthn Unboxing still needs a lot of care though, to not hit the v-table call :) 16:54
17:09 SunilJoshi left
timotimo yay, turning and into && brings me into run-time-error-land \o/ 17:10
17:10 SamuraiJack joined 17:11 MayDaniel joined
timotimo totally works now. thanks! 17:14
but the result is wrong. strange. 17:16
now i think i've got a correct version. still need to prettify.
17:19 sjn left 17:20 sjn joined 17:26 For-Odin joined 17:28 aindilis joined
timotimo can i have an optional parameter in nqp without using something like *%adverbs? 17:31
jnthn $foo? should work
moritz yes
timotimo cool :)
and then i can do if $foo.defined { ... }?
jnthn nqp::defined($foo) 17:32
17:32 ifim joined
dalek p/mop-cleanup: 56d2292 | jnthn++ | src/how/NQPClassHOW.pm:
Remove outdated comment.
17:43
p/mop-cleanup: 5826673 | jnthn++ | src/QAST/Operations.nqp:
Provide nqp:: ops for low-level iterator usage.

NQP will continue to support the $_.key/$_.value approach for hash iteration in normal code. However, off Parrot this will be provided through the NQPCORE.setting, and thus is not available to the MOP. Essentially, these ops are to saw a circularity that we got away with up until now, but isn't really sensible to burden other backends with.
p/mop-cleanup: 1bcaac4 | jnthn++ | src/QAST/Operations.nqp:
Add ops for get/set of a coderef's name.
p/mop-cleanup: 3b7e5a3 | jnthn++ | src/stage0/ (9 files):
Update bootstrap.
p/mop-cleanup: ce967e9 | jnthn++ | src/how/ (5 files):
A variety of circularity-sawing MOP updates.
17:45 zby_home_ joined
dalek p-jvm-prep: c148d02 | jnthn++ | / (2 files):
Type check and method cache related ops.
17:48
p-jvm-prep: fc4f7e4 | jnthn++ | lib/QAST/JASTCompiler.nqp:
Fix copy/paste fail.
p-jvm-prep: 08c659e | jnthn++ | / (3 files):
Implement and test setcodename/getcodename.
17:56 dakkar left
timotimo my "bounded levenshtein" optimisation is totally exploding in my face :( 17:57
tadzik hello #perl6 \o/
[Coke] wonders what time friday the contest is over? 17:58
n: say /:ignoremark ü/ ~~ "u" 17:59
p6eval niecza v24-17-gd343a2a: OUTPUT«===SORRY!===␤␤Action method mod_internal:sym<:m> not yet implemented at /tmp/_Awmp2R1VH line 1:␤------> say /:ignoremark⏏ ü/ ~~ "u"␤␤Unhandled exception: Check failed␤␤ at /home/p6eval/niecza/boot/lib/CORE.setting line…
[Coke] jnthn: how big a lift is :ignoremark for rakudo?
moritz depends on how easy it is to get an NFD representation of the string 18:00
then you can compile 'ü' to u\p{Mark}* # in p5 terminolog 18:01
18:02 SamuraiJack left, wk left 18:05 SamuraiJack joined
TimToady [Coke]: um, your ~~ is backwards 18:09
not that it would suddenly cause :m to be implemented the other way around... 18:10
18:13 kaare_ joined 18:16 japhb is now known as japhb_ 18:17 japhb_ is now known as japhb__, japhb__ is now known as japhb 18:19 japhb_ joined, am0c left 18:21 spider-mario joined
[Coke] TimToady: whoops, aye. 18:21
18:22 japhb_ left
kthakore lol ... TimToady how can ~~ be backwards? 18:23
PerlJam [Coke]: R is your meta-friend :) 18:24
18:26 b1rkh0ff left
jnthn [Coke]: Not too easy, I suspect 18:26
TimToady kthakore: ~~ is not a symmetrical operator since about 10 years ago 18:31
unfortunately p5 copied the original symmetrical design after we'd already changed it...
PerlJam kthakore was just attempting literal humor mrthinks 18:33
er, s/mrth/meth/
timotimo i hopeit won't be too hard to get a sensible list of candidates for the is trait operator 18:34
jnthn r: say &trait_mod:<is>.candidates 18:35
p6eval rakudo efac0f: OUTPUT«sub trait_mod:<is>(Mu:U $child, Mu:U $parent) { ... } sub trait_mod:<is>(Mu:U $type, :rw(:$rw)!) { ... } sub trait_mod:<is>(Mu:U $type, :nativesize(:$nativesize)!) { ... } sub trait_mod:<is>(Mu:U $type, :hidden(:$hidden)!) { ... } sub trait_mod:<is>(Attribute:D $at…
timotimo now get the list of all classes that one could derive from:)
jnthn Well, that's less simple... :P 18:36
timotimo hehe
it is required to improve the misspelt class error message 18:37
leaving the tram bbl
18:38 b1rkh0ff joined 18:55 FROGGS joined 18:58 SamuraiJack left
dalek p-jvm-prep: 5e9c7c3 | jnthn++ | lib/QAST/JASTCompiler.nqp:
Implement defor (just desugar it for now).
