Camelia, the Perl 6 bug

IRC log for #perl6, 2007-03-02

Perl 6 | Reference Documentation | Rakudo | Niecza | Specs

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

All times shown according to UTC.

Time Nick Message
00:01 TimToady j
00:03 nipra joined perl6
00:09 mako132 joined perl6
00:14 cj TimToady: hurm... I should check to see if I've got the right fonts installed then!
00:15 cj TimToady: what font packages do you recommend?  do you run debian?
00:17 TimToady cj: fedora
00:18 TimToady if it's a font issue, usually they'll put something consistent for the glyphs they don't have
00:19 TimToady but if you're getting gobbledygook, it's probably an encoding issue
00:20 TimToady make sure LANG=en_US.UTF-8 for instance
00:20 [particle] joined perl6
00:25 putter dinner, bbl
00:29 CardinalNumber joined perl6
00:29 Khisanth joined perl6
00:29 macli left perl6
00:39 diakopter joined perl6
00:48 CardinalNumber is now known as ProperNoun
01:09 nicodaemos joined perl6
01:12 CardinalNumber joined perl6
01:16 offby1 what a coincidence!  My cat is eating dinner too
01:19 IllvilJa left perl6
01:26 svnbot6 r15404 | lwall++ | neglected to convert nameroots to new sym form
01:26 svnbot6 r15404 | lwall++ | generalized syntax to allow "use" before file-scope package declarator
01:30 CardinalNumber is now known as ProperNoun
01:35 CardinalNumber joined perl6
01:45 Limbic_Region left perl6
01:52 CardinalNumber is now known as ProperNoun
02:10 meppl joined perl6
02:14 meppl joined perl6
02:18 putter let's see... aliases...
02:20 dmq joined perl6
02:31 rhizo is now known as PEARLDALEK
02:31 PEARLDALEK is now known as rhizo
02:37 bonesss joined perl6
02:37 bonesss joined perl6
02:43 rashakil joined perl6
02:44 mod_cure joined perl6
02:46 |PrinCo| joined perl6
02:48 * |PrinCo| #projectoX la nueva sala en español para ayuda sobre temas relacionados con windows y linux ;) todo los temas/probemas sera solucionados ;);)
02:48 |PrinCo| left perl6
02:50 REPLeffect_ joined perl6
02:52 mod_cure wonder when Perl6 will be out
02:52 tene mod_cure: christmas
02:52 tene we don't know the year, though
02:52 mod_cure i been waiting sinxe 2002 :)
02:54 rodi mod_cure: christmas has come like four times since then...
02:54 tene Well, it might be this year!
02:54 mod_cure u said that 4 years ago :)
02:55 rodi it was as true than as it is now!
03:07 diakopter "is this a rhetorical question?"
03:12 bucky joined perl6
03:13 bucky Randal Schwartz has had his conviction expunged
03:14 revdiablo Source? =)
03:14 revdiablo Cite your sources!
03:14 TreyHarris bucky: !!
03:14 bucky http://it.slashdot.org/article.pl?​sid=07/03/02/0117257&from=rss
03:14 lambdabot Title: Slashdot | Randal Schwartz's Charges Expunged
03:14 TreyHarris he said something about going to court today
03:14 bucky heh
03:15 revdiablo Neat
03:15 bucky yah
03:15 bucky 13 years of prosecutorial persecution
03:17 TreyHarris oops.  apparently this happened a month ago already
03:17 * TreyHarris has been out of the loop
03:17 TreyHarris i got spanked for mentioning it on another channel :)
03:18 bucky effective 1 Feb 2007
03:19 allbery_b heh
03:20 * allbery_b doesn't think he still has the scrollback from when randal popped in here himself and announced it
03:22 bucky BTW thanks to Orrin Hatch and the Patriot Act, the crime of computer intrusion is now punishable to up to a term of life in prison
03:24 cj TimToady: $ echo $LANG
03:24 cj C
03:24 cj d'oh
03:25 bucky heya cj
03:27 kdevin joined perl6
03:54 allbery_b hm.  wonder how much of a speedup pugs would get if there were a precompiled prelude va Data.Binary?
03:54 cj hey TreyHarris
03:54 cj how's life?
03:55 cj hey there bucky
03:55 cj what's going on?
03:55 * cj re-sets his $LANG... BRB
03:57 cj joined perl6
03:57 cj say something in kanji! :)
03:57 bucky cj jan kneschke had a UDF called UDF-mysql_udf_lua-0.9.3.tar.gz  do you know anything about it?
04:01 mako132 joined perl6
04:29 justatheory joined perl6
04:54 leed joined perl6
04:55 rod1 joined perl6
05:00 amnesiac joined perl6
05:11 REPLeffect_ left perl6
05:12 tene Has a trait to indicate caching on a function been specced yet?
05:12 tene "is cached" or "is memoized" or some such?
05:15 drupek12212157 joined perl6
05:17 TimToady well, if you just grep the specs for "cached" you might find something
05:19 tene Huh.  I thought I did that.
05:19 tene Thanks.  Apology.
05:23 TimToady cj: 漢字で何か
05:24 putter ah well.  downside is still no aliases.  upside is some pugs t/regex/  files now run.  three even pass.
05:24 svnbot6 r15405 | putter++ | yet_another_regex_engine - Some pugs t/regex/ test files now run.  A couple even pass.  See README.
05:25 cognominal joined perl6
05:28 putter TimToady: on reflection, it seem premature to attempt STD on yet_another, which still lacks ratchet, sigspace, etc.  maybe in a calendar week or three.
