»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'perl6: say 3;' or rakudo:, niecza:, std:, or /msg camelia perl6: ... | irclog: irc.perl6.org | UTF-8 is our friend!
Set by sorear on 25 June 2013.
00:00 eternaleye left
tadzik :) 00:00
maybe I should add a tl;dr at the top
I wrote a lot of boring stuff in there too, some people may just give up
but then again, those people probably won't have the patience to write a game either :P 00:01
timotimo aye
games are hard
tadzik I played the most hilarous board game today
I wonder if it's in english too, so I can show you
boardgamegeek.com/boardgame/71021/e...of-species this 00:02
ssutch_ i've wanted to make games my whole life, but have been too busy paying the bills and spending my ambition elsewhere
timotimo not enough ambition spent on Perl 6 recently :P 00:03
ssutch_ timotimo: agreed
i am inspired by games like thomas was alone 00:04
timotimo that's a beautiful game
tadzik it's pretty fun
timotimo that's the kind of game you can't just hack up in your garage, though
you need a professional voice actor with such a surly voice
ssutch_ true 00:05
00:05 eternaleye joined
timotimo though i've been told time and time again i have a nice voice 00:06
00:10 rindolf left 00:14 Sqirrel left, Sqirrel joined 00:18 Sqirrel left 00:20 kivutar left 00:25 kivutar joined 00:36 Sqirrel joined 00:40 Sqirrel left 00:41 Sqirrel joined 00:46 kivutar left 00:57 thou left 01:01 Sqirrel left 01:02 Sqirrel joined 01:08 dayangkun joined 01:16 Sqirrel left 01:17 Sqirrel joined 01:51 Sqirrel left 01:52 Sqirrel joined 01:57 Sqirrel left 02:08 xenoterracide left
JimmyZ .tell tadzik I updated lastest MoarVM and still got segfault 02:11
yoleaux JimmyZ: I'll pass your message to tadzik.
02:14 Sqirrel joined
BenGoldberg p: sub infix:<oof> is assoc('list') { $^a + $^b }; say [Roof] 1..4 02:14
camelia rakudo-parrot ffa572: OUTPUT«Too many positional parameters passed; got 4 but expected 2␤ in block at gen/parrot/CORE.setting:17009␤ in sub at gen/parrot/CORE.setting:17151␤ in block at /tmp/NarMS2B2c_:1␤␤»
BenGoldberg std: for ( 1 .. 2 ) { (state $) ^= 42 } 02:16
camelia std 09dda5b: OUTPUT«ok 00:01 128m␤»
BenGoldberg p: for ( 1 .. 2 ) { (state $) ^= 42 }
camelia rakudo-parrot ffa572: OUTPUT«===SORRY!=== Error while compiling /tmp/658dH691Kb␤Unsupported use of $) variable; in Perl 6 please use $*EGID␤at /tmp/658dH691Kb:1␤------> for ( 1 .. 2 ) { (state $)⏏ ^= 42 }␤»
BenGoldberg p6: my $y; for ( 1 .. 2 ) { $y := (state $ ) ^= 42 }; 0; 02:20
camelia ( no output )
BenGoldberg p6: my $y; for ( 1 .. 2 ) { $y := (state $ ) ^= 42 }; print 'alive';
camelia rakudo-parrot ffa572, rakudo-jvm ffa572, rakudo-moar ffa572, niecza v24-109-g48a8de3: OUTPUT«alive»
BenGoldberg p6: my $y; for ( 1 .. 2 ) { $y := (state $ ) ^= 42 }; print $y;
camelia niecza v24-109-g48a8de3: OUTPUT«(timeout)»
..rakudo-parrot ffa572, rakudo-jvm ffa572, rakudo-moar ffa572: OUTPUT«one(one(one(), 42), 42)»
BenGoldberg p6: my $y; for ( 1 .. 2 ) { $y := (state $ ) ^= 42 }; print so $y; 02:21
camelia niecza v24-109-g48a8de3: OUTPUT«Stack overflow: IP: 0x50322a, fault addr: (nil)␤Stacktrace:␤ at Niecza.Kernel.UnboxAny<T> (Niecza.P6any) <0x0002b>␤ <...>␤ at Niecza.CtxJunctionBool.Get (Niecza.Variable) <0x00117>␤ at Niecza.CtxJunctionBool.Get (Niecza.Variable) <0x0025…»
..rakudo-parrot ffa572, rakudo-jvm ffa572, rakudo-moar ffa572: OUTPUT«False»
02:23 Sqirrel left 02:25 xragnar_ joined, xragnar is now known as Guest45972, Guest45972 left, xragnar_ is now known as xragnar 02:40 woosley joined 02:45 Sqirrel joined 02:46 klapperl left 02:48 klapperl joined 02:50 Sqirrel left, FROGGS left, Sqirrel joined 02:54 hoverboard joined 03:00 Sqirrel left 03:02 Sqirrel joined 03:03 FROGGS joined 03:10 Sqirrel left 03:13 Sqirrel joined 03:14 xenoterracide joined 03:20 Sqirrel left 03:21 Sqirrel joined 03:58 ssutch_ left 04:00 Sqirrel left 04:02 Sqirrel joined 04:07 logie_ left 04:11 Sqirrel left 04:13 Sqirrel joined
dalek ast: feff03c | (David Warring [email@hidden.address] | integration/advent2013-day12.t:
adding advent 2013 day 12
04:15
04:17 Sqirrel left 04:31 BenGoldberg left 04:32 logie_ joined 04:34 Sqirrel joined 04:40 rurban1 left 04:59 ssutch_ joined 05:00 kaare_ joined 05:04 ssutch_ left 05:10 rurban1 joined 05:12 rurban2 joined 05:15 rurban1 left 05:16 rurban2 left 05:18 araujo left 05:24 bjz left, bjz joined 05:26 xenoterracide left 05:29 hoverboard left 05:38 Sqirrel left 05:39 Sqirrel joined 05:53 Sqirrel left 05:54 Sqirrel_ joined, logie_ left 06:02 dayangkun left 06:04 kaare_ left 06:19 darutoko joined 06:30 kaare_ joined 06:33 kaleem joined 06:39 [Sno] joined 06:48 kurahaupo joined 06:49 ssutch_ joined, dayangkun joined 06:53 ssutch_ left 06:57 lustlife joined, lustlife left 07:05 kurahaupo left, yakudza joined 07:13 rurban1 joined 07:16 brrt joined 07:18 bjz_ joined, rurban1 left, xragnar_ joined, xragnar left, xragnar_ is now known as xragnar 07:19 telex left, telex joined, hugme left, dmol joined, klapperl left, sjn left, dayangkun left, dayangkun_ joined, yves__ left, yves__ joined 07:20 klapperl joined, sjn joined 07:21 bjz left 07:25 woosley left 07:27 lustlife joined, woosley joined 07:34 ribasushi joined 07:36 hugme joined, ChanServ sets mode: +v hugme 07:39 bjz_ left 07:44 bjz joined
masak mornin, #perl6 07:48
07:50 ssutch_ joined
brrt hi masak 07:50
07:55 ssutch_ left 08:09 zakharyas joined
nwc10 good *, #perl6 08:14
08:14 rurban1 joined 08:19 rurban1 left 08:36 pdcawley joined 08:37 rindolf joined
FROGGS moin 08:37
08:40 fhelmberger joined
sergot_ morning 08:43
08:43 sergot_ is now known as sergot
sergot o/ 08:45
FROGGS hi sergot 08:48
cosimo simula67: i'll take a look, thanks 08:49
(it's been ages since I last touched perl6 code, unfortunately)
08:50 ssutch_ joined 08:52 terrencehan joined, pecastro_ left
JimmyZ jimmy@jimmy-virtual-machine ~/panda $ ../MoarVM/install/bin/perl6-m bootstrap.pl 08:53
==> Bootstrapping Panda
/bin/sh: 1: perl6-m: not found
==> Please make sure that /home/jimmy/MoarVM/install/languages/perl6/site/bin is in your PATH
jimmy@jimmy-virtual-machine ~/panda $ ls ../MoarVM/install/languages/perl6/site/
panda
.tell tadzik ^^ 08:54
yoleaux JimmyZ: I'll pass your message to tadzik.