18:58
p-jvm-prep: 334179b | jnthn++ | / (2 files):
Cheating nqp::where.

But good enough for what we'll need it for, for a while. :-)
p-jvm-prep: 18812aa | jnthn++ | lib/QAST/JASTCompiler.nqp:
Fix non-void while code-gen bug.
19:00 thou left 19:01 sahadev joined
jnthn Turns out the 1450 or so lines of nqp-mo.pm that we can now cross-compile boil down to about 88KB of JVM bytecode. :) 19:02
tadzik that's a big deal :) 19:03
arnsholt Cool. How much Parrot bytecode does it compile to? =)
jnthn It's not really fair to compare yet, because I haven't got all of the serialization stuff in place.
The PBC file comes out to 300 KB 19:04
19:04 For-Odin left
arnsholt Right 19:04
jnthn And the serialization blob itself is about 14KB.
But there's some fixup code to come.
Should still end up about half the size.
19:04 wk joined
arnsholt Nice! 19:05
jnthn Which ain't bad considering I know some of the code-gen is decidedly improvable. :)
arnsholt Even better =)
nwc10 hangon. nqp-mo.pm is only 1572 lines here
jnthn nwc10: Yeah, I'm ~100 off 19:06
nwc10: About 30 are lines we really need :)
And the others are Parrot-specific (the vtable mapping stuff)
nwc10 and then what - drink beer?
19:07 telex left
jnthn I'm afraid that there's about 2 or 3 uncommentable lines worth of easy win left. 19:07
And that's one of the LHF tasks.
The rest involves porting a good chunk of the bounded serialization.
nwc10 so drink gin? 19:08
or does it really need meths?
jnthn Well, it also needs me to get a good night's sleep so I can think clearly, which has been hard to come by this week. :(
19:08 telex joined
jnthn Time for some dinner here, bbl 19:09
arnsholt Porting bounded serialization sounds less low-hanging, yeah 19:14
19:33 cognominal left 19:39 japhb_ joined 19:40 snearch joined 19:42 cognominal joined 19:43 benabik left 19:47 arlinius left 20:12 sahadev left 20:19 zby_home_ left
[Coke] yawns. 20:21
20:21 cognominal left 20:23 eternaleye_ is now known as eternaleye 20:24 benabik joined 20:26 cognominal joined
dalek rl6-roast-data: 9981643 | coke++ | / (3 files):
today (automated commit)
20:30
[Coke] sorear: S04-declarations/my.niecza aborted 75 test(s) 20:31
I'll see if i can track that down after $dayjob
20:34 SmokeMac_ left 20:39 hoelzro|away is now known as hoelzro, cognominal left 20:41 Targen joined 20:44 hash_table left 20:46 cognominal joined 20:47 bluescreen10 left 20:53 arlinius joined 20:56 bluescreen10 joined 21:07 isBEKaml joined
isBEKaml hi 21:07
21:07 cognominal left
Ulti isBEKaml hi 21:09
isBEKaml hello, Ulti 21:13
21:14 cognominal joined
isBEKaml jnthn, moritz: If you're around, do we need to bump PARROT_REVISION in nqp? 21:15
masak would guess "no", but is happy to wait for someone else to confirm this 21:18
isBEKaml hello, masak... We still have to think up a name for this release. :-) We have Sonoma.pm available (japhb++'s home base) Any objections to using it? 21:20
masak sounds good to me. 21:21
21:22 thou joined
isBEKaml I'll wait for others to chime in too. We still have about 2.5 hrs to cutting a release. :-) 21:22
timotimo is there a nicer way to get user input in nqp than my $stdin := pir::getinterp__P().stdin_handle();
jnthn timotimo: probably not 21:25
isBEKaml: No, and if there hasn't been testing much on anything newer than what's in PARROT_REVISION, it's probably best not to
isBEKaml jnthn: Okay, I'll just bump the VERSION then. 21:26
pmurias jnthn: I have started stealing tests from nqp-jvm ;) 21:38
dalek kudo-js: 8133272 | (Paweł Murias)++ | / (5 files):
steal and pass t/qast_literal.t
spider-mario rust developers use the name “sigil” in their commit messages 21:39
I find it interesting.
(they’re actually applied to types, not variables)
(well, they *can* be applied to variables but it’s not part of their names) 21:40
(whereas it’s part of the type of the result of applying them to a value)
21:41 benabik left
jnthn pmurias: nice :) 21:41
21:44 hoelzro is now known as hoelzro|away
pmurias spider-mario: what do they do? 21:48
spider-mario they are used to make pointers
but there are different types of pointers
@, & and ~
@int means “managed (or shared) pointer to int”
& means borrowed pointer
and ~ means unique pointer 21:49
masak I like that.
they are using the sigils for a distinction they want to make in a very "core" place of the syntax. that's exactly how to thing of sigils, IMO. 21:50
spider-mario appearently, the recommended term for unique pointer is rather “owned pointer”
isBEKaml damn, is make -jx a bad idea for nqp?
masak well, "owned" rhymes better with "borrowed" and "shared".
