»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.
sorear good * #perl6 00:12
timotimo: fwiw, you make bugs by mailing to [email@hidden.address] 00:15
There is no format; anything in the mail goes into the bug
niecza bugs go to github.com/sorear/niecza/issues 00:18
I guess if I'm serious about this I should get an org acct eventually ... :P 00:19
japhb_ The niecza README says Mono "2.8.2 or later is recommended for best performance". What's the scale of the difference from 2.6.7? A few percent? Half again faster? Several times faster? 00:26
sorear Around twice as fast when I last measured it, which was rather long ago 00:27
You need to set MONO_ENV_OPTIONS=--gc=sgen to fully realize it
japhb_ sorear, Is this in addition to 'RUN_CLR=mono-sgen'?
japhb_ Also, has anyone tested recently whether Niecza still works with Mono 2.4.x? (2.4.3.1 in particular; that's what's packaged on RHEL/CentOS/SL 6) 00:29
sorear japhb_: no, equivalent.
No.
japhb_ OK
What does Niecza have for doing non-blocking socket IO these days? 00:33
sorear japhb_: still nothing p6y, I keep forgetting 00:38
japhb_ sorear, awww. :-)
Is Panda currently able to work with *both* a Rakudo and a Niecza installed? 00:39
timotimo so, is there an official guide on how to pronounce "niecza"? 00:46
sorear cz as in czech republic 00:48
TimToady knee eh cha
sorear (I have some regrets about this, but the cost-benefit analysis does not support a name change)
TimToady remembers remarking at the now-or-never point 00:49
.oO("We can't change this; we have 11 users already.")
sorear japhb_: if you wanted to invent a Perl 6 non-blocking socket I/O system and prototype it using msdn.microsoft.com/en-us/library/sy...elect.aspx , that would be *awesome* 00:50
japhb_ sorear, *chuckle* I saw what you did there. 00:51
sorear (note: the CLR, like the Win32 API on which it is modelled, only supports selecting on sockets, not general FDs)
japhb_ The part of me that 10 years ago would have ranted at that last statement has instead gone off and mixed itself something tasty as a distraction. 00:54
TimToady the right way forward will be a built-in event loop with several event sources that can feed it 00:55
then we don't have to care how the system splits these things up
japhb_ TimToady, an AnyEvent sort of beast, but standard in Perl 6? 00:56
TimToady yes
japhb_ Hmmmm. I will cogitate on this later -- any prior art (besides of course the Perl 5 AnyEvent module) that I should consider? 00:57
TimToady Erlang? 00:58
Go?
TimToady Smalltalk? :) 00:58
japhb_ Wow, that's a rather larger space of prior art than I was expecting, but OK, maybe it's time to learn (yet another) language or two. 01:00
japhb_ mentally weighs "learn new language", "research GC algorithms", and "JFDI" as options for next hack session.
colomon japhb: you should be able to use panda with rakudo and panda-niecza with niecza at the same time. 01:02
japhb_ colomon, OK. Where is panda-niecza? 01:04
colomon github.com/colomon/panda 01:05
japhb_ colomon, How much do the sources drift between panda and panda-niecza? And is there a plan to unify them again? 01:07
colomon drift is slight but significant 01:08
there is a plan to have a plan to unify them again.
japhb_ colomon, *chuckle* OK, understood, thank you. 01:11
Tene rakudobug? 01:18
Bah.
sorear TimToady: when you get a chance I'd like to finish our thing on Nil and @array vars from last night 01:34
TimToady: niecza currently treats 'my @foo' as something like 'my Array $foo ::= []'; I get the impression you're treating it as 'my Array $foo # will autovivify when written to', is this correct? 01:35
TimToady is not actually taking a position, merely trying to understand the semantic options 01:55
japhb sorear, is there a common set of benchmarks that you use when assessing the performance of niecza? 03:48
sorear japhb: only self-compilation 04:12
Fuad Hi all 04:22
sorear Fuad: Hello! Welcome to #perl6. 04:27
Fuad sorear: Thank you! 04:28
sorear Oops. 04:29
Just checked my logs, you aren't actually new here.