05:28 TimToady whatever, it's just the target
05:28 TimToady don't stress
05:29 putter :) k
05:34 cj bucky: not off the top of my head.  I don't even know what UDF might happen to be.  Ulf Wendel and Mike Zinner did something with lua at one point... you might google them and ask if they know...
05:34 cj TimToady: d'oh.  I blame putty.
05:34 * cj makes a late-night diaper run
05:34 TimToady putty does utf8 okay
05:35 cj hurm... let me twiddle some bits then...
05:35 bucky cj, you know.. User Defined Function
05:35 cj TimToady: oh, right... I probably need to install some fancy win32 fonts...
05:35 bucky cj, do you work with Brian Acker?
05:36 cj bucky: Aker, probably, yes?
05:36 bucky s/Acker/Aker
05:36 cj bucky: I work with him on occasion, yes.  But I work for Amazon these days, and not MySQL.  I switched over in June/July 06
05:36 bucky oh.. nm
05:37 cj I had a computer in his basement for a couple of years.  He's good at providing infrastructure :)
05:37 bucky seems like a cool guy http://krow.net/index.pl?node_id=32
05:37 lambdabot Title: Brian "Krow" Aker Presents: brian
05:38 cj TimToady: hurm... Maybe I'm still stuck on the LANG stuff... how do I change my locale from C to en_US.UTF-8?
05:39 cj TimToady: you did say that it would look like placeholder characters if I don't have the fonts, but have everything else set up correctly, didn't you?
05:39 cj there's probably a HOWTO around here somewhere, and I shouldn't be bothering you for your precious time :)
05:41 TimToady LANG is just an environment variable, so you'd set it in your .bashrc or equivalent
05:41 TimToady putty has a menu somewhere to tell it to interpret text as utf8
05:43 cj TimToady: woo!  That got closer... now I need fonts...
05:43 * cj googles kanji win32 fonts
05:44 cj bah.  I'll do this later.  diapers need getting :)
05:44 putter parting strawman hypothesis: once we have pugs oo and a regex engine, it might be worth doing a javascript implementation to exercise the oo.  the js 1.x spec is mechanically convertible to p6 code.  there's a years old one-day effort somewhere in misc.
05:46 putter the motivation is we don't have much p6 oo code to exercise the oo implementation.  whereas js has a testsuite and assorted code.  this might be an approach to shaking down the oo impl mechanically, rather than waiting for average users to trip over the inevitable bugs.
05:47 putter g'night &
05:52 Schwern joined perl6
05:58 Eidolos joined perl6
06:14 BooK_ joined perl6
06:24 cognominal joined perl6
06:26 rfordinal joined perl6
06:31 stevan_ joined perl6
06:36 rod2 joined perl6
06:46 bsb left perl6
07:00 dduncan left perl6
07:03 CardinalNumber joined perl6
07:19 the_dormant joined perl6
07:23 CardinalNumber is now known as ProperNoun
07:33 Aankhen`` joined perl6
07:36 gaal audreyt: in general, how closely are we willing to go with the techniques in "Haskell's Overlooked Object System"? (the 2005 paper by Oleg and Ralf)
07:42 cognominal joined perl6
07:45 iblechbot joined perl6
07:56 the_dormant joined perl6
08:14 aukjan|gone is now known as aukjan
08:16 aukjan is now known as aukjan|gone
08:17 forsaken joined perl6
08:24 jisom joined perl6
08:24 jisom left perl6
08:24 jisom joined perl6
08:37 Endymion joined perl6
08:38 Endymion hiho
08:40 marmic joined perl6
08:40 tene hi!
08:45 lichtkind joined perl6
08:52 lichtkind good morning your camels :)
08:53 mako132 joined perl6
09:13 CardinalNumber joined perl6
09:18 buetow joined perl6
09:26 svnbot6 r15406 | gaal++ | * minor cleanup
09:26 svnbot6 r15406 | gaal++ |   (More points-free would have merge_clean simply as "map . filter . (/=)";
09:26 svnbot6 r15406 | gaal++ |   I don't know if that's clearer.)
09:32 CardinalNumber is now known as ProperNoun
09:36 Ziggy6 joined perl6
09:37 idiotmax joined perl6
09:38 spinclad joined perl6
09:42 buetow joined perl6
09:57 meppl joined perl6
10:05 andara joined perl6
10:09 drrho joined perl6
10:29 IllvilJa joined perl6
10:35 svnbot6 r15407 | gaal++ | * fixed-point golf: get "HOW" and "WHICH" methods on all pure
10:35 svnbot6 r15407 | gaal++ |   classes automatically.
10:35 buubot Current winner is buu with 85 characters: ZOMG SCARY STUFF
10:37 tene ???
10:37 tene buubot--
10:37 buubot is a potvaliant land-lubber!appeared to work above, right?
10:56 ruoso joined perl6
11:10 fglock joined perl6
11:12 cmarcelo joined perl6
11:16 cmarcelo @tell gaal re: mutable classes, http://dev.pugscode.org/browser/docs/n​otes/mo-bootstrap-plan.pod?format=raw maybe helpful.
11:16 lambdabot Consider it noted.
11:34 wilx joined perl6
11:35 CardinalNumber joined perl6
11:45 cmarcelo @tell gaal in [15407], how "fix" know when to stop? it's like it is iterating until the function "converges"?
11:45 lambdabot Consider it noted.
11:46 gaal @massages
11:46 lambdabot cmarcelo said 29m 42s ago: re: mutable classes, http://dev.pugscode.org/browser/docs/n​otes/mo-bootstrap-plan.pod?format=raw maybe helpful.
11:46 lambdabot cmarcelo said 1m 34s ago: in [15407], how "fix" know when to stop? it's like it is iterating until the function "converges"?
