»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.
timotimo that was exhausting, i feel kind of drained >_< 00:03
00:21 b1rkh0ff left 00:24 Bucciarati joined 00:25 dayangkun joined 00:34 b1rkh0ff joined 00:36 thou_ left 00:41 stevan__ left 00:49 pjcj left 00:55 jaldhar joined 00:56 pjcj joined 01:07 For-Odin joined 01:09 thou joined 01:12 b1rkh0ff left, anuby joined 01:16 dayangkun left 01:20 hypolin joined 01:22 xilo_ joined 01:46 Nom- left, Nom- joined 01:49 jaldhar left 01:52 jaldhar joined 01:55 bbkr_ joined, bbkr left 02:00 xilo_ left 02:17 benabik left, For-Odin left, benabik joined 02:20 dayangkun joined 02:21 For-Odin joined 02:39 For-Odin left 02:43 sivoais left 02:45 bluescreen100 left 03:16 Pleiades` left 03:19 lue left 03:22 Pleiades` joined
timotimo To compile on the JVM backend, QAST::VM must have an alternative 'jvm' - does not compute 03:36
03:36 sivoais joined 03:37 orafu left 03:38 orafu joined 04:05 lue joined 04:15 For-Odin joined 04:22 preflex left, preflex_ joined 04:23 preflex_ is now known as preflex 04:28 For-Odin left 04:29 noxgirl joined 04:32 noxgirl is now known as autumn 04:40 For-Odin joined 04:41 jaldhar left 04:46 Pleiades` left 04:51 jaldhar joined 04:53 Pleiades` joined 04:58 swarles joined
swarles r: "foobar" ~~ /foo<.after bar>/ 04:58
p6eval rakudo 35b2a9: ( no output )
swarles r: say "foobar" ~~ /foo<.after bar>/
p6eval rakudo 35b2a9: OUTPUT«#<failed match>␤»
swarles oh. These moments are why I dislike booting into windows 04:59
sorear swarles: Are you deliberately running a regex that can never ever match?
swarles I was wondering if <after> was a directive
since there is <before>
sorear It is
r: say "bxr foobar" ~~ /b.r<.after foo>/ 05:00
p6eval rakudo 35b2a9: OUTPUT«#<failed match>␤»
sorear r: say "bxr foobar" ~~ /b.r<?after foo>/
p6eval rakudo 35b2a9: OUTPUT«#<failed match>␤»
sorear n: say "bxr foobar" ~~ /b.r<?after foo>/
p6eval niecza v24-18-gaf64300: OUTPUT«Match()␤»
sorear n: say "bxr foobar" ~~ /b.r<?after foo...>/
r: say "bxr foobar" ~~ /b.r<?after foo...>/
p6eval niecza v24-18-gaf64300: OUTPUT«===SORRY!===␤␤Unsuppored elements in after list at /tmp/XOmDccn2XF line 1:␤------> say "bxr foobar" ~~ /b.r<?after foo...⏏>/␤␤Unhandled exception: Check failed␤␤ at /home/p6eval/niecza/boot/lib/CORE.setting line 1443 (…
rakudo 35b2a9: OUTPUT«｢bar｣␤␤»
swarles o-o 05:01
05:04 rurban1 joined 05:06 Exodist joined
lue
.oO(Wait, where did "bar" come from?)
 05:09
05:09 dalek left
lue ah, I see. 05:09
05:09 dalek joined, ChanServ sets mode: +v dalek
lue r: say "bxr foobar" ~~ /b.r \s b.r<?after foo>/ 05:10
p6eval rakudo 35b2a9: OUTPUT«#<failed match>␤»
05:26 telex left 05:27 telex joined 05:37 telex left 05:38 telex joined 05:41 anuby_ joined 05:42 anuby left 05:49 am0c joined 05:58 kaleem joined, SamuraiJack joined 06:02 sevin left 06:08 b1rkh0ff joined 06:09 vividsnow joined 06:26 rurban1 left 06:29 daniel-s_ joined 06:35 daniel-s left, cibs left 06:36 vividsnow left 06:51 Patterner left 06:52 kaleem left 06:55 Psyche^ joined, Psyche^ is now known as Patterner 07:01 am0c left 07:35 swarles left 07:38 rindolf joined 07:43 rindolf left 07:44 am0c joined 07:47 kaleem joined 07:51 kaleem left 07:56 kaleem joined 07:57 Timbus joined 08:03 Xedus129 joined 08:07 cibs joined 08:10 FROGGS joined
FROGGS morning 08:10
arnsholt o/ 08:11
jnthn: I won't break all the things if I make the natives use a different compose protocol, right? 08:13
08:14 Xedus129 left 08:15 thou left 08:21 alec__ left 08:29 brrt joined 08:35 alec__ joined 08:42 For-Odin left 08:46 alec__ left
Su-Shee good morning everyone. 08:47
FROGGS good morning Su-Shee
08:50 hoelzro|away is now known as hoelzro 08:52 hoelzro is now known as hoelzro|away 08:55 sqirrel joined 08:59 alec__ joined 09:02 fgomez left 09:03 fgomez joined
jnthn arnsholt: For the attribute compose protocol, you can safely update CStruct to use it. For the various native types, yes, you can make them use a different compose protocol. 09:04
arnsholt: Given they don't use one at all yet 09:05
09:05 rindolf joined
arnsholt Excellent 09:06
'Cause then the native type compose protocol can be "pass the number of bits I'm supposed to be" 09:07
jnthn arnsholt: I was thinking more a hash of stuff 09:09
arnsholt: A key of which can be "bits"
arnsholt: Because we may need to pass unsigned too 09:10
arnsholt That's probably more future-proof, yeah
09:11 dayangkun left 09:22 fhelmberger joined 09:36 sivoais_ joined 09:37 dakkar joined 09:42 brrt left 09:43 Psyche^ joined 09:47 Patterner left, Psyche^ is now known as Patterner
masak good forenoon, #perl6 10:03
jlaire what's up in Perl 6 land? 10:04
sorear good afternoon masak 10:05
10:05 hypolin left
masak .oO( we may be bonkers, but we do have the same definition of "up" as everyone else ) :P 10:05
diakopter masak: howdy
jlaire mmkey 10:06
sorear masak: my up is your sideways.