Hi anyway.
shachaf sorear: Hello! Welcome to #perl6. 04:30
sorear shachaf: huh? 04:31
shachaf stops nonsensing. 04:32
sorear S09 seems quite low on tests.
Fuad sorear: yeah :) 04:45
dalek ecza: 1f35f89 | sorear++ | lib/ (2 files):
Implement Zen slices and @foo[1;2;3] access

Note that there are no actual shaped arrays yet; this is simply the
  @foo[1][2][3] equivalence per L<S09/Cascaded subscripting of multidimensional arrays>.
07:21
sorear TimToady: what is the intended semantics of my @array[**] ? 07:41
sorear wonders whether it would be more pragmatic to try and adapt niecza's existing Array and Hash classes to multi-D or to create new ones 07:42
moritz guesses that nobody here has much wisdom to share on that matter 07:45
and yes, I agree. Not many S09 tests
mostly because I haven't understood what most of S09 says
moritz -> GPW 07:47
sorear -> sleep 08:49
masak morgnung, #perl6 08:52
tadzik Guten Morgen, #perl6 08:53
jnthn GPW. Dag 2. 08:56
Oops. That wasn't German, was it... 08:57
jnthn clearly needs a second cup of coffee...
bonsaikitten jnthn: nee :) "Dag" is dutch, "Tag" would be proper german 08:59
masak yay! I just hit 100% preparedness for my talk. :) 09:39
with all of... 20 minutes left.
masak , master of procrastination
masak polishes some things 09:40
key_ hello to every one on this chanel, i am new to perl6, and also beginner in using IRC. i was wondering if templatelanguges (like template toolkit) could be compiled to parrot bytecode and than be used from within perl6 code? 09:44
tadzik possibly, yes 09:49
there's a Rosella library for templating
(Rosella is a library collection for Parrot) 09:50
key_ thank you 09:54
bacek ~~ 10:34
phenny bacek: 05 Mar 11:56Z <jnthn> tell bacek plesae don't change Makefile for now, I have a bunch of re-org of it in a branch, and I don't want to have to re-do that.
bacek: 05 Mar 11:57Z <jnthn> tell bacek kill_props_vtables branch should be fine
bacek: 05 Mar 11:58Z <jnthn> tell bacek oh, I see moritz++ already sorted stuff out wrt that branch :)
bacek jnthn, aloha. In which branch? bs?
moritz which reminds me, I still need to look at the new_props branch 10:37
it's just a simple argument reordering, right?
bacek moritz, yes
jnthn bacek: bs is merged 10:45
lunch & 10:47
bacek jnthn, ok, thanks. 10:48
bacek jnthn, moritz, can you look at gist.github.com/1985777? It's proposed changes in nqp Makefile. Life is too short to wait for useless rebuilds.. 11:26
moritz +1 if it works also with -j 12:03
jnthn++ is giving his talk right now :-) 12:04
rafl masak: are the slides for your macros talk available online? i didn't quite manage to attend earlier
bacek moritz, yes, it's step forward to parallel compilation. 12:06
moritz, I can start branch to do it bit more properly 12:11
moritz bacek: a branch would make it easier to apply and harder to forget
bacek moritz, deal :) 12:12
jnthn wonders if that Makefile change works on Windows... 13:03
If it's a branch, I'll try it.
