Perl 6 - the future is here, just unevenly distributed

IRC log for #perl6, 2011-04-02

Perl 6 | Reference Documentation | Rakudo

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

All times shown according to UTC.

Time Nick Message
00:06 Limbic_Region joined #perl6
00:16 cdarroch left #perl6
00:32 ymasory joined #perl6
00:43 demetro left #perl6
00:45 khisanth_ joined #perl6
00:46 Khisanth left #perl6
00:46 khisanth_ is now known as Khisanth
01:19 ajs joined #perl6
01:27 woosley joined #perl6
01:34 jaldhar joined #perl6
01:42 wooden joined #perl6
01:42 wooden left #perl6
01:42 wooden joined #perl6
01:57 kid51 joined #perl6
02:00 whiteknight left #perl6
02:16 kid51 This question was just posed in #parrot:  "can anyone recommend an article or document i can read to get started w/ nqp?"
02:22 sorear perhaps the squaak tutorial; IIRC that was rewritten to use nqp not long ago
02:23 sorear Why can't I define a subregex?
02:24 sorear method versus submethod and $~MAIN versus $~Regex are orthogonal in the implementation, but only three combinations are supported by the syntax
02:26 sorear kid51: ^
02:26 bubaflub joined #perl6
02:30 sorear hello bubaflub.
02:30 bubaflub hi sorear
02:30 sorear weclome to #perl6
02:30 bubaflub sorear: thanks
02:30 bubaflub someone in #parrot was asking for a resource on where to learn npq
02:30 bubaflub i pointed them to moritz++ 's blog
02:30 bubaflub any other suggestions?
02:30 MindosCheng Welcome to perl6!
02:30 sorear squaak
02:31 bubaflub i haven't done too much nqp myself
02:31 bubaflub great
02:31 bubaflub MindosCheng: thanks.
02:31 starcoder left #perl6
02:32 ymasory left #perl6
02:32 starcoder joined #perl6
02:36 bubaflub sorear: anything besides squaak?  or is it perl6-ish enough that general perl6 stuff will do?
02:36 arlinius joined #perl6
02:44 kid51 left #perl6
02:48 leprevost joined #perl6
02:49 ymasory joined #perl6
02:50 dayangkun joined #perl6
02:59 Limbic_Region left #perl6
03:09 wooden left #perl6
03:09 icwiener left #perl6
03:09 agentzh joined #perl6
03:21 dayangkun left #perl6
03:21 dayangkun joined #perl6
03:23 bubaflub left #perl6
03:27 Su-Shee left #perl6
03:28 Su-Shee joined #perl6
03:31 ymasory left #perl6
03:34 aesop left #perl6
03:47 aesop joined #perl6
03:51 mtk left #perl6
03:53 Chillance left #perl6
03:59 mtk joined #perl6
04:05 ajs left #perl6
04:12 bubaflub joined #perl6
04:13 leprevost left #perl6
04:13 lue left #perl6
04:13 leprevost joined #perl6
04:14 lue joined #perl6
04:22 bubaflub left #perl6
04:25 aesop left #perl6
04:27 leprevost left #perl6
04:27 donri left #perl6
04:35 ggoebel joined #perl6
04:44 aesop joined #perl6
04:52 justatheory left #perl6
04:53 aesop left #perl6
05:02 kanishka joined #perl6
05:05 aesop joined #perl6
05:08 mberends hi sorear, I have been following your gtk remarks and also read some documentation. I would with regret ignore gtk3 for now because it is not "ready" in the same way that Perl 6 implementations are not "ready". The Gnome developers decided at their latest conference in March 2011 to postpone the release of stable gtk 3.0 to September 2011 so that the team have time to improve it.
05:08 mberends The Ubuntu developers had already decided not to base their 2011.04 release on gtk3 because of shortcomings in its implementation whose details I do not know. It would not be a wasted effort to target gtk 2.x in the meantime because the changes, though numerous, are evolutionary, not revolutionary. To paraphrase, gtk3 is still gtk.
05:16 aesop left #perl6
05:17 mberends To minimize installation dependency hurdles I would recommend burning a LiveCD out of http://www.debian.org/CD/live/ not because lots of people use it, but because what it contains as base libraries is common to the greatest number of derived Linux distributions. It is a quick way to temporarily start up a perfectly clean and standardized reference platform to discover what you need to install to get your own project running on almost everyone else's
05:17 mberends computer. I just hope that your computer with 384MB ram is happy booting it.
