Perl 6 - the future is here, just unevenly distributed

IRC log for #perl6, 2005-07-20

Perl 6 | Reference Documentation | Rakudo

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

All times shown according to UTC.

Time Nick Message
00:02 larsen has quit IRC ("later")
00:48 mr_ank has quit IRC ("Download Gaim: http://gaim.sourceforge.net/")
00:48 vcv has quit IRC (Read error: 104 (Connection reset by peer))
00:49 mr_ank has joined #perl6
00:52 cm has joined #perl6
00:54 cognominal has quit IRC (Remote closed the connection)
00:56 ipv6guru has quit IRC (Read error: 60 (Operation timed out))
01:00 cognominal has joined #perl6
01:07 stef_ has joined #perl6
01:07 stef_ has quit IRC (Read error: 104 (Connection reset by peer))
01:12 sunym has joined #perl6
01:14 Darren_Duncan has quit IRC ()
01:15 meppl has quit IRC (Read error: 148 (No route to host))
01:15 castaway has joined #perl6
01:20 Nattfodd has quit IRC ("Leaving")
01:23 sunym has quit IRC (Read error: 104 (Connection reset by peer))
01:29 castaway_ has quit IRC (Read error: 110 (Connection timed out))
01:39 cm has quit IRC ()
01:43 justatheory has quit IRC ()
02:02 vcv has joined #perl6
02:05 sunym has joined #perl6
02:13 SamB has joined #perl6
02:14 justatheory has joined #perl6
02:14 justatheory has quit IRC (Client Quit)
02:23 rafl 19:32 < autrijus> is that going to need ghc team's help on ghc 6.4.1, or is it strictly a ghc maintainer's issue in the debian side?
02:24 rafl Amnesiac: I think the ghc team needs to support gcc-4.0 properly.
02:24 rafl Err.
02:24 rafl autrijus: I think the ghc team needs to support gcc-4.0 properly.
02:30 * mugwump still doesn't understand why MMD is a fundamentally different problem to SMD dispatch order
02:58 robkinyon has joined #perl6
02:58 vcv has quit IRC (Read error: 104 (Connection reset by peer))
02:58 svnbot6 r5713, putter++ | Added Rule::pattern() to Prelude.pm
03:00 QtPlatypus mugwump: MMD has a larger number of possable matches.
03:03 geoffb OK, so now having caught up on a week of #perl6 backlog, I've got a question or two.
03:03 geoffb What is the origin of the 'moose' jokes?
03:04 geoffb And does anyone know if there is a site (wiki, perhaps) that has lists of "recognized classic" books on various topics of computer science?
03:05 geoffb Like Knuth's Art of Computer Programming for mathematical underpinnings, or Foley & van Dam for general computer graphics . . . .
03:07 stevan geoffb: I think "moose" is a private joke between nothingmuch and himself :)
03:08 geoffb I keep seeing people mentioning various books in a way that suggests they are topic classics, but I've never heard of.  Which made me think, where do people learn what is good reading, outside of just hearing someone mention them?
03:08 geoffb stevan, thx
03:09 mugwump QtPlatypus: sure, but SMD also can be ambiguous; in fact in Damian's example I don't see what difference it makes having the first invocant there or not
03:10 mugwump nor why changing the program structure in the way that he has, that the results are somehow meaningful
03:10 stevan geoffb: what are you looking for in particular?
03:10 stevan what subjects
03:12 stevan mugwump: have you ever read/heard about generics functions in CLOS?
03:13 stevan they looks quite interesting
03:13 stevan kind of a weird variant of MMD
03:13 mugwump is there enough metamodel there to being prototyping these schemes?
03:13 dudley geoffb: I've often asked the same question. If you find some, add them to READTHEM.
03:13 geoffb stevan, well let me put it this way -- the bookshelves in my office contain my college textbooks, a few shelves of books on graphics and Perl, and a few more on various technology topics, such as HTML, TCP/IP, 386 assembly, and so on.  But really not much in the well rounded theory area.
03:13 stevan mugwump: you mean the MMD stuff? I have no been following the discussion
03:14 mugwump right.  Well, let me try
03:14 mugwump Luke is primarily concerned about the empty subclass test.  ie, deriving a class should not change which subs get called due to dispatch order weighting
03:14 stevan mugwump: please do it as an extension to the metamodel right now
03:14 stevan I am in the throws of a refactor
03:14 mugwump sure ... this is 10,000 mile stuff
03:15 geoffb Basically I'd like to be able to get a list of CS books that I could work through and have a pretty good grounding in most of CS, beyond just the stuff I *need* for work and hobbies.  Type Theory for instance (though of course this is one that READTHEM probably already covers)
03:15 stevan geoffb: what type of theory?
03:15 stevan geoffb: I used to buy old computer books off ebay
03:15 stevan got many "classics" for really cheap
03:15 stevan but it all depends on what you are interested in
03:15 mugwump Damian's position appears to be that this Manhattan distance metric, which weighs up the "best" one to call based on lots of factors, but this includes the number of specialisations
03:15 stevan geoffb: also, many of the classics are not as valid as they used to be
03:16 mugwump ... is the best of them.  and he claims to have done his homework
03:16 geoffb stevan, cost is always annoying with CS books, but the bigger issue is just knowing what's worth reading, and what's crap.
03:16 stevan for instance the "dragon" book on compiler design is a good book, but outdated
03:17 stevan geoffb: thats whats great about the ebay thing, I got some really good stuff for cheap and some real crap for just as cheap
03:17 geoffb stevan, point taken of course, but some books become "classics" the first year they are in print.  A modern dragon book would be a wonderful thing.
03:17 mugwump however, including the number of specialisations as a metric sounds like it breaks the "empty subclass" test, but Damian is adamant that the alternatives do, too.  I tried hard and don't follow it.
03:17 mugwump End of summary.
03:17 stevan mugwump: I will read the thread if I get a chance
03:17 robkinyon has quit IRC (Read error: 104 (Connection reset by peer))
03:18 geoffb (And while I'm wishing, it would be good to know which books were better in older editions than new ones, which sadly applies to a few in my library)
03:19 geoffb Anyway, I can't imagine that this hasn't been asked and answered thousands of times already . . . it's just a matter of finding the site that recorded everyone's collective book wisdom.
