Perl 6 - the future is here, just unevenly distributed

IRC log for #perl6, 2014-10-23

Perl 6 | Reference Documentation | Rakudo

| Channels | #perl6 index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:07 _slade_ joined #perl6
00:10 Celelibi joined #perl6
01:12 Mso150 joined #perl6
01:19 virtualsue joined #perl6
01:23 yeahnoob joined #perl6
01:24 Alina-malina joined #perl6
01:29 raydiak \o
01:35 colomon o/
01:37 * raydiak wishes for a better way to say "either this param or that param is required" than "die 'error' unless defined $a // $b"
01:40 raydiak work on anything cool lately colomon?
01:40 erkan joined #perl6
01:41 colomon not really, at least in the p6 world.  just doing bookkeeping sort of stuff there...
01:42 raydiak I just got back from camping...built a shelter out of trees, that was pretty cool
01:45 colomon very cool!
01:47 ilbot3 joined #perl6
01:47 Topic for #perl6 is now »ö« Welcome to Perl 6! | http://perl6.org/ | evalbot usage: 'perl6: say 3;' or rakudo:,  niecza:, std:, or /msg camelia perl6: ... | irclog: http://irc.perl6.org | UTF-8 is our friend!
01:47 colomon http://www.mlive.com/entertainment/bay-city/index.ssf/2013/03/bay_city_musicians_break_out_t.html # I wasn't there that day
01:47 colomon but that's the place and some of the people
01:48 raydiak nice
01:49 raydiak looks classy
01:49 raydiak my mother plays the fiddle
01:54 colomon what sort of music?
01:54 raydiak bluegrass mostly
01:54 colomon ah, cool.
01:54 raydiak think I used to hear her practice some calssical too from time to time
02:04 timotimo joined #perl6
02:08 oskie joined #perl6
02:10 esaym153 joined #perl6
02:15 oskie joined #perl6
02:23 Possum joined #perl6
02:43 Alula joined #perl6
02:45 noganex joined #perl6
02:50 chenryn joined #perl6
02:55 jack_rabbit joined #perl6
03:07 Possum joined #perl6
03:56 mauke_ joined #perl6
03:59 rindolf joined #perl6
03:59 ugexe joined #perl6
04:03 kaare_ joined #perl6
04:46 ggoebel11111118 joined #perl6
04:56 kaare_ joined #perl6
05:00 moritz star-m: say 'good morning, not just Vietnam'
05:00 camelia star-m 2014.09: OUTPUT«good morning, not just Vietnam␤»
05:04 masak \o/
05:04 masak m: say 'top of the morning to you, #perl6'
05:04 camelia rakudo-moar 315ec6: OUTPUT«top of the morning to you, #perl6␤»
05:12 chenryn joined #perl6
05:17 fhelmberger joined #perl6
05:22 _sri joined #perl6
05:23 darutoko joined #perl6
05:25 kurahaupo joined #perl6
05:25 _sri joined #perl6
05:26 masak interesting. Scala recently rethought their macros, too. http://scalamacros.org/paperstalks/2014-03-02-RethinkingScalaMacros.pdf
05:27 masak I should get back in touch with the Scala macro people.
05:28 _sri joined #perl6
05:29 _sri joined #perl6
05:31 SamuraiJack joined #perl6
05:33 masak this one looks highly relevant as well: http://scalamacros.org/paperstalks/2014-06-17-EasyMetaprogrammingForEveryone.pdf
05:35 masak Ven: re http://irclog.perlgeek.de/perl6/2014-10-22#i_9550354
05:37 masak Ven: I don't think comparisons of the kind "in Lisp, it works like this" are very useful without also considering all the forces that apply in the different languages. just the fact that Perl has sigilled variables and BEGIN time makes it a different environment with features that differ from a Lisp's.
05:37 grondilu raydiak: maybe die 'error' unless defined $a|$b
05:39 grondilu m: say defined (my $)|(my $)
05:39 camelia rakudo-moar 315ec6: OUTPUT«True␤»
05:39 grondilu m: say defined Junction
05:39 camelia rakudo-moar 315ec6: OUTPUT«False␤»
05:39 grondilu m: say defined my $
05:39 camelia rakudo-moar 315ec6: OUTPUT«False␤»
05:40 * grondilu would expected defined (my$)|(my$) to be False
05:40 * masak not
05:40 masak m: say so defined(my $) | defined(my $)
05:40 camelia rakudo-moar 315ec6: OUTPUT«False␤»
05:40 grondilu oh
05:40 masak a junction is an object, and objects tend to be defined.
05:40 masak unless they are type objects, which your junction wasn't.
05:41 grondilu ok.  Makes sense, but it's an easy trap to fell in.
05:42 grondilu m: say "ok" unless defined (my$)|my$
05:42 camelia rakudo-moar 315ec6: ( no output )
05:42 masak it's only easy to fall in if you haven't internalized the way things group during parse.
05:42 masak or, hm. no, that's not it.
05:43 masak the question is more "will `defined` ask about the definedness of the junction, or will it autothread?"
05:43 grondilu that's fine, I shouldn't have expected say $something to evaluate $something in boolean context I guess
05:44 grondilu after all, it does behave as expected when used after a if/unless
05:45 grondilu m: say 1 + 2|3
05:45 camelia rakudo-moar 315ec6: OUTPUT«any(3, 3)␤»
05:46 grondilu on second thought, why 'say defined (my$)|my$' says True?  Why not Juncion(...)?
05:46 grondilu I mean, it seems like it did indeed autothread, only not as expected.
05:48 grondilu I guess it meant "the junction is defined"
05:49 grondilu but then I don't see why adding 'so' would change anything
05:57 masak my parentheses do change things, though.
05:58 masak defined(my $) | defined(my $)
05:58 gfldex joined #perl6
05:58 grondilu ah yeah indeed
06:00 mauke_ joined #perl6
06:03 mauke joined #perl6
06:05 masak m: my $junction = any(Any, Any); say defined $junction
06:05 camelia rakudo-moar 315ec6: OUTPUT«True␤»
06:06 masak and `defined` evaluates the definedness of the junction. it doesn't autothread over the individual parts.
06:06 cognominal joined #perl6
06:07 yukko_ joined #perl6
06:08 d4l3k_ joined #perl6
06:09 tinita joined #perl6
06:10 erdic_ joined #perl6
06:10 felher_ joined #perl6
06:11 jnthn_ joined #perl6
06:11 anocelot_ joined #perl6
06:11 epochbell joined #perl6
06:11 amkrankr1leuen joined #perl6
06:11 cosimo joined #perl6
06:12 perlpilot joined #perl6
06:12 pecastro joined #perl6
06:14 mauke joined #perl6
06:14 renormalist joined #perl6
06:14 ilogger2 joined #perl6
06:16 geekosaur joined #perl6
06:19 lucs joined #perl6
06:19 * masak commutes &
06:26 woshty joined #perl6
06:32 [Sno] joined #perl6
06:40 Mso150 joined #perl6
06:42 epochbell joined #perl6
06:50 sergot hi o/
06:50 virtualsue joined #perl6
06:56 pochi_ joined #perl6
06:57 kst` joined #perl6
06:58 ggoebel11111119 joined #perl6
06:59 hoelzro joined #perl6
06:59 sjn_ joined #perl6
07:00 dana joined #perl6
07:00 cosimo_ joined #perl6
07:10 kjs_ joined #perl6
07:15 bjz joined #perl6
07:27 brrt joined #perl6
07:33 zakharyas joined #perl6
07:41 ghostlines joined #perl6
07:44 rurban joined #perl6
07:45 rurban left #perl6
07:47 Ven joined #perl6
07:47 Ven I was looking at my youtube feed, and I saw https://www.youtube.com/watch?v=ok1bcGvgqr0 :P
07:47 Ven I think I recognized masak++
07:48 FROGGS joined #perl6
07:49 masak yup, that seems to be me.
07:56 ptc_p6 joined #perl6
07:59 nebuchadnezzar joined #perl6
08:01 ptc_p6 joined #perl6
08:02 timotimo i saw that one already :)
08:02 timotimo oh, hold on
08:02 timotimo i saw that live and during the most important bit the stream cut out
08:02 masak a guy came up to me afterwards and told me about a few ways to improve the way points move from A to B.
08:03 masak the important optimization would be to keep all points fixed that exist in both pictures.
08:03 timotimo oh
08:03 timotimo yeah, that'd probably globally reduce the sum path length nicely
08:04 sqirrel joined #perl6
08:06 masak but I realized when he said that, that after that I would also like to greedily find paths with the shortest distance, and solve for those first
08:06 masak largely removing the need for simulated annealing
08:06 timotimo mhh
08:07 timotimo also, the animation would be less interesting if most pixels would just stay put
08:16 chenryn joined #perl6
08:23 bjz joined #perl6
08:24 dakkar joined #perl6
08:27 lizmat_ https://www.youtube.com/watch?v=bhAIyrk2ogo   # 1st jnthn talk at YAPC::EU
08:28 lizmat_ https://www.youtube.com/watch?v=CnCmU85mV7A  # lichtkind's talk
08:28 lizmat_ more seem to be on the way now
08:29 Ven yay :)
08:29 kjs_ joined #perl6
08:31 timotimo hooray!