05:29 aesop joined #perl6
05:39 woldrich Heh, I'd recommend email :)
05:39 kanishka left #perl6
05:39 noganex joined #perl6
05:39 mtk left #perl6
05:43 noganex_ left #perl6
05:53 sorear eh?
05:53 sorear (hi)
05:53 sorear (I'm back)
06:12 Mowah joined #perl6
06:14 mberends rakudo: my $d=42; say ++$d++; say $d
06:14 p6eval rakudo 4bf132: OUTPUT«43␤44␤»
06:15 mberends sane
06:16 mberends niecza: my $d=42; say ++$d++; say $d
06:16 p6eval niecza v4: OUTPUT«[31m===[0mSORRY![31m===[0m␤␤"++" and "++" are not associative at /tmp/foKy_STZ64 line 1:␤------> [32mmy $d=42; say ++$d++[33m⏏[31m; say $d[0m␤␤Unhandled exception: "++" and "++" are not associative at /tmp/foKy_STZ64 line 1:␤------> [32mmy $d=42; say ++$d++[33m⏏[31m;
06:16 p6eval ..say…
06:16 sorear stds++
06:16 sorear stds-- std++
06:17 mberends so is ++$d++ a misfeature?
06:18 diakopter pugs: my $d=42; say ++$d++; say $d
06:18 p6eval pugs: OUTPUT«*** Can't modify constant item: VInt 43␤    at /tmp/6lHe3DZPjU line 1, column 15-21␤»
06:20 gabiruh left #perl6
06:21 gabiruh joined #perl6
06:22 sorear mberends: i'd call rakudo's failure to handle :assoc<non> operators a bug
06:24 mberends yes, it is an unwanted surprise to users accustomed to the C style norms.
06:27 mberends rakudo: say "misfeature".WHAT ~~ "bug".WHAT # with loaded questions you can prove anything ;)
06:27 p6eval rakudo 4bf132: OUTPUT«Bool::True␤»
06:34 Exodist joined #perl6
06:35 Exodist whats the terminology for a... compund object would you call it? that thing based on quantum something.. Where a variable may be one thing, or it may be the other, depends on what is needed when it is used?
06:35 Exodist I ask here cause I know I read about it somewhere in the perl6 spec...
06:36 mberends junction?
06:37 Exodist thats it, thank you.
06:38 birdwindupbird joined #perl6
06:39 moritz good morning
06:40 Exodist got another 20 minutes till it is morning here
06:40 mberends hi moritz
06:41 sorear Exodist: the correct answer is "allomorphic"
06:41 sorear TimToady++ taught me that word
06:41 Exodist Is it just me, or has this channel shrunk significantly since I was last here (couple months I think, wow...)
06:42 moritz /names says 197 nicks... I don't think that's much less than it used to
06:42 moritz I kinda remember it having around 130
06:42 diakopter it was higher, but then a giant netsplit ate a bunch I think
06:43 sorear s/netsplit/channel takeover/
06:43 moritz giant net squid
06:43 Exodist octocat snezed
06:43 Exodist *sneased
06:43 sorear (on the plus side, they made enough noise that we were able to convince #freenode to drop the channel and set new metaops)
06:44 * Exodist probably needs to go to bed before he spells more words that way...
06:44 sorear sneezed.
06:44 Exodist oh, it was litterally taken over at some point?
06:45 sorear yes, some twerp stole 'audreyt' and /kicked out all the lurkers
06:46 Exodist ... I don't know whats more [blank stare] about that, the fact that the person did it, or the fact that he got audreyt's account....
06:46 woldrich what's wrong with people.
06:46 sorear au|irc doesn't use audreyt anymore
06:46 sorear audreyt was never registered
06:46 Exodist brb, dog needs out
06:46 Exodist then how did th account gain ops?