08:54 ssutch_ left
FROGGS ohh 08:56
08:57 rindolf left 08:58 toebu left 09:08 darutoko left 09:13 aborazmeh joined 09:15 rurban1 joined 09:17 rurban2 joined 09:19 rindolf joined 09:20 rurban1 left 09:21 rurban2 left 09:26 kurahaupo joined 09:33 darutoko joined 09:35 Ven joined 09:38 Alina-malina left
tadzik oh 09:47
yoleaux 02:11Z <JimmyZ> tadzik: I updated lastest MoarVM and still got segfault
08:54Z <JimmyZ> tadzik: ^^
tadzik should be an easy-ish fix, I'll open myself a bug
JimmyZ++
JimmyZ tadzik: one is about the game, another one is about panda :P 09:48
tadzik JimmyZ: well, I can't reproduce the one with the game
JimmyZ: can you get a backtrace maybe?
09:51 ssutch_ joined 09:56 ssutch_ left 10:00 Mouq left
JimmyZ tadzik: I don't get it yet, maybe tomorrow 10:01
10:02 Alina-malina joined 10:03 aborazmeh left 10:05 iolympian joined, aborazmeh joined
FROGGS hmmm, I think for modules with backend specific code it would be handy to have pragma "if" 10:06
colomon +1 10:09
FROGGS the Perl 5 version is just about a few lines, and I guess it could be the same for Perl 5 10:12
err, s/5$/6/
10:13 pecastro joined
colomon r: say $*VM 10:17
camelia rakudo-jvm ffa572: OUTPUT«("name" => "jvm", "config" => {"runtime.jars" => "/home/p6eval/rakudo-inst-2/languages/nqp/runtime/asm-4.1.jar:/home/p6eval/rakudo-inst-2/languages/nqp/runtime/asm-tree-4.1.jar:/home/p6eval/rakudo-inst-2/languages/nqp/runtime/jline-1.0.jar:/home/p6eval/raku…»
..rakudo-parrot ffa572: OUTPUT«("name" => "parrot", "config" => {"git_describe" => "RELEASE_6_1_0", "sha1" => "23928872782b93dff5b0a62749442cdb3c2fed81", "a" => ".a", "ar" => "ar", "ar_extra" => "", "ar_out" => "", "archname" => "x86_64-linux-gnu-thread-multi", "arflags" => "cr", "as"…»
..rakudo-moar ffa572: OUTPUT«("name" => "moar", "config" => {"ccoptiflags" => "-O1 -DNDEBUG", "uvrule" => "\$(AR) \$(ARFLAGS) \$\@ \$(UV_LINUX)", "tomobjects" => "3rdparty/libtommath/bn_error.o 3rdparty/libtommath/bn_fast_mp_invmod.o 3rdparty/libtommath/bn_fast_mp_montgomery_reduce.o …»
10:17 rurban1 joined
colomon … I guess that might be enough for most of the things I was thinking of. 10:18
10:21 salv0 left 10:22 rurban1 left 10:23 salv0 joined 10:25 dayangkun_ left, rindolf left 10:36 Alula left 10:37 Alula joined
colomon m: say "," x 4 10:40
camelia rakudo-moar ffa572: OUTPUT«,,,,␤»
10:45 mls_ is now known as mls 10:52 ssutch_ joined 10:53 woosley left 10:55 rindolf joined 10:56 ssutch_ left 11:02 SamuraiJack_ joined 11:03 salv0 left 11:11 terrencehan left 11:13 Ven left 11:18 salv0 joined
colomon m: say "+1".Int 11:18
camelia rakudo-moar ffa572: OUTPUT«1␤»
11:18 rurban1 joined 11:23 rurban1 left
lizmat r: say $*VM<name> 11:35
camelia rakudo-moar ffa572: OUTPUT«moar␤»
..rakudo-jvm ffa572: OUTPUT«jvm␤»
..rakudo-parrot ffa572: OUTPUT«parrot␤»
lizmat r: say $*VM.WHAT 11:36
camelia rakudo-parrot ffa572, rakudo-jvm ffa572, rakudo-moar ffa572: OUTPUT«(Hash)␤»
lizmat wonders whether that shouldn't be encapsulated in some object
11:36 brrt left 11:38 bash_ joined 11:41 dmol left 11:43 bash_ left 11:47 colomon left 11:53 ssutch_ joined
FROGGS lizmat: in case it will offer more than name=>value pairs or lists, yes 11:55
but until then a hash is all we need 11:56
11:56 woolfy1 joined, woolfy left, skids left
lizmat as long as the object supports postcircumfix {}, I guess 11:57
otoh, I'm reminded ot how DBI's tied hashes exploded :-(
11:57 ssutch_ left
timotimo today is the last chance to get some exciting things into the weekly :P 11:58
FROGGS hehe
timotimo well, at least this week's weekly
FROGGS timotimo: I am busy writing slides, so do not expect anything from me :o)
lizmat is also busy writing slides
timotimo slides are good, too 11:59
FROGGS I only have done like 33% :/
timotimo i don't even have any slides :(
lizmat has some more time available then FROGGS, as she only has to present hers on Mon 31 March
FROGGS I have to give the talks in about 45 hours 12:00
timotimo and i'll be attending :P 12:02
lizmat as will I :-) 12:03
FROGGS *g*
timotimo good timesn
er.
lizmat unfortunately, I won't be able to attend all of GPW
FROGGS yeah, I will attend all of it this year it seems
lizmat I have some other sad obligations to fulfill on late Thursday / Friday
FROGGS yeah :( 12:04
12:10 aborazmeh left
lizmat has jnthn merged the concurrency branch into nom already ? 12:10
FROGGS no, I've not seen it 12:11
timotimo he has not 12:13
FROGGS yeah, rakudo's moar-conc is 28 commits ahead
timotimo there's still a nasty concurrent GC bug where "work" gets lost
which, in my imagination, would lead to a bunch of objects getting removed, because they were not scanned
it doesn't happen very often, aiui 12:14
i think something like perf ought to help debug something like that, fwiw
12:14 iolympian left
timotimo perf can record and replay the scheduling decisions made for a given thread or the whole system 12:14
so you can turn a nondeterministic bug that depends on exact scheduling into a 100% reproducible test case that you can debug over and over again
i don't know the details, though
12:15 pecastro left
timotimo lwn.net/Articles/353295/ 12:15
it may not be exactly what i was advertising after all 12:16
"using simulated threads"
however, this was 2009 12:17
maybe it has advanced significantly since then
i've got a friend who may know about it; not sure if he's AFK at the moment or not 12:18
12:19 rurban1 joined 12:24 rurban1 left 12:31 colomon joined
[Coke] feather1 is at 90% of disk usage. 12:40
12:40 kbaker joined
timotimo the alot of disk usage 12:42
[Coke] tried to run throughthewindows with a just-built rakudo-moar. installed rakudo-moar... get the error "Could not find NativeCall".... 12:43
colomon TimToady: .Hash is supposed to coerce to Hash, right? instead of .hash?
[Coke]: right, I installed panda and used that to install NativeCall. and it all worked! 12:44
12:45 colomon left
lizmat colomon: rakudo only has .hash at the moment, not .Hash 12:50
12:51 dayangkun joined 12:54 ssutch_ joined
lizmat it has .Array though 12:54
I'm not sure what the reasoning for having .hash and not .Hash is / has been 12:55
FROGGS is .Array the same as .list?
lizmat r: say 1.Array.WHAT
camelia rakudo-parrot ffa572, rakudo-jvm ffa572, rakudo-moar ffa572: OUTPUT«(Array)␤»
lizmat r: say 1.list.WHAT
camelia rakudo-parrot ffa572, rakudo-jvm ffa572, rakudo-moar ffa572: OUTPUT«(List)␤»
FROGGS how obvious *g*
lizmat Array makes an Array, list makes a List 12:56
r: say 1.List.WHAT
camelia rakudo-parrot ffa572, rakudo-jvm ffa572, rakudo-moar ffa572: OUTPUT«No such method 'List' for invocant of type 'Int'␤ in block at /tmp/tmpfile:1␤␤»
12:56 iolympian joined 12:58 iolympian left, ssutch_ left 13:00 pecastro joined 13:01 denisboyun joined 13:03 guru joined, guru is now known as Guest30489, Guest30489 is now known as ajr_ 13:06 [Sno] left
[Coke] colomon: oh, whoops, I had assumed it was bundled for some reason that eludes me. 13:06
13:08 SamuraiJack_ left
[Coke] is panda smart enough at this point to deal with multiple installed backends? 13:08
lizmat good question: I assume, like with perl 5, it is the executor running panda that determines what the version / backend is 13:09
[Coke] lizmat: that would make sense. Danke.