08:36 vladtz maska: that a guy would have been me John vna Krieken
08:36 ghostlines joined #perl6
08:36 Ven MASKA!
08:36 Ven Looks like a norwegian rally cry :)
08:36 vladtz Oops!
08:36 ghostlines joined #perl6
08:38 arnsholt Bah. Another annoying thing with Python is that the reference grammar has some bits that aren't really useful for a top-down parser
08:39 Ven oh?
08:39 arnsholt Yeah
08:39 arnsholt The LHS of an assignment operation isn't an expression
08:39 arnsholt It has a slightly different syntax
08:39 Ven Ohh, I think I've heard about that one... haha...
08:40 arnsholt Not really a user-facing problem
08:42 Ven I really need to look at snake and v5's internals to see if we could ever bridge them to their Inline:: equivalent
08:44 Ven arnsholt: has snake "big blockers" like v5 and XS?
08:45 Mso150_q joined #perl6
08:51 brrt joined #perl6
08:56 masak vladtz: right! yes! sorry for forgetting the name ;)
08:58 vladtz masak: happens to me all the time!
08:59 masak :D
08:59 masak vladtz: I look forward to trying your suggestion. I will make sure to make some noise on the channel about it when I have something to show.
09:00 masak the past few days I've been engrossed in a small toy problem which could be described as "regular expressions on board game moves"...
09:00 Ven .oO( cameliaze )
09:00 Ven .oO( ... camareliaze! )
09:00 Ven carame*. sigh
09:00 masak I got up one hour early today to hack on it.
09:01 masak hexagonal coordinates are a beast.
09:01 masak baest: not you. :)
09:01 Ven How do you do that? I have to sleep on the bus to actually spend a nice day. otherwise I just get grumpy :P
09:01 masak Ven: I went to bed early.
09:01 masak that's usually the hard part.
09:01 Ven hehehe :P
09:02 Ven but then you could've hacked during that time, maybe
09:02 baest masak: ok, but I can be beastly :)
09:05 masak you're the best, baest. :)
09:05 masak Ven: in the case of yesterday -- no, I couldn't. I tried, but I was just too beat.
09:06 Ven surely you meant baet?
09:06 Ven :P
09:06 * masak .oO( beast, baest, best, beat... )
09:06 Ven 2spooky
09:07 * masak imagines a channel with only nicks of the form /^b<[ae]>+s?t$/
09:08 Ven now imagine messages like that. It's like speaking smurf...
09:09 vladtz so what the smurf is wrong with hex coordinates?
09:11 masak nothing wrong with them. just different.
09:11 masak each cell has six neighbours.
09:13 masak the whole thing can be modeled by taking the ordinary eight neighbors (four side-neighbors, four corner-neighbors) of a square, and then removing two (corner-neighbors).
09:14 vladtz Yes, I remember from the tim ewhen I was doing sims on BCC and FC cristals.
09:14 masak it feels a little asymmetrical because you remove two (opposite) corner-neighbors but keep the other two.
09:15 masak it's all an artifact of wanting to store a rhombic board in a rectangular array.
09:15 timotimo I remember having written some code to do hexagonal grids for cellular automaton display
09:16 masak right now what I'm trying to do is find mirror symmetries across the six possible symmetry lines intersecting in a cell. I haven't done that before, so it's killing me a little. :)
09:17 moritz group theory FTW!
09:17 timotimo I did the thing where I cut alternating corners every row so that it still fills a rectangular area kind of tightly
09:17 moritz masak: have you ever worked with a 3D hex grid?
09:17 masak (I need to find the symmetries because if I know that the pattern is symmetric, then I can save a lot of computation by matching against 6 rotations instead of 12 rotations-and-flips)
09:18 nine_ colomon: moritz setup up camelia on my private webserver. It's mostly idling around so we might as well use those 4 cores for some Perl 6 :)
09:18 masak moritz: no -- I fail to see how it extends cleanly to 3D.
09:18 rmgk_ joined #perl6
09:18 masak moritz: are the hex cells just stacked cylindrically?
09:19 abraxxa joined #perl6
09:19 moritz masak: https://en.wikipedia.org/wiki/Sphere_packing
09:21 brrt joined #perl6
09:22 moritz there are actually two intuitive-ish ways to extend hex to 3D
09:23 moritz face-centered cubic and body-cenetered cubic
09:30 masak interesting.
09:34 ecocode joined #perl6
09:39 chenryn joined #perl6
09:45 erkan joined #perl6
10:22 datums_nb joined #perl6
10:28 denis_boyun joined #perl6
10:28 chenryn joined #perl6
10:41 colomon 198 passing modules this morning
10:44 FROGGS colomon: is that good or bad?
10:44 colomon FROGGS: it's great
10:44 colomon highest number ever
10:44 FROGGS then I am happy :D
10:44 colomon one year ago it was around 100, I believe
10:45 arnsholt Ven: The only big blockers for Snake I'm aware of ATM is that implementing a compiler is a lot of work =)
10:45 Ven arnsholt: definitely :)
10:46 arnsholt I'm sure there're going to be shenanigans I have to deal with in the object model and such, but the parsing and semantics of Python should definitely be implementable
10:46 arnsholt Part of the problem for v5 is that Perl 5 is notoriously, insanely hard to parse
10:47 Ven well, the big *blocker* is XS, though
10:50 dalek Perlito: 1c9d7b1 | (Flavio S. Glock)++ | / (2 files):
10:50 dalek Perlito: Perlito5 - js - split() with multiline regex
10:50 dalek Perlito: review: https://github.com/fglock/Perlito/commit/1c9d7b12d9
10:54 nine_ The same blocker exists for Python support. Alternative Python implementations struggle with it as well.
10:58 arnsholt Oh, yeah. I'm not even considering C extension modules to Python
11:02 timotimo the pypy people seem to be doing fine! :P
11:04 nine_ Yeah, I guess they actually have some users somewhere...
11:04 timotimo oi
11:04 timotimo i take offence to that
11:04 timotimo skeinforge comes with pypy nowadays
11:05 timotimo pypy is pretty fucking awesome, IMO.
11:05 nine_ "Skeinforge is a tool chain composed of Python scripts that converts your 3D model into G-Code instructions for RepRap."
11:05 timotimo not only reprap
11:05 nine_ Like I said...some users somewhere. There's always some nieche where it's possible.
11:05 timotimo pretty much every 3d printer slices with skeinforge
11:09 dayangkun joined #perl6
11:14 chenryn joined #perl6
11:20 leont joined #perl6
11:22 * colomon doesn't understand why File::Spec::Case won't work with panda on linux.
11:23 tadzik linux may be a little sensitive about this
11:23 tadzik that may be the case
11:24 yeahnoob joined #perl6
11:25 FROGGS :P
11:25 FROGGS tadzik: can you be a bit more specific please?
11:25 tadzik :D
11:25 tadzik complaint filed
11:28 colomon File::Spec::Case works fine if you run it uninstalled by hand in Linux.
11:28 colomon it's just panda that fails.
11:28 colomon under panda, I mean
11:28 colomon and it worked fine there two weeks ago
11:29 smls joined #perl6
11:29 colomon To compile on the MoarVM backend, QAST::VM must have an alternative 'moar' or 'moarop'
11:29 colomon build stage failed for IO::Select: Failed building lib/IO/Select.pm]
11:29 colomon ?????
11:29 brrt joined #perl6
11:32 FROGGS hmmm
11:33 FROGGS is there pir code in it?
11:35 colomon yes
11:35 colomon pir::loadlib__Ps('select');
11:35 colomon $!pmc := pir::new__Ps('Select');
11:35 colomon (according to ack pir)
11:35 denis_boyun joined #perl6
11:38 smls m: my Int @a; @a[4]++; say @a[0]; say @a
11:38 camelia rakudo-moar 315ec6: OUTPUT«(Int)␤(Any) (Any) (Any) (Any) 1␤»
11:39 dalek doc: 12d18c1 | (Kamil Kułaga)++ | lib/Language/variables.pod:
11:39 dalek doc: added @*INC
11:39 dalek doc: review: https://github.com/perl6/doc/commit/12d18c1dc3
11:39 smls Is that *supposed* to say "(Any)" there?
11:46 moritz no, (Int)
11:49 * masak submits rakudobug
11:51 Ven joined #perl6
11:51 Ven masak: just reading the backlog now..(it's better to .tell me)
11:52 masak got it.