06:47 Exodist *really brb this time
06:47 sorear "effectively"
06:47 Teedev joined #perl6
06:47 diakopter the account 'audreyt' was never registered, to clarify what sorear said
06:48 sorear Exodist: have you been gone long enough to miss the new implementation? :>
06:51 Exodist sorear, most likely, I started a new job bac in december, I think thats when I was last on
06:51 Exodist (back)
06:51 sorear I had already started for a while in December
06:51 Exodist how did the account get ops if it was never registered
06:52 sorear Exodist: you don't need to register an account to give a nickname persistant ops
06:52 Exodist sorear, I think I stopped paying attention though a month or 2 before that as my faux go at it fizzled (back as my new language attemp is ramping up, though its not perl6)
06:52 Exodist sorear, my misunderstanding, I assumed that since it wasn't registered it was never used, and thus should not have had ops.
06:53 Exodist and that the person came in and convinced someone to op them
06:53 sorear niecza: "foobar" ~~ / [ foo | \w+ ] { say "got here: $/.pos" } /
06:53 p6eval niecza v4: OUTPUT«got here: Cursor()<instance>.pos␤»
06:53 sorear niecza: "foobar" ~~ / [ foo | \w+ ] { say "got here: $/.pos()" } /
06:53 p6eval niecza v4: OUTPUT«got here: 6␤»
06:53 sorear look! LTM!
06:54 Exodist been so long since I looked at perl6 grammars stuff that I am a bit rusty, but I remember LTM being a hickup, so congrats
06:59 sorear "hickup"?
06:59 dayangkun_ joined #perl6
06:59 Exodist bad word use on my part, they were difficult.
06:59 diakopter hiccup
07:00 Exodist bas language abuse on my part :-)
07:00 Exodist *bad
07:00 Exodist wow, I really should probably call it a night.
07:00 dayangkun_ left #perl6
07:26 Teedev left #perl6
07:51 nymacro joined #perl6
07:51 tadzik self.awake :)
07:51 sorear hello tadzik
07:52 tadzik hmm, I wouldn't actually mind a longer timeout
07:52 tadzik hello sorear
07:52 Mowah left #perl6
07:52 woosley left #perl6
07:55 sftp left #perl6
07:56 sftp joined #perl6
07:58 dalek niecza: 891ede5 | sorear++ | lib/Kernel.cs:
07:58 dalek niecza: Incremental method table generation for single inheritance
07:58 dalek niecza: review: https://github.com/sorear/niecza/commit/891ede56ed
07:58 dalek niecza: 6128027 | sorear++ | lib/ (2 files):
07:58 dalek niecza: New multiple dispatch model
07:58 dalek niecza:
07:58 dalek niecza: Proto and dispatch subs are now inferred at MRO construction time,
07:58 dalek niecza: instead of always deferring set calculation to call time.  This
07:58 dalek niecza: opens the door to lexical multis and type-based dispatch.
07:58 dalek niecza: review: https://github.com/sorear/niecza/commit/6128027c29
07:59 * sorear asleep soon; looks forward to trying to get those things in tomorrow
08:24 Teedev joined #perl6
08:24 Teedev left #perl6
08:29 birdwindupbird left #perl6
08:36 birdwindupbird joined #perl6
08:46 frettled joined #perl6
08:51 marcio_ferreira joined #perl6
09:05 dayangkun left #perl6
09:07 dayangkun joined #perl6
09:28 masak joined #perl6
09:28 masak yo zebrists.
09:28 phenny masak: 01 Apr 23:10Z <Tene> tell masak about http://ethanschoonover.com/solarized
09:28 masak Tene: that page is beautiful, and perfectly captures the kind of color esthetic/awareness I'd like to see in my color picker.
09:29 masak just like CRUD screens are thoughtless and user-unaware, most of today's color pickers are tailored to bits and bytes, not to brains.
09:30 Tene I thought you'd like that.  I'm always glad when I get things like that right.
09:30 Tene :)
09:30 masak :)
09:30 masak I'm pretty good at mentally deciphering a six-digit hex number into an actual color. but on some deeper level, I don't *care*.
09:30 * moritz finds the text on that page slightly difficult to read, due to lack of contrast
09:31 masak moritz: ssh! it's scientific, so it can't be wrong! :P
09:32 masak but yes, I agree. backgrounds needn't be #fff, but text should be #000, or so near as making no difference.
09:33 masak I thought I'd try making a Yapsi release today. I just need to run a small errand first.