03:19 stevan geoffb: a great cheap way to learn some of this stuff is to hunt down papers on the net
03:19 stevan geoffb: if you give me a topic list, I can try to compile something for you based on what I have and know
03:20 stevan it by no means comprehensive knowledge, but I can tell you what I know/like
03:21 b6s has joined #perl6
03:21 geoffb I don't even know the topics I'm missing.  :-)  But how about this.  I'll list the ones I have and consider classic, and you can do the same, and we'll compare and maybe feed to READTHEM.
03:21 stevan ok
03:22 geoffb Compilers: The dragon book (yes, I know, dated): Aho, Sethi, Ullman: Compilers: Principles, Techniques, and Tools
03:23 geoffb Queueing Theory: Tanner: Practical Queueing Analysis
03:24 geoffb Algorithm Implementation: Sedgewick: Algorithms in C (1st Ed.)
03:24 geoffb Team/project management: McConnell: Rapid Development
03:25 geoffb Cryptography: Schneier: Applied Cryptography
03:25 stevan Practical Cryptography is a good one too
03:25 stevan I think that is also schiener
03:25 geoffb Mathematical Basis of CS: Knuth: Art of Computer Programming
03:26 geoffb Assembly Language Thinking: Abrash: Zen of Assembly Language and Zen of Code Optimization (details are dated, but concepts less so)
03:27 stevan yuk assmebly :P
03:27 geoffb Graphics Intro: Foley, van Dam, Feiner, Hughes: Computer Graphics: Principles and Practice
03:27 stevan Concurrent Programming in Erlang # totally changed how I think of programming
03:28 * stevan is trying to recall from his head since he is upstairs and the books are down
03:28 geoffb Graphical Design: Tufte: The Visual Display of Quantitative Information
03:28 geoffb :-)
03:28 stevan Tufte Rules :)
03:29 geoffb Game Design: Koster: A Theory of Fun for Game Design
03:29 stevan The Pattern on the Stone : Daniel Hillis # not just CS, but really interesting stuff
03:29 stevan The Invisible Computer : Donald Norman
03:29 geoffb Computer Architecture: Hennessey and Patterson: Computer Architecture: A Quantitative Approach
03:30 stevan The Humane Interface : Jef Raskin # chages the way I think about UIs
03:30 stevan ML for the Working Programmer : Paulson # I am still working my way through this book
03:31 geoffb Raw Unix Coding: Stevens: Advanced Programming in the UNIX Environment
03:31 stevan The Pragmatic Programmer # a modern classic IMO
03:31 geoffb TCP/IP Fundamentals: Comer & Stevens: Internetworking with TCP/IP
03:31 stevan ANSI Common Lisp : Paul Graham
03:31 stevan geoffb: I was going to say that one too :)
03:31 geoffb TPP: agreed
03:31 geoffb heh
03:32 stevan Smalltalk 80: Golberg and Robinson # a must have for understanding pure OO
03:32 jdv79 you guys opening a library or something?
03:33 stevan :)
03:33 geoffb Advanced C++: Meyers: Effective C++/More Effective C++ (these books actually are why I won't program in C++ :-)
03:33 geoffb jdv79, see scrollback
03:34 geoffb Pragmatism: Bentley: Programming Pearls and More Programming Pearls
03:34 jdv79 oh wow - put it on a wiki
03:35 geoffb Regex: Friedl: Mastering Regular Expressions
03:35 geoffb jdv79, READTHEM at a minimum, Wiki if someone knows a good place.  :-)
03:35 stevan Reusable Software : The Base Object-Oriented Component Libraries : Bertrand Meyer # this book is really really really good book about writing good large scale APIs
03:36 * stevan has to take the dog out, bbiab
03:37 geoffb stevan, nod
03:37 mugwump ?eval my @suggestions=<this that t'other>; {any($_)&all($_)}(@suggestions)
03:37 evalbot6 't\'other'
03:37 geoffb Perl 5 & Networking: Stein: Network Programming with Perl
03:38 geoffb It's sad how many books on my shelf I haven't actually read enough of to make a good recommendation
03:39 geoffb Perl 5 Stretching: Dominus: Higher-Order Perl and Conway: Object Oriented Perl
03:41 geoffb Numerical Analysis: Conte and De Boor: Elementary Numerical Analysis: An Algorithmic Approach
03:41 stevan The Art of the MetaObject Protocol # just came today, but I can already tell it will be a classic
03:41 stevan ok, thats as good as I can do for now, more later &
03:42 geoffb stevan++ # thanks for your input!
03:42 jdv79 has quit IRC (Remote closed the connection)
03:42 * geoffb has more than a half dozen books on firewalls and internet security and can't remember which ones were any good . . . .
03:49 geoffb #haskell has a few suggestions (perhaps already covered in READTHEM):
03:49 geoffb Abelson, Sussman, Sussman: Structure and Interpretation of Computer Programs
03:50 geoffb Pierce: Types and Programming Languages
03:51 geoffb Fowler: Analysis Patterns
03:53 geoffb "Richard P. Gabriel's writings for software engineering"
03:55 geoffb Van Roy & Haridi: Concepts, Techniques, and Models of Computer Programming
03:55 justatheory has joined #perl6
04:00 Amnesiac has quit IRC ("Leaving")
04:03 justatheory has quit IRC ()
04:08 masak has joined #perl6
04:16 geoffb This is getting interesting . . . The shapr on #haskell has given me a number of places to look for lists . . . so now I'm chasing those . . . .
04:23 geoffb Hmmm,  I've now gone from a lack of info to an embarrasment of riches.
04:24 geoffb What I need now is to categorize
04:25 * geoffb ruminates and masticates thoughtfully
04:28 kakos has joined #perl6
04:30 vcv has joined #perl6
04:33 dudley I was already complaining about the size of my books-to-read stack. This certainly isn't helping.
04:33 geoffb heh.  And it will get worse.
04:33 geoffb All though I think categorization and some mild prioritization will help
04:33 * dudley considers growing an extra head
04:33 geoffb and believe me, I KNOW HOW YOU FEEL
04:51 castaway You need my virtuallibrary site.. ;)
04:51 vcv has quit IRC (Read error: 104 (Connection reset by peer))
04:53 geoffb castaway, do tell
04:54 castaway well its still in the making.. but the point is a site where you can list your own library, look at those of others, discuss books / categories of books, add notes to items, create your own categorisations, etc etc
04:55 castaway see http://theorbtwo.perlmonk.​org/kwiki/index.cgi?VBFAQ
04:58 castaway (and I'm looking for helpers, if you want another project to work on ,)
04:58 geoffb like I need a few more things to do . . .