13:10 iolympian joined 13:12 colomon joined
[Coke] Building panda (Warning: this may take a while) 13:12
13:17 iolympian left 13:18 skids joined 13:20 rurban1 joined 13:24 rurban1 left 13:25 [Sno] joined
[Coke] looks like panda installs a single binary, so I'm not sure it will work if you run it with multiple rakudos. 13:26
(old panda used to end up in parrot/PARROT_REV, current panda seems more global)
nativecall warning on a lot of tests: t/03-simple-returns.t .. /bin/sh: PREFIX: command not found 13:27
13:29 kivutar joined
[Coke] ah well. got as far as it complaining about not being able to open ./sdlwrapperlib.dylib - can't install libsdl via ports on this mac for some reason. 13:29
13:30 simcop2387 joined
[Coke] (tried installing SDL via libsdl.org, still can't find SDL.h 13:32
[Coke] gives up for now. 13:34
13:34 lustlife left 13:35 lustlife joined 13:40 lustlife left 13:41 molaf joined 13:43 kaleem left 13:51 salv0 left 13:52 lustlife joined
jnthn waves from Vienna airport, which not only has free wifi, but also a place to sit comfortably with a laptop and plug it in! 13:53
Dammit, why can't Copenhagen learn from this.
[Coke] wonders if he can get someone in NY to hire jnthn so he can complain about the local airport! 13:54
13:54 ssutch_ joined
jnthn Convincing me to take a job outside Europe would be quite a challenge. :) 13:58
13:59 ssutch_ left
jnthn timotimo: It's a parallel GC bug, not a concurrent one. We don't have a concurrent GC. Thank heavens... :) 14:03
14:04 btyler joined
lizmat jnthn: even if they would have decent beer ? 14:05
moritz wouldn't it be awesoem to have the GC run in a separate thread? :-)
14:06 rurban1 joined
timotimo oh i meant that 14:06
14:06 salv0 joined
[Coke] is reminded, he discovered a pretty good (to him) porter this weekend... which his wife liked, so he swapped with her. Coke needs to go back and get more of that porter. :) 14:08
14:08 bluescreen10 joined 14:09 logie_ joined
lizmat assumes porter is a beer ? and not a person porting code ? 14:09
jnthn moritz: Depends. 14:10
moritz or maybe even in several separate threads
jnthn moritz: For interactive applications, yes, as you can get pause times really low
moritz: For non-interactive/server ones, not really, as you have lower throughput typically than stop-the-world.
14:13 zakharyas left
jnthn lizmat: Yes, a porter is a kind of dark beer. 14:14
lizmat: Beer is only *one* factor :P
14:15 thou joined 14:18 jnap joined
nwc10 jnthn: just curious - what's your external IP? 14:18
eg from www.whatismyip.com/
hunting IP addresses for test purposes :-) 14:19
lizmat nwc10: 80.100.131.131 but the site got the location all wrong :-) 14:20
jnthn nwc10: 193.43.158.234
nwc10 thanks 14:21
jnthn waits to get packeted...
:P
14:21 FROGGS left
timotimo how long until the daily roast, [Coke]? 14:23
nwc10 jnthn: awesome. It actually is useful. It doesn't have a City associated with it in our lookup system 14:24
I didn't expect life to be that easy :-)
14:25 salv0 left 14:26 zakharyas joined 14:34 kaare_ left 14:36 zakharyas left 14:37 FROGGS joined 14:42 jnap left 14:44 salv0 joined
dalek kudo-star-daily: e71320b | coke++ | log/ (5 files):
today (automated commit)
14:44
kudo-star-daily: 830ca37 | coke++ | log/ (5 files):
today (automated commit)
rl6-roast-data: 45b19fa | coke++ | perl6_pass_rates.csv:
today (automated commit)
[Coke] timotimo: there's yesterday's.
timotimo thanks :)
[Coke] today's is up to S05 on the jvm, which is the last impl to run. 14:45
timotimo oke :)
[Coke] want them faster? I need to setup the JVM to run tests through the eval server again. (and hopefully it will JustWork™) 14:46
timotimo nah, i don't need them that bad 14:47
i'll just procrastinate writing the p6weekly post some more :)
i'd love for the number of failing tests to go down on moar :\
jnthn Well, the PR I just merged claims to get rid of a few more. 14:48
timotimo the ones in ord_and_chr i suppose?
[Coke] jnthn: when did it merge? 14:49
timotimo hmm, quite a few exceptions failures
[Coke]: 5 minutes ago
[Coke] here's today's moar run: "total", 29203, 50, 579, 1326, 31158, 28960
FROGGS still 50 :/
[Coke] so, tomorrow, then. 14:50
timotimo yeah
jnthn [Coke]: like 2 minutes ago :)
timotimo: yes
timotimo jnthn: have you looked at that already?
jnthn timotimo: the exceptions ones are mostly NYIs
lizmat don't we need a version bump to see those changes ?
jnthn timotimo: Looked at what?
timotimo lizmat: doesn't coke run against master?
jnthn: the exception failures
probably just a case of creating a type dexception instead of mvm_exception_throw_adhoc
jnthn timotimo: No, but as I said, they're just places where we didn't write the code to throw a typed exception yet.
14:50 zakharyas joined
lizmat timotimo: I have no idea 14:51
timotimo should be an easy thing to do for me right now :)
should be lots of stealable code nearby
14:51 jnap joined
FROGGS [Coke]++ is smoking rakudo HEAD and the specified versions of nqp/moar/parrot 14:51
jnthn timotimo: yes, shouldn't be hard. 14:52
[Coke] timotimo: not explicitly, but I do run with a fresh nqp checkout that is pre-fetched, so we should get it implicitly.
... or froggs could be right. :)
I think FROGGS is right - not explicit, so get the last specified version. 14:53
e.g. rakudo.moar's nqp is: 21a89d7
FROGGS github.com/coke/perl6-roast-data/b...do.moar.sh
[Coke] (which is 2014.03). FROGGS++ for remembering what I did better than I did. :)
FROGGS *g*
timotimo [Coke]: sounds good; but i suppose we still need to bump moar then :) 14:54
[Coke] timotimo: aye.
(this is as oppposed to the rakudo-star-daily one, which gets master for everything but parrot, IIRC.)
14:55 ssutch_ joined 14:57 kurahaupo left
timotimo jnthn: where should i look to find how to throw a typed exception properly from inside the perl6_ops.c? 14:57
would that be get_thrower?
jnthn get_thrower, yeah. 14:58
15:00 ssutch_ left
[Coke] gets a headhunter bite off of -parrot- 15:00
weird.
jnthn wonders if that's the same headhunter bite he got off MoarVM/Rakudo today :) 15:02
15:03 kst joined 15:05 hoelzro_ is now known as hoelzro
hoelzro morning #perl6! 15:06
FROGGS hi hoelzro
15:06 araujo joined
hoelzro timotimo: I haven't had a chance to look at it yet (still waiting on internet at home =/), but I should get a chance soon 15:06
ahoy FROGGS
timotimo which "it" is that exactly? 15:07
hoelzro timotimo: the kate syntax file 15:09
timotimo oh
hoelzro you messaged me on like Saturday =)
timotimo i see :)
why would you have to wait for internets at home? :P
isn't that the right exercise to do when you're out of internets?
hoelzro heh
well, I need to install things at home to be able to work on the syntax stuff )
also, no internet time is usually good for getting misc. housework done 15:10
since I have nothing better to do
timotimo OK :) 15:11
15:15 xenoterracide joined
timotimo didn't someone fix the bug with pack "mA"? 15:17
or rather, the test? 15:18
FROGGS timotimo: me
it was "mÄ"
timotimo i forgot to re-fudge :) 15:19
FROGGS ahh
:o)
colomon what's the current status of web frameworks on rakudo? 15:21
15:22 ajr_ left
lizmat what do you mean with web framework? Plack? Dancer? Catalyst? 15:24
colomon Dancer or Catalyst, that sort of thing 15:25
Bailador.
but I don't know if it's up-to-date?
it is passing all tests.
lizmat then it is up to date to some extent, I would guess :-) 15:26
colomon lizmat: that might mean it doesn't actually have any tests. ;0
;)
lizmat ah, eh, hmmmm 15:27
15:27 jnap1 joined 15:28 Sqirrel_ left, jnap left 15:30 guru joined, guru is now known as Guest53214 15:31 Sqirrel joined 15:32 kaleem joined 15:33 cognominal left, cognominal joined
timotimo hmm. i'm changing a wrong exception type to an X::AdHoc with the right message 15:35
is that an improvement? :P
jnthn No, that happens automagically, I think... :)
timotimo jnthn: how will we teach the binder on moar to do typed exceptions? currently we have the $error[0] = "a string message" "solution". can we just put a typed exception into that instead of a string? 15:39
tadzik Bailador has some tests :)
that means it's working to at least some extent
colomon tadzik: Just installed it using panda
jnthn timotimo: We could if we teach the thing that consumes it whant to do, yes
tadzik nice 15:40
timotimo sounds easy enough. i'd like to do it if you'll accept that code :)
jnthn have a crack at it.