You're gravitationally stuck to the side of the Earth. It's so funny.
diakopter jlaire: be more specific? :)
10:07 anuby_ left
jlaire diakopter: I haven't followed Perl 6 in a little over a year, what have I missed? :) 10:07
moritz lots of fun
improved error speed, error messages, cool features, modules, doc.perl6.org
jnthn debugger ;)
moritz a perl 6 coding contest, an advent calendar 10:08
nwc10 most recent fun is that jnthn has started on getting NQP to also target the JVM
jnthn moritz: we...give errors more speedily? ;)
moritz much sanity and much insanity :-)
r: class A is Ing { }
p6eval rakudo 35b2a9: OUTPUT«===SORRY!===␤'A' cannot inherit from 'Ing' because it is unknown.␤Did you mean 'Int'?␤␤at /tmp/3NdcQfaXhJ:1␤------> ␤»
jlaire niiice
nwc10 aye. Is nice. 10:09
moritz see? we even have spelling correction for class names
FROGGS r: class A is Ant { }
moritz timotimo++ did it
p6eval rakudo 35b2a9: OUTPUT«===SORRY!===␤'A' cannot inherit from 'Ant' because it is unknown.␤Did you mean one of these?␤ 'Any'␤ 'Int'␤ 'int'␤␤at /tmp/e0Rb5FQ5IP:1␤------> ␤»
nwc10 github.com/jnthn/nqp-jvm-prep/blob...ocs/LHF.md
jlaire nwc10: that's very interesting, too
FROGGS r: say "abc".subt(0,2)
p6eval rakudo 35b2a9: OUTPUT«No such method 'subt' for invocant of type 'Str'␤ in block at /tmp/HT_7grMRkH:1␤␤»
FROGGS hmmm
it doesnt handle methods yet 10:10
moritz and it can't easily
FROGGS it cant? ohh, sad
moritz because at the time the error is translated into a typed exception, we only know the name of the class, we don't have the class itself
masak why is that? why do we lose the class itself? 10:11
moritz and since types can be lexical, we might find the wrong type, or none at all, if we try to look it up
masak: because parrot only thorws error strings, with no payload attached
well, it throws a number and string
masak oh hahahah Parrot
10:11 xinming_ left
masak suffice it to say that our goals and ideals today differ from those ten years ago... ;) 10:12
10:12 hoelzro|away is now known as hoelzro 10:13 xinming joined 10:14 odoacre joined
arnsholt The disadvantage of coming back to a project after a long time: "Your branch and 'origin/dyncall-sized-num' have diverged, and have 2 and 509 different commit(s) each, respectively." 10:17
moritz so, you merge 10:18
arnsholt Yeah. I'm just not sure what those 2 commits I hadn't pushed are supposed to do
Or why I didn't push them
masak moritz: I read arnsholt's thing and thought "so, you rebase". :) 10:23
10:27 SmokeMachine joined 10:30 SmokeMachine left
FROGGS arnsholt: you could push to another repo/branch to see these commits, before merging 10:36
dont know if one can look at commits locally
moritz sure one can :-) 10:38
FROGGS the question is: how?
moritz you can do almost anything locallz
s/z/y/
FROGGS git is a bit to powerful for my HEAD
moritz FROGGS: gitk for example
FROGGS some sort of kde frontend? 10:39
jnthn tk, I think
moritz right 10:40
tk, despite the name
the kde thing would be kgit :-)
(no idea if that exists)
FROGGS nope
doesnt exist
11:09 Su-Shee left 11:11 Su-Shee joined 11:22 kaleem left 11:48 wk left 12:17 brrt joined 12:34 vividsnow joined 12:37 rindolf left, rindolf joined 12:39 dayangkun joined
arnsholt ARGH! Python, why U no understand how Unicode works? 12:53
</vent>
12:53 kivutar joined
hoelzro arnsholt: what did it do wrong? 12:55
arnsholt hoelzro: String literals are still ASCII even if you specify coding=utf-8 =)
hoelzro maintains a Python file named grievances.py for this purpose
arnsholt Just had to vent
hoelzro oh
arnsholt =D#
hoelzro arnsholt: Python 2, though? 12:56
arnsholt Yah 12:59
hoelzro well, at least they fixed *that* in Python 3
Su-Shee now they just have to port everything to python 3 ;) 13:01
hoelzro heh
I'm pretty sure that the 2-3 split is in my file ;)
Su-Shee I still simply have no need for python..
not even with my open data stuff, because there's R. 13:02
arnsholt Yeah, I'm 1) a slave to my university's IT department for this, and 2) IIRC NLTK is still Python 2 only
hoelzro Su-Shee: the only Python I've written lately is the Perl 6 lexer for pygments
arnsholt *sigh* I really need to learn me some R
Su-Shee arnsholt: like more or less every other library/framework
hoelzro writing Python was ok, but it kinda gets on my nerves
Su-Shee arnsholt: it's a-ma-zing. seriously.
13:03 wk joined
Su-Shee arnsholt: also, "learn"? you just _use_ it, it's wonderful. 13:03
the only thing I 13:04
hey! enter key1
arnsholt PCRS =)
Su-Shee the only thing I'm going to use python for will probably the NLTK.
brrt Su-Shee: agreed with the R thing
Su-Shee (which should have been perl a long time ago) 13:05
brrt R is awesome
(and what is NLTK)
Su-Shee brrt: have you tried shiny? the web framework?
brrt: natural language toolkit
arnsholt I'm a PhD student in a field that at least pretends to care about quantitative analysis, so R is definitely something I should invest in ^_^
brrt oh, yes, that thing, is pretty nice
and, no
Su-Shee arnsholt: we're telling you it's a matter of a weekend. :)
arnsholt Excellent 13:06
Su-Shee brrt: try the examples and marvel :) www.rstudio.com/shiny/
arnsholt I'll try to get around to it at some point, then
Probably some time around May =)
brrt oh, that is fr'king awesome indeed
you have any idea what pain it has been to statistics in a web context?
probably you do 13:07
Su-Shee brrt: I've just started with R and had within two days a) a shiny example based on stupid simple open data and b) three charts :)
arnsholt Oh, neat!
brrt ive used R mostly for offline analysis
Su-Shee brrt: I know web very well so I know why shiny is cool.
arnsholt That is pretty much what it says on the tin. Shiny, shiny! 13:08
brrt for what its worth, p6 can easily do what R does
Su-Shee brrt: then all it's need is a smoothly to use framework ;)
brrt: I like the ease of R.
brrt at my university we thought some very non-programmer folks R
and it went pretty well 13:09
Su-Shee yes, I had to look up some weird list handling/merging/greping stuff but that's all
brrt: I totally see why.
13:09 SamuraiJack left
Su-Shee brrt: the R book I have: "programming R"? happens on maybe 20 pages of 400. the rest: "just use this and that" 13:10
brrt we used 'statistics using R'
Su-Shee brrt: also, I'm a political scientist by education, so I see programming still from a "humanities department" point of view sometimes :)
brrt: I have the R cookbook and R in action 13:11
brrt: I did some chart stuff for a comparison in D3.js - which is as nice as a js lib can get - and it's still at least double the amount of code...
arnsholt: also, in case you're interested.. R defines "lots of data, large dataset" in the tb range ;) 13:12
arnsholt Always good to know 13:13
I'm in NLP, so datasets can occasionally be kinda silly in size
(Even though that's rarely a problem for me personally)
brrt you know what is really, really, really a shame
that we can't call python code from perl
(i.e., NLTK code) 13:14
Su-Shee brrt: no, you know what really the shame is? that python guys went and wrote those kinds of frameworks and we perl folk didn't...
and nltk is old.