04:58 geoffb :-)
04:58 castaway heh
04:59 geoffb I'll think about it, but I've got too much on my plate as it is.
05:00 geoffb I am right now actively thinking about great book categorization, so that may be useful to you at some point
05:00 geoffb (or not)
05:02 castaway cool
05:06 * castaway needs more hours in the day to work on this ,)
05:07 * geoffb needs more hours in the day period
05:07 castaway indeedy
05:16 * geoffb is creating a mini-database of information about the "great CS books" he has on hand, to get a feel for categorization and such
05:18 castaway have fun ;)
05:18 geoffb actually, it is.
05:18 * castaway nods
05:18 geoffb And a nice break from harder things.  :-)
05:18 castaway I'm still squeezing the bugs out of my book importing mechanism :(
05:20 castaway its odd, the entire import code fails without telling me why, sticking bits in subs and calling them works fine *sigh*
05:20 geoffb bleah
05:20 castaway yeah, its beginning to really annoy me
05:34 autrijus greetings λ camels
05:34 geoffb greetings, autrijus
05:35 coral autrijus: surviving the storm?
05:36 autrijus yup.
05:36 geoffb And there was much rejoicing.  :-)
05:36 autrijus :)
05:37 Amnesiac has joined #perl6
05:50 saorge has quit IRC ("Leaving")
06:00 masak autrijus: is the cyclone still... cycloning?
06:01 * masak has no cyclones outside, just regular bad weather
06:01 masak ...ideal for programming
06:02 autrijus no... it fizzled now
06:02 autrijus which means I get to go outside for food
06:02 autrijus bbiab
06:03 pasteling has quit IRC (Client Quit)
06:04 pasteling has joined #perl6
06:27 pasteling has quit IRC (Client Quit)
06:28 pasteling has joined #perl6
06:33 vcv has joined #perl6
06:54 malaire has joined #perl6
06:54 vcv has quit IRC (Read error: 104 (Connection reset by peer))
06:56 iblechbot has joined #perl6
07:07 masak i think in everything one does one has the choice to err on the side of being correct or on the side of being pragmatic
07:08 masak what i like about perl is that the language doesn't make that choice for me
07:08 geoffb oh yes, definitely
07:08 * masak is programming java presently
07:08 masak java makes that choice for me
07:08 geoffb painfully so
07:08 masak :(
07:09 masak i can't sneeze in java without unintentionally defining ten classes
07:11 masak and, oh, tracing the cause of a NullPointerException in the servlet logs is not as exciting as it may sound
07:12 * masak likes perl where the errors are mostly caused by bugs, not by language design
07:12 geoffb so agree with you there
07:33 larsen has joined #perl6
07:55 Aankhen`` has joined #perl6
08:03 iblechbot has quit IRC (Read error: 104 (Connection reset by peer))
08:09 Amnesiac has quit IRC ("Leaving")
08:11 osfameron has joined #perl6
08:16 Aankh|Clone has joined #perl6
08:16 Aankhen`` has quit IRC (Connection reset by peer)
08:20 elmex has joined #perl6
08:26 Aankh|Clone is now known as Aankhen``
08:33 vcv has joined #perl6
08:33 malaire has quit IRC ("Chatzilla 0.9.68a [Firefox 1.0/20041107]")
08:45 test has joined #perl6
09:14 pasteling Someone at 69.110.115.185 pasted "A concept for finding great" (75 lines, 1.3K) at http://sial.org/pbot/11869
09:14 vcv has quit IRC (Read error: 104 (Connection reset by peer))
09:17 geoffb dangit
09:17 geoffb castaway, ping
09:18 masak neat nopaste
09:18 geoffb well, for anyone who's interested, here's the correct nopaste (that last one was premature):
09:18 pasteling "geoffb" at 69.110.115.185 pasted "A concept for finding great books to read" (82 lines, 1.5K) at http://sial.org/pbot/11870
09:19 geoffb thanks, masak
09:19 castaway pong?
09:20 lumi has quit IRC (Read error: 113 (No route to host))
09:20 geoffb Thought you might be interested in that paste, castaway
09:20 * castaway looks
09:21 pasteling "geoffb" at 69.110.115.185 pasted "A concept for finding great books to read" (95 lines, 1.9K) at http://sial.org/pbot/11871
09:21 geoffb oh, forgot to include help text before, see most recent paste
09:21 castaway heh. and the code?
09:22 geoffb It's ~200 lines of perl 5.
09:22 geoffb I'll commit a p6 version hopefully tomorrow, if I'm awake enough.
09:22 castaway plus how you get the info in there in the first place :)
09:22 geoffb Right now, I'm pretty tired
09:22 castaway okiedokie
09:22 geoffb YAML.  And manual entry.  :-)
09:24 geoffb Book suggestions came from this channel (stevan, mostly), #haskell (shapr), and the c2.com wiki book lists
09:24 geoffb many still to be added to the YAML, of course -- I got quite a few suggestions
09:25 castaway nifty, anyhoo :)
09:25 geoffb thx.
09:25 geoffb And a fun few hours of hacking
09:25 castaway hopefully that'll be possible on the library when I'm done (sometime next century..)
09:25 castaway indeedy
09:25 geoffb heh
09:28 geoffb OK, time to annoy wife by coming to bed at 2:30 AM
09:28 castaway oh, and add in a 'pick one at random' command .)
09:28 castaway g'night :)
09:28 geoffb castaway++ # good idea!
09:28 geoffb g'night to you too
09:29 Aragone is now known as Arathorn
09:31 elmex_ has joined #perl6
09:36 pasteling "geoffb" at 69.110.115.185 pasted "The pick command (castaway++)" (42 lines, 808B) at http://sial.org/pbot/11872
09:36 geoffb Must . . . leave . . . keyboard
09:37 elmex has quit IRC (Read error: 60 (Operation timed out))
09:38 castaway heh, get to bed!