15:41 [Sno] left
jnthn flight & 15:43
timotimo well, it looks like it just calls nqp::die($error[0]) 15:44
i'm not sure i'll even have to teach it about typed exceptions especially
r: try nqp::die(X::Item.new(aggregate => "cow", index => 1)); say $!.WHAT; say $!; 15:45
camelia rakudo-moar ffa572: OUTPUT«(X::AdHoc)␤This type cannot unbox to a native string␤ in block at /tmp/tmpfile:1␤␤»
..rakudo-parrot ffa572: OUTPUT«(X::AdHoc)␤Cannot index Str with 1␤ in block at /tmp/tmpfile:1␤␤»
..rakudo-jvm ffa572: OUTPUT«(X::AdHoc)␤This representation can not unbox to a native str␤ in block at /tmp/tmpfile:1␤␤»
timotimo oh, i was wrong.
r: try nqp::throw(X::Item.new(aggregate => "cow", index => 1)); say $!.WHAT; say $!;
camelia rakudo-parrot ffa572: OUTPUT«(X::AdHoc)␤Not a throwable object␤ in block at /tmp/tmpfile:1␤␤»
..rakudo-moar ffa572: OUTPUT«(X::AdHoc)␤Can only throw an exception object␤ in block at /tmp/tmpfile:1␤␤»
..rakudo-jvm ffa572: OUTPUT«(X::AdHoc)␤throw needs an object with VMException representation␤ in block at /tmp/tmpfile:1␤␤»
timotimo okay, that's much more low-level
but i suppose i can just create the object and .throw it
hm, though, being in the BOOTSTRAP.nqp, i'll probably have to have some magic to look up the actual classes 15:46
15:46 kivutar left 15:47 kaare_ joined 15:56 ssutch_ joined 15:57 sqirrel_ joined 16:00 ssutch_ left
colomon Bailador working on my Mac, tadzik++ 16:00
tadzik nice! 16:02
timotimo but can it do websockets? 16:06
tadzik nope
well volunteered! P:
timotimo i don't even know how that works.
PerlJam sri does ;> 16:07
(or just copy what mojo does)
tadzik I think it's an upgrade from regular sockets 16:08
btyler tadzik: I grabbed ThroughTheWindow on my mac, it worked without anything special. runs very smooth (or did for about 1009 pixels :)
tadzik so, a layer over (under?)
btyler: awesome, thanks :)
btyler: did you like it?
btyler its a lovely little game. I certainly did that as a kid in the car
16:09 fhelmberger left
hoelzro tadzik++ 16:12
16:13 denisboyun left
tadzik glad to hear that :) 16:17
16:17 kaleem left 16:23 dmol joined
lizmat wonders whether "coccoon" is just the American spelling of "cocoon" 16:28
or do I miss some other joke with "coccoon" ?
16:28 Sqirrel left
hoelzro lizmat: doesn't look familiar to me =/ 16:28
lizmat so, just a misspelling, right? 16:29
hoelzro probably
lizmat thanks, was afraid I missed some pop culture reference 16:30
but even the movie is spelled "Cocoon"
tadzik Racoon
geekosaur just one of those words that gets misspelled a lot
16:32 Sqirrel joined
colomon tends to get confused about which letters are doubled and which are not in words like that. 16:34
geekosaur you just gotta remember it, there's no actual rules :/
english are weird
(well, there are rules, they just have so many exceptions that they might as well not exist)
16:35 molaf_ joined, molaf left 16:37 SamuraiJack_ joined 16:38 zakharyas left
tadzik colomon++ # improving bailador 16:39
lizmat cycling& 16:40
16:42 telex left
colomon tadzik: I needed it so it didn't conflict with my Dancer instance on my Linux box. :) (That would be the smoke test results page.) 16:42
16:44 telex joined
[Coke] lunch break - anyone have a suggestion on how to get sdl working on os x 10.9 with nativecall? (specifically TTW's build) 16:45
16:46 kivutar joined
colomon [Coke]: I got it working (on 10.8) by building sdl from source. then everything was groovy 16:46
[Coke] colomon: I tried installing libsdl via ports (boom) and via libsdl.org (whiff) - if I change the sniffer to use sdl-config instead of sdl2-config, I get further into the build... but then it dies (version mismatch, I assume?) 16:47
... weird. the sdl-config is giving me a macports one... 16:48
btyler [Coke]: I got it working on 10.9 with zero hassle, just 'brew install sdl2' 16:49
and 'panda install NativeCall'
[Coke] I assume brew is some hip new replacement for port ?
btyler yeah, essentially
homebrew. keeps everything in /usr/local so you don't need to install stuff as root 16:50
[Coke] realizes he was literally installing 'libsdl' not 'libsdl2', and tries that.
16:51 larks left, larks joined
geekosaur brew is the macports alternative for people who hate seatbelts 16:52
hoelzro and like driving in the fastlane =P
geekosaur violates security, generally written without any error checking, naturally people love it to death
yes, it;s really fast until it trashes you
enjoy it
hoelzro I happen to agree
I have a love/hate relationship with brew 16:53
btyler I'm just an uninformed user, but it has worked well for my needs
haven't used/tried macports
hoelzro and the sad thing is, people look at it and say "why can't all Linux package managers be like this?"
it works 95% of the time
but it breaks stuff a *lot*
at least it has for me
tadzik hrm, so why do people miss it on linux? 16:54
I like basically every linux package manager I ever used
geekosaur because linuc package managers enforce rules, while homebrew doesn't give a flying fuck abotu rules
so it's easy and fast and painless right up until it screws you hard
hoelzro I think that people like the fact that they can submit PRs for package formulae 16:55
which by itself isn't a bad idea
but there's no review process
it's anarchy
tadzik :o wwatt
so you install a package, and hope no one put malware in there 10 seconds ago?
geekosaur pretty much
[Coke] hoelzro: so, it's cpan? 16:56
hoelzro bah dum tssssh
[Coke] :)
16:56 ssutch_ joined
geekosaur so over in #haskell there's one person who used to strongly recommend homebrew, to the extent of immediately jumping on and insulting anyone who dared to mention macports 16:57
over the past 2 months, he's had to help people untangle problems with the recommended ghc brew recipe 16:58
he's stopped recommending homebrew
[Coke] ignores the weird warning at the end of installing libsdl2, and is able to play the game! 17:00
tadzik yay :)
colomon \o/ 17:01
17:01 ssutch_ left
retupmoca .tell jnthn I golfed my Compress::Zlib::Raw compilation issue from the other day: gist.github.com/retupmoca/9744166 - I have not tested to see if this is NativeCall specific 17:01
yoleaux retupmoca: I'll pass your message to jnthn.
vendethiel geekosaur: homebrew has had problems with haskell for a very long time ...
[Coke] tadzik: when i was a kid, I imagined it was spiderman. made it easier to explain why he was jumping from building to building. :)
tadzik ah :)
I was mostly driving through countryside 17:02
[Coke] speaking of haskell, anyone want to help me get pugs working on the new roast box?
tadzik well, being driven :P
but everyone imagined something similar: ain't that something cool
colomon mostly read on car trips of significant length
tadzik I was getting carsick when I read. I maybe still do, but I play it safe and don't try 17:03
now I mostly sleep, when I'm supposed to keep the driver awake :P
17:03 raydiak joined
timotimo out of lack of need for self-preservation? :P 17:05
i should start with the p6weekly post, actually 17:06
17:07 Guest53214 left
colomon Installing Bailador from a local directory 'Bailador' 17:08
hmmm
17:11 kurahaupo joined 17:21 Alula left, BenGoldberg joined 17:24 Rotwang joined, Rotwang left, Rotwang joined
timotimo not too many changes to report for the weekly, i'm afraid 17:26
or am i missing something?
do we have a star release manager for this month yet? 17:29
oh, we didn't bump the parrot version for this month's release 17:31
17:34 Alula joined
dalek rl6-roast-data: 371b329 | coke++ | / (6 files):
today (automated commit)
17:35
17:35 SamuraiJack_ left, SamuraiJack joined
tadzik colomon: if there's a local dir with a maching name, panda will install from there 17:36
I should maybe change that
colomon It's a great feature for testing, I guess.
but it's really important to know that it's there. ;) 17:37
tadzik it should probably be more explicit though
like, require ./Bailador
PerlJam tadzik: +1
17:37 guru joined
tadzik yay, new bug 17:37
PerlJam As long as that doesn't affect "panda install ."
timotimo or installing a module named "."