11:52 Ven yes, it's nothing more than a "uh, I expect the levels to be separated here", at least I think. I have some kind of feeling that *SOMETHING* will go wrong with nested macros :P. But nothing more
11:53 Ven and yes, scala macros used to be available as an "external" plugin named "macro-paradise" (or -ize)
11:54 masak surely -ise, unless you're talking about converting things to a parade :P
11:54 Ven (the big "news" stuff explained here is the quasiquoting)
11:54 Ven (I mean in the .pdf you linked)
11:54 masak *nod*
11:55 Ven I actually heard paulp say the compiler should use it internally. And paulp knows scala pretty good..
11:56 Ven (paulp has a lot to say about scala, a lot of bad stuff. He has a talk "we're doing it all wrong" that explains it. very, very interesting. Makes you think about the "it's all about trade-offs" phrase)
11:56 masak is it online somewhere?
11:56 Ven give me 5 minutes
11:56 Ven masak: https://www.youtube.com/watch?v=TS1lpKBMkgg
11:56 masak ok, will watch.
11:57 Ven (if you don't know scala, this probably won't talk too much to you. Unless you plan to work on p6-jvm, I guess...)
11:58 Ven Though I'm currently more in the dependent-scala clan (as opposed to the macro-scala one :P)
12:00 Ven masak: http://stackoverflow.com/a/4443972/1737909 you might be interested in this. this is more about DT than macros though
12:08 colomon raydiak: ping?
12:10 colomon .tell raydiak in Math::ThreeD's Build.pm, $?FILE.path.absolute.parent.child('gen-libs.p6') is wrong.  The absolute part needs to come last, because it generates a Str instead of a Path.
12:10 yoleaux colomon: I'll pass your message to raydiak.
12:11 Alula joined #perl6
12:12 rurban joined #perl6
12:14 ecocode joined #perl6
12:14 PZt joined #perl6
12:18 dalek Perlito: 2c65ebf | (Flavio S. Glock)++ | / (2 files):
12:18 dalek Perlito: Perlito5 - standard tests for split() - failed 9/81
12:18 dalek Perlito: review: https://github.com/fglock/Perlito/commit/2c65ebfa6f
12:20 Ven masak: well, I didn't manage to find the stackoverflow answer. still: http://stackoverflow.com/questions/12935731/any-reason-why-scala-does-not-explicitly-support-dependent-types/12937819#12937819 quite nice
12:24 dalek roast: 958f352 | usev6++ | S04-phasers/end.t:
12:24 dalek roast: Add tests for RT #111766
12:24 synopsebot Link: https://rt.perl.org/rt3//Public/Bug/Display.html?id=111766
12:24 dalek roast: review: https://github.com/perl6/roast/commit/958f3520e7
12:26 Ven bartolin++ # so many tests!
12:26 Ven (I hope I don't get that wrong)
12:27 masak bartolin++ # awesome
12:27 bartolin Ven: yes, I'm usev6 on github.
12:28 perlpilot So ... any good ideas for the release neame?
12:28 perlpilot *name
12:32 Ven PerlJam: Jammin' good
12:32 PerlJam Ven: typically releases are named after Perl Mongers groups.  :)
12:32 Ven :P
12:32 Ven just open your own pm group then...
12:33 FROGGS PerlJam: just tell 'em you created one :P
12:33 PerlJam If one of you guys starts Salzburg.pm, I'll name the release after it  ;)
12:35 dalek ecosystem: 5315084 | colomon++ | META.list:
12:35 dalek ecosystem: Remove Squerl from Ecosystem
12:35 dalek ecosystem:
12:35 dalek ecosystem: It depends on Perl6::Sqlite, which is not in the ecosystem.
12:35 dalek ecosystem: review: https://github.com/perl6/ecosystem/commit/531508487c
12:38 colomon m: say ($*IN & $*OUT).t
12:38 camelia rakudo-moar 315ec6: OUTPUT«Cannot find method 'isatty': no method cache and no .^find_method␤  in method t at src/gen/m-CORE.setting:15171␤  in sub AUTOTHREAD_METHOD at src/gen/m-CORE.setting:3789␤  in any  at src/gen/m-Metamodel.nqp:2719␤  in block <unit> at /tmp/YpZ85sTFQP…»
12:38 colomon awkward
12:40 colomon (multitasking) https://github.com/tony-o/perl6-http-server-async/blame/master/META.info tony-o switch the dependency to "Pluggable" : "*" on Sept 17th.  Panda seems to be choking on that.  lizmat_ FROGGS tadzik ?
12:46 colomon m: say $*IN.t
12:46 camelia rakudo-moar 315ec6: OUTPUT«Cannot find method 'isatty': no method cache and no .^find_method␤  in method t at src/gen/m-CORE.setting:15171␤  in block <unit> at /tmp/N7MqaGGHc9:1␤␤»
12:46 FROGGS colomon: we don't understand that syntax/structure yet
12:46 colomon FROGGS: fine answer.  :)
12:49 dalek Perlito: 679d546 | (Flavio S. Glock)++ | / (3 files):
12:49 dalek Perlito: Perlito5 - standard tests for split() - tweak
12:49 dalek Perlito: review: https://github.com/fglock/Perlito/commit/679d5465ff
12:50 colomon tony-o: just sent you a pull request
12:52 colomon m: say $*IN.t # lizmat_ , any idea how to fix this?
12:52 camelia rakudo-moar 315ec6: OUTPUT«Cannot find method 'isatty': no method cache and no .^find_method␤  in method t at src/gen/m-CORE.setting:15171␤  in block <unit> at /tmp/iJTH9T3Yer:1␤␤»
12:53 colomon m: nqp::getstdin().isatty
12:53 camelia rakudo-moar 315ec6: OUTPUT«Cannot find method 'isatty': no method cache and no .^find_method␤  in block <unit> at /tmp/iLP9hlAEO3:1␤␤»
12:56 wtw joined #perl6
12:57 guru joined #perl6
12:57 dalek roast: d657053 | usev6++ | S06-macros/unquoting.t:
12:57 dalek roast: Adapt test for RT #122746
12:57 synopsebot Link: https://rt.perl.org/rt3//Public/Bug/Display.html?id=122746
12:57 dalek roast: (was failing for rakudo.jvm)
12:57 dalek roast: review: https://github.com/perl6/roast/commit/d6570538ae
12:59 Alula joined #perl6
12:59 JimmyZ joined #perl6
13:00 dalek rakudo/nom: 2572884 | duff++ | docs/announce/2014.10.md:
13:00 dalek rakudo/nom: update release announcement
13:00 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/2572884418
13:05 bartolin in the release announcment under %PEOPLE% you have usev6 and bartolin. thats one and the same person (me). (Also I don't mind if you use my full name, see https://github.com/usev6)
13:06 PerlJam aye, I didn't peruse the list yet, just included the output from the contributors script
13:06 bartolin okay, just wanted to note ;-)
13:06 PerlJam Thanks :)
13:08 PerlJam bartolin: do you have a preferred email address?
13:08 PerlJam (I'm adding you to the CREDITS file)
13:08 masak PerlJam++
13:09 PerlJam bartolin: if not, that's fine btw
13:09 kill joined #perl6
13:09 amkrankruleuen joined #perl6
13:10 bartolin PerlJam: thanks, for adding me to the CREDITS file ;-)
13:11 dalek rakudo/nom: 59eaee5 | duff++ | CREDITS:
13:11 dalek rakudo/nom: Add bartolin to CREDITS
13:11 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/59eaee5e08
13:14 dalek rakudo/nom: 8bcdfd2 | duff++ | docs/announce/2014.10.md:
13:14 dalek rakudo/nom: Add the current output of contributors.pl :)
13:14 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/8bcdfd2f5f
13:14 PerlJam The release process feels like it's gotten more complicated that I remember.
13:14 [Coke] PerlJam: hasn't gotten any faster, that's for sure. :)
13:15 dalek Perlito: c256085 | (Flavio S. Glock)++ | / (3 files):
13:15 dalek Perlito: Perlito5 - standard tests for split() - cleanup
13:15 dalek Perlito: review: https://github.com/fglock/Perlito/commit/c256085be3
13:15 [Coke] I eventually want to have all the annoying bits automated. :P
13:15 PerlJam amen.
13:15 masak PerlJam: personally I find running all the tests on all the backends to be the hard part.
13:16 [Coke] "if only we had some automated way to do that." :P
13:16 masak ;)
13:16 PerlJam yeah, that bit seems the same though. (tedious)
13:16 masak is there already a this month's moar release?
13:16 PerlJam masak: no.
13:17 bartolin btw, I'm going to fudge two failing tests for rakudo.jvm in S17-supply/lines.t and S17-supply/words.t
13:22 Hor|zon joined #perl6
13:24 Ven Can we stop using Big-O stuff as a reason to justify everything? We're in 2014 now :(
13:24 Ven (and by that, I mean: lol, we have multiple levels of memory.)