11:46 gaal cmarcelo: re: fix, lazy evaluation
11:47 |newbie| joined perl6
11:47 gaal I can't find a good explanation online, I remember there was one
11:47 gaal it's pretty !ing, I agree :)
11:48 gaal but assuming you grok the Y combinator in principle, think of it as a magical implementation of it
11:48 audreyt fix f = let x = f x in x
11:48 lambdabot audreyt: You have 4 new messages. '/msg lambdabot @messages' to read them.
11:49 audreyt gaal++ # let fix = fix fix in fix!
11:49 audreyt fix is not that magical compared to mfix
11:49 audreyt mfix is mind-bogglingly magical
11:49 gaal y my nexy Q for you was whether MutClass should use mfix. :)
11:50 gaal if indeed we were to use a monadic class for mutable classes
11:50 audreyt nay they are orthogonal
11:50 audreyt currently we still use const self
11:50 gaal so despite the notes cmarcelo++ linked me to, where does the mutability come from?
11:51 audreyt but ^Str.add_method is the same as ^Scalar.add_method
11:51 audreyt it comes after we get Pugs.Prim ported
11:51 audreyt then we augment the "MI" data type
11:51 audreyt where it only contains a table of methods
11:51 cmarcelo gaal: ok.. in some sense it was like I was imagining.. but, what does "!ing" mean? :P
11:51 audreyt it will carry a mutable table of methods
11:51 audreyt shocking
11:52 gaal which reminds me, does ByteString have unicode-aware toUpper etc. functions? I was looking for them
11:52 gaal cmarcelo: think Chess notation, where "!" means a surprising and stong move
11:52 audreyt gaal: no, and it doesn't need to
11:52 audreyt hint: use perl5embed
11:53 gaal oi
11:53 gaal I knew you'd say that :)
11:53 audreyt the p5 unicode core is _nice_
11:53 audreyt it's just the surface semantic is borderline braindead
11:53 audreyt but we avoid that
11:53 audreyt the engine can totally be reused
11:53 audreyt I consider it saner than linking libICU
11:53 audreyt and our p5 bridge already is SvUTF8_on()
11:54 audreyt so should be straightforward to just call uc() on p5 land
11:54 audreyt please make some very easy shorthand for that :)
11:54 gaal is p5 unicore exposed?
11:54 audreyt just call &uc
11:54 audreyt it should Transparently Work (tm)
11:54 CardinalNumber is now known as ProperNoun
11:54 gaal that will also work, but doesn't it entail some overhead?
11:54 audreyt not more than calling a normal p5 function
11:54 audreyt I wouldn't care that much
11:55 audreyt also makes augmenting code easier
11:55 audreyt you really don't want to write XS at this stage, trust me :)
11:55 gaal okay.... the plan is to implement that stuff in Perl 6 evenually, yes?
11:55 ruoso joined perl6
11:55 audreyt not until several milestones from now
11:55 audreyt so no
11:55 audreyt yes re "eventually"
11:55 chris2 joined perl6
11:55 audreyt no re "plan"
11:56 audreyt of course if you find it fun by all means do it... but I don't think it's sensible to move that much runtime into p6
11:57 audreyt C-land stuff is still sensible for Unicode
11:57 gaal well, I mean, how tightly coupled do we want to be with perl5?
11:57 audreyt and p5 is as good as any
11:57 audreyt for the Pugs Haskell+Perl dual runtime
11:57 audreyt I think forever
11:57 gaal okay
11:58 audreyt I don't see a good way to do DBI.xs compat
11:58 audreyt beyond what we already have
11:58 audreyt not convinced that "Ponie for Pugs" can work
11:58 gaal now, re the Moose thing, picking up last night's thread...
11:58 audreyt but glad to be proven wrong by someone eventually :)
11:58 audreyt ok. the idea is that when you declare
11:58 audreyt "class Foo { ... }"
11:58 audreyt in p6-land
11:58 gaal does the glue write Moose bridges automoosically?
11:58 audreyt it should generate a corresponding perl5 symtable
11:58 audreyt *perl6::Foo
11:59 audreyt most likely
11:59 audreyt yes, that I think is the plan. it runs Moose bridgecode
11:59 gaal there's some old stuff in DrIFT/Perl6Class
11:59 audreyt yes, aware of that
11:59 audreyt I'm still evaluating the possibility of dyld
12:00 moritz hey, what about a C to perl 6 translator, and use that on the p5 source code </kidding> ;)
12:00 audreyt that is it generates a genuine Haskell data structure
12:00 audreyt use ghci to compile it
12:00 audreyt then byteload it
12:00 audreyt and drift it
12:00 audreyt and eval that in p5 land
12:00 audreyt the gain is that compact structures gets to be really compat
12:00 aukjan|gone is now known as aukjan
12:00 audreyt but that is optional and not relevant to getting 6.28.0 released
12:00 audreyt maybe 6.28.1 or .2
12:01 gaal .. is that the libGHC idea mentioned in that note?
12:01 audreyt yes./
12:01 gaal ok let me see if I understand the picture
12:02 gaal Parse "class {...}" -> Hs code -> ??? -> P5 Moose code -> P5bridge
12:02 audreyt or rather, class {...} generates a series of MO calls
12:03 audreyt that in turn generates Hs code and p5 moose code
12:03 gaal which are compiled to Hs code?
12:03 gaal ah y
12:03 audreyt yes
12:03 audreyt but the 6->5 bridge I intend to do after 5->6
12:03 gaal runtime drifiting, heh
12:03 audreyt instance Boxable Eval PerlSV -- I want to get this written first
12:04 Juerd Anyone in here going to the European hackathon today?