17:37 guru is now known as Guest71490 17:38 Guest71490 is now known as ajr_
tadzik oh yes, 'install .' is a common usecase 17:39
even bootstrap relies on that
PerlJam tadzik++ 17:40
17:46 denisboyun joined
timotimo hm, the lead moar has over jvm isn't all unicode database, is it? 17:48
17:49 hoverboard joined
simula67 hi 17:49
17:49 rindolf left
timotimo greetings, simula67 17:51
simula67 Is there any windows graphical text editor ( like Notepad++ ) with syntax highlighting for Perl 6 ?
timotimo is GVim an acceptable candidate? 17:52
simula67 I need to try that again. Last time I tried it I remember it did not allow copy pasting to and from other apps 17:55
17:55 IllvilJa left
simula67 might as well use plain vim in that case 17:55
17:57 ssutch_ joined
raydiak simula67: gvim does allow access to the os clipboard, it just doesn't use it as the default one for its copy/paste 17:58
retupmoca simula67: are you familier with vim registers? On my linux machine, the * and + registers sync to my clipboard 17:59
raydiak simula67: if you put "set clipboard=unnamed" in your _vimrc, it'll use the os clipboard by default 18:00
18:01 ssutch_ left, denis_boyun__ joined 18:02 denisboyun left
simula67 raydiak: thanks, i will try that when i switch to windows. I need a way to hack on this while at work :) 18:04
raydiak simula67: works on linux and mac too :) (except w/.vimrc instead of _vimrc) 18:07
simula67 Pacman gives me this :: gvim and vim are in conflict. Remove vim? [y/N] n 18:10
raydiak gvim package under arch provides both console and gui vim, iirc
18:14 sqirrel_ left
raydiak yes, gvim provides vim (just checked)...though fwiw, you don't need gvim to use the clipboard 18:15
tadzik timotimo: it seems like the font changes after the 2nd paragraph of p6weekly 18:18
simula67 raydiak: yes, i just installed it. looks good. let me try on windows as well. thanks 18:19
raydiak simula67: under linux you might need to use unnamedplus instead of unnamed, depending on which clipboard/cut buffer you use. you're welcome and good luck 18:20
tadzik timotimo++ # nice post
timotimo: mind linking to the blog? People might not know who the tadzik fellow is 18:21
18:23 btyler left 18:25 rindolf joined 18:26 IllvilJa joined 18:28 btyler joined
hoelzro timotimo: wrt that syntax highlighting bug you found, it seems that y is being recognized as a keyword 18:36
18:41 zakharyas joined 18:49 SamuraiJack left 18:52 btyler left
vendethiel for GSoC, who decides on who's in / who's out ? 18:52
tadzik org mentors
to some extent, at least. They're given a number of slots by google, and are to fill them to best of their abilities 18:53
vendethiel oh, so I could've a chance actually
18:56 ssutch_ joined 18:59 darutoko left
timotimo tadzik: but i linked to your blog 19:06
[Coke] wonders if we could autogen github.com/vim-perl/vim-perl/blob/.../perl6.vim from STD.p6
tadzik ah, it's just visible on hover :) ok
timotimo wtf, why is it all the stuf bold? 19:07
hoelzro [Coke]: that's not a bad idea
hell, autogen pygments + perl6.vim + perl6-kate
hoelzro .oO( Perl6 Cate? )
19:09 MrMeek joined
BenGoldberg std: sub infix:<ow> is assoc('list') { $^a + $^b }; say [Row] 1..4 19:13
camelia std 09dda5b: OUTPUT«ok 00:01 130m␤»
vendethiel r: 0 R 1 19:14
camelia rakudo-parrot ffa572, rakudo-jvm ffa572, rakudo-moar ffa572: OUTPUT«===SORRY!=== Error while compiling /tmp/tmpfile␤Two terms in a row␤at /tmp/tmpfile:1␤------> 0 R⏏ 1␤ expecting any of:␤ postfix␤ infix stopper␤ …»
19:14 Alina-malina left, Alina-malina joined, Alina-malina left, Alina-malina joined
vendethiel r: 0..2 R+ 3..5 19:14
camelia rakudo-parrot ffa572, rakudo-jvm ffa572, rakudo-moar ffa572: OUTPUT«===SORRY!=== Error while compiling /tmp/tmpfile␤Operators '..' and '..' are non-associative and require parenthesis␤at /tmp/tmpfile:1␤------> 0..2 R+ 3.⏏.5␤ expectin…» 19:15
vendethiel r: say (0..2) R+ (3..5)
camelia rakudo-parrot ffa572, rakudo-jvm ffa572, rakudo-moar ffa572: OUTPUT«6␤»
vendethiel I don't even know what does this do :P
BenGoldberg The R meta-operator reverses the arguments to the operator it's attached to
r: say 5 r/ 3
camelia rakudo-parrot ffa572, rakudo-jvm ffa572, rakudo-moar ffa572: OUTPUT«===SORRY!=== Error while compiling /tmp/tmpfile␤Two terms in a row␤at /tmp/tmpfile:1␤------> say 5 ⏏r/ 3␤ expecting any of:␤ postfix␤ infix stopper…» 19:16
BenGoldberg r: say 5 R/ 3
camelia rakudo-parrot ffa572, rakudo-jvm ffa572, rakudo-moar ffa572: OUTPUT«0.6␤»
jnthn gets home
yoleaux 17:01Z <retupmoca> jnthn: I golfed my Compress::Zlib::Raw compilation issue from the other day: gist.github.com/retupmoca/9744166 - I have not tested to see if this is NativeCall specific
BenGoldberg r: sub infix:<ow> is assoc('list') { $^a + $^b }; say [Row] 1..4
camelia rakudo-parrot ffa572: OUTPUT«Too many positional parameters passed; got 4 but expected 2␤ in block at gen/parrot/CORE.setting:17009␤ in sub at gen/parrot/CORE.setting:17151␤ in block at /tmp/tmpfile:1␤␤»
..rakudo-moar ffa572: OUTPUT«Too many positional parameters passed; got 4 but expected 2␤ in block at src/gen/m-CORE.setting:16808␤ in sub at src/gen/m-CORE.setting:16950␤ in block at /tmp/tmpfile:1␤␤»
..rakudo-jvm ffa572: OUTPUT«Too many positional parameters passed; got 4 but expected 2␤ in block at gen/jvm/CORE.setting:16754␤ in sub at gen/jvm/CORE.setting:16896␤ in block at /tmp/tmpfile:1␤␤»
FROGGS vendethiel: you have submitted a GSoC application?
19:16 zakharyas left
vendethiel FROGGS: no, I havn't, because I thought only google decided (basically randomly) and I'm never lucky 19:17
[Coke] next time talk to us! 19:18
FROGGS no, the mentors are supposed to vote the proposals up/down, and then the top X are taken (where X is number of slots we got)
actually mentors + org admins 19:19
vendethiel Well, is it too late already ?