13:28 FROGGS Ven: what are you talking about?
13:28 brrt Ven - please elaborate :-)
13:28 brrt as in: interesting point, why?
13:28 Ven http://www.yarchive.net/comp/slow_memory_effects.html
13:29 Ven Big-O notation assumes a uniform memory model
13:29 brrt right
13:29 brrt but
13:29 brrt but
13:29 brrt big-o notation takes problem scope into account
13:30 brrt and that's vital whenever problems get big
13:30 Ven I've been told about two major cases: matrix multiplication and linear programming. You might have a lower big-o complexity (Strassen/Coppersmith-Winograd) but in the real-world, that didn't work out
13:30 brrt well, that depends on the size of your problem. that's what big-o says
13:30 nine_ Big-O notation is a tool, not a law.
13:31 brrt and you may argue that your problem is not determined by complexity but by memory access
13:31 brrt and that's where you become the engineer rather than the scientist
13:31 Ven nine_: this is what I'm arguing for :P
13:31 lizmat colomon: wrt to $*IN.t  , I think I copy pasted that from the old implementation, so my questiion becomes: has this ever worked?
13:32 colomon lizmat: yes, it definitely worked at some point.
13:32 colomon might have been pre-nom
13:32 lizmat eh, ever on moar ?
13:32 colomon no idea
13:33 lizmat seems to work on parrot
13:33 lizmat so, maybe a case of NYI on jvm and moar ?
13:33 lizmat .tell jnthn could it be that $!PIO.isatty is NYI on jvm and moar ?
13:33 yoleaux lizmat: I'll pass your message to jnthn.
13:33 colomon oh, interesting.
13:34 brrt also Ven: "Even odder, for small tables, it may be better to do linear searches,
13:34 brrt touching only a few cache lines, than to hash at all." - this is no news and wasn't news when the e-mail was posted
13:34 colomon I was assuming it was the special case of $*IN / $*OUT that was the issue.
13:34 dalek roast: 19cb59e | usev6++ | S (3 files):
13:34 dalek roast: Fudge for rakudo.jvm release
13:34 dalek roast: review: https://github.com/perl6/roast/commit/19cb59eb91
13:34 lizmat nope
13:34 ajr joined #perl6
13:34 colomon lizmat++
13:35 azawawi joined #perl6
13:35 azawawi hi
13:39 azawawi http://paste.debian.net/128317/    # How can I speed @list.grep( { $_ ~~ /$pattern/ }  ) ?
13:41 azawawi http://paste.debian.net/128319/   # without the <$pattern>... use this please
13:41 Ven can't you .grep(/$pattern/) ?
13:41 azawawi Ven: yeah it was a mistake
13:42 azawawi for 1400 file names (i.e. strings), that grep takes like 10 secs
13:42 azawawi which is very slow
13:43 Ven right. needs some parallelization :P
13:43 azawawi Ven: same performance with .grep(/$pattern/)
13:44 Ven yes, I guessed so. I was just saying :P
13:44 Ven well, maybe a *little bit* faster as the pattern isn't recompiled
13:46 xfix joined #perl6
13:47 PerlJam azawawi: .grep seems to be implemented in terms of .map, so to make .grep faster you'd have to make MapIter faster.
13:49 PerlJam you could probably make your own specialized grep that doesn't use map that is faster.
13:50 * azawawi tries a manual grep :)
13:53 azawawi PerlJam: the bottleneck seems to be in the $file_name ~~ /$pattern/
13:53 colomon is $file_name a string?
13:53 azawawi yeah
13:54 colomon hmmm…. try my $compiled-pattern = m/$pattern/  and then grep($compiled-pattern)  ?
13:54 * colomon doesn't know if that's really compiled
13:54 colomon actually, that seems unlikely to help?
13:54 azawawi tried it... same performance...
13:54 colomon try it with an explicit pattern instead of a variable?
13:55 azawawi it is way faster
13:55 colomon so probably .grep isn't the issue
13:55 azawawi /$pattern/ seems to be the bottleneck
13:55 colomon yeah
14:02 azawawi m: my @s = 'aaa'..'zzz'; say @s.elems; my $p = 'aa'; my $t = time; my @m = @s.grep(/$p/); say @m; say "{time - $t} secs";
14:02 colomon seems like there must be a way around that with EVAL, but I dunno what it is
14:02 camelia rakudo-moar 315ec6: OUTPUT«17576␤aaa aab aac aad aae aaf aag aah aai aaj aak aal aam aan aao aap aaq aar aas aat aau aav aaw aax aay aaz baa caa daa eaa faa gaa haa iaa jaa kaa laa maa naa oaa paa qaa raa saa taa uaa vaa waa xaa yaa zaa␤14 secs␤»
14:02 azawawi m: my @s = 'aaa'..'zzz'; say @s.elems; my $t = time; my @m = @s.grep(/aa/); say @m; say "{time - $t} secs";
14:02 camelia rakudo-moar 315ec6: OUTPUT«17576␤aaa aab aac aad aae aaf aag aah aai aaj aak aal aam aan aao aap aaq aar aas aat aau aav aaw aax aay aaz baa caa daa eaa faa gaa haa iaa jaa kaa laa maa naa oaa paa qaa raa saa taa uaa vaa waa xaa yaa zaa␤2 secs␤»
14:03 azawawi 14 secs for /$p/
14:03 azawawi 2 secs for /aa/
14:04 thou joined #perl6
14:08 immortal joined #perl6
14:08 immortal joined #perl6
14:11 masak azawawi: that... sounds like it could be optimized somewhat.
14:14 brrt btw, do we (perl) have a room for fosdem 2015 yet
14:14 dalek Perlito: 634b582 | (Flavio S. Glock)++ | / (3 files):
14:14 dalek Perlito: Perlito5 - js - create internal function p5tr() - wip
14:14 dalek Perlito: review: https://github.com/fglock/Perlito/commit/634b582416
14:15 brrt seems we do
14:20 azawawi masak: true
14:21 azawawi masak: it seems that /$pattern/ is getting recompiled every time
14:26 FROGGS m: my @s = 'aaa'..'zzz'; say @s.elems; my $p = anon regex { aa }; my $t = time; my @m = @s.grep($p); say @m; say "{time - $t} secs";
14:26 camelia rakudo-moar 315ec6: OUTPUT«17576␤aaa aab aac aad aae aaf aag aah aai aaj aak aal aam aan aao aap aaq aar aas aat aau aav aaw aax aay aaz baa caa daa eaa faa gaa haa iaa jaa kaa laa maa naa oaa paa qaa raa saa taa uaa vaa waa xaa yaa zaa␤2 secs␤»
14:27 FROGGS azawawi: ---^
14:27 kaare_ joined #perl6
14:27 FROGGS colomon: you don't need to EVAL anything there
14:28 colomon FROGGS: will that work if you stick a variable in the anon regex definition?  That's the issue at hand.
14:28 * colomon admits he totally did not know that anon regex syntax.  FROGGS++
14:31 FROGGS m: my @s = 'aaa'..'zzz'; say @s.elems; my $aa = 'aa'; my $p = anon regex { $aa }; my $t = time; my @m = @s.grep($p); say @m; say "{time - $t} secs";
14:31 FROGGS wait for it...
14:31 camelia rakudo-moar 315ec6: OUTPUT«17576␤aaa aab aac aad aae aaf aag aah aai aaj aak aal aam aan aao aap aaq aar aas aat aau aav aaw aax aay aaz baa caa daa eaa faa gaa haa iaa jaa kaa laa maa naa oaa paa qaa raa saa taa uaa vaa waa xaa yaa zaa␤14 secs␤»
14:31 FROGGS m: my @s = 'aaa'..'zzz'; say @s.elems; my $aa = 'aa'; my $p = anon regex { "$aa" }; my $t = time; my @m = @s.grep($p); say @m; say "{time - $t} secs";
14:31 camelia rakudo-moar 315ec6: OUTPUT«17576␤aaa aab aac aad aae aaf aag aah aai aaj aak aal aam aan aao aap aaq aar aas aat aau aav aaw aax aay aaz baa caa daa eaa faa gaa haa iaa jaa kaa laa maa naa oaa paa qaa raa saa taa uaa vaa waa xaa yaa zaa␤2 secs␤»
14:32 FROGGS there you go :o)
14:33 FROGGS m: my @s = 'aaa'..'zzz'; say @s.elems; my $p = 'aa'; my $t = time; my @m = @s.grep(/ "$p" /); say @m; say "{time - $t} secs";
14:33 camelia rakudo-moar 315ec6: OUTPUT«17576␤aaa aab aac aad aae aaf aag aah aai aaj aak aal aam aan aao aap aaq aar aas aat aau aav aaw aax aay aaz baa caa daa eaa faa gaa haa iaa jaa kaa laa maa naa oaa paa qaa raa saa taa uaa vaa waa xaa yaa zaa␤2 secs␤»
14:33 FROGGS azawawi: see this one ---^
14:34 erkan joined #perl6
14:36 JimmyZ looks like /$p/ needs to be optimized
14:36 PerlJam Why does that make a difference?