[Coke] misses au! 13:24
[Coke] n: say 70/90.perl 13:47
p6eval niecza v15-4-g1f35f89: OUTPUT«0.77777777777777779␤»
[Coke] n: say 70/90
p6eval niecza v15-4-g1f35f89: OUTPUT«0.77777777777777779␤»
au /o 13:49
phenny au: 02 Mar 22:58Z <[Coke]> ask au given feather.perl6.nl/~coke/pugs.diff, all the method seem to work: I can do "3.Int.Num.Rat" -> 3/1 - however, when I say "Int" with this patch, I get an error about multis. pre-patch it returns "::Int"; any pointers?
au tries the patch 13:50
[Coke]: commit to a branch e.g. wip/typefix next time? 13:51
au does exactly that 13:52
dalek gs.hs/wip/typefix: 0285f84 | au++ | Pugs/src/Pugs/Prim.hs:
* WIP: Type methods for Int/Num/Rat from coke++
tadzik nom: Int.^methods.grep(*.name ne "").sort 14:03
p6eval rakudo 3c5838: OUTPUT«(signal SEGV)»
moritz tadzik: [X] Achievement unlocked 14:10
tadzik :)
it's all the _decontainerize() thingy 14:11
moritz it's in decontainerize from find_best_candidate 14:11
tadzik nom: Int.^methods.grep(*.name ne "").eager 14:12
p6eval rakudo 3c5838: OUTPUT«(signal SEGV)»
[Coke] au: you're back! 14:13
moritz perl6: say so "\n" ~~ /\v/
p6eval pugs b927740: OUTPUT«*** No such subroutine: "&so"␤ at /tmp/joxO82VRHP line 1, column 5 - line 2, column 1␤»
..rakudo 3c5838, niecza v15-4-g1f35f89: OUTPUT«True␤»
moritz perl6: say so "\r" ~~ /\v/ 14:14
p6eval pugs b927740: OUTPUT«*** No such subroutine: "&so"␤ at /tmp/TXLG8zNONa line 1, column 5 - line 2, column 1␤»
..rakudo 3c5838, niecza v15-4-g1f35f89: OUTPUT«True␤»
moritz perl6: say so "\r" ~~ /\h/
p6eval rakudo 3c5838, niecza v15-4-g1f35f89: OUTPUT«False␤»
..pugs b927740: OUTPUT«*** No such subroutine: "&so"␤ at /tmp/TS5s0QKXXK line 1, column 5 - line 2, column 1␤»
[Coke] do we announce git issues on github.com/perl6 projects here? 14:16
(should we?)
masak my autopun lightning talk contains 13 autopuns :) 14:17
[Coke] au: any pointers on where to look to resolve that issue? I could not figure out with casual inspection where "Int" was originally defined.
moritz masak: and is it self-referential? 14:18
masak the talk? I'm too confused to know. 14:19
there's a high chance it is.
au [Coke]: just a sec... 14:21
fsergot o/ 14:22
[Coke] au++ 14:23
tadzik \o
[Coke] karma au 14:24
aloha au has karma of 34.
[Coke] n: my $a="au"; $a++ for 1..34; say $a
p6eval niecza v15-4-g1f35f89: OUTPUT«cc␤»
sisar rant: there is no easy way to make a private gist public :( However, there is a workaround: chris.dzombak.name/blog/2011/05/mak...ist-public 14:32
jnthn heh, I thought it was about .gist in Perl 6 :P 14:34
moritz sisar: you can just share the URL
dalek gs.hs/wip/typefix: cb4d7e9 | au++ | Pugs/src/Pugs/Prim.hs:
* Line up return types of Int/Rat/Num and declare op0 fallbacks.
au [Coke]: please checkout wip/typefix and see if it passes a bit more tests, and merge back to master if there's no regression
moritz sisar: also wrt github I'd not rant, but simply write the friendly github staff with a feature request 14:35
sisar moritz: will do 14:37
moritz: hae a look: gist.github.com/1986576
*have
[Coke] au++ # roger.
sisar line 1 and line 10 are exactly equal
moritz std: for map { $_, $_}, 1..3 { .say;} 14:38
p6eval std 1ad3292: OUTPUT«ok 00:01 112m␤»
sisar moritz: weird, ain't it ?
tadzik nom: for map { $_, $_}, 1..3 {.say;}
p6eval rakudo 3c5838: OUTPUT«1␤1␤2␤2␤3␤3␤»
tadzik nom: for map {$_, $_}, 1..3 { .say;}
p6eval rakudo 3c5838: OUTPUT«1␤1␤2␤2␤3␤3␤»
tadzik oh, huh 14:39
[Coke] au: so far, so good. I'll re-fudge roast later today, and merge things back. Thank!