FROGGS yes
deadline was on friday
vendethiel Oh,really? Gotta tell a friend then, he's still getting messages from a mentor
FROGGS vendethiel: btw, you can still pick a topic and we do the mentoring, you just won't get payed :o) 19:20
vendethiel FROGGS, I can't, it needs to be official for my school to count it 19:21
FROGGS you can do it just for fun and profit, err, fun and learning perhaps :o)
vendethiel I have 8 months to do a 6-month internship, I've been told that "GSoC might be okay but not long enough" so I'd have had to find something else on the side, but everything they have internally is 6-8 months 19:22
[Coke] if you want to set it up as an independent study or something, I'm sure you could find a mentor to work with you.
or get your advisor to join the community. :)
vendethiel Yeah, though I can't really do it in GSoC's period of time because I'll be very busy with said internship
well, I'm not that introverted not to participate in a big open source project (I love it), I just don't really know where to start and, as everybody, I have a limited free time 19:23
although I've seen "low-hanging fruits" in p6weekly by timotimo++ ? 19:24
19:24 sqirrel_ joined
timotimo once, aye 19:27
vendethiel that's a good incentive 19:30
(is that the good word for it ? eh)
colomon vendethiel: just trying to write code in p6 can be really useful. or at least, it seems like I turn up some sort of compiler issue every time I do…. ;)
vendethiel colomon: to try and code what ? 19:31
colomon vendethiel: whatever catches your fancy!
vendethiel colomon: nothing does, that's my problem
up to the point I'm thinking of making a "Programming Ideas" website ...
colomon vendethiel: maybe you should go back to the classics. program "hunt the wumpus" or something. 19:32
vendethiel googles
19:32 spider-mario joined
vendethiel I already did a game of life, but that kind of stuff is not ... "real-lify", you don't really hit problems 19:33
though, is there an ascii-based-games framework for perl6 yet ?
19:34 btyler joined 19:35 jnap1 left, jnap joined
FROGGS vendethiel: perhaps this? github.com/masak/Adventure-Engine 19:36
vendethiel haha, so that's not a project either
FROGGS vendethiel: why don't you improve it or do your own and just get some ideas? 19:37
vendethiel I don't feel like it's interesting to do it on my own 19:38
FROGGS bah
:P
vendethiel I can do "on my own" project ideas, but framework I prefer to leave state of the art
FROGGS well, it sounds more like choosing between making state of the art in distant future or producing something smaller in near future 19:39
19:40 sivoais joined
vendethiel FROGGS, yeah, I know 19:41
it's just that it feels "distant" 19:42
lizmat is back from cycling 19:47
19:48 bluescreen10 left
lizmat timotimo++ # weekly update 19:48
timotimo the primary property of this week's update is that it's weekly :| 19:49
dalek ecs: a30bb46 | jonathan++ | S22-package-format.pod:
Minor language tweaks.
19:52
lizmat jnthn++ 19:53
jnthn Here are a few other things that crossed my mind as I read it: gist.github.com/jnthn/9747719 19:54
lizmat jnthn: wrt to emulates: the way I see it, is that if compunit Foo emulates Bar, it will install an alias 19:55
that will cause loading of compunit Bar to become a noop
no more and no less
module Foo will need to make sure it is otherwise compatible 19:56
e.g. by exporting the same subs / providing the same method API
19:59 molaf__ joined
lizmat jnthn: does that make sense? 20:00
wrt to development of CompUnit / CompUnitRepo
FROGGS jnthn: the production readiness is a problem for installed dists, yes
lizmat the idea is that a developer has a CompUnitRepo::Local as its first @INC entry
FROGGS not for dists in your dev environment though, since we only care about names (like in provides) there, but not about auth, ver, or production state 20:01
lizmat that special version will use the current compunit load semantics where the requested name directly maps to a directory/filename
*without* looking at auth / ver
FROGGS yeah, like we do for rakudo/nqp things 20:02
20:02 molaf_ left
jnthn lizmat: Ah, so emulates just means "yes, I herby confirm that the module does the right things to implement the other thing"? OK, then I'm comfortable. 20:02
lizmat the CompUnitRepo API accepts a compunit name to be searched
FROGGS these are bound to a filesystem lookup
jnthn s/implement/provide a compatible implementation of/
lizmat jnthn: indeed
FROGGS but as I said, the installed non-production dists need some pondering 20:03
lizmat wrt to "production" / "not production": the way I see that
is that a "not production" distribution will never be returned by a recommendation manager 20:04
jnthn lizmat: Does -Ilib count as installing a CompUnitRepo::Local?
lizmat yes, that's the idea
FROGGS jnthn: yes
jnthn OK, that makes sense.
lizmat so, "not production" is a barrier against automatic installation / update
FROGGS lizmat: what about smokers, that pick up recent releases and test them? these should smoke non-production stuff as well 20:05
lizmat but it is *not* a barrier against specific installation
FROGGS and other ppl might want to have bleading edge too
lizmat it's just as with the -TRIAL now in Perl 5 distributions
it will never appear on the module list, so it won't be used for automatic updates
FROGGS lizmat: but how do I say that I want latest Foo::Bar, even a dev version?
say I have it already installed 20:06
and then?
how does my use statement look like?
lizmat
.oO( pondering )
FROGGS it is tricky
smokers will prefer an env var, but common users probably not
20:06 kurahaupo left
lizmat the use statement, in that context, is only sensitive to version 20:07
jnthn lizmat: Ah, so production is reifying a convention (though from the other angle) that has already appeared in the Perl 5 community?
lizmat yes
that was my idea
jnthn lizmat: That makes me more comfortable with its existence.
FROGGS jnthn: I did that often
lizmat a dev version has a different version number
the "production" flag is only applicable to installers 20:08
jnthn lizmat: I was a little worried it might be a solution looking for a problem, but trusted that you'd have a better grasp on the set of problems out there than I do ;)
lizmat an installer has the option of installing "production" distributions only
FROGGS I released SDL 2.540 stable, and then 2.541_1 to 2.541_9 as dev releases to hunt down bugs, and then rereleased the _9 as next stable
lizmat or install non-production distributions also 20:09
FROGGS ohh, »the "production" flag is only applicable to installers«, interesting idea
lizmat the "use" statement only knows about what's installed
FROGGS sure
lizmat and what's installed really is determined how CompUnitRepo's handle requests for compunits 20:10
20:10 rindolf left
lizmat I could envision a CompUnitRepo::Cloud 20:10
FROGGS so when I choose to install Foo-0.1-TRIAL, it will later be considered like stable releases in use statements
lizmat that would just download any module ad-hoc without even really installing
FROGGS: yes, that's my idea 20:11
FROGGS lizmat: I like it
lizmat of course, during runtime, you can introspect the distriibution info and find out whether the code is actually production ready
jnthn: does that answer your questions? 20:13
jnthn lizmat: Yes, sufficiently well. :)
lizmat: I suspect implementation and usage will give some clarity too :) 20:14
lizmat hoping to find time for that after I get back from Cluj
FROGGS I will also have more time for implementing latest changes after gpw 20:15
would be awesome to get the CPAN indexer ready soonish, I guess that is the next thing I tackle when my slides are in a good state 20:16
segomos FROGGS: a CPAN indexer?
lizmat way cool
FROGGS segomos: yes
segomos i'd be interested in building it.. in perl6 or ? and for what purpose?
lizmat yes, we're going to have Perl 6 distributions on CPAN soon :-)
FROGGS segomos: in the first step it just records all Perl 6 releases, so we can plug panda to CPAN
segomos: in the PAUSE code (Perl 5) 20:17
lizmat
.oO( with a very strong Perl 4 odour )
FROGGS if you say so :o)
lizmat 20+ year old code has that smell :-) 20:18
FROGGS true
segomos how are you differentiating perl6 on CPAN?
lizmat is glad she doesn't have to maintain her Perl code of 20 years ago
segomos: it's all in S22
segomos i will read s22 then
FROGGS segomos: the release tarball must contain a META6.json, and then it is put in a Perl6/ subdir in your author's directory 20:19
lizmat basically, PAUSE checks whether there is a META6.json file in the distribution
what FROGGS says :-)
FROGGS and then a Perl 6 indexer carries on, instead of the Perl 5 one which handles namespace right very very strightly
rights*
ownership, even 20:20
segomos i'll read it, it sounds like something i'd contribute to
FROGGS segomos: the problem/task is not that difficult, but the codebase is 20:21
lizmat and there is no dev environment
and we don't want to break Perl 5 distribution uploads to CPAN :-)
FROGGS true, very true 20:22
segomos i need to learn how to get the information from CPAN too, i'll read a little and come back and ask some more questions
FROGGS that is where we hook in: github.com/andk/pause/blob/master/...ch.pm#L455
segomos: the idea is to create a huge json blurb by a cron job, containing basic distribution information from the META6.json files 20:23
segomos the most i've looked at cpan is 'cpanm <module>' and that's pretty much my total interaction with it
20:23 rurban1 left
lue FROGGS: how long has that # XXX been there? :) 20:23
FROGGS lue: FROGGS 3 days ago 20:24
let indexing for Perl6 be a noop
[Coke] oooh, cf regex support \b! my day is made! 20:25
20:26 ssutch__ joined
FROGGS PCRE FTW! 20:26
timotimo jnthn: what does your tuit supply look like while the gpw happens? 20:27
20:28 rurban1 joined
dalek ecs: 4591626 | (Elizabeth Mattijsen)++ | S22-package-format.pod:
Some clarifications, after questions from jnthn++
20:28
20:29 ssutch_ left
hoelzro wrt using STD.pm6 to generate syntax files, what did you have in mind? introspecting it to find certain keywords, or...? 20:38
tadzik I was offline for the last 1.5 hours :o 20:42
and didn't even notice
lizmat and you thought we were quiet :-)
tadzik yes!
lizmat well, I have been for many weeks, but not today :-)
20:43 Sqirrel left 20:46 ssutch__ left
timotimo tadzik: i really missed you! 20:46
jnthn timotimo: I should have some during the next week. 20:47
timotimo sounds good
can you tell me how to get the exception classes into the BOOTSTRAP.nqp for the binder to use? 20:48
jnthn You don't
You install things to create such exceptions in P6EX or so
Search for X::Multi::NoMatch for an example 20:49
But here you'd have it construct the exception and return it, rather than actually throw it.
timotimo ah, looks simple
and then i call .throw on it when it's time to go explodey-plode? 20:50
jnthn And have the text fallback for the "compiling th esetting" case.