09:33 Tene masak: and your thoughts on the reverse?
09:33 masak Tene: sorry, could you rephrase that question?
09:33 Tene masak: given light text on a dark bg, does the bg need to be #000?  does the text need to be #fff?
09:34 masak hm.
09:34 masak I... don't know. would need to experiment a bit with that.
09:34 * moritz is not good at design, but good at identifying bad design
09:35 masak Tene: but I generally have that kind of setup in dark environments at night, and then I think colors need not be bright at all. a #777 text would probably be fine.
09:35 Tene 'k
09:35 masak I think I'd definitely prefer the bg to be #000 in that case.
09:37 masak &
09:40 bacek left #perl6
09:41 orafu left #perl6
09:45 orafu joined #perl6
09:55 _twitch joined #perl6
10:08 benabik_ joined #perl6
10:08 benabik left #perl6
10:08 benabik_ is now known as benabik
10:14 agentzh left #perl6
10:21 MayDaniel joined #perl6
10:24 LarsTorben joined #perl6
10:24 LarsTorben !!! DAS neue LINUX www.snowlinux.de THE NEW linux !!!!!
10:24 LarsTorben left #perl6
10:25 araujo left #perl6
10:26 masak LarsTorben: you might want to consider a PR strategy for your product that doesn't piss people off.
10:28 sorear masak: you need to set your client to ignore things said by people who immediately quit after talking
10:29 dalek niecza: f6aa8d6 | sorear++ | / (3 files):
10:29 dalek niecza: Use new dispatcher API when compiling protoregexes
10:29 dalek niecza: review: https://github.com/sorear/niecza/commit/f6aa8d6e58
10:29 dalek niecza: 58a9396 | sorear++ | lib/Cursor.cs:
10:29 dalek niecza: Remove old protoregex handlers
10:29 dalek niecza: review: https://github.com/sorear/niecza/commit/58a9396ae5
10:30 bacek joined #perl6
10:35 dalek niecza: 2eec5c9 | sorear++ | test2.pl:
10:35 dalek niecza: Test multi regex without proto regex
10:35 dalek niecza: review: https://github.com/sorear/niecza/commit/2eec5c953c
10:36 masak sorear: interesting notion. I wonder if my client does that.
10:51 MayDaniel left #perl6
10:54 allbery_b is now known as geekosaur
10:54 benabik left #perl6
10:56 benabik joined #perl6
11:00 benabik left #perl6
11:01 benabik joined #perl6
11:10 * moritz finds it hilarious that spammers are too silly to even include a http:// in their URLs
11:15 karupanerura joined #perl6
11:15 Chillance joined #perl6
11:19 masak moritz: he was in a hurry! no time for protocol. :P
11:21 whiteknight joined #perl6
11:36 kanishka joined #perl6
11:37 birdwindupbird left #perl6
11:37 birdwindupbird joined #perl6
11:38 birdwindupbird left #perl6
11:38 birdwindupbird joined #perl6
11:39 birdwindupbird left #perl6
11:39 birdwindupbird joined #perl6
11:42 kanishka left #perl6
11:43 envi joined #perl6
11:49 birdwindupbird left #perl6
11:49 birdwindupbird joined #perl6
11:50 birdwindupbird left #perl6
11:50 birdwindupbird joined #perl6
11:53 Patterner left #perl6
11:54 lateau joined #perl6
11:55 Psyche^ joined #perl6
11:55 Psyche^ is now known as Patterner
11:56 lateau left #perl6
11:56 lateau joined #perl6
11:58 ggoebel left #perl6
12:01 kanishka joined #perl6
12:04 benabik_ joined #perl6
12:04 masak left #perl6
12:05 benabik left #perl6
12:05 benabik_ is now known as benabik
12:06 kanishka left #perl6
12:19 dayangkun left #perl6
12:22 Axius joined #perl6
12:22 kanishka joined #perl6
12:31 ccc joined #perl6
12:39 cjk101010 joined #perl6
12:52 felher How would one let rakudo run all tests in perl6/roast and get an overview? Is there already something or should one just script something like: >> for i in **/*.t; do rakudo $i 2>&- 1>&- || echo "test $i failed"; done <<
12:53 moritz TEST_JOBS=3 make spectest
12:53 moritz tha's the list of known tests
12:53 moritz erm, tests that are known to pass
12:53 moritz and if you want to run ALL tests, use 'make spectest_full'
12:55 felher moritz: Ah, it clones roast. K, thnx!