*Thanks! 14:40
sisar tadzik: any insight ?
au np, glad to help :)
-> &
tadzik sisar: nope
masak std: -> & {} 14:43
p6eval std 1ad3292: OUTPUT«ok 00:01 109m␤»
tadzik seen slunski 14:44
aloha slunski was last seen in 15405 days 14 hours ago .
tadzik I guess that's aloha's for "never"
masak seen sluszky
aloha sluszky was last seen in 15405 days 14 hours ago .
masak seen druszky
aloha druszky was last seen in 15405 days 14 hours ago .
masak seems so :)
moritz [Coke]: should pugs' "make spectest" pass? 14:45
dalek gs.hs/so: e0ec661 | moritz++ | Pugs/src/Pugs/Prim.hs:
try to implement &so
14:47
moritz [Coke]: I have no idea if that worked or not, but if it did, you can merge that branch :-)
masak moritz++ 14:52
au . o O { Travis CI for spectest } 14:53
PerlJam au: don't they have to "enable" perl 6 support? 14:54
au mm? a perl5 "test.pl" that builds and runs spectest and outputs TAP should work 14:57
flussence ooh, I have one of those 14:58
au s/test.pl/make test/
[Coke] moritz: there is no spectest target, but the tests mostly pass yes. might be a dozen failures or so. 15:01
[Coke] moritz++ #thanks. 15:01
masak autopun talk delivered. :) 15:14
colomon masak++
masak there was the appropriate amount of rejoicing ;)
tadzik it was awesome!
masak thanks. 15:15
sisar masak: upload slides ! 15:20
sisar has submitted a feature request to github about gist 15:21
masak sisar: ok, I'll try. 15:28
oh, Fuad was here this morning.
rafl: I'll put up the macro slides for you too. 15:31
sisar rafl ? 15:33
oh, never mind
masak he asked for them in the backlog.
sisar our Niecza $developer: gist.github.com/1986826. What is wrong ? 15:36
masak my two talks: github.com/masak/gpw-talks 15:38
grab them while they're hot.
I'll put them up properly on feather.perl6.nl/~masak/ once I get proper FTP access.
sisar masak++ 15:39
moritz masak: you can scp to feather, the ssh port is open here 15:40
oh nice, "age: in 14 minutes"
jnthn :D
Slides from the future 15:41
tadzik The Distant Future
moritz that's so fitting :-)
masak moritz: yeah, my VM's clock is still off.
moritz is in the conference office, at the place where we did the p6 hackathon
tadzik seen bbkr 15:42
aloha bbkr was last seen in #perl6 4 days 14 hours ago joining the channel.
PerlJam "dive-through example"? I like that. 15:43
moritz masak++'s autopun talk was really hilarious
PerlJam indeed masak++
masak :)
PerlJam I didn't get the "je ne sais quoi" one though ;-)
moritz having a whole room full of laughing geeks is a nearly self-amplifying 15:44
after that talk, everybody would have laughed about "chicken" too :-)
fsergot tadzik: What about your talk? :)
tadzik PerlJam: I guess that means "I don't know" or such
fsergot: wasn't too bad, it almost filled 15 minutes out of 20 :) 15:45
luckily, audience++ came to the rescue
I'm quite satisfied with it; wasn't perfect, but I see an improvement over previous talks I did
PerlJam masak: I *really* like slide 14 of your macros talk
jnthn My slides are at jnthn.net/papers/2012-gpw-meta-programming.pdf btw 15:46
moritz fwiw my slides are at moritz.faui2k3.org/files/p6-features/
jnthn Suddenly...slides!
moritz and I was lucky to get enough questions from the audience to fill my 40 minutes
tadzik mine is on tjs.azalayah.net/index.html if anyone wants to take a look 15:49
masak PerlJam: thanks :)
PerlJam: implementing macros means having a keen sense of compile/run moding. 15:50
moritz std: 'a' ~~ /a*+/ 16:10
p6eval std 1ad3292: OUTPUT«ok 00:01 111m␤»
masak yeah, *+ means *? 16:11
hm, or was it *: ?