I've toyed with it end of the 90ies already.
brrt well, most people were busy writing advertisement delivery software i guess 13:15
Su-Shee still.. we had a decade to deliver.. 13:16
brrt true
13:23 vividsnow left
arnsholt Sigh. This return to Python reminds me of my semester of fighting Emacs =) 13:28
Su-Shee I pulled the modem cable to close emacs. then I heard of vi. 13:30
masak .oO( because closing vi is famously simple ) :P 13:32
13:36 sqirrel left
Su-Shee by then I had a tutorial :) 13:36
so I was well equipped to deal with a real unix. I knew ls, rm, mv and vi. ;) 13:37
13:42 cedrvint left, cedrvint joined 13:46 zoomby left 13:55 cedrvint left, cedrvint joined 13:56 cedrvint left 14:00 vividsnow joined
hoelzro is it possible to add arbitrary adverbs for strings? 14:04
in the style of Q:exec{...} 14:05
14:10 bluescreen10 joined
tadzik hello #perl6 14:11
hoelzro ahoy tadzik 14:12
FROGGS hi tadzik
tadzik what a day
14:13 skids left
timotimo good * 14:15
FROGGS hi timotimo 14:16
14:16 PacoAir joined 14:36 SunilJoshi joined
timotimo calling nqp nqp-jvm-cc.nqp --setting=NULL -e "nqp::say('hi from JVM')" gives me "To compile on the JVM backend, QAST::VM must have an alternative 'jvm'". all tests pass, though. what am i doing wrong? 14:36
jnthn timotimo: Using a too-old NQP 14:37
nwc10 oh, he beat me :-)
jnthn The cross-compiler tracks NQP master very closely at the moment.
timotimo OK
14:38 spider-mario joined
jnthn digs back into $dayjob after a walk 14:40
arnsholt Oh, now I see. Finally figured out my diverging histories 14:45
timotimo github has a nice view for that, btw 14:46
arnsholt Oh? 14:47
I didn't really want to push stuff onto github when I didn't know what kind of shape the code was in =)
14:51 matija joined
timotimo right, but you could have deleted it immediately ;) 14:53
FROGGS is there an nqp op that tells me if a variable is readonly? 14:54
14:54 SunilJoshi left
arnsholt Heh. make -j 4 test install had unintended consequences (durr) 14:56
FROGGS I got told not to use -j4 for install
was*
arnsholt Don't know about install, but being test install, it used one job to run the tests and the remaining three to install =D 14:59
15:03 kaare_ joined 15:07 spider-mario left 15:08 SunilJoshi joined 15:11 kivutar left, spider-mario joined
arnsholt adds printf statements to the NQP C code 15:11
timotimo hum. i'm blindly copying the MAIN from rakudo-debugger, but i have no idea what most of it means. what are the command line options for, for instance? can't find all of them in man perl6 or parrot --help :| 15:14
moritz there's also docs/running.pod in rakudo
timotimo not helpful :( 15:15
github.com/jnthn/rakudo-debugger/b...g.nqp#L445 - talking about those
15:16 stevan_ joined
moritz setting is for providing a setting explicitly 15:16
PerlJam timotimo: -c is compile only, -s is for an alternate setting, -I is to add to the module include path, -M is to load a module
moritz I: add to include path
PerlJam timotimo: (at a guess)
:-)
moritz M is for loading modules
and -c... what PerlJam++ said 15:17
timotimo and what is this mysterious "s"?
PerlJam timotimo: s == string
timotimo huh!
moritz timotimo: and docs/running.pod does talk about -c
the =s is a left-over from getopts
timotimo i see. never used that
now it makes sense, though 15:18
i thought that part was to make the compiler load a setting called "s"
PerlJam timotimo: it's very similar to Perl 5's Getopt::Long
timotimo never used perl5 either :P
PerlJam aye, but the docs for that are readily available.
timotimo good point. thanks! 15:19
15:19 brrt left
arnsholt jnthn: prod? 15:20
15:28 rurban1 joined 15:31 vividsnow left 15:35 magoo joined 15:36 magoo left
dalek kudo/nom: 14c1a6f | (Tobias Leich)++ | src/core/Str.pm:
RT #116224, Str.match and Str.subst set $/

In case the pattern is a regex match() and subst() will set $/. If it cant match, $/ will be a #<failed_match>. Setting $/ is protected by "try" because it might be readonly.		 15:50
15:50 sqirrel joined 15:51 stevan_ left, xenoterracide left
dalek ast: 7c6bc80 | (Tobias Leich)++ | S05-substitution/subst.t:
RT #116224 test for $/
 15:53
15:53 xenoterracide joined, skids joined, am0c left 15:56 vividsnow joined
dalek ast: 0c9b948 | (Tobias Leich)++ | S05-substitution/subst.t:
unskipping working tests
 15:56
16:02 rindolf left 16:04 awwaiid joined 16:05 Juerd left 16:07 Util left, hugme left 16:08 dalek left, PerlJam left, [Coke] left 16:09 p6eval left, pmichaud left, sergot left, tadzik left 16:10 Juerd joined, pmichaud joined, dalek joined, ChanServ sets mode: +v dalek 16:11 sergot joined 16:13 PerlJam joined 16:14 Util joined, hugme joined, ChanServ sets mode: +v hugme 16:15 [Coke] joined
dalek rl6-roast-data: 07ef9a8 | coke++ | / (4 files):
today (automated commit)
 16:15
16:17 p6eval joined, ChanServ sets mode: +v p6eval
FROGGS r: $/ = 42; say "abc".subst(/(.)/, {$0 x 2}); say $/ 16:17
p6eval rakudo 35b2a9: OUTPUT«4242bc␤42␤» 16:18
16:18 hash_table joined
FROGGS what? 16:18
16:18 stevan_ joined
FROGGS r: $/ = 42; say "abc".subst(/(\w)/, {$0 x 2}); say $/ 16:18
p6eval rakudo 35b2a9: OUTPUT«4242bc␤42␤»
moritz r: $/ = 42; say "abc".subst(/(\w)/, -> $/ {$0 x 2}); say $/
FROGGS r: $/ = 42; say "abc".subst(/(\w)/, {$0 xx 2}); say $/
p6eval rakudo 35b2a9: OUTPUT«42 42bc␤42␤»
rakudo 35b2a9: OUTPUT«aabc␤42␤»
FROGGS r: $/ = 42; say "abc".subst(/(\w)/, {$0 x 2}, :g); say $/ 16:20
p6eval rakudo 35b2a9: OUTPUT«424242424242␤42␤»
FROGGS hmmm
jnthn arnsholt: pong
FROGGS r: $_ = 'a b c'; s:g[ (\w) ] = $0 x 2; say $_
p6eval rakudo 35b2a9: OUTPUT«aa bb cc␤»
FROGGS r: $_ = 'a b c'; s:g[ (\w) ] = $0 x 2; say $/ 16:21
p6eval rakudo 35b2a9: OUTPUT«｢c｣␤ 0 => ｢c｣␤␤»
16:21 sivoais left
FROGGS n: say "a b c".subst(/(\w)/, {$() x 2}, :g); 16:22
p6eval niecza v24-18-gaf64300: OUTPUT«aa bb cc␤»
16:23 sivoais_ left, sivoais joined
FROGGS r: say 'a'.subst(/(.)/,{$0~$0}) 16:23
p6eval rakudo 35b2a9: OUTPUT«use of uninitialized value of type Any in string context in block at /tmp/vNqlx_Om0m:1␤␤use of uninitialized value of type Any in string context in block at /tmp/vNqlx_Om0m:1␤␤␤»
FROGGS rakudo didnt get recompiled 16:25
:/
moritz 35b2a9 looks pretty recent to me
FROGGS but it works locally 16:26
timotimo different nqp perhaps? ;)
FROGGS I checked out NQP_REVISION 16:27
moritz bah, feather is SLOOW
FROGGS not my fault?