13:00 kid51 joined #perl6
13:11 sftp left #perl6
13:31 Axius left #perl6
13:32 sftp joined #perl6
13:34 araujo joined #perl6
13:42 Mowah joined #perl6
13:45 arlinius left #perl6
13:48 tadzik brainstorming. Modules go to lib/, execs go to bin/, where do generic things go, like examples? res/, misc/?
13:51 Axius joined #perl6
13:56 Axius left #perl6
13:59 Axius joined #perl6
14:01 woldrich examples/ ?
14:01 tadzik well, something more general I suppose
14:01 woldrich that's where I put mine. And most other people as well, as far as I can tell
14:02 tadzik seen supernovus
14:02 aloha supernovus was last seen in #perl6 74 days 16 hours ago saying "Well, that set of optimizations seems to have helped, went from 1m16.339s with the old page generation method to 0m41.936s with the new one.".
14:02 woldrich doc/ ?
14:02 kid51 left #perl6
14:02 tadzik woldrich: where would you expect such things to be after you installed a module? in ~/.perl6/where?
14:03 woldrich I wouldn't expect it to be installed at all ...
14:03 tadzik int eresting;
14:04 woldrich str ange?
14:04 tadzik !at-all
14:04 tadzik I see that supernovus puts those in docs/
14:04 ajs joined #perl6
14:05 ajs Hmm... so it an April fools post is published on a blog no one reads, is it still funny? http://essays.ajs.com/2011/04/clamch​owder-ne-perl-6-to-offer-shell.html
14:05 ajs ;-)
14:05 ajs *if
14:05 woldrich I'd prefer doc, without the 's' suffix. :p
14:05 woldrich important stuff, this
14:05 tadzik I guess I'll implement it and see who complains
14:05 tadzik yeah, doc/ feels more unixy
14:06 tadzik I wonder how creepy would it be to install /doc.+/
14:07 woldrich I assume you know about the good old File::ShareDir?
14:08 tadzik I don't think so
14:08 woldrich might be worth reading the docs, then
14:10 arlinius joined #perl6
14:13 lateau left #perl6
14:30 kanishka left #perl6
14:30 hudnix left #perl6
14:33 risou joined #perl6
14:36 Axius left #perl6
14:36 jimmy1980 left #perl6
14:41 hudnix joined #perl6
14:44 czp joined #perl6
14:49 czp left #perl6
14:55 JimmyZ joined #perl6
14:56 ccc left #perl6
15:08 marcio_ferreira left #perl6
15:15 justatheory joined #perl6
15:21 jimmy1980 joined #perl6
15:23 sorear good * #perl6
15:23 tadzik hello sorear
15:25 marcio_ferreira joined #perl6
15:26 tadzik loliblogged! http://ttjjss.wordpress.com/2011/0​4/02/about-meta-info-and-friends/
15:26 Axius joined #perl6
15:36 * sorear reads the TPF GSoC 2011 ideas list
15:46 moritz tadzik: the META.info spec looks fine... with some small nits as always :-)
15:47 moritz tadzik: for one github doesn't like your links to json.org - not sure if that's your fault or github's
15:48 moritz tadzik: the second is that I'd rename s:g/repo/source/ because when we support links to tarballs, "repo" will sound oh so wrong
15:48 tadzik I don't know either
15:48 tadzik right
15:48 tadzik moritz: wanna commit bit?
15:49 moritz sure
15:50 tadzik OTOH, s/repo/source/ will be oh-so-painful to make everyone migrate on it now
15:50 tadzik so repo-* should stay supported, yet deprecated
15:50 tadzik moritz: done
15:51 tadzik (the commit bit)
15:52 * sorear studies Perl6MultiSub
15:57 moritz tadzik: what do you think of making source-type optional, and use the schema to guess the type?
15:57 tadzik moritz: wfm
16:00 TimToady sorear: you mean rakudo's obsolescent Perl6MultiSub?
16:00 PerlJam what does "optional" mean exactly?  When it can't be guessed, it's required?
16:00 sorear TimToady: how so?