14:36 treehug88 joined #perl6
14:40 timotimo could be it thinks it has to continue working after $aa gets changed so it could be it has to (or thinks it has to) close over that
14:41 timotimo maybe that ought to grab the value of $aa only one at the start of parsing,  but consist you can change $aa inside a code block, too
14:41 PerlJam m: my @s = 'aaa'..'zzz'; say @s.elems; my $p = 'aa'; my $r = rx/ $p /; my $t = time; my @m = @s.grep($r); say @m; say "{time - $t} secs";
14:41 timotimo really depends on what the specs guarantee (or allow)
14:42 camelia rakudo-moar 315ec6: OUTPUT«17576␤aaa aab aac aad aae aaf aag aah aai aaj aak aal aam aan aao aap aaq aar aas aat aau aav aaw aax aay aaz baa caa daa eaa faa gaa haa iaa jaa kaa laa maa naa oaa paa qaa raa saa taa uaa vaa waa xaa yaa zaa␤14 secs␤»
14:42 timotimo can you add a $please. =such?
14:42 timotimo dann phone keyboards
14:43 [Coke] m: my Int @a; @a[4]++; say @a[0]; @a[2]=3;  say @a
14:43 camelia rakudo-moar 315ec6: OUTPUT«(Int)␤(Any) (Any) 3 (Any) 1␤»
14:44 [Coke] RT #123037 - seems "ok" that the undefined elements are (Any) instead of (Int) there. If you want to know what the constraint is, that's not how you should be checking. What's the harm?
14:44 synopsebot Link: https://rt.perl.org/rt3//Public/Bug/Display.html?id=123037
14:44 guru joined #perl6
14:48 azawawi FROGGS: Perfect :)
14:48 azawawi FROGGS++
14:48 FROGGS [Coke]: imagine:  my Kitty @a; ... ; @a[42].=meow()
14:49 ajr joined #perl6
14:49 sjn jnthn's performance talk at YAPC is out \o/ https://www.youtube.com/watch?v=bhAIyrk2ogo
14:50 FROGGS \o/
14:50 sjn severa other from the main track too D
14:50 sjn :D
14:50 sjn several*
14:50 sjn FINALLY
14:51 azawawi farabi6 is way faster now :)
14:52 FROGGS *g*
14:52 FROGGS nice :o)
14:53 azawawi https://github.com/azawawi/farabi6/commit/da163bc610b726dd5fe6dc4e7271ded5dea414b6   # FROGGS++
14:55 [Coke] FROGGS: doesn't change my answer.
14:55 [Coke] m: my Junction @a; @a[4] = 1|3; say @a;
14:55 camelia rakudo-moar 315ec6: OUTPUT«(Any) (Any) (Any) (Any) any(1, 3)␤»
14:55 [Coke] ok . THAT's a bug. :)
14:55 colomon FROGGS++
14:56 FROGGS [Coke]: the harm is that you cannot rely on your trait when Any sneaks into the autoviv'd slots
14:56 azawawi FROGGS: but it would be nice that /$pattern/ is optimized though
14:57 FROGGS azawawi: I'd like to hear jnthn on that topic, I bet he can explain why it is how it is right now...
14:57 FROGGS azawawi: I can kinda imagine why it should be that why, but I can't explain it
14:57 azawawi oh well :)
14:58 azawawi at least it is faster now :)
14:58 FROGGS that way*
14:58 FROGGS yeah :o)
14:58 azawawi the other problem is https://github.com/MoarVM/MoarVM/issues/149 :)
14:59 azawawi you add a Ctrl-C handler... you get 100% cpu usage for free :)
14:59 FROGGS *g*
14:59 azawawi FROGGS: thanks for your help. I got to go home now :)
15:00 azawawi have a nice weekend everyone
15:00 FROGGS azawawi: see ya :o)
15:00 * azawawi &
15:03 lizmat seems p5p is starting a project to rewrite the regex compiler
15:04 FROGGS lizmat: what's the motivation?
15:04 lizmat if I could only convince those people to use their brainpower more towards the future than towards the past
15:04 FROGGS :o)
15:05 lizmat "Recent bugs and developments in the regex compiler have made me think that the code we have is too close to unmaintainable.  The process of a compiling a pattern is inefficient, difficult to work with, extend or optimise, and IMO error prone"
15:06 FROGGS lizmat: I have that feeling about other code bases as well
15:08 geekosaur so how many times have they rewritten the regex compiler to be maintainable now?
15:08 geekosaur (and how many of those were successful?)
15:09 lizmat at least once, to my knowledge
15:09 brrt tell them we have JIT-table regexes
15:09 brrt ;-)
15:09 FROGGS :P
15:10 lizmat alas, I'm the wrong person to tell the initiator of this project just that  :-(
15:10 geekosaur I can only think of the one successful one, the others ran up against the sudden sinking realization that all those horrible special cases / hacks were relied up by too many other things
15:10 moritz geekosaur: well, since perl 5.10 it doesn't use the stack for recursion anymore. I'm sure that was a non-trivial rewrite
15:10 geekosaur redesigning p5's regex engine comes very close to redesigning p5's entire core
15:11 lizmat well, no, I don't think so: P5's regex engine is quite separate from the rest of the system
15:11 lizmat a pitfall that P6 mended
15:12 chenryn joined #perl6
15:12 geekosaur I recall their being "optimizations" everywhere that led to pretty much everything having "private" access to the regex engine's internals
15:13 geekosaur maybe ilya's rewrite fixed that, I was kinda on my way out of the p5 community at that point, but my impression was untangling the mess was well nigh impossible
15:15 geekosaur (more to the point, your pitfall is an API level thing, I am talking about hidden-action-at-a-distance stuff that p6 is going out of its way to avoid...)
15:15 rurban1 joined #perl6
15:17 dalek Perlito: dc8529e | (Flavio S. Glock)++ | / (2 files):
15:17 dalek Perlito: Perlito5 - js - internal function p5tr() - wip
15:17 dalek Perlito: review: https://github.com/fglock/Perlito/commit/dc8529e670
15:21 lizmat geekosaur: I hadn't realised that IlyaZ has done a rewrite as well
15:21 geekosaur that may date me, it would have been in the early to mid 90s sometime
15:23 PerlJam mid to late 90s by my recollection, but whatever  :)
15:23 yeahnoob joined #perl6
15:24 geekosaur it was certainly being discussed before that. given the state of the rgeex engine at the time, I can well imagine it took that long to actually happen :)
15:25 * geekosaur was sidelined by illness in 1995 and went radio silent for a couple years, so would have missed the denouement
15:26 lizmat geekosaur++ for getting better and back :-)
15:27 molaf joined #perl6
15:27 PerlJam Didn't someone (dmerphq?) add quite a few regex features and optimizations in the last 10 years or so?
15:28 PerlJam (I haven't paid too much attention to p5p since around 2000)
15:29 lizmat PerlJam: yes, he did
15:29 PerlJam And there was no rewrite associated with that?
15:29 lizmat only partial, afaik
15:30 lizmat PerlJam: you're doing the release, right?
15:30 PerlJam lizmat: yep.
15:30 lizmat are we in release freeze?
15:30 PerlJam If you know of a better way to ping jnthn or someone to do a moarvm release, please do  :)
15:32 lizmat afk for a bit&
15:40 * PerlJam forgot how long it takes parrot to do things.
15:43 PerlJam So, when running tests on nqp-p, I get a failure on t/nqp/46-charspec.t  ... Unrecognized character name LATIN SMALL LETTER O at line 10, near ", LATIN SM"
15:44 PerlJam I'm guessing that's because of ICU somehow?
15:47 FROGGS[mobile] joined #perl6
15:49 Hor|zon joined #perl6
15:49 JimmyZ no icu?
15:52 yeahnoob joined #perl6
15:54 [Coke] FROGGS[mobile]: what do you mean, "rely on your trait" ?
15:56 FROGGS[mobile] it means that I want to have Kitties in the slots of an Array[Kitty]
15:57 FROGGS[mobile] so I would expect that I can call methods on a Kitty type object on an autoviv'd slot
15:58 ugator joined #perl6
15:59 [Coke] ... ok, fair enough.
16:03 smls FROGGS[mobile]: I think that already works.