I hope
moritz I have no idea, I haven't had a chance to run any diagnostics yet
FROGGS 35b2a9 is from before my commit
moritz load average: 13.41, 8.26, 3.92 16:28
FROGGS uhhh
16:28 SamuraiJack joined
timotimo swappy death? 16:29
moritz yes
it's a pugs process
FROGGS \o/ not my fault \o/
moritz which consumes a lot of memory on a $veryBigRange.pick test
moritz killed it 16:30
FROGGS ... with fire
[Coke] that was me, trying to figure out a failure from yesterday's run.
moritz [Coke]: and my fault
[Coke]: in 53c8fe6a824937a1769cc7eeadb4409cde0d17eb I accidentally deleted a pugs fudge 16:31
along with the rakudo fudge
dalek ast: 44c5f4d | moritz++ | S32-list/pick.t:
restore pugs fudge
 16:32
moritz triggers a fresh pugs rebuild 16:33
16:34 tadzik joined 16:40 SunilJoshi left 16:41 SunilJoshi joined
FROGGS r: $/ = 42; say "abc".subst(/(\w)/, {$0 x 2}); say $/ 16:41
p6eval rakudo 14c1a6: OUTPUT«aabc␤｢a｣␤ 0 => ｢a｣␤␤»
FROGGS r: $/ = 42; say "abc".subst(/(\w)/, {$0 x 2}, :g); say $/
p6eval rakudo 14c1a6: OUTPUT«aabbcc␤｢c｣␤ 0 => ｢c｣␤␤»
masak there's something oddly eerie about how removing a line of fudge can slow down the server many of us are running screen/irssi on.
FROGGS yay
masak it feels like a violation of the levels of reality somehow.
moritz a well-placed ulimit would have prevented that 16:42
FROGGS .oO( and then when moritz repairs it you see the cat walking twice from right to left ... ) 16:43
[Coke] moritz: I wasn't running the test under ulimit by hand. the daily spec does, however.
moritz [Coke]: that's a good first step 16:44
16:45 rurban1 left 16:46 SunilJoshi left, thou joined 16:47 SunilJoshi joined
[Coke] let me know if there's something else that needs adding to the daily runs. 16:47
moritz I'd just like to encourage users to put a ulimit of maybe 1.5G into their .bashrc or whatever 16:48
because it's usually the process you don't expect to blow up that kills you
and when does a legitimate process on feather need > 1.5G anyway? 16:49
16:49 FROGGS left
timotimo 1.5G should be enough for everybody! 16:51
moritz on feather, yes :-) 16:52
16:57 awwaiid left 16:58 stevan_ left 17:12 SunilJoshi left 17:16 fgomez left, sizz_ left, fgomez joined 17:18 hash_table left, Chillance joined 17:19 MayDaniel joined 17:22 hash_table joined, sizz joined 17:24 FROGGS joined
FROGGS r: my $s = "abc"; say $s ~~ s[d] = "e"; say $s 17:33
p6eval rakudo 14c1a6: OUTPUT«True␤abc␤»
FROGGS n: my $s = "abc"; say $s ~~ s[d] = "e"; say $s 17:34
p6eval niecza v24-18-gaf64300: OUTPUT«False␤abc␤»
FROGGS I have no idea how to fix that
... since that wrong True comes from 17:37
r: my $s = "abc"; say $s.ACCEPTS( "abc" )
p6eval rakudo 14c1a6: OUTPUT«True␤»
17:38 hoelzro left, rafl left 17:42 rafl joined 17:52 SamuraiJack left 17:55 integral joined, integral left, integral joined 17:57 vividsnow left, hoelzro joined 17:58 dakkar left 17:59 hoelzro left, integral left, jlaire left, sjn left 18:00 jlaire joined 18:04 Nom- left 18:05 sjn joined
masak FROGGS: yeah, that's kinda the problem. 18:06
18:09 stevan_ joined 18:12 kaleem joined, Nom- joined 18:13 wk left, stevan_ left 18:14 hoelzro joined 18:15 vividsnow joined, integral joined, integral left, integral joined 18:17 kaleem left 18:21 rindolf joined
rindolf Hi all. 18:22
FROGGS hi rindolf 18:23
rindolf FROGGS: what's up?
Elections tomorrow.
18:24 kaleem joined
FROGGS rindolf: need to lull my kids to sleep in a bit, and thinking about an rakudo internal problem 18:25
masak: here the modified string is stored: github.com/rakudo/rakudo/blob/nom/...s.pm#L5008 18:26
masak: using a special variable it might be solvable... is there a better way?
brb
masak FROGGS: I might have misunderstood the issue -- but to me it feels less about "how to implement this" and more about "how to make the semantics consistent between ~~ and s[]" 18:27
18:31 kaleem left 18:39 fhelmberger left
tadzik moritz: did you, by any chance, see the Death Star PR videos? 18:48
18:50 Timbus left, autumn left 18:51 sivoais left, MikeFair_ left 18:52 MikeFair_ joined
arnsholt jnthn: I've made a first attempt at implementing compose in P6int.c. From the looks of it, the function gets passed a list like the rest of the REPRs expect 18:54
Is that expected or unexpected?
jnthn arnsholt: What's passing it the list, ooc? 19:08
dinner, bbs
feel free to gist me a patch...
I need to spec these compose things a bit more...
19:13 rindolf left
arnsholt jnthn: I haven't gotten farther than seeing the problem, really 19:24
Had to do $dayjobby stuff =) 19:25
masak 'Unknown variable $dayjobby. Did you mean: $dayjob?' :P 19:33
19:34 rindolf joined, japhb_ joined
arnsholt =D 19:34
19:36 lacx joined
lacx ello perl peeps X) 19:36
masak y0 lacx 19:37
welcome to our little cave ^^ 19:38
19:38 rindolf left, vividsnow left
masak .oO( we've heard rumors that there are things beyond these shadows on the wall. can you confirm or deny this? ) 19:38
lacx hey masak X) 19:39
can neither confirm nor deny..
masak dang.
lacx were living in the Matrix anyway!