09:49 test has quit IRC ("Chatzilla 0.9.68.5 [Firefox 1.0.6/20050717]")
09:51 masak geoffb++ # neat program
10:13 pupilzeng has joined #perl6
10:19 scook0_ has joined #perl6
10:25 ipv6guru has joined #perl6
10:31 iblechbot has joined #perl6
10:32 scook0 has quit IRC (Read error: 113 (No route to host))
10:42 nothingmuch has joined #perl6
10:44 vcv has joined #perl6
10:47 sunym has quit IRC ()
10:55 nothingmuch has quit IRC (Read error: 104 (Connection reset by peer))
10:57 kolibrie has joined #perl6
10:57 ipv6guru has left
10:58 ipv6guru has joined #perl6
11:39 Nattfodd has joined #perl6
11:39 vcv has quit IRC (Read error: 104 (Connection reset by peer))
11:39 Nattfodd has quit IRC (Read error: 104 (Connection reset by peer))
11:39 Nattfodd has joined #perl6
11:39 Nattfodd has quit IRC (Read error: 104 (Connection reset by peer))
11:39 Nattfodd has joined #perl6
11:39 Nattfodd has quit IRC (Read error: 104 (Connection reset by peer))
11:39 Nattfodd has joined #perl6
11:40 Nattfodd has quit IRC (Success)
11:40 Nattfodd has joined #perl6
11:40 Nattfodd has quit IRC (Read error: 104 (Connection reset by peer))
11:41 Nattfodd has joined #perl6
11:41 Nattfodd has quit IRC (Read error: 104 (Connection reset by peer))
11:41 Nattfodd has joined #perl6
11:41 Nattfodd has quit IRC (Read error: 104 (Connection reset by peer))
11:47 ipv6guru has quit IRC (Read error: 60 (Operation timed out))
11:54 nothingmuch has joined #perl6
11:55 nothingmuch this is fun:.
11:55 nothingmuch i have a test suite to run forth code
11:55 nothingmuch with ingy's Test::Base
11:55 nothingmuch i run it against gforth
11:56 nothingmuch when it starts passing I know my tests are OK
11:56 nothingmuch then I switch the backend to harrorth
11:56 nothingmuch when they pass again I know harrorth is OK
11:56 Nattfodd has joined #perl6
11:57 wilx O_O
11:57 wilx P6 Forth backend?
11:57 nothingmuch actually perl 5 forth
11:57 nothingmuch but it's only an intermediate impl
11:57 nothingmuch i'm really doing forth in haskell
11:57 castaway funky!
11:57 nothingmuch to learn haskell
11:57 castaway you're just crazy :)
11:58 nothingmuch and the perl5 one is supposed to help me define the minimal prelude
11:58 nothingmuch such that most of my forth machine is actually written in forth itself
11:58 nothingmuch with 3 levels of bootstrapping:
11:58 nothingmuch the primitives, which are provided by the system
11:58 nothingmuch and then the words to get the ':' and ';' words
11:58 nothingmuch they are compiled via a bootstrapper
11:59 nothingmuch and then the rest of the prelude words are compiled using ':' and ';'
11:59 nothingmuch castaway: actually I think this is very sane =)
12:01 castaway where do you get time to do all this?
12:05 nothingmuch well, i'm babysitting compilations right now
12:05 nothingmuch as the day will progress I will become busier and busier
12:07 * castaway usually feels quite guilty about fiddling with toy projects at work, even if the work computer is busy..
12:07 nothingmuch well, i really have nothing to do
12:07 nothingmuch when everybody goes home at 20:00-21:00 (hard week), then I can really start working
12:07 nothingmuch because they will need results for tomorrow
12:07 nothingmuch based on their work today
12:08 castaway I see
12:09 stevan morning nothingmuch, castaway :)
12:09 nothingmuch hola stevan
12:10 castaway hey stevan
12:10 Limbic_Region has joined #perl6
12:10 elmex_ has quit IRC ("leaving")
12:12 * castaway nudges Limbic
12:13 nothingmuch wow, my forth backend is more complete than I thought =D
12:13 * Arathorn wonders if he's got himself caught in a second-degree warnock on p6l
12:14 wilx warnock?
12:14 osfameron that sounds like something out of the rules of Mornington Crescent..
12:14 * Limbic_Region falls over at the nudging, hits the floor, and promptly falls back asleep in a more comfortable position
12:14 Arathorn http://en.wikipedia.org/wiki/Warnock
12:15 * osfameron pushes Limbic_Region down a treacle well
12:15 Arathorn treacle treacle treacle treacle...
12:15 stevan Arathorn: which post?
12:15 castaway Hrrmm
12:16 Arathorn the most recent one from [email@hidden.address] trying to grapple with Larry's commentary on ::
12:16 Arathorn i'm just being impatient, tho' (or i'm worried that I wrote the mail too late last night for it to make any coherent sense at all ;D)
12:17 Arathorn the most amusing thing about playing Mornington Crescent is how deserted the station is irl... i can only assume that people tend to get caught out in Warnock Dilemmas and Register Alligators in the outer reaches...
12:17 * Arathorn waffles off
12:23 nothingmuch me reads mail after not having done so for about 24 hrs
12:29 Arathorn nothingmuch: you're sending mails from 24 hours ago, fwiw
12:30 castaway ,)
12:31 nothingmuch ?
12:31 nothingmuch oh, right
12:31 nothingmuch postfix died
12:31 nothingmuch queue was full
12:31 nothingmuch =)
12:31 Arathorn ah - fair enough
12:31 nothingmuch mutt hasn't finished opening
12:31 Arathorn was trying to piece together whether they were just old or misdated
12:32 nothingmuch they are old
12:32 nothingmuch i replied to luke, i think
12:32 nothingmuch can't remember what else
12:32 pdcawley has quit IRC ()
12:58 vcv has joined #perl6
12:59 wilx Huh.
12:59 wilx Could somebody post somewhere some examples of P6 compiled into JS?
13:00 wilx I'd really like too see it but I do not have Pugs and I kinda am not able to build it on Windows.
13:01 osfameron tried PxPerl ?
13:01 osfameron autrijus posted link to http://m19s28.vlinux.de/iblech/stuff​/not_perm/pil2js-demo/05-sub.t.html on his journal
13:01 Arathorn wilx: http://m19s28.vlinux.de/iblech/stuff​/not_perm/pil2js-demo/mandel.p6.html
13:01 wilx Nope, what is it?
13:02 osfameron wilx: PxPerl is a windows binary of Pugs
13:02 Limbic_Region has quit IRC ("CGI:IRC (EOF)")
13:02 osfameron FireFox-- # a script on this page is causing Mozilla to run slowly popup
13:02 osfameron (which you can't cancel)
13:02 wilx Yup, same here.