12:04 gaal Juerd: not unless I can borrow a teleporter
12:05 audreyt a time machine will also do
12:05 Juerd I'd lend you mine, but I'd need another teleporter for that :)
12:05 Juerd Otherwise it turns inside out during.
12:06 Juerd And eventually teleports everything surrounding it, but not itself.
12:06 audreyt if a p5p hacker gets online via dial-up, does that make a teleporter?
12:06 Juerd This happened already in some parts of the world. They now drive on the wrong side of the road, as a result.
12:07 moritz Juerd: currently I am in such a country... it's not the only thing that changed ;)
12:08 gaal audreyt: so actually the 5->6 is pretty critical now, if we're to delegate much of stringy Prim to it
12:09 gaal I'm not familiar with the existing 5->6 bridge at all
12:09 gaal what does "si" in MO.si stand for?
12:10 gaal hmm okay, I'm called off for lunch now...
12:10 gaal (single invocant?)
12:10 cmarcelo gaal: single inheritance..
12:10 gaal ah, okay :)
12:10 fglock joined perl6
12:10 gaal *waves* &
12:11 cmarcelo but si.hs is a file used for testing... :o)
12:11 cmarcelo see you.
12:13 gaal (this is a pretty minor detail but we probably want a p5prelude too, to implement things like &capitalize without hitting the bridge muliple times)
12:19 fglock I've been thinking on a low-level p6lib (possibly in C) that could be used by pugs, v6.pm, and also be converted to pmcs too
12:23 rindolf joined perl6
12:23 rindolf Hi all!
12:23 rindolf Does anyone know what is the channel for the hackathon today?
12:25 moritz @seen all
12:25 lambdabot I haven't seen all.
12:25 moritz ;)
12:25 moritz rindolf: you're out of luck ;)
12:26 rindolf is now known as all
12:26 all is now known as rindolf
12:26 rindolf Wonder who took the "all" nickname.
12:32 pasteling "cmarcelo" at 200.245.119.10 pasted "error doing 'svk pull' @ feather" (17 lines, 586B) at http://sial.org/pbot/23256
12:51 yhsuan joined perl6
12:54 elmex joined perl6
12:55 shay hi folks
12:56 shay cmarcelo, I know that glibc bug
12:56 shay give me a sec
13:01 shay oh, it's not the same error :/
13:05 audreyt cmarcelo: maybe pull file:///data/svn/pugs
13:11 Limbic_Region joined perl6
13:13 * avar says hi to feather's webserver
13:14 cmarcelo hmm.. should //mirror/pugs point to file:///... too?
13:17 cmarcelo it exists already. its //mirror/local :o)
13:21 the_dormant joined perl6
13:25 avarab joined perl6
13:31 iblechbot joined perl6
13:35 gnuvince_ joined perl6
13:48 avar joined perl6
14:06 Limbic_Region salutations all
14:12 drupek12212157 joined perl6
14:15 cmarcelo ola
14:16 Lorn joined perl6
14:31 bernhard joined perl6
14:35 weinig|bbl is now known as weinig
14:48 diakopter joined perl6
14:50 penk joined perl6
15:05 fglock joined perl6
15:08 rindolf Hi all!
15:09 rindolf So where's the Hackathon IRC channel?
15:11 amnesiac joined perl6
15:14 moritz rindolf: is that a parrot hackathon?
15:15 rindolf moritz: http://conferences.yapceurope.org/hack2007nl/
15:15 lambdabot Title: European Perl Hackathon 2007 - The Netherlands - Home
15:15 rashakil_ joined perl6
15:15 rindolf moritz: they also work on Perl 6 and Parrot.
15:15 kanru joined perl6
15:17 forsaken joined perl6
15:22 * bernhard greetings from the hackathon in Arnhem
15:23 fglock bernhard: is there an IRC channel for the hackathon?
15:23 bernhard joined perl6
15:23 fglock bernhard: is there an IRC channel for the hackathon?
15:25 bernhard I think we use #hackathon on MagNet irc.perl.org
15:30 vel6608 joined perl6
15:34 rindolf Hi bernhard
15:36 bernhard Hi
15:47 cj bernhard: what are you hackath on?
15:57 rashakil__ joined perl6
15:59 forsaken joined perl6
16:01 aukjan is now known as aukjan|gone
16:05 cognominal joined perl6
16:09 gaal moose
16:09 svnbot6 r15408 | andara++ | [runpugs]
16:09 svnbot6 r15408 | andara++ | -refactoring Server.pm to catch race hazard in preloaded session management.
16:09 svnbot6 r15408 | andara++ | -adding testing framework.
16:10 turrepurre joined perl6
16:17 bonesss joined perl6
16:20 rindolf Hi gaal
16:38 PerlJam The perl.com article on perl 6 has two comments to the effect of "we want lots more perl 6 articles"
16:38 svnbot6 r15409 | andara++ | [runpugs] -cleaning up obsolete files
16:38 PerlJam I imagine that's a common sentiment
16:40 TimToady 'course, even this one's a little bit wrong.  A file starting with 'package' should be considered Perl 5, not Perl 6.  Wanted 'module" there...
16:42 gaal hi rindolf. I'm not really here.
16:42 gaal what perl.com article?
16:42 kolibrie gaal is tele-porting?
16:42 TimToady http://www.perl.com/pub/a/2007/03​/01/perl-6-parameter-passing.html
16:42 lambdabot Title: perl.com: The Beauty of Perl 6 Parameter Passing, http://tinyurl.com/2x2opj
16:43 rod1 joined perl6
16:43 gaal oh, nice. reading...
16:43 * kolibrie too
16:43 TimToady just be sure to s/package/module/ in your head
16:44 TimToady other than that, a good article
16:45 gaal human parsers are usually transparently error-correcting. to the chagrin of copyeditors...