16:01 TimToady the old MD didn't go via proto
16:01 TimToady and there was no way to ref to &foo for the collective list
16:01 TimToady *refer
16:02 tadzik PerlJam: I meant "the module manager may make use of it, may not"
16:02 TimToady that's why we respec'ed it to use 'dispatch' routines several months ago
16:02 tadzik Panda displays it on `panda info`, ignores it in every other case
16:02 sorear well, I'm actually just interested in the dispatch algorithm
16:02 sorear esp. the minutae of candidate list sorting
16:02 TimToady ah, that parts pretty good
16:02 TimToady *'
16:03 PerlJam tadzik: This META.info is *only* for modules or could it be used for installed programs too?
16:03 tadzik PerlJam: oh, it's for everything. Ufo uses it too
16:03 TimToady I don't know how many of the minutae have been carried over to 6model yet...
16:04 TimToady though certainly 6model is better formed at the higher levels
16:06 nymacro left #perl6
16:07 marcio_ferreira left #perl6
16:08 * moritz has a git fail
16:08 moritz I can't seem to push my commit
16:08 moritz willt ry something else
16:09 moritz tadzik: pushed
16:14 donri joined #perl6
16:15 JimmyZ left #perl6
16:17 jimmy1980 left #perl6
16:20 justatheory left #perl6
16:21 jedai_ joined #perl6
16:22 * TimToady wanders off to tear off more bits of his house...
16:22 jimmy1980 joined #perl6
16:22 jimmy1980 left #perl6
16:23 [Coke] oooh, destruction. That part is fun.
16:23 jimmy1980 joined #perl6
16:23 jimmy1980 left #perl6
16:23 jimmy1980 joined #perl6
16:24 jedai left #perl6
16:24 birdwindupbird left #perl6
16:25 risou left #perl6
16:26 sbp left #perl6
16:26 sbp joined #perl6
16:27 ymasory joined #perl6
16:34 tadzik moritz: why the new branch?
16:38 tadzik moritz+= want to discuss my GSoC proposal a bit in a while?
16:48 pmurias joined #perl6
16:49 pmurias sorear: mildew had multimethods implemented mostly in Perl6
16:49 phenny pmurias: 01 Apr 00:29Z <sorear> tell pmurias AFAICT GSoC doesn't like 'exploratory programming' much.  We don't know if the CL-backend will still have decent performance after you make 1 + "2" work
16:49 Axius left #perl6
16:51 pmurias sorear: see https://github.com/perl6/mu/blob/master/v​6/Mildew-Setting-SMOP/MildewCORE.setting
16:59 _twitch left #perl6
17:00 Rotwang joined #perl6
17:02 sbp left #perl6
17:02 sbp joined #perl6
17:16 tylercurtis joined #perl6
17:18 felher Where is defined how sub-routines like print/say work? (Are they even sub-routinges at all?). I noticed that they behave differently than they did in perl5 (e.g. say; doesn't print $_ anymore).
17:22 karupanerura left #perl6
17:23 tylercurtis In S32/IO. Perl 6 gets rid of all (I think) of the implicit argument stuff from previous versions of Perl. Instead, there's a general implicit invocant syntax for calling methods on $_ more conveniently.
17:23 tylercurtis So, .say insteadd of say;
17:25 enigma_ymee joined #perl6
17:32 icwiener joined #perl6
17:33 pmurias left #perl6
17:43 lionheart joined #perl6
17:46 enigma_ymee left #perl6
17:54 sorear tylercurtis \o/
17:58 tylercurtis sorear: Hi.
18:04 lionheart left #perl6
18:08 MayDaniel joined #perl6
18:09 sorear tylercurtis: how are things for you?  I'm excited, MMD in Niecza is *really* close now
18:18 sbp left #perl6
18:18 sbp joined #perl6
18:18 sbp left #perl6
18:18 sbp joined #perl6
18:20 felher tylercurtis: Oh, k, thnx. :)
18:27 felher tylercurtis: btw: seems to be S16/IO.
18:27 envi left #perl6
18:28 colomon sorear++
18:46 icwiener left #perl6
18:47 icwiener joined #perl6
18:53 tylercurtis sorear: Things are good for me. I think this quarter my courses will be significantly less demanding time-wise, so I might actually start working on Perl 6 or Parrot-related things once more.