13:03 Arathorn i can cancel it, fwiw - but i think iblech has breaking the main CPU-intensive loop into a background thread in his pipeline
13:04 osfameron oh, actually, I can cancel it but then it asks me again
13:04 osfameron for fuck's sake, I told you to piss off already you stupid bag of shite
13:04 * osfameron considers going back to Opera
13:04 castaway tsktsk
13:04 castaway Opera++
13:05 castaway (took 3min40 but didnt complain ,)
13:05 osfameron wilx: PxPerl is at http://www.pixigreg.com/
13:06 Arathorn my opera was miles faster than firefox, weirdly
13:06 Arathorn (under opera 8)
13:06 Arathorn must see how Safari 2 runs with its new JS engine...
13:06 wilx Hmm, it doesn't seem to work, the mandel example, under IE6.
13:07 Arathorn works fine for me
13:07 Arathorn what does the javascript console say?
13:08 Arathorn bear in mind that it'll take a few minutes to draw ;)
13:08 Arathorn (and IE's buffering's a bit bashful)
13:08 nothingmuch omfg, quicksilver is absolutely insane
13:08 svnbot6 r5714, iblech++ | * PIL2JS: Renaming: Former PIL2JS.Box.Proxy is now PIL2JS.Box
13:08 svnbot6 r5714, iblech++ |   to better reflect their behaviours.
13:08 svnbot6 r5714, iblech++ | * PIL2JS: Repaired/fully implemented &infix:<=:=>.
13:08 svnbot6 r5714, iblech++ | * PIL2JS: my @a = (0, 1, 2); @a[1] = 42 works now (was dieing:
13:08 svnbot6 r5714, iblech++ |   "cannot modify constant").
13:08 svnbot6 r5714, iblech++ | * PIL2JS: Binding improved:
13:08 svnbot6 r5714, iblech++ |   * "3 := 4" is now a real error (and not a silent JavaScript error).
13:08 svnbot6 r5714, iblech++ |   * "my $a; my $b := $a; $a := $b" is an error now, too
13:08 svnbot6 r5714, iblech++ |     ("cannot create a bind cycle").
13:08 svnbot6 r5714, iblech++ |   * Binding array and hash elems really work now.
13:09 nothingmuch if you type a URL, and you ask for it's subitems, it gets the page, takes it apart, and gives you the interesting bits
13:09 nothingmuch without ever opening the browser
13:11 Arathorn . o O ( wget -r )
13:11 nothingmuch o I hear perl6? http://lambda-the-ultimate.org/node/view/852
13:11 nothingmuch Arathorn: quicksilver is like a launchbar thing
13:11 nothingmuch you hit a hot key
13:11 nothingmuch type some text
13:11 nothingmuch and it guesses what item you mean
13:12 nothingmuch if it's a URL, you can open it
13:12 nothingmuch or you can view it's subitems in it, by pressing the right arrow
13:12 Arathorn ah, cool
13:13 nothingmuch do you hav OSX?
13:13 nothingmuch http://quicksilver.blacktree.com/
13:13 Arathorn i have a mac mini buried in my desk somewhere
13:13 nothingmuch install it then... it's lots of fun
13:13 iblech has joined #perl6
13:13 * Arathorn will give it a whirl next time he's feeling fruity :)
13:13 nothingmuch i've mapped it to the fn key
13:14 nothingmuch when I just press it it launches, if I press and hold it doesn't get in the way
13:14 castaway buried *in* your desk?
13:14 osfameron quicksilver and subethaedit are the 2 main reasons I wanted to try OSX
13:14 Arathorn yeah, i wanted it for subethaedit and omnigraffle
13:14 nothingmuch i tried it a while ago (qs), but it didn't support hebrew
13:14 Arathorn (and quartz composer)
13:14 nothingmuch now it does
13:14 osfameron though I found the rest of it annoying enough that I haven't really bothered
13:14 * castaway mutters something about SEE
13:14 nothingmuch Arathorn: omnigraffle isn't as good as visio (ack) IMHO
13:14 nothingmuch not that i've used anything other than these two
13:14 Arathorn visio *still* can't do wretched dropshadows
13:14 Arathorn because it's stuck with 1996-vintage GDI
13:15 nothingmuch oh, the pretty factor
13:15 nothingmuch i don't know about the looks, i must admit i never really cared
13:15 Arathorn well, unfortunately the pretty factor is often quite important for public-facing diagrams
13:15 osfameron Visio also sometimes decides that though you explicitly moved smart connectors, it's now going to put them back to its default path-finder algorithm path
13:16 Arathorn omnigraffle 4 is a big improvement over 3, too
13:16 Arathorn (although i don't dispute that visio has much better stencil libraries)
13:17 * nothingmuch doesn't use those programs often
13:17 meppl has joined #perl6
13:17 nothingmuch i'd rather sketch and scan
13:26 robkinyon has joined #perl6
13:26 robkinyon has left
13:40 Maddingue has joined #perl6
13:41 Maddingue has quit IRC (Read error: 104 (Connection reset by peer))
13:41 Maddingue has joined #perl6
14:08 Limbic_Region has joined #perl6
14:08 vcv has quit IRC (Read error: 104 (Connection reset by peer))
14:08 svnbot6 r5715, iblech++ | PIL2JS: New file runjs.pl (compiles Perl 6 to JavaScript and runs it):
14:08 svnbot6 r5715, iblech++ | perl5/PIL2JS/runjs.pl t/01/01-tap.t
14:25 autrijus iblech: I made runjs.pl happier to invoke
14:25 svnbot6 r5716, autrijus++ | * runjs.pl - allow it to be invoked from any directory,
14:25 svnbot6 r5716, autrijus++ |   using FindBin.pm to find the location of prelude etc.
14:25 svnbot6 r5716, autrijus++ | * Also, `--` is no longer mandatory -- this should Just Work:
14:25 svnbot6 r5716, autrijus++ |     % perl runjs.pl -e "say 'Hello, World'"
14:25 cventers_ ?eval say $?PUGS_VERSION
14:25 evalbot6 Perl6 User's Golfing System, version 6.2.8, July 13, 2005 (r5715) bool::true
14:25 iblech autrijus: Sure, thanks :)
14:25 autrijus iblech++ # idea==>iblech==>code
14:26 iblech autrijus: :pil.yaml my %hash = (a => 1) -- look at the context of &infix:<,>, it's TCxtVoid. That's a bug, right?