16:50 svnbot6 r15410 | fglock++ | PCR - updated TODO
16:54 gaal I'd have mentioned variable declaration and calling converntions were merged; and explained the recommendation against using Perl for numeric computations.
16:55 gaal re-&
16:56 IllvilJa joined perl6
16:56 Ziggy6 left perl6
16:56 Ziggy6 joined perl6
16:58 lichtkind joined perl6
17:00 fglock joined perl6
17:01 PerlJam TimToady: I don't know ... I thought the article didn't give enough explanations.
17:01 TimToady I didn't call it an excellent article :)
17:03 PerlJam I was going to spend some time playing with parrot/cardinal this weekend but maybe I could cobble together a perl6 article for perl.com or TPR or any other venue that'll have it.
17:04 PerlJam (Becaise I tend to agree with the commenters that we need more perl 6 articles :-)
17:04 TimToady we do need multiple voices explaining things on various levels
17:04 PerlJam s/ais/aus/
17:04 thepler joined perl6
17:06 PerlJam The hardest part (for me) would be picking a narrow enough topic to write on.  There's just *so* *much* ... stuff  in perl 6  :-)
17:06 andara left perl6
17:07 kolibrie my problem is forgetting that people are not familiar with Perl 6 syntax
17:08 TimToady and part of the problem is that you're writing for two very different audiences simultaneously
17:08 TimToady those who know Perl 5 syntax, and those who don't.
17:10 offby1 I would gently suggest: if you cannot please both, then write for the new users, not the old.  Assume that, five years from now, more people will be writing perl6 than perl5 (i.e., plan to succeed)
17:10 offby1 this from a perl5 user.
17:11 drupek12212157 joined perl6
17:13 PerlJam offby1++  good advice I think.
17:13 TimToady sure, I agree completely, but...
17:13 TimToady the downside of it is
17:14 TimToady if the P5 people perceive that they're being written off, they will also feel free to ignore you
17:14 [particle] say "{ newton(8, \&f, verbose => 1, epsilon => .00005) }";
17:14 TimToady and then you get what we see continually in p5p--piecemeal solutions to individual RFC problems
17:14 [particle] can \&f be &f?
17:15 TimToady why is there a \ in the first place?
17:15 [particle] does &f invoke, or is it just a sigil?
17:15 [particle] i think it's an error
17:15 TimToady the & never invokes in P6
17:15 [particle] it's too perl5y
17:15 PerlJam I wonder if we shouldn't submit corrections to the author (or the editor) before too many people read that article :)
17:16 TimToady also, the say "{...}" is a bit strange since say ... does the same.
17:16 * [particle] sometimes thinks of  as a twigil in p5 :)
17:16 [particle] pidgin perl 6
17:16 PerlJam [particle]: then you've got something backwards as the \ normally comes first ...  \$thing  # $ is the "twigil"  here
17:17 [particle] pj: i know, it's just that \ is used so frequently in p5 that it's twigil-like
17:18 [particle] and \ will be used so infrequently in p6, i feel
17:18 [particle] in fact, if you need a new twigil, \ could be a candidate
17:18 TimToady still and all, I like the way he's doing it better than the way other people aren't doing it.  :)
17:18 PerlJam :-P
17:18 [particle] agreed! i plan on writing a lot of pidgin p6
17:19 PerlJam TimToady: will you be around #perl6 later tonight and this weekend?
17:19 TimToady yes, for the most part.  have to sneak some $job in there as well though.
17:20 PerlJam If my kids don't distract too much, I'll start writing something and I'll want an expert opinion  :-)
17:20 PerlJam speaking of $job ... back to work for me!
17:20 [particle] is there a difference between {foo(...)}.say and foo(...).say?
17:20 TimToady I'll be glad to give feedback
17:21 TimToady the first one will try to print a closure instead of calling it?
17:21 TimToady {foo(...)}().say would be the same
17:21 offby1 ooh! closures print?
17:21 TimToady "try to"
17:21 offby1 :-(
17:22 TimToady ?eval {1+1}.say
17:22 evalbot_r15410 OUTPUT[<SubBlock(<anon>)>␤] Bool::True
17:22 TimToady ?eval {1+1}().say
17:22 evalbot_r15410 OUTPUT[2␤] Bool::True
17:23 TimToady closures never autocall themselves unless in a calling context of some sort.  use as bare statement, use before (), use in regex, etc.
17:24 TimToady much like &foo never autocalls itself without context.
17:32 TimToady now I'm wondering whether our current module names are sugar for the general adverbial names
17:33 TimToady use Foo:v<1.2.3>:auth<JRANDOM>
17:33 TimToady use :root<perl5>:DBI:6.0:auth<JRANDOM>:foo<bar>:baz
17:33 TimToady assuming :6.0 desugars to :v<6.0> or some such
17:34 TimToady that also would solve the problem of :root<C#> and :root<C++> not being identifiers
17:36 TimToady use :DBI<6.0>:root<perl5> shows some other possibilities
17:37 [particle] i like having the root up front
17:38 TimToady use:root<perl5> DBI(:1.2.3, :auth<TBUNCE>)
17:38 [particle] hrmm
17:39 [particle] with a default root of perl6
17:40 [particle] are parens required there? i'll have to check up on that syntax again
17:40 TimToady or do the trick we were doing with :sym<foo> and make use:perl5 mean use:root<perl5> for identifiers
17:40 TimToady we're talking new syntax here
17:40 TimToady I'm just trying to think about what really modifies what
17:40 [particle] i'm thinking about existing syntax, so see how close the new syntax is
17:41 [particle] i think the root modifies use, as the import routine will differ based on library language
17:41 TimToady use:<C#> Foo:(...)