X)
19:40 rindolf joined
lacx I reckon If I say too much Agent Smith might come looking for me.. 19:40
masak .oO( "Mr. Lacxerson..." )
lacx hehe X)
masak lacx: how may I serve you? you're here for Perl 6, I presume. 19:41
lacx yea hehe, I actually came in with a serious question.. one you guys have heard a lot no doubt.. 19:42
is there an ETA for an official (stable) release of Perl6?
masak well, you're in the right place. 19:43
please specify the length (in minutes) of the answer.
lacx Ive got all day/night X) 19:44
masak excellent. complete story it is.
lacx lolzz cheers X)
masak by Christmas.
BAM!
:P
srsly though, it might be a while. 19:45
of course, I'm assuming you know that we have releases going on monthly of two compilers, and have been for the past four-five years. 19:46
so clearly *that's* not what we mean by "official (stable) release" here.
lacx yea I know that X)
masak right.
so, what *do* we mean? :)
any suggestions?
(yes, this is an interactive story. feel free to chip in.)
lacx im looking into it from the perspective of a rather llarge company/corporate client.. 19:47
masak right.
lacx i.e. heavy use of Perl (5) right now..
trying to take the decision whether to start a rather large project in Perl 5, or wait for Perl 6..
masak so, essentially, you want something that is as trustworthy of Perl 5, and with the same amount of features, and preferably not *much* slower.
diakopter that's about 3 years away, imho 19:48
masak lacx: from your last sentence, I would suggest Perl 5 immediately.
diakopter (with current levels of resources)
lacx yea stability and trust is key
masak lacx: not because I don't like Perl 6 -- I do -- but because I think you would conserve resources right now sticking with Perl 5.
lacx fair enough X) 19:49
cheers for your help btw all
masak lacx: that said, you might want to keep an eye on Perl 6. let's say diakopter estimates right, and we're three years away. then you can spend three years learning about Perl 6.
that will help you when we finally officially make our release by Christmas 2016 :D
(again, assuming that's when it will happen) 19:50
lacx hehe, I was also interested about what happens to Perl 5 when Perl 6 is out..
diakopter "we're not just usable; we're LTLTA"
lacx are they developed/supported equally?
masak when Perl 6 is finally out, all Perl 5 software will spontaneously stop and refuse to ever run again. :P
diakopter heh; Perl 5 has hundreds of times more resources
lacx will there be an EOL/time when Perl5 is just obsolete?
masak lacx: Perl 5 isn't going away anytime soon. 19:51
lacx hehe ;)
ok thts good to know X)
masak lacx: I count on being able to get Perl 5 jobs until I retire if I want.
and I'm not that old.
diakopter masak: you're old!
masak :P
lacx hehe oops [agewar]
lol X) 19:52
thanks for the reassurance masak X)
masak no worries.
stick around a bit, though.
I want to demo some cool stuff for you.
lacx ok cool X)
cheers 19:53
masak r: say "Introducing... $*PERL<name>!"
p6eval rakudo 14c1a6: OUTPUT«Introducing... rakudo!␤»
masak r: say "hi, lacx!"
p6eval rakudo 14c1a6: OUTPUT«hi, lacx!␤»
19:54 eternaleye_ is now known as eternaleye
masak r: sub postfix:<!>($n) { [*] 1..$n }; say 5! 19:54
p6eval rakudo 14c1a6: OUTPUT«120␤»
masak r: say 5! # look ma, no cheating
p6eval rakudo 14c1a6: OUTPUT«===SORRY!===␤Two terms in a row␤at /tmp/58bieX4dn5:1␤------> say 5!⏏ # look ma, no cheating␤ expecting any of:␤ infix or meta-infix␤ infix stopper␤»
masak hm.
std: say 5! 19:55
p6eval std 7deb9d7: OUTPUT«===SORRY!===␤Negation metaoperator not followed by valid infix at /tmp/DyXdT_OBxE line 1:␤------> say 5!⏏<EOL>␤ expecting infix or meta-infix␤Parse failed␤FAILED 00:00 42m␤»
masak jnthn: rakudobuggable?
lacx: you know about floating-point numbers, right?
lacx hehe nice X)
yea
masak lacx: and how sometimes you divide 1 by 3 and then multiply by 3 and you get something like 0.999998437 19:56
diakopter I'll float your point
19:56 vividsnow joined
lacx yeaa 19:56
masak lacx: well, Perl 6 fixes that.
r: .say for 0, 0.1 ... 1.0
p6eval rakudo 14c1a6: OUTPUT«0␤0.1␤0.2␤0.3␤0.4␤0.5␤0.6␤0.7␤0.8␤0.9␤1␤»
masak it's *exact*. 19:57
lacx wickedd
diakopter taylor series would get a bit.. .unwieldy
masak :P
diakopter: you can always opt out of the exactness.
lacx i do like the inclusion of strict typing X) 19:58
diakopter pragma preserve_not_all_the_informations;
masak lacx: oh, let's show a bit of strict typing.
r: subset Positive of Real where { $_ > 0 }; sub f(Positive $p) { say "yay!" }; f(10) 19:59
p6eval rakudo 14c1a6: OUTPUT«yay!␤»
masak r: subset Positive of Real where { $_ > 0 }; sub f(Positive $p) { say "yay!" }; f(-5)
p6eval rakudo 14c1a6: OUTPUT«Constraint type check failed for parameter '$p'␤ in sub f at /tmp/Tjhmgtj2Rj:1␤ in block at /tmp/Tjhmgtj2Rj:1␤␤»
lacx niceee X) 20:00
must be a huge rehaul..?
masak r: role Woofing { method woof { say "woof!" } }; role Barking { method bark { say "BARK!" } }; class Dog does Woofing does Barking {}; given Dog.new { .woof; .bark }
p6eval rakudo 14c1a6: OUTPUT«woof!␤BARK!␤»
masak lacx: Perl 6 is *the* large rehaul.
when people think of Second-System Syndrome nowadays, they think of us and shiver. :D 20:01
except when they think about Duke Nukem Forever, but I hear that shipped...
lacx hehe it could have been so good!
so OOP is built into the core now? 20:03
masak yeah, it's very central.
in fact, I would venture to say that we advance the state-of-the art on MOPs a bit.
(a MOP is an OO API to the OO system) 20:04
nwc10 The thing that I'm most curious about, is how gradual typing is going to pan out
masak nwc10: you mean performance-wise? 20:05
nwc10 yes
masak I'd certainly spend more time putting types in my Perl 6 code if I knew there was a performance win to be had from it :)
jnthn Error wise is also interesting.
nwc10 is it going to be "have your cake and eat it", in terms of flexibility between dynamic language, and static speed
jnthn r: sub foo(Int $x) { }; foo('lol') 20:06
p6eval rakudo 14c1a6: OUTPUT«===SORRY!===␤CHECK FAILED:␤Calling 'foo' will never work with argument types (str) (lines 1, 1)␤ Expected: :(Int $x)␤»
jnthn That's a compile time error.
masak \o/
lacx: have you heard about multi subs?
benabik I think it's working out reasonably well performance-wise based on the usages of things like int, num, and str in Rakudo's setting.