18:53 sorear yay!
19:01 icwiener left #perl6
19:03 GinoMan joined #perl6
19:12 cjk101010 left #perl6
19:12 pmurias joined #perl6
19:13 pmurias sorear: have you decided if you will apply for GSoC this year?
19:18 justatheory joined #perl6
19:18 sorear pmurias: I haven't seen anything I really want to apply for
19:21 jevin left #perl6
19:21 pmurias sorear: and doing a niecza related project?
19:23 sorear EINSUFFICIENTGOALS
19:26 * pmurias doesn't really know what to apply for too
19:33 justatheory left #perl6
19:33 justatheory joined #perl6
19:39 icwiener joined #perl6
19:45 VXZ joined #perl6
19:46 sorear rakudo: multi sub foo(:bar(:$quux)!) { }; foo(:quux); say "alive"
19:46 p6eval rakudo 4bf132: OUTPUT«alive␤»
19:46 sorear rakudo: multi sub foo(:bar(:$quux)!) { }; foo(:bar); say "alive"
19:46 p6eval rakudo 4bf132: OUTPUT«No applicable candidates found to dispatch to for 'foo'. Available candidates are:␤:(Any :quux(:bar($quux))!)␤␤  in main program body at line 22:/tmp/g7rVbbaDBi␤»
19:47 sorear I broke rakudo's MMD engine ...
19:55 * moritz submits a bugreport
20:00 sorear moritz: the bug (which I noticed in the code while trying to read it) is around perl6multisub.pmc line 259
20:01 sorear moritz: if a parameter has 2 or more names, it is not safe to make the assumption that (the whole signature can only match if the first name is present)
20:01 sorear it's a slightly bogus optimization
20:01 moritz agreed
20:17 marcio_ferreira joined #perl6
20:20 jevin joined #perl6
20:22 pmurias sorear: you have seen mildew's MMD implementation?
20:22 sorear pmurias: yes, but I don't really understand how the mildew bootstrap works
20:23 pmurias some classes are defined in C+DSL
20:23 pmurias other ones that use those are written in Perl 6
20:23 pmurias mildew is written in Perl 5 so it doesn't bootstrap
20:23 sorear also, current versions of mildew are incompatible with current versions of gcc
20:24 sorear even with -O0 gcc 4.4 wants >2GiB to compile the .c file resulting from MildewCORE
20:25 pmurias the performance from the code generated by mildew was horrible so there is not much use trying to get it to work
20:26 pmurias sorear: there is a second backend in mildew that generates much smaller C files (arrays of bytecode, rather than giant switches)
20:28 pmurias sorear: the tricky part in the boostrap of the mildew object system is that the 'use adhoc-signatures' makes the signatures be simpler objects implemented in C rather full signatures (which are a class written in Perl6)
20:29 sorear ahah
20:29 sorear did mildew/smop do gather?
20:30 pmurias don't remember
20:30 pmurias it had everything that was need for that for sure
20:30 pmurias coroutines etc.
20:31 sorear How did coroutines work with 'polymorphic eval'?
20:33 pmurias polymorphic eval is basically 'while 1 {invoke the current couroutine}'
20:34 pmurias * coroutine
20:34 pmurias perl5 interop worked using Coro
20:34 dalek niecza: d9d73c9 | sorear++ | / (2 files):
20:34 dalek niecza: Parse "multi method" correctly, add basic type-MMD tests
20:34 dalek niecza: review: https://github.com/sorear/niecza/commit/d9d73c9af0
20:34 dalek niecza: 39cf462 | sorear++ | lib/Kernel.cs:
20:34 dalek niecza: Implement MMD longname extractor
20:34 dalek niecza: review: https://github.com/sorear/niecza/commit/39cf4627c9
20:34 pmurias continuations would have require adding a .clone method to coroutines
20:34 sorear sounds a lot like the for (;;) { th = th.code(th); } at the heart of niecza
20:35 pmurias yes, seems to be the same thing
20:35 pmurias is .code polymorphic?