17:43 TimToady use:perl5 DBI:(version => 1.2.*, auth => 'TBUNCE')
17:43 TimToady thinking of use:<C#> as more the infix:<*> syntax here
17:45 cmarcelo joined perl6
17:45 [particle] use DBI of perl5 is :(version<1.2.3>)
17:45 TimToady couple things wrong with that
17:46 TimToady perl5 is not a type
17:46 TimToady and "is" basically means : already
17:46 [particle] i like use:perl5 better anyway
17:47 wilx` joined perl6
17:47 [particle] use:<perl5> DBI is {version<1.2.*>, author<TBUNCE>}
17:47 cj what benefit do you see the <...> providing?
17:47 TimToady anyway, it bugged me that my grammar had to have cplusplus
17:47 cj ah.  special character escaping?
17:47 TimToady <...> quotes
17:48 TimToady :<C#>
17:48 cj maybe optional?
17:48 TimToady yes, optional for ident
17:48 TimToady :perl5
17:49 TimToady use:cobol COLLATE_BUSINESS_FORMS
17:50 TimToady maybe allow Foo-1.2.3 as sugar, but nothing with - beyond that maybe
17:50 TimToady Or just make Foo:1.2.3 work right
17:51 [particle] :v1.2.3 wouldn't be horrible
17:51 TimToady would still have to be a special case adverbial form
17:52 [particle] hrmm
17:52 TimToady :16<dead_beef> might require brackets, and :16.42 could be version
17:52 [particle] don't we have a colon number form for base/radix conversion?
17:53 TimToady but it would still have to special-case the .
17:53 TimToady just wondering if people will get used to the notion of reading :1.2 as a version
17:54 [particle] no, but :v1.2 they might
17:54 [particle] or v:1.2
17:54 penk joined perl6
17:54 TimToady :v<1.2>
17:54 TimToady hmm, that's just a Pair
17:55 [particle] that will work today
17:55 TimToady but
17:55 TimToady it wants a special syntax to force Version object construction
17:55 TimToady or we fall into the v-strings fallacy again
17:55 [particle] hrmm
17:56 [particle] can it be typed that way?
17:56 [particle] have use accept a param v of type Version
17:56 TimToady in context it's a version, but I'm thinking about version literals
17:56 justatheory joined perl6
17:57 TimToady v1.2.3 would map nicely to :v1.2.3 to use a literal as an option
17:58 TimToady but have to think about wildcard version literals
17:58 TimToady can get away with v1.2.* at least
17:59 TimToady but anything more complicated is going to require brackets, and then it starts going strange
17:59 * offby1 wonders how IRC will have affected the design of perl6 as compared to perl5
17:59 TimToady v({ matchme()}) is just a function call
17:59 TimToady and :v(...) is just a pair again.
18:00 TimToady offby1: this seems like apples and oranges to me
18:00 TimToady IRC is not a language exactly
18:01 TimToady what aspect of IRC are you thinking of?
18:01 offby1 just that you're thinking out loud with so many people now
18:01 offby1 I assume that wasn't the case back in perl5
18:02 TimToady oh, I misparsed your sentence entirely
18:02 offby1 I should have used more parens :-)
18:02 TimToady IRC wasn't used in the design of Perl 5 at all.
18:02 offby1 didn't think it would have been
18:02 [particle] versions with idents will definitely be trouble :v1.2.3.beta1
18:02 TimToady I only started using IRC in the AP era (after pugs)
18:03 [particle] "no such method: beta1"
18:03 * offby1 hopes TimToady, [particle], PerlJam, et al don't have aching heads from having so many ideas bounced off them
18:03 TimToady so at some point you go :v<1.2.3.beta1>
18:03 [particle] right
18:04 TimToady but then it's back to a pair, not a version :(
18:04 [particle] so, a V operator?
18:04 [particle] prefix:<V>
18:05 TimToady V '1.2.3.beta1'
18:05 TimToady hmm
18:05 lichtkind joined perl6
18:05 TimToady I don't doubt someone will want V for something list infixish
18:05 [particle] that's what multis are for
18:06 TimToady just worrying about market confusion
18:06 [particle] call dibs
18:06 TimToady it's not pretty enough
18:06 [particle] true
18:07 TimToady huffmanly, probably Version('1.2.3.beta1')
18:07 [particle] that's boring
18:07 TimToady and :v<1.2.3.beta1> where a pair will be interpreted as a version
18:08 [particle] that works
18:08 TimToady so Version() only used in normal code for weird version names
18:08 [particle] where versions are expected, the syntax is compact. for other reasons, you're penalized appropriately
18:08 TimToady and :v contexts really just feed the :v argument to Version()
18:08 [particle] s/reasons/uses/
18:09 [particle] yop
18:09 [particle] it's got my vote
18:09 * TimToady wonders if v1.2.3.beta1 would still become an unfortunate faq
18:10 [particle] just give them enough rope
18:11 [particle] this is only a problem in locales with . as a num sep
18:12 [particle] is whitespace following a list sep , required?
18:12 [particle] eval? (1,2,3).say
18:13 [particle] ?eval (1,2,3).say
18:13 evalbot_r15410 OUTPUT[123␤] Bool::True
18:13 drupek12212157 joined perl6
18:14 [particle] so v1,2,3,beta1 has *other* problems
18:18 TimToady we ignore the num sep cultural problem, basically, and practice cultural imperialism on that point (no pun intended)
18:19 [particle] it's unlike you not to intend a pun
18:19 [particle] the low-coffee diet is affecting your humour
18:20 TimToady well, it was unintended only up to the point that I noticed it.  :)
18:20 TimToady at any rate, we simply can't have the meaning of comma flapping in the cultural breeze.