(Yes, I know some of it is circularity sawing.) 20:07
masak r: multi sub fib(0) { 0 }; multi sub fib(1) { 1 }; multi sub fib($n where { $n > 1 }) { fib($n-1) + fib($n-2) }; say fib(5)
p6eval rakudo 14c1a6: OUTPUT«5␤»
masak heh. 20:08
lacx masak: no I dont X)
masak r: multi sub fib(0) { 0 }; multi sub fib(1) { 1 }; multi sub fib($n where { $n > 1 }) { fib($n-1) + fib($n-2) }; say fib(8)
jnthn A lot of use of natives is for performance and memory reasons.
p6eval rakudo 14c1a6: OUTPUT«21␤»
20:08 Vlavv_ joined
masak lacx: basically, you can re-use the same name for a sub, but with different parameter lists. Perl 6 will call the right one. 20:08
(see above)
it gives a very Haskell feel to the language sometimes.
lacx Oh yu mean Polymorphism? X) 20:09
cool
masak polymorphism is when objects can be of different types.
but yeah, the idea is sort of same-ish.
tadzik I can never remember the meaning of "polymorphism"
it always feel like something completely natural, I can't even think of a counterexample to it
masak tadzik: apples and pears both going into a bowl of fruit. that's polymorphism.
tadzik: or when you call a method on an AST node without knowing exactly what type the AST node is, just that all types have the method. 20:10
tadzik yeah. Because they're like, common descendants of something
masak .oO( condescent of man )
tadzik: yes. but with roles, it's no longer restricted to the inheritance hierarchy. 20:11
r: role Fruit {}; class Apple does Fruit {}; class Pear does Fruit {}; my Fruit @bowl = Apple.new, Apple.new, Pear.new; say @bowl 20:12
p6eval rakudo 14c1a6: OUTPUT«Apple<918830095> Apple<918853156> Pear<918865463>␤»
lacx Ooops im thinking of overloading I think masak
flussence polymorphism is when you have a bunch of different stuff, and then you do OOP code to it and it all feels very enterprisey and vague
masak lacx: yeah. multis do overloading.
lacx cool X)
masak lacx: in Java or C# there's no special keyword for it. in Perl 6, you have to do 'multi [sub]' or 'multi method'.
arnsholt Also, in Java (and I think C#), overloaded methods are resolved at compile time. Perl 6 multis are resolved at run-time 20:13
masak flussence: polymorphism is replacing switch statements by method dispatch.
jnthn r: multi infix:<+>('bear', 'deer') { 'beer' }; say 1 + 1; say 'bear' + 'deer'
p6eval rakudo 14c1a6: OUTPUT«2␤beer␤»
flussence that's a better definition :)
lacx yea I know Java X)
masak arnsholt: surely in some sense even in Java, methods are bound at runtime.
arnsholt: and in C# they can be made to, if you do some crap with 'virtual'. 20:14
lacx well they can be, using Reflection I think..?
masak no no, not reflection.
by default.
dang, we should have a Java eval bot in here. :)
arnsholt The type-signature bit is resolved at compile-time in Java
masak (but it'd be quite some one-liners)
arnsholt Which can have hilarious and confusing consequences
masak :< 20:15
well, it's 2013. it's fine to ridicule Java, I guess...
tadzik kekekeke
a bad example of a language shipped with this nice JVM thing, eh? ;)
I still prefer it to C++
arnsholt So if you have meth(A a) and meth(B b), where A is a superclass of B, which variant is called depends on the type used in your code on the variables, not the run-time type of the argument
tadzik blergh 20:16
lacx will you be able to use Reflection with Perl 6?!
tadzik I really like how Go does this
felher Not to forget about type erasure :) <3 Java ...
tadzik it feels nice
is Reflection another buzzword I don't get which describes stuff I know? :)
arnsholt lacx: You can do all kinds of crazy stuff via the MOP, yeah
tadzik it sounds similar to introspection 20:17
flussence and we can certainly do that: github.com/flussence/perl6-XMMS2/b...p6xmms2#L7
jnthn tadzik: it is :)
tadzik shakes fist at buzzwords
masak lacx: the MOP is what allows you to do reflection.
lacx awesome X)
tadzik occam's buzzwords
masak r: class Dog { method woof {} }; say Dog.^methods
p6eval rakudo 14c1a6: OUTPUT«woof␤»
lacx tadzik: en.wikipedia.org/wiki/Reflection_(c...ogramming)
masak lacx: see that '^'? it means "I want to do reflection now". 20:18
lacx its a very powerful principle
tadzik flussence: whoa, that's cool :)
hm, so it's more than introspection
masak flussence: heh, nice. I did something very similar in crypt.
flussence
.oO( that bit of code seems ripe for making into a role )
 20:19
lacx good for frameworks/modular systems tadzik
tadzik well, it's something that's begging to be abused
20:19 sqirrel left
jnthn tadzik: But it's tasty, tasty abuse. Like Grammar::Tracer. :) 20:19
tadzik :D
yeah :)
G::Tracer is supercool 20:20
felher arnsholt: @method(A a)... : If this would use the runtime types, there is no need for such beautiful patterns like the visitor pattern... Wouldn't that be awkward ;) 20:21
lacx anything and everything is open for abuse really..
masak that's the spirit!
lacx: I gave a talk about that on the last YAPC:EU.
20:22 Rotwang joined
lacx masak: will have to give it a go X) 20:22
masak you can download Perl 6 from here: perl6.org
arnsholt felher: Yeah, that'd be a real shame! ^_^
lacx how are you going to promote Perl 6 usage? or just gonna let it be organic..?
masak lacx: well, we have to finish it first... :) 20:23
jnthn imho, the best promotion is people building nice things and getting stuff done in it :)
20:24 bruges joined
lacx masak: I mean, a lot of ppl might be dependent on Perl5 code and might be unwilling to make the switch? 20:24
Rotwang hi
masak lacx: they don't have to.
Rotwang: oh hai
Rotwang I have a question about building rakudo, what is the appropriate channel for that? 20:25
masak lacx: especially people with lots of existing code I have lots of sympathy for.
arnsholt This is a good place for all things Rakudo
masak lacx: maybe they would be happier just interfacing with newer Perl 6 code in some way.
nwc10 End Of Life on anything open source is semi-meaningless. It's not like anyone can forbid you from supporting it yourself, or paying someone *else* to support it for you
lacx for sure jnthn
20:25 bruges_ left
nwc10 not like closed source, where you don't get those choices - support is only from the original vendor 20:26
Rotwang rakudo version: 2012.12, parrot version: 4.10.0, nqp version: 2012.12, 64 bit linux, build hangs on: 20:28
./perl6 --setting=NULL --optimize=3 --target=pir --stagestats --output=src/gen/CORE.setting.pir src/gen/CORE.setting
Stage start : 0.000
how can I debug it?
nwc10 is it swapping?