Yeah
You can use nqp::isstr to see if you have a string one, and if not then you know to .throw
timotimo so if nqp::isstr($error[0]) { nqp::die(... 20:51
yeah
i was about to write that, but i got distracted by Firefly on the telly :)
colomon You can't take the sky from me. 20:59
21:02 dayangkun left
timotimo r: sub test(int $foo) { }; my str $test = "hi"; test($test); 21:03
camelia rakudo-parrot ffa572, rakudo-moar ffa572: OUTPUT«===SORRY!=== Error while compiling /tmp/tmpfile␤Calling 'test' will never work with argument types (str)␤ Expected: :(int $foo)␤at /tmp/tmpfile:1␤------> est(int $foo) { }; my str $test = "hi"; …»
..rakudo-jvm ffa572: OUTPUT«java.lang.ClassCastException: java.lang.String cannot be cast to org.perl6.nqp.sixmodel.SixModelObject␤ in sub test at /tmp/tmpfile:1␤ in block at /tmp/tmpfile:1␤␤»
timotimo oh, well.
don't need to turn that into a typed exception if it won't ever fly in the binder
jnthn Um...wat, jvm? :)
Well, on Parrot and Moar it successfully ges it at compile time. 21:04
timotimo right
21:05 sqirrel_ left
timotimo i don't seem to find typed exceptions for most things the binder would complain about >_> 21:06
actually ... i kind of see none 21:10
jnthn Yeah, the binder isn't very typed-execptioned 21:12
21:12 LLamaRider joined
LLamaRider lue: I just came to your P6 type hierarchy repo, while searching for a pretty dependency graph of the built-in type/class system. Did anything come of that? 21:13
lue LLamaRider: the perl6/doc repo does a much better job of having a typegraph :)
LLamaRider does that mean a graph exists in there and your repo is obsolete? 21:14
I will do some digging then
21:15 cognominal left
LLamaRider oh there are nice graphs on the individual class pages, cool! 21:16
lue afk
timotimo where is that repository?
may want to take it down if it's outdated or update it 21:17
FROGGS github.com/lue/Perl-6-Type-Hierarchy
LLamaRider this is lue's repo: github.com/lue/Perl-6-Type-Hierarchy
doc.perl6.org/images/ is exactly what I was dreaming of, though it is missing a large image of everything 21:18
FROGGS m: say ObjAt 21:19
camelia rakudo-moar ffa572: OUTPUT«(ObjAt)␤»
FROGGS what is that?
timotimo what, where?
21:19 dayangkun joined
lizmat it's a type ? 21:20
FROGGS n: say ObjAt
camelia niecza v24-109-g48a8de3: OUTPUT«(ObjAt)␤»
FROGGS seems so...
seen the .svg of it and now I wonder what it does
m: say ObjAt.^methods
camelia rakudo-moar ffa572: OUTPUT«WHICH Str gist␤»
lizmat it's declared in BOOTSTRAP
tadzik can someone read my "rakudo on android" comment (ttjjss.wordpress.com/2014/03/23/per...mment-302) and see if I got it mostly right?
jnthn Specified in S02 I think. 21:21
Related to boject hashing.
FROGGS tadzik: sounds about right to me, but what do I know :o) 21:22
tadzik :P
21:22 guru joined
tadzik FROGGS: I don't think I get your comment on panda bug 21:22
21:22 ajr_ left
FROGGS humm? 21:22
tadzik github.com/tadzik/panda/issues/73
21:22 guru is now known as Guest24017
lizmat S02:620 21:22
synopsebot Link: perlcabal.org/syn/S02.html#line_620
lizmat oops: S02:640
synopsebot Link: perlcabal.org/syn/S02.html#line_640
21:23 cognominal joined, Guest24017 is now known as ajr_
FROGGS tadzik: well, the solution would be to put an absolute path in the installed scripts, right? 21:23
tadzik I think the problem in the bug is happening much sooner 21:24
FROGGS lizmat: thanks
tadzik panda, when it's bootstraping panda, runs 'perl6 ./bin/panda da da da yada'
jnthn tadzik: I think Android runs Dvalik
tadzik fsvo perl6
jnthn tadzik: Not a JVM
FROGGS tadzik: ahh, troo
jnthn tadzik: So the question is when JSR292 gets implemented in Dvalik
timotimo X::TypeCheck::Binding doesn't have a slot for the name of the bound thing
tadzik and that vo perl6 is wrong :P
jnthn bitbucket.org/jpilliet/android-292 is some work towards that 21:25
tadzik jnthn: Dalvik :)
timotimo the current AdHoc does tell which name was bound
jnthn oh wwow :)
tadzik well, I wrote "android jvm implementation", isn't that what dalvik is?
geekosaur not exactly 21:26
there's a step in building an android program that translates jvm bytecode to dalvik bytecode
jnthn Not quite. It runs something other than JVM bytecode, but JVM bytecode can easily be turned into that, as I understand it.
timotimo except if you use invokedynamic
tadzik aha
I'll just say "Dalvik (Android's VM)" 21:27
geekosaur also it seems recent android has an experimental new runtime in it (you can access via developer mode, but there are a lot of nasty warnings on using it)
FROGGS a usable rakudo on android would be a killer... but I guess we are not that near as we would hope :o)
hoelzro dalvik doesn't support invokedynamic =( 21:28
tadzik uh, I overworkedout
FROGGS what about ubuntu phone thingy? has somebody such a phone?
21:28 skids left
timotimo how do i proceed? 21:29
21:29 dmol left, jnap1 joined
tadzik was it ever released, in an installable state? 21:29
oh, I can probably try running rakudo-moar on my n900 :) 21:30
FROGGS yes, you can download it for two phone models
tadzik hoelzro can try running it on Jolla
hoelzro hmm
21:30 jnap left
FROGGS timotimo: the problem is that you would have to extend X::TypeCheck::Binding? 21:30
hoelzro that's a cool idea
no reason I couldn't 21:31
tadzik linking to QML may be a bit of a pain
I know Go needed to get a C++-nativecall to do that
hoelzro =( 21:32
Go has a C++ nativecall?
timotimo ok 623 - right exception type (X::TypeCheck::Binding)
yay
FROGGS: i think i could do that, aye
maybe with just an optional parameter?