16:03 smls m: my Int @a; @a[4]=4; say @a[0];
16:03 camelia rakudo-moar 315ec6: OUTPUT«(Int)␤»
16:04 smls It's just when you .Str / .gist / .perl the whole array, that the "(Any)" show up
16:05 FROGGS[mobile] ahh, intersting
16:08 moritz m: my Int @a; @a[4] = 4; say EVAL @a.perl
16:08 camelia rakudo-moar 315ec6: OUTPUT«Type check failed in .new; expected 'Int' but got 'Any'␤  in block  at src/gen/m-CORE.setting:9581␤  in method reify at src/gen/m-CORE.setting:8300␤  in method reify at src/gen/m-CORE.setting:8136␤  in method gimme at src/gen/m-CORE.setting:8619␤…»
16:08 moritz that's how you know it's a bug :-)
16:09 PerlJam moritz++
16:14 dalek rakudo/nom: b2e81b9 | duff++ | docs/announce/2014.10.md:
16:14 dalek rakudo/nom: Add release name
16:14 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/b2e81b989a
16:27 Hor|zon joined #perl6
16:33 Mso150 joined #perl6
16:38 * [Coke] wonders if PROFIELR_OUTPUT is intentional.
16:42 japhb Just skimming scrollback, I see a LOT of talk of performance issues and optimization over the last week.  To all those dealing with particular performance problems, pretty please can some of them result in PR's (or commits, of course) to perl6-bench?
16:42 yukko joined #perl6
16:43 japhb I can't keep up with this channel well enough to track all the things we're not yet benchmarking, but if even half the people with a performance itch scratch their way to a benchmark, we'd have much broader coverage.
16:43 japhb :-)
16:43 Isp-sec joined #perl6
16:44 japhb (Myself I'm focused on how to visualize stress test results at the moment.)
16:47 masak japhb++
16:58 guru joined #perl6
17:03 Mso150 joined #perl6
17:08 SamuraiJack joined #perl6
17:15 smls joined #perl6
17:15 smls m: my @a; @a[*-1]; say "alive"
17:15 camelia rakudo-moar 315ec6: OUTPUT«Unsupported use of [-1] subscript to access from end of Array; in Perl 6 please use [*-1]␤  in block <unit> at /tmp/ePbngOP0Nw:1␤␤»
17:16 smls ;)
17:16 vendethiel- camelia wtf?
17:16 smls Should ideally die/warn with "Out of bounds" instead...
17:17 grettis joined #perl6
17:17 smls vendethiel-: It's because it evaluate the WhateverCode first, and then checks the *result* of that to catch said Perl5-ism
17:18 Hor|zon joined #perl6
17:18 masak that bug has been submitted somewhere
17:18 masak it's in RT
17:18 smls k
17:23 smls .oO( Newest discipline in the Perl Olympics: Make Rakudo fail with an error message that tells you to 'use instead' exactly what you were alrady using... )
17:25 ajr_ joined #perl6
17:25 vendethiel- .oO( compiler golfing )
17:27 bronco_creek joined #perl6
17:29 bartolin smls: thats RT #111924
17:29 synopsebot Link: https://rt.perl.org/rt3//Public/Bug/Display.html?id=111924
17:31 moritz m: use fatal; my @a; try { @a[*-1] }; say $!.backtrace.full
17:31 camelia rakudo-moar 315ec6: OUTPUT«  in method throw at src/gen/m-CORE.setting:12125␤  in sub fail at src/gen/m-CORE.setting:13687␤  in sub postcircumfix:<[ ]> at src/gen/m-CORE.setting:2647␤  in sub postcircumfix:<[ ]> at src/gen/m-CORE.setting:2725␤  in block  at /tmp/rFsSLa2j9z:1…»
17:31 bronco_creek joined #perl6
17:34 bronco_creek Newbe question: when I try to write to  a file with $foo.say($string_thing) newlines are never written
17:34 bronco_creek I installed on Windows with the latest rakudo/moarvm msi
17:34 moritz bronco_creek: maybe it just writes a line feed, not a carriage return?
17:35 bronco_creek Even if I explicitly have $foo.say("$string_thing\n") it gets chopped
17:36 moritz bronco_creek: try it with \r\n
17:36 bronco_creek I'll give that a try.  Thanks moritz.
17:42 bronco_creek Moritz: that worked.  .say seems to work differently to a file than to STDOUT.  Thanks again.
17:43 PerlJam sounds like a bug in say to me.
17:44 raydiak good morning #perl6
17:44 yoleaux 12:10Z <colomon> raydiak: in Math::ThreeD's Build.pm, $?FILE.path.absolute.parent.child('gen-libs.p6') is wrong.  The absolute part needs to come last, because it generates a Str instead of a Path.
17:45 colomon o/
17:45 gfldex joined #perl6
17:46 raydiak so if I don't do .absolute first, I guess .parent automatically prepends a '../' then ?
17:46 raydiak I did that b/c I wanted to be working w/the full path, not a relative version of it
17:46 FROGGS joined #perl6
17:46 colomon raydiak: I don't know.  But what you've got there now definitely does not work.
17:47 colomon m: say $?FILE
17:47 camelia rakudo-moar 315ec6: OUTPUT«/tmp/zWnoU31zrk␤»
17:47 raydiak dang, thanks...used to work
17:47 dalek Perlito: 5291a2c | (Flavio S. Glock)++ | / (2 files):
17:47 dalek Perlito: Perlito5 - js - simple tr()
17:47 dalek Perlito: review: https://github.com/fglock/Perlito/commit/5291a2c22b
17:47 colomon yes, I believe this is fallout from lizmat++'s IO changes
17:48 raydiak wonder why it returns a Str instead of a Path (which can be used as a Str)...think it's a bug I should report?
17:49 colomon lizmat?
17:50 lizmat you rang?
17:50 lizmat it's a change of spec
17:50 raydiak why Str instead of Path?
17:50 lizmat because making an IO::Path is rather expensive
17:51 lizmat and in many cases (maybe like the one above), one is only interested in a Str
17:51 avuserow masak: backlogging, saw you mentioning hexagonal grids. Curious how you are representing them and if you've seen http://keekerdc.com/2011/03/hexagon-grids-coordinate-systems-and-distance-calculations/
17:52 avuserow hexagon grids are kind of interesting because there are a bunch of okay ways to represent them, all with pros and cons
17:52 raydiak lizmat: hm...okay, thanks, I won't report it as a bug then :)
17:52 lizmat I'm not saying it is without bugs
17:52 lizmat but please check the spec in S16 / S32:IO
17:52 bronco_creek Bye for now o/
17:53 raydiak will do
17:53 lizmat if you want to turn a Str into a Path, use .IO
17:53 dalek Perlito: a4b08f3 | (Flavio S. Glock)++ | / (2 files):
17:53 dalek Perlito: Perlito5 - js - simple tr() - tweak emitter
17:53 dalek Perlito: review: https://github.com/fglock/Perlito/commit/a4b08f3a1a
17:53 lizmat if you're just about filename mangling
17:54 colomon lizmat++
17:54 lizmat you probably don't need .IO at all (except maybe at the end)
17:54 lizmat you can call all of the IO::Spec functions for your situation using $*SPEC
17:55 lizmat afk for a bit again&
17:55 raydiak thanks for the tips lizmat++
17:56 guru joined #perl6
17:57 raydiak and thanks for the heads-up colomon++
17:57 masak avuserow: hadn't seen that page. thanks.
17:57 masak avuserow: I've always been partial to the representation under "A better course..."
17:58 vendethiel- *g*
17:58 masak avuserow: first I thought the link would take me to http://www.redblobgames.com/grids/hexagons/ , which I did know about.
18:00 masak avuserow: this morning, I was toying with the xyz representation, just to see if it made my mirrorings fall out nicer.
18:03 vendethiel- I was reading stuff up on ocaml, and it really seems people only use the non-oo part of the language
18:03 * masak .oO( they'd walk a mile for a caml )
18:03 Mso150_g joined #perl6
18:06 mauke_ joined #perl6
18:20 ingy seen jasonmay
18:20 vendethiel- \o, ingy. you're missing a dot
18:20 ingy ?
18:20 raydiak .seen vendethiel-
18:20 yoleaux I saw vendethiel- 18:20Z in #perl6: <vendethiel-> \o, ingy. you're missing a dot
18:21 ingy .seen jasonmay
18:21 yoleaux I haven't seen jasonmay around.
18:26 [Coke] hio, ingy
18:27 Mso150 joined #perl6
18:28 rurban joined #perl6
18:35 masak ahoj, ingy
18:35 zakharyas joined #perl6
18:38 sqirrel_ joined #perl6
18:38 FROGGS_ joined #perl6
18:41 vendethiel .oO( ahoy, ingj )
18:43 Hor|zon joined #perl6
18:45 jack_rabbit joined #perl6
18:48 timotimo Jo/
18:50 dalek nqp: b3bfe61 | duff++ | tools/build/MOAR_REVISION:
18:50 dalek nqp: Bump MOAR_REVISION for release
18:50 dalek nqp: review: https://github.com/perl6/nqp/commit/b3bfe61485
18:50 dalek nqp: 8f76ab4 | duff++ | VERSION:
18:50 dalek nqp: bump VERSION to 2014.10
18:50 dalek nqp: review: https://github.com/perl6/nqp/commit/8f76ab4e6d
18:51 moritz PerlJam: you should comment out the two flappy test files from t/spetest.data for the release
18:52 PerlJam moritz: which two would those be?