14:26 autrijus must be a bug, yes
14:26 autrijus I noticed it a while ago
14:27 nothingmuch perl runjs.pl
14:27 nothingmuch that's 3 languages in one system
14:27 nothingmuch =D
14:29 * autrijus is still doing relatively uninteresting PDF::Template hackery for $work
14:30 * iblech (tries to) implement context in PIL2JS :)
14:31 autrijus in old parrot, Cxt is an extra implicitly-passed param to each sub
14:32 autrijus in new parrot, Cxt is deduced from the signature of the return continuation.
14:32 autrijus the new parrot had not yet come to pass, though
14:32 autrijus but there is a lot in favour of unifying Params and Cxt into the same struct
14:32 nothingmuch autrijus: darcs pull harrorth
14:32 autrijus Dan likes that, and I think lwall is happy too
14:33 autrijus nothingmuch: god it
14:33 autrijus s/god/got/
14:34 nothingmuch autrijus: under mis is a new distribution like thing
14:34 nothingmuch if you feel like playing with Forth, write some tests ;-)
14:34 Arathorn morning autrijus: further to lwall's response to my post on p6l, i'm concluding that variables in the default namespace should be $*Main::foo (also can be written as $*::Main::foo and $::('*')::Main::foo - but not as $::Main::foo) - does that sound sane to you?
14:34 autrijus Arathorn: yes.
14:34 Arathorn woo
14:35 autrijus but I agree with lwall that this is very hard to explain ;)
14:35 autrijus I'd rather we had used three different symbols :)
14:35 * Arathorn nods.
14:36 Arathorn especially the assymetry of $::('*::Main')::foo being valid, but $::*::Main::foo being invalid
14:36 Arathorn and the latter not actually having any welldefined meaning currently
14:36 autrijus yeah. I think if you can think of a saner treatment, there's a good chance that it will be adopted
14:36 autrijus the current situation is messy.
14:36 * Arathorn hopes that lwall will decree that $::foo is dead, or actually a way of deferencing a type-space object, or just equivalent to $::('foo')
14:37 nothingmuch do anonymous classes have a sigil?
14:37 Arathorn having $::foo as an alias for $main::foo as in p5 would be utterly horrible.
14:38 autrijus nothingmuch: do anonymous closures have a sigil?
14:38 nothingmuch i don't know
14:38 autrijus Arathorn: actually I wouldn't mind it meaning $*Main::foo...
14:38 nothingmuch they can either fit in a scalar ($foo) or a function name (my &moose = sub  { }), right?
14:38 autrijus right.
14:39 nothingmuch and in $foo they are a code ref
14:39 autrijus so you can have
14:39 autrijus my $bar = class {}
14:39 autrijus or
14:39 autrijus my ::bar = class {}
14:39 nothingmuch and ::bar is a relative namespace, that exists (or rather, is findable) just for the duration of the current scope, right?
14:39 jhorwitz has joined #perl6
14:39 autrijus yes.
14:39 Arathorn autrijus: but wouldn't $::('foo') being so different to $::foo be *really* counterintuitive?
14:39 nothingmuch hmm
14:40 autrijus Arathorn: I think $::() is the horrible party.
14:40 nothingmuch what's the clash between :: as a sigil and :: as a namespace separator?
14:40 nothingmuch ::::somepackage::class?
14:40 autrijus nothingmuch: no clash just messy
14:40 Arathorn well, the 'correct' way to do it would be to then have :: as a trailing sigil after the interpolation construct, no?
14:40 autrijus trailing sigil?
14:41 Arathorn i.e. $($foo)::bar (rather than $::($foo)::bar
14:41 * nothingmuch ponders a horrible change: the # sigil
14:41 Arathorn if that wasn't then completely ambiguous
14:41 nothingmuch #class = ...
14:41 nothingmuch # class = ...  (this is a comment)
14:41 ToniEisner has quit IRC (Remote closed the connection)
14:41 nothingmuch doe
14:41 nothingmuch s the opaque type have a sigil?
14:42 autrijus opaque type?
14:42 nothingmuch yep, an object
14:42 nothingmuch you can have a blessed hash ref in an object.
14:42 nothingmuch or an opaque
14:42 nothingmuch or ...
14:42 nothingmuch i assume that $?SELF contains a reference
14:42 nothingmuch what is opaque's sigil?
14:42 nothingmuch or is it by design unsigillable?
14:43 Arathorn ?eval $foo="foo"; $($foo);
14:43 evalbot6 Error:  unexpected "(" expecting "::"
14:43 autrijus I think it is unsigilable.
14:44 autrijus ?eval $foo="foo"; $::($foo);
14:44 evalbot6 Error: Undeclared variable: "$foo"
14:44 autrijus ?eval my $foo="foo"; $::($foo);
14:44 evalbot6 \'foo'
14:44 autrijus there you go
14:44 * Arathorn nods
14:44 autrijus ?eval my $foo="foo"; say $::($foo);
14:44 evalbot6 foo bool::true
14:44 nothingmuch what about a unicode sigil?
14:44 Arathorn but, what ambiguities would be introduced by using $($foo):: instead?
14:44 autrijus nothingmuch: you need an ascii version for it anyway.
14:44 * nothingmuch goes to look at his pallette
14:44 nothingmuch autrijus: $@!^ or something like that ;-)
14:45 Arathorn i.e. is $($something) a valid alternative syntax for $$something ?
14:45 autrijus I don't think $@!^ would work.
14:45 nothingmuch autrijus: i was just bashing my keyboard, please don't try to parse it =)
14:46 justatheory has joined #perl6
14:46 autrijus nothingmuch: ah. please keep bad ideas like that to yourself :)
14:46 nothingmuch hehe
14:46 autrijus Arathorn: I suspect it can work. at least that's one :: less
14:47 nothingmuch we have some unused parenthesis
14:47 Arathorn thing is, that :: is still a prefix for disambiguating PackageNames, right?