20:37 sorear yes
20:38 sorear niecza requires a rigid format for the frames; they aren't general MOP-objects
20:42 pmurias niecza seems to have a lot in common with smop, with more things hardcoded (so faster) and loads more opcodes instead of having things in C
20:44 pmurias the main reason i stoped working on smop/mildew wasn't any big design problems but that writing a compiler+vm by myself didn't seem a fun/sensible thing
20:45 pmurias with niecza using .net you don't have to worry with things such as writing a fast garbage collect
20:45 pmurias using reference counting was a design problem
20:45 pmurias as it's really a pain to debug
20:45 sorear I was under the impression reference counting was a sacrifice to allow p5interop
20:48 pmurias one of the plans was to integrate p5 and smop really closely (add a ->RI pointer to SV*)
20:49 pmurias sorear: reference counting is not nessary for p5interop but helps
20:50 pmurias sorear: and use p5 types for strings etc.
20:50 justatheory left #perl6
20:51 pmurias it might be fun to try that with niecza and a (rewritten) smop at some point
20:55 Patterner left #perl6
20:55 cjk101010 joined #perl6
20:56 arlinius left #perl6
20:56 Psyche^ joined #perl6
20:56 Psyche^ is now known as Patterner
20:56 pmurias sorear: the other reason besides p5 integration for using reference counting is that writing a fast gc is a hard thing
21:01 GinoMan left #perl6
21:02 sftp left #perl6
21:03 arlinius joined #perl6
21:08 MayDaniel left #perl6
21:09 sftp joined #perl6
21:11 starcoder left #perl6
21:11 starcoder joined #perl6
21:20 jevin left #perl6
21:22 jevin joined #perl6
21:37 aesop left #perl6
21:44 zer0o joined #perl6
21:44 bacek left #perl6
21:49 sorear moritz: Is it a known bug that Rakudo doesn't implement "constrained multis are called in order and are never ambiguous"?
21:51 plobsing I know that bug. I complained about it previously. That's one of the reasons the multi-ackerman example on rosettacode is broken.
22:00 pmurias left #perl6
22:17 aesop joined #perl6
22:18 Tene phenny: tell masak to look at http://vimeo.com/channels/stanfordvis#19185801
22:18 phenny Tene: I'll pass that on when masak is around.
22:20 starcoder is now known as starfoo
22:23 aesop left #perl6
22:24 Mowah left #perl6
22:24 cjk101010 left #perl6
22:34 sorear plobsing: okay.
22:35 * sorear is not quite following how the sorting process is intended to work
22:35 sorear consider the dag A -> B -> C, A -> D
22:35 sorear there are two total orders consistant with this, thus two possible outcomes from "topological sorting"
22:35 sorear 1. A, [B,D], C
22:36 sorear 2. A, B, D, C
22:36 sorear actually I miscounted
22:36 aesop joined #perl6
22:36 sorear 3. A, B, [C,D]  4. A, D, B, C  5. A, B, C, D
22:36 sorear Rakudo produces #1, but they're all wrong
22:38 sorear if the capture in question rejects A, then B and D are tied candidates; however, #2, #3, #4, and #5 will erroneously consider one to have won
22:38 sorear if the capture in question rejects both A and B, then C and D are the tied candidates, but #1, #2, #4, and #5 will fail to detect this
22:39 * sorear pokes TimToady, jnthn, pmichaud
22:40 Rotwang left #perl6
22:40 cognominal sorear is back to BASIC?
22:43 starfoo is now known as starcoder
22:48 sorear cognominal: I don't get it
22:50 cognominal http://en.wikipedia.org/wiki/PEEK_and_POKE
22:51 cognominal I probably betray the generation I belong to
22:53 cognominal and my last line is a literal translation from French and sounds wrong in English :(
23:00 spooneybarger joined #perl6
23:03 sorear cognominal: other than the tense (should be 'betrayed'), that line is completely fine
23:04 zby_home_ left #perl6
23:04 dalek niecza: 2bdc3ff | sorear++ | / (2 files):
23:04 dalek niecza: A few more MMD tests
23:04 dalek niecza: review: https://github.com/sorear/niecza/commit/2bdc3ffc26
23:07 spooneybarger left #perl6
23:29 cognominal Indeed, according to the american heritage dictionary, except for the "deceive" denotation, the various meanings are the same as French.
23:54 jasonmay left #perl6

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

Perl 6 | Reference Documentation | Rakudo