I think the latter.
TimToady yes 16:12
dalek kudo/nom: 3ce390d | jonathan++ | src/Perl6/ (2 files):
Start collecting LEAVE, KEEP and UNDO phasers; not doing anything with them yet.
16:30
kudo/nom: a5b3ede | jonathan++ | src/core/Block.pm:
Keep a list that knows the order we should run LEAVE/KEEP/UNDO, since they should be considered as a single queue.
kudo/nom: 0dbe16a | jonathan++ | src/core/Block.pm:
LEAVE/KEEP/UNDO should run in opposite order.
kudo/nom: d4bb162 | jonathan++ | src/ (4 files):
Incorporate various of the patches from mls++, with various updates and tweaked phasers lookup, to get LEAVE essentially working. May not be perfect, but passes a bunch of tests.
kudo/nom: d31ad04 | jonathan++ | t/spectest.data:
Turn on a few tests that we now pass.
dalek ast: d48e69f | jonathan++ | S04-phasers/descending-order.t:
Fix a test that missed expectation of ENTER phaser.
16:30
ast: c95f0d1 | jonathan++ | S04-phasers/enter-leave.t:
Fudge for Rakudo.
tadzik \o/
dalek atures: fcacff9 | jonathan++ | features.json:
Rakudo now implements ENTER and LEAVE.
16:32
moritz nom: &x() 16:33
p6eval rakudo 3c5838: OUTPUT«===SORRY!===␤Symbol '&x' not predeclared in <anonymous> (/tmp/z6RSDHBBtm:1)␤»
masak submits rakudobug
dalek atures: 872a7d3 | tadzik++ | features.json:
Rakudo now has Set, Bag and friends
16:35
masak un-submits rakudobug
moritz 's fault
masak there's plenty of real bugs waiting for us :) 16:36
PerlJam jnthn: slide 45 says "the case class" instead of "the base class" 16:43
dalek ast: a553a91 | jonathan++ | S04-phasers/next.t:
Fudge for Rakudo.
PerlJam jnthn: slide 46 says: multi trait_mod:(Mu:U $type, ...) Looks like a missing <is> 16:44
jnthn: sorry, just now got around to checking out your slides for gpw :)
masak yes, moritz++ noticed that omission during the talk as well. 16:45
jnthn jnthn++ also noticed it during the talk
:)
dalek kudo/nom: ccd0ec1 | jonathan++ | t/spectest.data:
Run another test.
16:47
masak social & 16:51
jnthn socializing & 16:53
sorear good * #perl6 17:01
sorear [Coke]: 70/90.perl means 70/(90.perl) => 70/'90' => 70/90 17:02
sisar: I suspect that you did a git pull, then tried to run niecza again without 'make'. Usually this works, but if there have been additions to the C# api, it won't 17:04
PerlJam why is is "ParametricRoleHOW" instead of just RoleHOW? 17:06
s:2nd/is/it/
sorear because there are several kinds of role types 17:07
sorear you can't have a plain RoleHOW 17:07
PerlJam seems like an artificial distinction to me. 17:08
Why not a ParametricClassHOW too ?
or a ParametricGrammarHOW?
etc.