18:52 moritz PerlJam: S17-procasync/print.t and S17-lowlevel/lock.t
18:52 PerlJam will do, thanks.
18:53 timotimo can't wait to read what these perl6 folk have been doing in the last month
18:57 Guest1419 joined #perl6
18:58 kurahaupo joined #perl6
18:59 amkrankr1leuen joined #perl6
19:03 [Sno] joined #perl6
19:04 FROGGS_ joined #perl6
19:04 Mso150 joined #perl6
19:07 mauke_ joined #perl6
19:07 timotimo so ... is there a simple way to figure out what class in the method resolution order gets their method selected?
19:08 timotimo hmm, maybe that's what find_method is for
19:08 kurahaupo_ joined #perl6
19:09 timotimo aah, .package is the thing i was looking for
19:10 timotimo hah, "set_composalizer"? %)
19:13 rmgk joined #perl6
19:13 PerlJam Anyone have any guidance on what to do about and NQP test that only fails with nqp-p?
19:13 moritz PerlJam: I don't think there's a fudging mechanism
19:14 moritz PerlJam: you'd have to comment out the test for everybody
19:14 moritz PerlJam: or fix it :-)
19:15 PerlJam yeah, that's where I'm at, but I don't think I have will to fix it right now.
19:15 timotimo that should be fine
19:15 timotimo what test is it?
19:15 PerlJam t/nqp/46-charspec.t  the 3rd test
19:16 timotimo oh, charspec? that's unicode trait stuff?
19:16 PerlJam nqp-p barfs with Unrecognized character name LATIN SMALL LETTER O at line 10, near ", LATIN SM"
19:16 timotimo o_O
19:17 timotimo that error report surely is wrong?
19:17 moritz that's such an exotic character!
19:19 timotimo m: Bool(1, 0)
19:19 camelia rakudo-moar 315ec6: ( no output )
19:19 timotimo m: Bool(1, 0).perl.say
19:19 camelia rakudo-moar 315ec6: OUTPUT«Bool::True␤»
19:20 timotimo m: Bool(0, 1).perl.say
19:20 camelia rakudo-moar 315ec6: OUTPUT«Bool::True␤»
19:20 timotimo that's kinda strange
19:20 j4jackj joined #perl6
19:21 dalek perl6-roast-data: f80760f | coke++ | / (5 files):
19:21 dalek perl6-roast-data: today (automated commit)
19:21 dalek perl6-roast-data: review: https://github.com/coke/perl6-roast-data/commit/f80760f05a
19:22 PerlJam timotimo: think of it as Bool( (0,1) ).perl.say and it's less strange I think
19:23 timotimo yeah, but that's not something we do
19:23 PerlJam m: my @a = "a".."z";  Bool(@a).perl.say;   # another example
19:23 camelia rakudo-moar 315ec6: OUTPUT«Bool::True␤»
19:23 timotimo m: enum Boor <Trew Fawls FailNotFawnd>; say Boor(1, 2, 3).perl
19:23 camelia rakudo-moar 315ec6: OUTPUT«Boor::Fawls␤»
19:24 timotimo we don't just turn the arg list into a list in a coercer
19:26 PerlJam perhaps we're missing a constraint somewhere
19:26 timotimo yeah, see Enumeration.pm
19:26 timotimo this dance is kind of interesting
19:26 timotimo nqp::atpos(nqp::p6argvmarray, 1).at_pos(0)
19:26 timotimo is the argvmarray perhaps an array with the positionals and then the nameds as elements?
19:27 timotimo er
19:27 timotimo like a hash of nameds and a list of positionals?
19:27 timotimo well, the invocant, the list of positionals and the hash of nameds perhaps
19:28 timotimo m: say True(0)
19:28 camelia rakudo-moar 315ec6: OUTPUT«[31m===[0mSORRY![31m===[0m Error while compiling /tmp/perAltexPH�Variable '&True' is not declared�at /tmp/perAltexPH:1�------> [32msay [33m�[31mTrue(0)[0m�»
19:28 timotimo m: say True
19:28 camelia rakudo-moar 315ec6: OUTPUT«True␤»
19:28 timotimo m: say Bool(True)
19:28 camelia rakudo-moar 315ec6: OUTPUT«True␤»
19:28 masak m: say Bool(Bool)
19:28 camelia rakudo-moar 315ec6: OUTPUT«False␤»
19:28 dalek nqp: ce6e3cb | duff++ | t/nqp/46-charspec.t:
19:28 dalek nqp: comment out failing test for now
19:28 dalek nqp: review: https://github.com/perl6/nqp/commit/ce6e3cb80b
19:29 masak m: say Int(Int)
19:29 camelia rakudo-moar 315ec6: OUTPUT«(Int)␤»
19:29 masak m: say Str(Str)
19:29 camelia rakudo-moar 315ec6: OUTPUT«use of uninitialized value of type Str in string context  in block <unit> at /tmp/nWcMzAQQzZ:1␤␤␤»
19:29 psch joined #perl6
19:30 lizmat m: my Int @a; @a[4] = 4; say @a.perl   # FWIW, I have a fix for this issue, but will wait until after the release
19:30 camelia rakudo-moar 315ec6: OUTPUT«Array[Int].new(Any, Any, Any, Any, 4)␤»
19:30 araujo joined #perl6
19:30 psch evening, #perl6 o/
19:30 masak lizmat++
19:30 masak psch: \o
19:30 * PerlJam still can't get over how slow computers continue to seem sometimes.
19:31 psch i noticed i appeared as a contributer \o/
19:32 psch i purposefully didn't create a PR for CREDITS, because someone somewhen said the first commit should be adding oneself there
19:32 psch and i'm still not quite in a position to send the CLA
19:32 psch on that note: PerlJam++ for the release
19:33 masak adding oneself to CREDITS is more of a tradition than a requirement, I guess.
19:33 PerlJam I think it makes the contributors list a little nicer in that it shows everyone by name
19:33 PerlJam (if they're in the CREDITS)
19:36 psch fwiw, i don't mind appearing by name, but i also don't need to
19:37 psch and considering that PerlJam probably moved past that point and might have tagged already, it's probably not sensible anymore to do now :)
19:38 PerlJam psch: nah, there's still time
19:38 timotimo here's an interesting question:
19:39 timotimo how do we generate the one-argument cases for infix operators that get meta'd by R?
19:39 timotimo i kind of think R should be a no-op in that case
19:39 travis-ci joined #perl6
19:39 travis-ci NQP build failed. Jonathan Scott Duff 'comment out failing test for now'
19:39 travis-ci http://travis-ci.org/perl6/nqp/builds/38855989 https://github.com/perl6/nqp/compare/8f76ab4e6d1c...ce6e3cb80b80
19:39 travis-ci left #perl6
19:39 lizmat .oO( another out of disk space? )
19:39 timotimo no
19:40 timotimo a single test case in 46-charspec.t is failing
19:40 timotimo on parrot
19:41 lizmat Question: my @a; .say for @a   # which method is called to iterate over @a in this case?
19:42 Ven joined #perl6
19:42 Akagi201_ joined #perl6
19:43 moritz currently (@a, ).map # iirc
19:47 guru joined #perl6
19:47 * lizmat checks
19:48 lizmat nope  :-(
19:48 immortal joined #perl6
19:48 immortal joined #perl6
19:49 moritz m: my @a = 1; die() for @a;
19:49 camelia rakudo-moar 315ec6: OUTPUT«Died␤  in block <unit> at /tmp/X0fhR537cy:1␤␤»
19:50 moritz m: my @a = 1; try { die() for @a }; say $!.backtrace.full
19:50 camelia rakudo-moar 315ec6: OUTPUT«  in method throw at src/gen/m-CORE.setting:12125␤  in sub die at src/gen/m-CORE.setting:759␤  in block  at /tmp/K3rfiKp1wy:1␤  in block  at src/gen/m-CORE.setting:8300␤  in method reify at src/gen/m-CORE.setting:8223␤  in block  at src/gen/m-COR…»
19:50 lizmat moritz++
19:50 moritz lizmat: my theory is that it calls .map
19:50 moritz lizmat: and on the return value, calls sink
19:51 moritz which then goes through the gimme/reify/reify chain
19:51 cognominal joined #perl6
19:51 dalek nqp: e600d5c | duff++ | t/nqp/46-charspec.t:
19:51 dalek nqp: oops, forgot to change the plan
19:51 dalek nqp: review: https://github.com/perl6/nqp/commit/e600d5cd4f
19:52 Mso150 joined #perl6
19:52 psch grml
19:52 psch accident PR from the wrong branch
19:53 ecocode joined #perl6
19:54 kjs_ joined #perl6
19:55 psch just to make sure: U: lines don't exclusively apply to git/svn nicks, but irc as well?