14:47 nothingmuch ⎩
14:47 Arathorn turning $::($foo) into $($foo):: and ::PackageName into PackageName:: just so that $::foo is free to be $*Main::foo again seems rather wrong :/
14:47 nothingmuch ‰ ‱  =)
14:47 nothingmuch those two look sigilized
14:48 autrijus Arathorn: the thing is that &Foo and ::Foo can both be referred to as Foo
14:48 Arathorn sure
14:48 autrijus Arathorn: and as such each need one sigil
14:48 autrijus also I think the proposal is to turn $::($foo) into $($foo)
14:48 autrijus ::PackageName would still exist as ::PackageName
14:49 Arathorn okay - but then you get the contradiction that ::PackageName is a disambiguated package name - but $::PackageName is actually $*Main::PackageName
14:49 Arathorn which is counterintuitive too.
14:49 autrijus yes. now if we adopt the ^ sigil for packages
14:49 autrijus ^PackageName would mean the package
14:49 autrijus and there is much less ambiguities
14:49 Arathorn and :: would only ever be a trailing separator/prefix for typespace
14:49 Arathorn s/prefix/suffix/
14:49 Arathorn (gah)
14:50 autrijus ^Foo::Bar  then is the Bar type in the Foo namespace
14:50 Arathorn $client on phone - brb
14:50 elmex has joined #perl6
14:50 autrijus I don't know, but this might be sane
14:50 autrijus however it clashes with $^x ;)
14:51 * autrijus stops speculating about mere sytnax
14:55 * Arathorn fumes at $client
14:55 PerlJam has joined #perl6
14:56 Arathorn i really like ^ (or something), fwiw
14:56 Arathorn $Foo="Foo"; ^($Foo)::Bar then makes a whole lot more sense
14:56 Arathorn as does $($Foo)::Bar
14:57 Arathorn and $::foo is free to be $*Main::foo (or perhaps $OUR::foo?) again
14:57 chip anybody around who's familiar with the parameter unpacking rules?
14:58 * Arathorn will try suggesting an alternative typespace sigil than :: to lwall if his p6l response gets dewarnocked
14:58 chip e.g.  sub quicksort ( [ $pivot, *@data ], ?$reverse, ?$inplace ) {...}
14:59 nothingmuch what is ... defined as?
14:59 nothingmuch sub &nofix:<...> { die "yada yada" }; # ?
15:00 nothingmuch hola chip... how's paypal coming along?
15:00 * Arathorn has no clue about parameter unpacking, sorry :/
15:00 nothingmuch chip: slightly
15:00 chip nothingmuch: Something else to draw donations is pending.  I'll give a hint: It's the most powerful force in the universe
15:00 autrijus nothingmuch: no, "fail" not "die"
15:01 nothingmuch are you going to nuke all your cash to create fusion-dough ?
15:01 nothingmuch autrijus: err, yep... but what about nofix: ?
15:01 chip nothingmuch: Is the full parameter semantic available inside the unpacking?   sub foo ( [ $a, ?$b, +$c, *%x, *@y ], $foo )
15:01 autrijus nothingmuch: it is just a nullary prefix, silly.
15:01 nothingmuch autrijus: so how do i write that? sub nullarry:<...> ?
15:01 autrijus no
15:01 nothingmuch chip: i would think that no
15:01 autrijus sub term:<...> {}
15:01 autrijus see A12
15:01 nothingmuch oh
15:01 chip nothingmuch: which is NOT to say I don't need the paypal contributions, by the way!
15:01 nothingmuch perhaps optionality
15:02 nothingmuch but not namedness
15:02 chip nothingmuch: "Help Save The World" by spreading the word
15:02 nothingmuch brb, i need to save someone
15:02 chip nothingmuch: S6 only shows optionality, so I guess you're right
15:02 autrijus chip: unpacking hash gives positional bindings.
15:02 autrijus err
15:02 chip autrijus: hello!
15:02 autrijus chip: unpacking an array gives positional bindings.
15:03 autrijus unpacking a hash gives named bindings.
15:03 autrijus I don't think you can have both
15:03 autrijus chip: hey :)
15:03 chip autrijus: are the array and hash bindings required to work left-to-right in the overall function signature?
15:03 chip autrijus: is this legal:   sub foo ( [ $x ], ?$y = $x )
15:03 autrijus I think it is.
15:04 autrijus although if you have powerful arguments for disabling it
15:04 autrijus then I think lwall will consider that :)
15:04 chip well, it's just that name binding is looking less and less like something I want to address.  That's just putting a cap on it.
15:05 chip i.e. get_params will probably never be aware of the lexical pad
15:05 autrijus sure. we can deal.
15:05 autrijus iblech has perl 6 named bindings implemented in JavaScript of all things. ;)
15:05 chip autrijus: JS is the flavor of the month, isn't it?  *sigh*
15:06 autrijus chip: until ctx5 is back to business, I'm afraid nothing much we can do on PIR targetting :)
15:06 Arathorn it's actually kind of cool, though, as complementary to parrot stuff
15:06 autrijus "back to business" as in "usurps as the new trunk"
15:06 chip Does P6 (or Pugs) have any need to specify the specific PMC to be used for *@a or *%b?
15:06 chip autrijus: I see.  Well, I'll be reviewing the differences soon then.
15:06 autrijus chip: hm? the user can request a PMC.
15:07 autrijus sub foo (FooBarArray *@a) {...}
15:07 Arathorn (unless you're planning to persuade macromedia to embed parrot in Flash Player rather than ECMAScript, or persuade M$ to replace JavaScript in IE with it...)
15:07 chip sub foo (MyArrayType *@a) ... ?
15:07 autrijus I suspect so
15:07 autrijus I'm fine if you'd like to disallow that too
15:07 autrijus I don't think it will be used often.
15:08 chip It's OK with me if the :slurpy PMC has to be pre-created, but leo's gonna have to change something
15:08 chip he likes to have the get_params be the very first opcode in the sub.  Oops.
15:08 autrijus I'm fine with the PMC being dynamically created too.
15:10 Aankhen`` has quit IRC (No route to host)
15:14 masak has left
15:14 Khisanth hmm p6 no longer has glob() ... I guess that does get rid of one problem
15:17 Arathorn autrijus: were you surprised by lwall saying that $OUR::foo is the right way to do package MyPackage; $MyPackage::foo="foo"; say $OUR::foo # if you don't happen to know what namespace scope you're currently in
15:18 Arathorn despite $foo not having been declared as a lexically-scoped global using 'our', but a package-variable style symbol table entry?
15:18 autrijus no, not at all. OUR:: is a hackathon addition.
15:18 autrijus and it was designed to do that.