IIRC it needs about 1G
Rotwang yes!
lacx masak: my company has lotss! 20:29
Rotwang nwc10: that was brilliant [;
so how much ram do I need to build rakudo?
nwc10 Rotwang: it might be obvious, but that question has been asked before
for a 64 bit system, last I measured, I think it peaks about 1G 20:30
masak lacx: yes, and so do many other companies. it's not self-evident at all that all that Perl 5 code should be re-written.
lacx Rotwang: perms?
jnthn 64-bit builds will happily swallow around 1GB.
nwc10 for 32 bit, the peak is something like .5G
jnthn And...yes, that.
Rotwang lacx: perms?
arnsholt I think 1G to 1.2G on my OS X last time I checked
nwc10 if you don't have that much, then the build will swap
arnsholt Rotwang: How long did you leave it for?
nwc10 it does complete, given time. Maybe a *lot* of time.
arnsholt It can easily take two minutes 20:31
masak om nom nom swap
tadzik hehe, two minutes
nwc10 good point
Rotwang arnsholt: a few hours
tadzik you guys have fast machines
nwc10 it can take a couple of minutes *without* swapping.
tadzik true that
arnsholt blushes
tadzik Rotwang: is that the freshest build around?
I mean, git HEAD and all that
arnsholt Right, at the hours mark, you're probably swapping
Rotwang tadzik: no, rakudo version: 2012.12, parrot version: 4.10.0, nqp version: 2012.12 20:32
20:32 bluescreen100 joined
Rotwang ok it took ~900MB of memory 20:32
thanks guys, now I see what is going on 20:33
actually I feel a bit ashamed that I didn't check memory consumption
at first
arnsholt It's an easy thing to overlook
Rakudo is really greedy when it comes to memory 20:34
jnthn It's less greedy than it used to be. At least the trend is in the right direction. :)
arnsholt Yeah, the trend is definitely good
jnthn But yeah, needs more dieting :)
arnsholt And compiling compilers is pretty resource-intensive. As anyone who's built GCC from source knows 20:35
20:36 bluescreen10 left
felher indeed. 20:36
20:41 kaare_ left
tadzik heh, I extended my laptop's RAM from 2 to 4 GBs for Rakudo's sake :) 20:44
it's a lot better these days
when nom was in the early days I was actually shutting down all the other programs to prevent it from swapping that never ended 20:45
flussence mine can do it (just about) in 1GB
20:47 xenoterracide left, bluescreen__ joined 20:50 bluescreen100 left 20:53 vividsnow left 20:55 rindolf left, rindolf joined 20:59 snearch joined 21:00 MikeFair_ left 21:02 Rotwang left
lue hello o/ 21:09
masak lue! \o/ 21:11
21:14 awwaiid joined
lue ooc, has any more discussion occurred on S32::Temporal that I missed? 21:15
21:15 kivutar joined
masak not from my quick scan of the backlog. 21:17
21:17 Rotwang joined
masak lue: re plugging in custom calendars -- I'm not too hot on the idea. it sounds like overgeneralization to me. 21:17
at some point, you gain exactly nothing from plugging into the existing DateTime framework. 21:18
I believe that point comes already when you start talking about lunar calendars or hexadecimal calendars or Klingon calendars.
lue Yeah. I can't recall what exactly I meant by that. But I agree that plugging in custom calendars is not something we need to worry about. 21:19
moritz we could rephrase the quetion 21:20
if you want to plugin in a different calendar, what would the framework do for you?
sorear what about lunisolar calendars? 21:21
lue
.oO(DateTime::Mayan)
moritz what about them?
sorear can we please put slightly more effort into supporting the Hebrew calendar than the Klingon calendar? :p
moritz you mean, none and nearly none? 21:22
sorear I was hoping for the other way around
moritz nearly none and none? 21:23
sorear yes
21:23 vividsnow joined
flussence
.oO( supporting the Hebrew calendar is easy... just fork a PHP process )
 21:24
moritz does adding a sentence about a calendar not being supported as "slightly more work"?
21:25 xenoterracide joined
masak it's nice to see that temporal bikeshedding is still alive and well ;) 21:26
anyway, I plan to add something like a .delta method into the spec when I find the tuits. 21:27
and then I might as well go and implement it, too :)
japhb_ sorear, I was going to make a joke that we may want to support Klingon calendars first because "Jews won't attack you with a bat'leth if you don't comply", until I remembered that the only person I know who speaks fluent Klingon is Jewish. There went that theory. :-)
flussence (oh wait, I misremembered that part of PHP - they have core functions for Hebrew *strings*, not *dates*...) 21:28
lue also votes for a more useful Duration object (i.e. one I can use to represent arbitrary lengths of time and use in arithmetic operations with DateTimes)
masak .oO( mmm, Hebrew dates ) 21:29
lue: yes, Duration is part of CPAN's DateTime.
lue: part of the reason the temporal spec doesn't have it yet is that civic time is so messy, and I'm not even sure Durations can be made to follow standard mathematical laws. 21:30
maybe they can, to the point where you can say "things either behave sanely, or throw an exception".
japhb_ masak, if by "standard mathematical laws" you mean associativity and suchlike, I'm pretty sure the answer is no. 21:31
(Once upon a time, /me tried to write an app that had Correct Time Behavior. OMG what a rat's nest.) 21:32
masak right.
it's fun to query old Perl hands about horror stories about time. they always have a few.
flussence rn: say now R- now
p6eval rakudo 14c1a6: OUTPUT«0.0215852␤»
..niecza v24-18-gaf64300: OUTPUT«0.00015306472778320313␤»
moritz rn: say 0.0215852 / 0.00015306472778320313 21:33
p6eval niecza v24-18-gaf64300: OUTPUT«141.02007897320871␤»
..rakudo 14c1a6: OUTPUT«141.020078973208718135␤»
japhb_ rn: say now - INIT now
p6eval niecza v24-18-gaf64300: OUTPUT«Unhandled exception: Cannot use value like Instant as a number␤ at <unknown> line 0 (ExitRunloop @ 0) ␤ at /home/p6eval/niecza/lib/CORE.setting line 293 (Any.Numeric @ 6) ␤ at <unknown> line 0 (ExitRunloop @ 0) ␤ at /tmp/nsvCTMTS9z line 1 (mainline @ 3…
..rakudo 14c1a6: OUTPUT«0.020357␤»
moritz nr: say now - BEGIN now
p6eval rakudo 14c1a6: OUTPUT«0.0414784␤»
..niecza v24-18-gaf64300: OUTPUT«0.053300857543945313␤»
flussence REPL doesn't seem to like that version 21:34
japhb_ interesting Niecza bug there
lue Huh. Can I not use supersede with classes? 22:01
22:01 donaldh joined
masak 'night, #perl6 22:04
sorear night masak.
felher o/ masak
lue o/ masak
22:04 Rotwang left 22:05 skids left
lue how can I supersede a class? Rakudo doesn't seem to let me. 22:06
sorear You can't. It doesn't make sense. 22:09
lue according to S12, I can, so it seems to me a rakudobug. 22:11
(under Open vs Closed Classes) 22:13
22:15 japhb_ left, japhb_ joined
japhb_ moritz, was there a reason that the irclogs don't show the colors produced by p6eval error messages, or is it simply a matter of lacking tuits? 22:18
sorear I thought they did/used to !? 22:25
japhb_ Hmmm, not looking like it to me. Maybe because I have JS off, if it's a client-side post-processing? 22:27
FROGGS I see colors 22:29
lue irclog colors work for me (even when I turned off JS for perlgeek.de) 22:31
japhb_ FROGGS, within the error messages, as in: irclog.perlgeek.de/perl6/2013-01-21#i_6357210 ?