19:55 * moritz wonders how slow it is to read 400k lines in rakudo, and json-ing them
19:56 lizmat reading the lines with :eager should be pretty fast now
19:56 PerlJam psch: pretty sure.
19:56 moritz lizmat: sorry, but :eager is cheating
19:57 moritz to-json('/usr/share/dict/ngerman'.path.lines.classify(*.substr(0, 2).lc)); # is the code I'm timing
19:57 lizmat moritz: please s/path/IO/
19:57 psch PerlJam: https://github.com/rakudo/rakudo/pull/325 then
19:58 moritz lizmat: I'm going with open(), I'm know (how) that one works :-)
19:58 avuserow masak: after making a somewhat bizarre choice for a hexagon grid, I agree with their better course. (My method was to have every other row have its y value offset by 0.5 IIRC, and then pretend it's a square grid. It has interesting properties of its own but isn't great.)
19:59 masak heh.
19:59 masak yeah, I can see how that one's possible.
19:59 masak I do have a soft spot for integers, though.
19:59 lizmat moritz: getting lost in reify land, so keeping the Int @a.perl fix to just .perl
20:00 masak they're God-made, after all.
20:00 lizmat instead of fixing it more generally for 'for @a'
20:00 lizmat afk for a bit&
20:00 raydiak avuserow: I've done the same thing for a triangular grid before
20:00 avuserow so one of these days, I was wanting to make a custom data structure (as opposed to storing it in a hash, ewww) that would allow hex grids to be stored better
20:00 avuserow also interesting is how the game Civilization 5 does hex grids -- answer primarily being "poorly"
20:00 masak I just store mine in a 2s array.
20:00 masak 2D*
20:01 avuserow they have multiple coordinate systems that their game scripts have to convert between (think embedded lua for some game code)
20:01 travis-ci joined #perl6
20:01 travis-ci NQP build passed. Jonathan Scott Duff 'oops, forgot to change the plan'
20:01 travis-ci http://travis-ci.org/perl6/nqp/builds/38857941 https://github.com/perl6/nqp/compare/ce6e3cb80b80...e600d5cd4fa6
20:01 travis-ci left #perl6
20:01 avuserow raydiak: a triangular grid sounds interesting, have any reading on it?
20:02 avuserow masak: I would want a 2D array but with negative indexes, because my use cases involve extending it after the fact in potentially any direction
20:02 raydiak avuserow: nah I just kinda did what seemed right :) it was a 3d terrain-roaming demo I wrote in lua
20:02 PerlJam travis-ci joins just long enough to say something, then leaves?  that feels weird.
20:03 raydiak avuserow: but a triangle grid is basically a hex grid with an extra vertex in the middle of each hex
20:03 moritz PerlJam: lots of bots do that
20:03 moritz PerlJam: like the github notification bot
20:03 masak avuserow: I'm pretty sure you can build that with today's Rakudo.
20:04 PerlJam maybe so
20:04 Hor|zon joined #perl6
20:04 avuserow masak: then maybe I'll try to translate my proof-of-concept P5 implementation to P6... if I stop being so tired after $dayjob
20:05 avuserow oh wait, that implementation does the silly 0.5 offsets... which may actually be more interesting to implement
20:06 fhelmberger joined #perl6
20:06 masak :)
20:07 masak meanwhile, my move pattern matcher now correctly identifies symmetries in patterns.
20:07 psch another thing on my mind is that after bringing my cliparser branch up-to-date by rebasing it onto nom yesterday or so it stops building
20:07 moritz meh, where did the icycle graph / flame graph go?
20:08 moritz I just made a --profile, and there's no such thing anymore
20:08 moritz should be in the call graph tab, right?
20:08 psch apparently partly caused by 1e037277, which is 'Call DYNAMIC when fetching @*INC'
20:09 psch although reverting the parts that break, i.e. lines 42 to 48 doesn't help
20:09 psch i think this might all be too imprecise for feedback, so here's what i'm doing in that branch
20:09 avuserow masak: one final cute thing from my hexagon bookmark stash: http://cssdeck.com/labs/responsive-hexagonal-grid
20:10 treehug88 joined #perl6
20:10 psch i'm inheriting from HLL::CommandLine::Parser and HLL::CommandLine::Result in perl6-land (i.e. src/Perl6/CommandLine) and overwriting the corresponding bits from HLL::Compiler in Perl6::Compiler to use those instead of the HLL equivalents
20:10 masak avuserow: heh.
20:11 psch with the ModuleLoader bits reverted i get an 'Error during STable deserialization (probably from loading two classes that cannot be loaded together)' or some such
20:11 dalek rakudo/nom: 6b24ed7 | (Pepe Schwarz)++ | CREDITS:
20:11 dalek rakudo/nom: Add self to CREDITS.
20:11 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/6b24ed78a3
20:11 dalek rakudo/nom: 4cd5ad5 | duff++ | CREDITS:
20:11 dalek rakudo/nom: Merge pull request #325 from peschwa/credits-psch
20:11 dalek rakudo/nom:
20:11 dalek rakudo/nom: Add self to CREDITS.
20:11 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/4cd5ad5278
20:11 psch i remember having had the same error when my CommandLine::Parser didn't inherit from HLL::CommandLine::Parser
20:12 Isp-sec joined #perl6
20:14 psch https://github.com/peschwa/rakudo/compare/cliparser-with-grammar#diff-4f7fa503a091c26cdd0dc446b6a12c6fR57 this is where i'd suspect to be a problem, but i don't see it
20:15 psch somewhere around there, not that line... :)
20:15 psch jvm error with a NPE, also in ModuleLoader
20:16 dalek rakudo/nom: 93f4dae | duff++ | docs/announce/2014.10.md:
20:16 dalek rakudo/nom: update contributors
20:16 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/93f4dae1b8
20:16 dalek rakudo/nom: 0594d13 | duff++ | t/spectest.data:
20:16 dalek rakudo/nom: comment out flappy tests
20:16 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/0594d13f0b
20:16 dalek rakudo/nom: 66d3b95 | duff++ | tools/build/NQP_REVISION:
20:16 dalek rakudo/nom: [release] bump NQP revision
20:16 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/66d3b95902
20:16 dalek rakudo/nom: a372592 | duff++ | VERSION:
20:16 dalek rakudo/nom: [release] bump VERSION
20:16 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/a37259208a
20:18 Hor|zon joined #perl6
20:22 ecocode joined #perl6
20:26 travis-ci joined #perl6
20:26 travis-ci Rakudo build failed. Jonathan Scott Duff '[release] bump VERSION'
20:26 travis-ci http://travis-ci.org/rakudo/rakudo/builds/38860399 https://github.com/rakudo/rakudo/compare/4cd5ad5278d0...a37259208a0e
20:26 travis-ci left #perl6
20:29 masak 'night, #perl6
20:29 psch good night, masak
20:29 PerlJam good night masak
20:42 virtualsue joined #perl6
20:51 Ven joined #perl6
21:03 * psch sleep &
21:08 * lizmat confirms spectest clean on moar
21:21 itz "Coloring in colorForth has semantic meaning" <=- seems like we missed a trick ;)
21:24 PerlJam Well ... my machine is still spectesting and I have to go take my son to a baseball game.  I'll have to finish the release later tonight.
21:24 PerlJam bbl &
21:26 * lizmat goes to sleep&
21:31 woolfy left #perl6
21:33 woolfy joined #perl6
21:40 erkan joined #perl6
21:40 erkan joined #perl6
21:52 cognominal itz: to do that you must have an environment wich comes with a editor that know the language. Have payed with forth  more than 25 years, I would not be surprised to be possible except I don't know any forth with graphic style as sematic.
21:53 immortal joined #perl6
21:53 j4jackj joined #perl6
21:53 cognominal itz: to do that, you must have an environment which comes with an editor that knows the language. Having played with forth  more than 25 years, I would not be surprised to be possible except I don't know any forth with graphic style as semantic.
22:09 colomon joined #perl6
22:14 erkan joined #perl6
22:20 Hor|zon joined #perl6
22:58 ingy O HAI vendethiel [Coke] masak! o/
22:58 ingy busy day in #inline :)
23:20 chenryn joined #perl6
23:45 immortal joined #perl6

| Channels | #perl6 index | Today | | Search | Google Search | Plain-Text | summary

Perl 6 | Reference Documentation | Rakudo