18:22 TimToady at best maybe we can recognize 1,2 and give a warning if things go wrong later.  but I don't know that 1,2 would cause things to go wrong later soon enough for the compiler to notice
18:22 [particle] i18n is still on the horizon, anyway
18:23 TimToady I think we're already past the event horizon on that one, though a lot of p5p still thinks they can continue to send messages to the outside universe as long as they like
18:24 TimToady but the universe has changed, and Unicode is the new simplicity
18:24 TimToady and yes, I smile when I say that...
18:28 [particle] dingbat
18:28 drupek12212157 joined perl6
18:28 TimToady to zapf and beyond!
18:34 TimToady but in the overall scheme of things Unicode really is a simplification in the same sense that Perl was a simplification of Unix, a complex set of trees, but a smaller forest
18:37 bonesss joined perl6
18:38 diakopter joined perl6
18:39 stevan_ joined perl6
18:44 TimToady but in general people can only perceive the complexity of the bit of forest they can see from where they stand
18:45 TimToady "can't see the forest for the trees" as some folks put it
18:45 TimToady most of p5p is perpetually in that state
18:46 TimToady shower &
18:46 [particle] they're picking the bark off of all the trees so they can see more of the forest
18:53 gnuvince_ joined perl6
18:57 drupek12212157 joined perl6
19:00 * fglock trying to see the tree - concentrate on a single problem and fix it
19:02 explorer joined perl6
19:02 [particle] with pcr, mp6, kp6 etc?
19:02 forsaken joined perl6
19:03 fglock [particle]: yeah - I'm trying to figure out a plan
19:04 fglock kp6-perl5 is powerful, but it's almost another language, incompatible with cpan modules
19:04 fglock v6.pm OTOH, is too perl5 in many cases
19:05 [particle] heh
19:06 fglock pcr should really be written in p6, but it would be much better if this were coordinated by pmichaud or audreyt, IMO
19:10 fglock other kp6 backends (C, C++) seem to be dead ends, though some code could be reusable by parrot
19:11 fglock C would still be nice to have for fast small programs
19:12 [particle] it'd be nice to have a shared regex lib to embed in a browser
19:13 [particle] p6cre
19:13 fglock I think pmichaud was working on one, before starting PGE
19:14 fglock mm - PCR-on-MP6-on-C ?
19:16 fglock I wonder if it's reasonable to require the host program to provide the object system and memory allocation
19:16 [particle] you might target .net then
19:17 fglock a C lib would probably require a huge API specification
19:17 [particle] yes
19:19 fglock yes, but then it seems reasonable to use PCR-on-p6 -> dotnet
19:26 fglock re kp6, I noticed that mp6 can be implemented with even less features (and be much more portable)
19:27 [particle] is it worth doing?
19:27 fglock namespaces and pads are better represented as objects than native things - this gives much more p6-ish closures
19:28 fglock [particle]: parrot already does this, to some extent
19:29 fglock and the Haskell implementation
19:29 fglock it's probably the way to go
19:31 fglock but then, this is one of the things that make kp6-perl5 incompatible with cpan
19:31 drupek12212157 joined perl6
19:36 fglock a compiled kp6 program would look like an ast, with Closure objects pointing to native code
19:37 fglock an OO version of m-expressions (is there a name for this?)
19:39 fglock it's just one indirection level, but this kills the performance in p5
19:40 ruoso fglock, may I talk to you in private?
19:40 fglock sure
19:41 rindolf joined perl6
19:45 fglock the solution would be a smart compiler that would desugar as much as possible to native p5/.net/jvm/parrot
19:45 ruoso fglock, did you receive my privmsg or freenode just didn't deliver?
19:45 fglock and keep the difficult parts as plain oo
19:46 fglock ruoso: I dit - I think I can't msg back - can you try google chat?
19:46 ruoso you mean gtalk?
19:46 [particle] /nickserv identify <user> <pass>
19:46 fglock one sec...
19:49 fglock ruoso: I sent a ping
19:49 ruoso nop... but if you have gtalk you can add ruoso at jabber.org and talk to me directly
19:49 eric__ joined perl6
19:59 putter joined perl6
20:01 putter fglock: re what to work on, so how do you see the future of PCR?
20:01 * putter ponders  [ $2:=(a) ]*
20:04 jisom joined perl6
20:04 the_dormant joined perl6
20:05 jisom left perl6
20:05 jisom joined perl6
20:09 TimToady $2 ends up with an array of a, I expect
20:10 putter ah, ok.  tnx
20:12 dduncan joined perl6
20:12 drupek12212157 joined perl6
20:25 drupek12212157 joined perl6
20:26 simcop2387 joined perl6
20:28 fglock putter: ping
20:32 GabrielVieira joined perl6
20:43 putter pong
20:43 fglock putter: re PCR
20:44 * putter listens attentively
20:44 fglock :)
20:44 putter ;)
20:44 fglock it will likely keep the API, but some things can improve
20:44 nothingmuch BAM!
20:44 nothingmuch oh crap, i lost my moment... i was going to make everyone bounce
20:45 * putter notes loud noises rendered as text are somewhat less startling than the original
20:45 fglock the compiler needs an ast processing step
20:46 fglock access to p5 lexicals could be improved
20:47 putter has it been run against parrots rx_* tests?
20:47 fglock some things need to be rethought, like the haskell backend
20:47 fglock yes, some of them pass
20:47 putter :)
20:48 drupek12212157 joined perl6
20:48 putter other thoughts?