'night, #perl6 21:51
isBEKaml: good luck with le release.
spider-mario good night
isBEKaml masak: night, masak. :)
jnthn 'night, masak
isBEKaml non-zero exit status in t/nqp/46-charspec.t. is that expected? 21:55
jnthn isBEKaml: yes, if you didn't build Parrot with ICU
isBEKaml jnthn: I don't have icu. 21:56
jnthn: so I'll ignore it.
22:03 slavik1 left 22:05 kaare_ left 22:13 MayDaniel left
rindolf jnthn: hi. 22:20
22:22 PacoAir left
dalek p: f7c5f75 | isBEKaml++ | VERSION:
Bump VERSION to 2013.01
22:24
jnthn hi, rindolf
rindolf jnthn: what's up?
Hi all.
jnthn rindolf: Not much, just wondering why in 2013 getting printer drivers installed can take 30 minutes...to fail. 22:27
rindolf jnthn: sucks. :-(
jnthn: are you using Windows?
jnthn rindolf: On the desktop, yeah.
rindolf jnthn: ah, OK. 22:28
diakopter jnthn: usually if you hunt carefully there's a way to install just the drivers without all the accompanying software
jnthn diakopter: Yeah, the added fun is that it's a network printer... 22:29
Quite why I've ended up with a network printer *for home use*, I don't know. :)
On the upside, a bunch of boring admin work I've put off for months is now done... 22:30
22:35 pmurias left 22:38 Chillance left 22:39 leprevost joined 22:40 leprevost left
isBEKaml nqp release done. Where do I upload nqp tarball? It can't be github now, right? 22:42
timotimo hm. i wonder how to decide when to use levenshtein guessing for the "class A is NoSuchClass" case and when to give the regular candidate list for trait-mod "is" 22:45
maybe i should look for the place where that error message originates first 22:46
22:46 rindolf left
timotimo that reminds me that i have no idea how to figure stuff like that out :| 22:47
22:49 spider-mario left 22:53 slavik1 joined
timotimo r: class A is NoSuchClass { } 22:57
p6eval rakudo efac0f: OUTPUT«===SORRY!===␤Cannot call 'trait_mod:<is>'; none of these signatures match:␤:(Mu:U $child, Mu:U $parent)␤:(Attribute:D $attr, :rw(:$rw)!)␤:(Attribute:D $attr, :readonly(:$readonly)!)␤:(Attribute:D $attr, :box_target(:$box_target)!)␤:(Routine:D $r, …
timotimo oh, this is in c code >_< 22:58
jnthn timotimo: Earlier today I showed a possible extra trait_mod:<is> candidate to add to catch situations where things go wrong. 23:05
timotimo decommute&
jnthn timotimo: An alternative is to look in World where traits are applied 23:06
isBEKaml jnthn: I can't seem to run tools/contributors.pl. :( 23:10
I get this error: List form of pipe open not implemented at tools\contributors.pl line 16.
line 15 actually.
23:13 snearch left
isBEKaml 2013.01 rakudo release draft: gist.github.com/4551934 23:18
Please proof read and let me know if anything's amiss. 23:19
23:20 bluescreen10 left
jnthn The next release of Rakudo (#60), assuming the world doesn't end today, is 23:22
scheduled for February 21, 2013.
...I think the reference to the end of the world doesn't need to be in this month's release ;)
isBEKaml jnthn: it's down below. ;) 23:23
jnthn isBEKaml: There's...another doomsday conspiracy? ;) 23:24
isBEKaml jnthn: Mayans never tire themselves out. :P 23:25
lue more critically: line 4: January 2013 release of Rakudo Perl #60 "Sonoma"
line 53: The next release of Rakudo (#60),
jnthn ay, good catch 23:26
lue++
isBEKaml lue++ 23:27
jnthn++
lue Also a bit of an accuracy thing, the Mayans never predicted diddly-squat (it's the modern-day nutjobs we're laughing at)
isBEKaml updated gist 23:28
ponbiki .win 23 23:31
jnthn sleep & 23:32
isBEKaml 'night, jnthn 23:33
lue: Right, I have updated my gist to say "conspiracy theorists" :) 23:37
lue :)
23:39 benabik joined 23:41 hash_table joined
geekosaur what, the nutjobs haven't come up with a new prediction yet? 23:44
TimToady are you predicting that? 23:47
timotimo oh, jnthn went to bed :| 23:52
irclog.perlgeek.de/perl6/2013-01-16#i_6340095 - and following. so would this implementation, with a multi trait_mod:<is>(Mu:U $type, *%fail) be the best way? or should i try the levenshtein distance calculation instead? 23:53
23:54 hash_table left, stopbit left