FROGGS timotimo++
why not, yes
there are other exceptions that have an optional attribute and therefor two different messages 21:33
tadzik hoelzro: yeah
FROGGS (like with suggestions)
tadzik since 1.something
hoelzro I wonder how they pull that off... 21:34
tadzik and now it can use qml and things
wait
I have a gif just for that
FROGGS a C++ nativecall meme?
tadzik goo.gl/WxAgDl
it speaks in your head, even 21:35
FROGGS hehehe
tadzik++
that might be a response I could give at the gpw on wednesday *g* 21:36
21:36 kaare_ left
tadzik gifs work quite well on presentations 21:36
I tried :P 21:37
FROGGS I just need a rediculous shirt
timotimo ./perl6-m -e 'my class A {}; (-> &c, $m { A.new()(); CATCH { default { $m } } } )(A, "")' 21:39
Type check failed in binding &c; expected 'Callable' but got 'A'
masak heh, I remember that one :) 21:40
it used to blow up horribly.
managed to Null PMC access on the $m not being properly initialized because the CATCH caught the type mismatch error. 21:41
timotimo ouch
masak in the *middle of the signature binding*.
timotimo well, at least we get the name of the parameter on moar now
moar is getting better than the other backends day by day :P 21:42
masak better yet; we don't get a Null PMC access anymore :)
timotimo: which RT # was that, ooc?
timotimo no clue
masak where di you get the code from? 21:43
did*
jnthn I think that one mighta been fixed all the way back in the PAST -> QAST transition.
masak I did a talk about it last summer, but it might've been fixed by then. 21:44
jnthn Which killed PAST::Handler in favor of nqp::handle
In which case we're talking a couple of years ago.
dalek kudo/nom: aa99985 | (Timo Paulssen)++ | src/vm/moar/ops/perl6_ops.c:
turn an adhoc into X::ControlFlow::Return
21:47
kudo/nom: af8d426 | (Timo Paulssen)++ | src/ (2 files):
teach the moarvm binder X::TypeCheck::Binding and X:: in general
timotimo anything wrong with that? :)
lizmat pulls and spectests 21:48
timotimo thank you, liz
masak ah, found it: rt.perl.org/Ticket/Display.html?id=114134 21:51
lizmat timotimo: hmm seems to hang in spectest, rerunning without parallel testing 21:53
timotimo: hangs on t/spec/S02-literals/listquote.rakudo.moar ..................... 1/23 21:54
timotimo er ... huh?
did i accidentally push something i didn't mean to?
21:55 ajr_ left
timotimo i don't think i did 21:55
lizmat testing again without your changes 21:57
21:57 kbaker left
[Coke] so, rakudo-jvm craps out for me when running the roast suite still. Can someone else try it on their machine to see if they can dup it? I can step you through how to do it. 21:59
21:59 ajr joined 22:00 ajr is now known as Guest99590, Guest99590 is now known as ajr_
[Coke] (er, using the eval server) 22:00
lizmat timotimo: should I be able to run without your last commit ?
[Coke] starts aborting as of S10-packages/basic.rakudo.jvm 22:01
not sure if that particular test file is doing it, or if it's just exhausted. :) 22:02
22:04 lustlife left
timotimo hm, the very last one? 22:04
i think it should work with the commit before this one
it *does* seem to hang, wtf? 22:05
22:05 Rotwang left
FROGGS rm -rf install ? 22:06
lizmat feels like an exception being thrown while throwing the exception
FROGGS ohh
lizmat e.g. because it can't find the exception ?
timotimo: trying again without the last commit 22:07
22:13 cognominal left
lizmat timotimo: tests ok without the last commit 22:13
timotimo weeeeird 22:15
i'll review it over and over
lizmat timotimo: in the first else, you're referring to %ex, but if that nqp::gethllsym failed, that wouldn't work, no? 22:16
timotimo that's why i nqp::isnull it
lizmat but if it *is* nwp::isnull, how can you refer to it in the else ? 22:17
$error[0] := nqp::atkey(%ex, 'X::TypeCheck::Binding...
timotimo er
i think i forgot to ! it :)
jnthn That's !good...
timotimo er
no, if it's not nqp::isnull, the else will fire
lizmat hmmm... guess I read it wrong :-) 22:18
timotimo am i infinilooping by causing the binder to fail when it's trying to report a bind fail? 22:19
lizmat feels like something like that 22:20
been through many of those while working on the setting
22:20 Rotwang joined 22:22 denis_boyun__ left
lizmat looking at the diff, it can only be the "$error[0] := nqp::atkey(%ex, 'X::TypeCheck::Binding" 22:23
or the "@error[0].throw();"
since the other code paths are basically the old situation 22:24
could nqp::isstr be false when it really is a string? 22:25
so that it tries to throw a string ?
lizmat goes back to writing slides
fwiw: in spectestiing, it seems to hang on: 22:27
ok( ?((1 | 3) < 3), '(...) < 3 no parsefail');
lizmat goes really back to writing slides 22:28
22:31 REPLeffect joined
lizmat timotimo: seems that $error[0] := nqp::atkey(%ex, 'X::TypeCheck::Binding')($oval.WHAT, $nom_type.WHAT, $varname); is throwing an exception 22:34
22:34 spider-mario left 22:35 hoverboard left
jnthn lizmat: On...JVM? 22:35
lizmat no, on Moar
jnthn Ah, OK
Carry on :)
(The NQP-implemented binder ain't used on JVM...)
masak 'night, #perl6 22:38
lizmat night masak 22:39
timotimo: feels to me it is in the use of .WHAT 22:43
timotimo huh, something like "can only call .WHAT on a sixmodelobject? except this is moarvm where it should work :) 22:45
22:46 colomon left
lizmat changing it to: $error[0] := nqp::atkey(%ex, 'X::TypeCheck::Binding')($oval.HOW.name($oval), $nom_type.HOW.name($nom_type), $varname); 22:46
22:46 LLamaRider left
lizmat makes it not hang anymore, but bomb out the test with: 22:46
Type check failed in binding $a; expected 'Str' but got 'Str'
in sub infix:<<> at src/gen/m-CORE.setting:4129
22:46 colomon joined
lizmat which is: proto infix:«<»($a?, $b?) is pure { * } 22:51
so that explains the $a
lizmat stops with debugging and slide writing and calls it a day 22:54
gnight, #perl6! 22:56
22:57 btyler left
lizmat timotimo: final thought for the day: could it be that the throwing now also breaks out of any other exception catching? 22:59
I couldn't otherwise explain the failure in t/spec/S02-literals/listquote.t, test #6 23:00
sleep&
23:02 grondilu joined
grondilu r: say sqrt sqrt abs cos pi/2 # was expecting zero 23:02
camelia rakudo-parrot ffa572, rakudo-moar ffa572: OUTPUT«8.84596494520137e-05␤»
..rakudo-jvm ffa572: OUTPUT«8.845964945201365E-5␤»
23:02 skids joined
grondilu (well I guess that would be asking a lot numerically) 23:04
23:04 Rotwang left
grondilu defines a half_pi role and make a multi case return zero in that case. 23:08
r: role half_pi {}; multi r(half_pi) { 0 }; multi r(t) { sqrt sqrt abs cos t }; say r($_) for 0, pi/4, pi/2 but half_pi 23:10
camelia rakudo-parrot ffa572, rakudo-jvm ffa572, rakudo-moar ffa572: OUTPUT«===SORRY!=== Error while compiling /tmp/tmpfile␤Invalid typename 't' in parameter declaration.␤at /tmp/tmpfile:1␤------> pi {}; multi r(half_pi) { 0 }; multi r(t⏏) { sqrt sqr…» 23:11
grondilu r: role half_pi {}; multi r(half_pi) { 0 }; multi r(\t) { sqrt sqrt abs cos t }; say r($_) for 0, pi/4, pi/2 but half_pi
camelia rakudo-parrot ffa572, rakudo-moar ffa572: OUTPUT«1␤0.917004043204671␤0␤»
..rakudo-jvm ffa572: OUTPUT«1␤0.9170040432046712␤0␤»
grondilu I wonder if it would not be opportunate to define a pi role so that cos(pi) and similar return exact values. 23:16
23:20 hoverboard joined
TimToady seems more like a units type 23:20
yoleaux 23 Mar 2014 19:16Z <Mouq> TimToady: Should :q be able to escape its starter? This is the same as asking if, e.g. "< \< >" should produce "<" or "\<", considering that "< < >" is a syntax error
23 Mar 2014 19:42Z <Mouq> TimToady: And is "In addition, the double angles allow for comments beginning with #. These comments work exactly like ordinary comments in Perl code" still valid and just NYI?
TimToady .tell Mouq yeah, should probably allow :q to \ an opener, and if double angles allow #, it's more because that's how shells work than how Perl works :) 23:22
yoleaux TimToady: I'll pass your message to Mouq.
23:29 kivutar left 23:37 kivutar joined 23:43 skids left 23:48 hoverboard left 23:57 skids joined
skids r: role A { }; role B does A { }; class C does B { }; C.^roles.say; C.^roles(:all).say; 23:58
camelia rakudo-jvm ffa572, rakudo-moar ffa572: OUTPUT«(B)␤(B)␤»
..rakudo-parrot ffa572: OUTPUT«(B)␤too many named arguments: 1 passed, 0 used␤ in any roles at gen/parrot/Metamodel.nqp:2585␤ in block at /tmp/tmpfile:1␤␤»