20:48 fglock improving the compiler and adding functionality are orthogonal
20:49 fglock but only to some extent
20:49 fglock like:  <rule> <rule>  - the compiler needs to detect this and arrange for array storage
20:50 fglock the TODO list is huge
20:51 fglock it's bigger than just failing tests
20:51 putter bigger in what way?
20:52 fglock several XXX and ??? places in the source are not detected as test failures, which means we need more tests
20:53 putter oh yeah
20:53 putter for a lot of the alias stuff, there isn't a single test case anywhere.  :(
20:55 putter perhaps we should write a "we need more regex tests regards X" wish list?
20:55 putter what else might be done to improve things?
20:56 putter "things" == "big picture", aside from TODO
20:56 fglock I've been looking for a DFA implementation
20:56 putter to generate test cases?
20:56 fglock documenting the regex AST
20:57 putter makes sense, what else?
20:58 fglock creating a benchmark directory, for tests regarding algorithms
20:58 fglock some things are already in the TODO list, like adding the p5regex backend to the distro
20:59 putter re benchmark. ??   like optimization tests?  "dont search where you know there isnt a character you need" kinds of tests?
20:59 elmex joined perl6
21:01 fglock yes - there are several things, like - how much faster is it to compile to p5regex than pure-perl5 - when is it worth to change mode
21:01 fglock and - is it worth to use XS here
21:02 putter that ones easy... "Is it worth using XS?"  "No, until proven otherwise". ;)
21:03 fglock reimplement the compiler in p6, and compile it back to p5
21:04 fglock then, pugs could probably use the p6 version directly
21:04 putter re ->p5, yeah.  I was wondering if we could special case the nodes, pragma "straight to p5", so we don't take the moose performance hit.
21:05 fglock remove the Parse::Yapp dependency - we could use a p6 version
21:05 fglock re Moose, we can use mp6
21:05 fglock mp6 == no dependencies
21:06 putter mp6 faster than moose?
21:06 fglock sure - it's much simpler
21:06 putter (I should have said "hypothetical rumored moose performacne hit")
21:06 putter k
21:06 putter hmmm
21:07 rashakil joined perl6
21:07 [particle] anything that hits a moose is likely to suffer performance problems
21:07 fglock of course, mp6 is uglier than a moose
21:08 putter unless performance is measured in running-away-from-irritated-moose velocity
21:08 putter re uglier, oh my, that's a scary thought
21:08 putter ;)
21:09 [particle] what's uglier, pugs or moose?
21:09 rindolf fglock: mp6 == mod_perl6?
21:10 fglock "MiniPerl6"
21:10 [particle] somebody here has been smoking crack
21:10 TimToady pugses are ugly enough that they wrap around the backside to cute.  not sure mooses make it that far around...
21:10 [particle] great, now i'm picturing mooses backsides
21:10 fglock MiniPerl6 - imagine a pugs without hair
21:10 putter lol
21:11 gaal @google pretty moose
21:11 lambdabot http://www.unexplained-mysteries.c​om/forum/index.php?showtopic=88674
21:11 lambdabot Title: The ghost moose of Maine. - Unexplained Mysteries Discussion Forums
21:11 putter oh my
21:12 gaal "The sight of him is so frightening, he once scared a bicyclist up a tree."
21:12 putter "a moose version of Moby Dick"
21:12 putter "what will your language earn thee in Nantucket market"
21:14 drupek12212157 joined perl6
21:14 gaal .oO( St. Antler's Fire )
21:14 putter groan
21:15 fglock OTOH, PCR didn't get to the point of full-spaghetti yet (unlinke v6-pm emitter) - so it's probably fixable
21:15 putter "point of full-spaghetti" concept++
21:16 gaal fixed point of full gnocci
21:16 putter "well modularized software is like raviolli"
21:16 fglock some software architecture work would be nice
21:17 fglock gnocci is like little modules
21:18 [particle] better than java beans
21:18 putter some systems are just a shell with no meat
21:18 [particle] i prefer mine hand rolled
21:19 gaal there's that chapter or three where he just talks about taxonomy of whales and stuff
21:19 gaal that's like the meta model but tedious
21:19 putter lol, hard
21:20 putter bbiab
21:20 gaal bb after sunrise &
21:21 fglock weekend starts now for me
21:24 TimToady is that good or bad?
21:25 fglock I miss being online - but going to the beach is great
21:26 fglock converting PCR to use the compiler tools in kp6 should be easy - but it would be better if mp6 were in cpan
21:26 fglock not sure if publishing mp6 is a good thing
21:27 [particle] there's one way to find out
21:29 fglock ok - I've got to close the place here and go home &
21:31 fglock left perl6
21:35 Aankhen`` joined perl6
21:41 drupek12212157 joined perl6
21:56 rafl joined perl6
21:57 Caelum is now known as merlyn
21:57 merlyn is now known as Caelum
22:03 BooK joined perl6
22:03 BooK left perl6
22:06 devogon joined perl6
22:08 drupek12212157 joined perl6
22:19 Psyche^ joined perl6
22:28 Psyche^ is now known as Patterner
22:37 putter the room echos emptily
22:38 stevan_ joined perl6
22:38 tene rooms full of stuff usually don't echo very well.
22:38 putter :)
22:44 putter &
22:46 weinig is now known as weinig|bbl
22:48 khisanth_ joined perl6
22:51 khisanth_ is now known as Khisanth
22:57 |newbie| joined perl6
23:07 ruz joined perl6
23:16 drupek12212157 joined perl6
23:27 Limbic_Region joined perl6
23:31 drupek12212157 joined perl6
23:44 drupek12212157 joined perl6
23:55 kdevin joined perl6
23:57 drupek12212157 joined perl6

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

Perl 6 | Reference Documentation | Rakudo | Niecza | Specs