22:31 vividsnow left
FROGGS japhb: no, there are no colors, just here: irclog.perlgeek.de/perl6/2013-01-21#i_6356683 22:32
flussence I see lots of mojibake
japhb_ FROGGS, oh interesting. I wonder if Rakudo isn't producing colors in exactly the same way as niecza, and thus not getting correct irclog treatment. (I'm guessing this because Rakudo's colored error support is much newer.) 22:34
FROGGS japhb: looks like the colored text is the one after the --------> and at the time you posted there is no such text
lue Huh. Yeah, I don't see *rakudo's* colors.
FROGGS irclog.perlgeek.de/perl6/2013-01-21#i_6359196
here 22:35
flussence
.oO( now look what you've done! the ircbot's gone off sulking )
22:35 _ilbot joined
japhb_ Yeah, looks like the long arrow was a trigger of some sort. 22:35
flussence: and yes, it would be nice to correct the mojibake as well. 22:36
lue That seems to be iso-8859-1 encoded text put through UTF-8 (I can't be sure though, because firefox won't let me switch to ISO-8859-1 on perlgeek :/) 22:37
22:38 hash_table left
japhb_ Why would iso-8859-1 text be coming out of error messages? We're using Unicode characters for things like the eject marker, so encoding at any point in iso-8859-1 makes little sense. 22:39
22:39 xinming left
benabik Looks much worse to me in iso-8859-1 22:39
FROGGS shel escapes for the colors?
shell*
lue well, I'm thinking UTF-8 ==> ISO ==> UTF-8, irc to storage to browser 22:40
jnthn Maybe it's the ANSI escape stuff
22:40 rindolf left, xinming joined
FROGGS lue: then these carets wouldnt be there: irclog.perlgeek.de/perl6/2013-01-21#i_6358267 22:41
22:41 aindilis left
lue that's true. jnthn's idea seems more right then. 22:42
22:42 aindilis joined
lue std's not getting all its colors either: irclog.perlgeek.de/perl6/2013-01-21#i_6359200 22:42
FROGGS WE WANT ALL COLORS! 22:43
FROGGS draws signs now
donaldh jnthn: I'm exploring nqp-jvm-prep 22:44
trying an nqp loop I get 'No support for result child yet' 22:45
lue r: say "\x[c3,82]"
p6eval rakudo 14c1a6: OUTPUT«Ã␤»
lue r: say [c3,82].decode
p6eval rakudo 14c1a6: OUTPUT«===SORRY!===␤Undeclared routine:␤ c3 used at line 1␤␤»
lue r: say [0xc3,0x82].decode
p6eval rakudo 14c1a6: OUTPUT«No such method 'decode' for invocant of type 'Array'␤ in block at /tmp/9l3R3bJDSh:1␤␤»
donaldh jnthn: Is that something it would be worth me trying to hack on? 22:46
japhb_ BTW, are people with terminal-based IRC clients seeing the ANSI colors in their chat window? I'm sortof wondering if p6eval should convert ANSI ==> mIRC, then have the irclogs understand mIRC codes coming from p6eval. That way we should see the correct look everywhere.
flussence I get them 22:47
(in irssi)
FROGGS is using XChat, so, no colors
tadzik no colours in weechat either
japhb_ is using xchat as well.
lue FROGGS: I use XChat too, and I recently got a perl script for colors
(ANSI, that is)
jnthn donaldh: That one is probably quite easy to add a test for and not too bad to hack on.
japhb_ lue, ah, now that would be useful.
Standard plugin?
donaldh jnthn: tests are QAST fragments, right? 22:48
jnthn donaldh: yeah
lue I believe I got it from here: lwsitu.com/xchat/ansi_color.pl
jnthn donaldh: You'd need to allocate a temporary to store the result in and the get it on the stack at the end.
It's probably a 5-10 line of code fix tops.
22:49 snearch left, matija left
jnthn donaldh: What code did you try, out of curiosity? 22:49
donaldh jnthn: once I figure out the opcodes :-)
nqp/examples/loops.nqp
FROGGS r: huh
p6eval rakudo 14c1a6: OUTPUT«===SORRY!===␤Undeclared routine:␤ huh used at line 1␤␤» 22:50
FROGGS cool
FROGGS haz colors now
22:51 xinming left, xinming joined
jnthn donaldh: There's a helper fresh that can do the temporary allocation and load_ins/store_ins. They all take a type, and do the appropraite thing :) 22:53
donaldh jnthn: thanks
22:54 Pleiades` left 22:56 Pleiades` joined
japhb_ r: test 23:01
p6eval rakudo 14c1a6: OUTPUT«===SORRY!===␤Undeclared routine:␤ test used at line 1␤␤»
dalek p-jvm-prep: 65301a3 | (Jonathan Stafford)++ | / (5 files):
implements stat
 23:03
p-jvm-prep: 9e4a767 | jonathan++ | / (5 files):
Merge pull request #9 from thecabinet/stat

implements stat
23:18 zby_home_ joined
japhb_ r: test 23:18
p6eval rakudo 14c1a6: OUTPUT«===SORRY!===␤Undeclared routine:␤ test used at line 1␤␤»
23:21 spider-mario left
donaldh “” 23:22
jnthn 'night, #perl6 23:24
sorear night jnthn
23:25 MayDaniel left, zby_home_ left 23:32 Chillance left 23:39 donaldh left
japhb_ r: test 23:41
p6eval rakudo 14c1a6: OUTPUT«===SORRY!===␤Undeclared routine:␤ test used at line 1␤␤»
23:43 Timbus joined
japhb_ Patch to ansi_color.pl xchat script so that it works when you're highlighted and when you're PM'ing with p6eval: gist.github.com/4590650 23:45
23:47 autumn joined 23:50 b1rkh0ff left 23:53 benabik left 23:55 benabik joined, sivoais joined
lue japhb_: thanks 23:59

Powered by Cro and the Raku® Programming Language.

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

Thank you!