Seems like the simpler thing is to say that they're all parametric with the possibility of a null parameter 17:09
sorear PerlJam: All explicitly defined roles *are* ParametricRoleHOW 17:44
When you say MyRole[Int], that's CurriedRoleHOW - still not applied because no ::?CLASS available 17:45
sorear once ::?CLASS is available (at compose time), you have a ConcreteRoleHOW 17:45
PerlJam sorear: mostly I'm just quibbling over the name. I see no benefit to calling it "ParametricRoleHOW" versus just "RoleHOW" and it keeps the nice (perhaps foolish) consistency with the other HOWs. 17:56
TimToady but which one should be the unmarked form? you could just as easily argue that it should be the concrete one, sinceit's "closest" to the user 17:57
one could also argue that the unmarked from should be the base role from which the other roles derive 17:58
PerlJam I'd go the other way since HOWs are more abstract anyway
sjn Silly question... are there any obvious wikis we could use as a "notepad" for the Perl 6 hackathon in Oslo? 18:05
is the Perl Foundation wiki meaningful place for stuff like that? 18:06
(is it necessary to have a wiki?) 18:07
sorear sjn: the TPF wiki is the obvious one, but I have some misgivings about anyone using it for any purpose 18:44
sjn: the traditional solution here is to use the pugs/mu repo as a wiki 18:45
awwaiid could use a github repo wiki 18:47
[Coke] github++ - easy enough to create a temp wiki wthere. 19:06
sorear I wonder if it makes any sense to have a reified part/lazy part split on a multidimensional array. 19:24
[Coke] n: say (70/90).perl 19:32
p6eval niecza v15-4-g1f35f89: OUTPUT«<7/9>␤» 19:33
[Coke] sorear++
sorear [Coke]: hi. do you still make test summaries broken down by synopsis? 19:44
[Coke] that is a very manual process, so I only do it when someone asks and I have spare time. 19:47
sorear ah. 19:48
(I'm not asking right now)
[Coke] ping me after the week of the 20th. ;) 20:05
sorear Why then?
colomon recovered from St Patrick's day hangover? ;) 20:06
sorear o/ colomon
colomon \o
[Coke] a $dayjob induced brainfreeze 20:07
big project due that monday, in the UK to show off bits of it that week.
sorear ah.
colomon sorear: I'm hoping to have my 64-bit Windows machine in good enough shape to try niecza on it in another day or two.
sorear eeeexclent 20:08
sorear is still trying to figure out what the sanest representation of multi-d arrays and hashes is 20:09
colomon it needed a new motherboard, so I'm needing to reinstall a development environment from scratch on it.
[Coke] I just got a 64bit macbook to play with, will try various perlsixes on that. 20:10
(again, not just yet. ;) 20:12
colomon Once I've got the Windows machine running, I'm thinking I might reinstall linux on my big linux box and see if I can get everything working on that, too. 20:13
colomon expects he can't usefully help sorear++ figure out multi-d arrays and hashes, but is willing to be a sounding board if needed 20:17
fglock updates perlcabal.org/~fglock/perlito5.html 20:50
80% done, 80% to go
TimToady you can iterate that...
colomon has working SVN on his windows box!!! 20:55
[Coke] \o/ 20:58
perl6: say 0.Int 21:02
p6eval pugs b927740: OUTPUT«*** No such method in class Int: "&Int"␤ at /tmp/m_kXnSGXXJ line 1, column 5 - line 2, column 1␤»
..rakudo 3c5838, niecza v15-4-g1f35f89: OUTPUT«0␤»
colomon has working git (and github access) on his windows box! 21:05
PerlJam colomon: \o/ (*that* is cause for celebration :) 21:06
dalek p/new_make: ab98638 | bacek++ | 3rdparty/libtommath/Defines.mak:
Add included makefile for libtommath.
21:09
p/new_make: 8eddacf | bacek++ | src/6model/Defines.mak:
Start splitting 6model rules into separate makefile.
p/new_make: 492179a | bacek++ | tools/build/Makefile.in:
Remove bits from Makefile.in which were moved into separate makefiles.
p/new_make: 2a429c3 | bacek++ | src/6model/Defines.mak:
Add todo marker
sjn sorear: right, next question would be "is there a point doing it?" (or maybe that sho+uld have been the first one:-P) 21:14
sjn .u PILE OF 21:34
phenny sjn: Sorry, no results for 'PILE OF'.
sjn .u PILE
phenny sjn: Sorry, no results for 'PILE'.
sjn aw. no unicode 6.0? :-) 21:35
dalek p/new_make: c9ddedd | bacek++ | tools/build/Makefile.in:
Add includepath for 3rd party libraries
21:53
[Coke] suggests niecza and rakudo get giant ascii art in their startup screen. ;)( 22:05
er, ;)
PerlJam giant unicode art 22:10
:)
benabik The spec should require all implementations to output Camilla when starting the REPL. :-D 22:11
PerlJam benabik++
[Coke] that'd be an excellent repl prompt. »ö« > say ?0 22:30
benabik Re-ran pugs spectest: gist.github.com/1941582 22:31
(now with less complaining about locales!)
[Coke] benanbik: when I run it the way it is, I get no complaints. 22:39
[Coke] I can try it with yours. if I still get no errors, I'll change it. 22:40
benabik [Coke]: Likely your system has the ISO-8859-1 locale. I imagine this is somewhat system-dependent.
I changed it to what my locale normally is. I have no idea how portable it is. 22:41
[Coke] benabik: with that change, I get more warnings. 22:42
benabik I'm not sure why it sets LC_ALL anyway. 22:43
[Coke] because without it, you get a ton more errors. 22:44
because it defaults to UTF8, IIRC.
s/errors/warnings
benabik LC_ALL should be set to something sane by the platform. Attempting to blindly set locale information is fraught with problems. 22:45
I think the only one you can be sure exists is "C". And I'm not even sure of that.
[Coke] I believe it is to compensate with the encoding of the spec files we're reading. 22:46
benabik The spec tests aren't in UTF8?
[Coke] I'm assuming not, but to be fair, au suggested the patch to avoid all the warnings. 22:49
benabik Is it trying to avoid the decodeUTF8 messages? Perl gives me piles of errors because en_US.ISO-8859-1 doesn't exist in OS X. gist.github.com/1941582/da3846f2f5...53bd18ffa2 22:50
[Coke] Yes. 22:51
(trying to avoid the decodeUTF8)
benabik Hm.
Ah-hah. I have a en_US.ISO8859-1 22:52
Although using it still gives me some decodeUTF8 errors.
benabik Nevermind. It still gives me all the same decodeUTF8 errors. Blah. 22:54
benabik can't win. 22:57
[Coke] If you can find one that works for you, lemme know what it is, we can put in an alternate.
benabik It seems odd to me that the spec tests are in Latin-1 anyway. 22:58
[Coke] seems like we should standardize on utf8 (if that's indeed the problem.) 22:59
benabik .u » 23:07
phenny U+00BB RIGHT-POINTING DOUBLE ANGLE QUOTATION MARK (»)
[Coke] # 03/06/2012 - niecza++ ; pugs (32.14%); rakudo (0%) 23:14
"niecza", 20306, 9, 749, 1540, 22604, 23759
"pugs" , 6527, 38, 2838, 1645, 11047, 23568
"rakudo", 0, 20384, 0, 0, 20384, 23803
benabik Rakudo 0?
[Coke] (Can't exec "./perl6") 23:15
I'll see about fixing that and redoing today's run.
benabik It looks to me like Pugs is trying to convert a string from UTF8 that's already been decoded. 23:19
.u 9829
phenny U+9829 CJK UNIFIED IDEOGRAPH-9829 (頩)
benabik .u 2665 23:21
phenny U+2665 BLACK HEART SUIT (♥)
benabik decodeUTF8': bad data: '\9829' means it's reading the character U+2665 as UTF8 data. 23:22
Vim insists that the file is encoded in UTF-8, so I'm guessing there's something odd happening between the file and decodeUTF8 in Pugs. 23:23