15:19 Arathorn ah, I was wondering why it wasn't in the AES's at all
15:19 Arathorn isn't it really weird to put non-lexically-scoped variables in something called OUR::, though? :/
15:20 * Arathorn was expecting PACKAGENAME:: or ::($?PACKAGENAME)::
15:20 autrijus sure you can also do that.
15:20 Arathorn right
15:23 * Arathorn has a cassandraesque vision of legions of future newbies all doing perl6 -e '{ $*Main::foo = "foo"; } say $OUR::foo' and getting suitably confused about OUR:: including non-lexicals, but 'our' defining lexicals only
15:23 osfameron me has a cassandraesque ision of doing the same thing himself
15:24 * autrijus purrs.
15:25 * osfameron strokes autrijus's ears
15:25 Arathorn . o O ( is that a catbert purr, or a right-now-it's-easier-to-be-feline purr? )
15:25 * autrijus meows happily
15:26 autrijus Arathorn: anyway, our's scope is lexical but the thing it points to is packagewide
15:26 autrijus whilst my's scope is lexical but the thing it points to is lexical
15:26 Arathorn okay, i do understand it correctly then
15:27 * Arathorn will plaintively try asking that OUR:: gets called PACKAGENAME:: or something if it's going to contain packagewide variables rather than just lexical globals
15:27 Arathorn in a preemptive strike against the strife of newbies & not-so-newbies to come.
15:28 chip wtf is "*Main::" ?
15:28 autrijus chip: it is the new name of main::
15:28 autrijus I have no idea why larry changed main to Main.
15:28 autrijus the * indicates toplevelness, that's fine
15:29 chip * is toplevel _and_ global?
15:29 autrijus the two are the same thing.
15:29 Arathorn *Main is a specialcase shorthard for *::Main
15:29 chip autrijus: / is Unix fs toplevel, but / isn't implicitly in everyone's $PATH
15:29 Arathorn if that makes it any clearer
15:29 chip toplevel isn't automatically global
15:30 autrijus chip: right. to clarify: in perl6 "/" is always the end of your $PATH.
15:30 autrijus better?
15:30 chip clear but disturbing.  Grr.
15:30 Arathorn which allows most builtins to live in /, presumably
15:30 autrijus yes.
15:30 autrijus which is I presume the idea.
15:31 autrijus chip: in p5 it's called CORE if I'm not mistaken
15:31 autrijus which is sort of like /bin/
15:32 chip autrijus: for hysterical raisins, all of CORE::, CORE::GLOBAL::, and UNIVERSAL:: have aspects of *
15:32 autrijus yes.
15:32 Arathorn i guess that's why it's been normalised, then..
15:32 autrijus I'm aware of the hysterical raisins... that's why we are raised as hystericals...
15:32 chip "Why Be Normal"
15:32 autrijus "A B Normal"
15:33 chip SQL bumper sticker: "Why Be Normalized?"
15:33 * Arathorn is a bit askeered of people going and using $Main::foo, and then getting royally shot in the foot if they happen to be enclosed in another namespace scope called Main
15:34 chip Arathorn: well, anybody who creates a namespace C<My::Own::Main> is just asking for trouble anyway.
15:34 Arathorn true.
15:35 Arathorn peeps will prolly all go and use $Main::foo and forget about it, then - and it does make sense to ucfirst main, to fit in with all the other ucfirst typespace names
15:36 wilx has quit IRC (Read error: 104 (Connection reset by peer))
15:39 * Arathorn is perfectly normal... orthogonal, in fact :/
15:39 * Arathorn sighs and returns to actionscript-infused $work :|
15:40 integral hmm, why Main and not MAIN?
15:40 vcv has joined #perl6
15:41 autrijus because p6 refuses to use ALLCAPS...
15:41 Arathorn are there any uc'd typespace identifiers in p6?
15:41 autrijus not afaics
15:41 integral autrijus: except for MY, CALLER, OUR, etc I thought?
15:41 Arathorn Capitalised Nouns look more teutonic and reliable, whereas SHOUTING looks like COBOL..
15:41 Arathorn oh, true.
15:42 Arathorn (but they're really-really-special-cases, whereas Main is just another normal package scope?)
15:42 autrijus integral: those are pseudopackages
15:43 autrijus you can't enter them
15:43 autrijus package MY; # boom
15:43 integral ah
15:43 autrijus package Main; # just fine
15:43 integral so it's not just a tied package? :-P
15:44 putter has joined #perl6
15:45 wilx has joined #perl6
15:46 autrijus mmm tied package ;)
15:46 saorge has joined #perl6
15:47 Aankhen`` has joined #perl6
15:48 pasteling "putter" at 66.30.119.55 pasted "[PATCH] parrot: creates t/p6rules/rx_grammar.t" (76 lines, 3.4K) at http://sial.org/pbot/11880
15:49 autrijus putter: ?
15:49 putter Hi all.  Could someone with parrot commit bits add http://sial.org/pbot/11880  ?  Thanks.  All PGE's tests are passing... can't have that...
15:50 putter rx_grammar.t is a failing (as in segfault) test.
15:50 putter It's an incomplete grammar for p6 regexps.  Once it runs, PGE will give us a regexp parse tree in the form of a Match tree.
15:51 putter Which can then be used to generate... whatever.
15:52 putter Sound plausible?
15:54 autrijus yes. are you on irc.perl.org #parrot?
15:54 autrijus if not, try bringing it there to chip+leo and ask for a commit bit
15:54 autrijus :)
15:54 autrijus (chip is here in #perl6; leo isn't)
15:54 putter Not at the moment.  Tried there last night, but (almost) noone around, and nothing seems to have happened.
15:55 autrijus oh. it's at this moment a very happening place :)
15:55 * putter desperately trying to avoid having a parrot commit bit again.  silly, I know.
15:56 putter ok, will try there now.  tnx.
15:57 autrijus np :)
16:12 Limbic_Region http://moon.google.com/ # maximum magnification for a chuckle
16:16 putter putter maligned PGE.  fixed grammar works!  :)  backporting to rules...
16:21 autrijus nice!
16:32 b6s has left
16:45 stevan hola autrijus
16:46 stevan question on the metamodel bootstrapping when you have a sec
16:47 * putter spoke too soon... (wasn't entering grammar at root)  looks like there is still a pge issue...
16:47 stevan hey putter :)
16:48 autrijus sure... I'm chipping with chip in #parrot about the insanity of by-value Pair binding
16:48 autrijus you know, the one that makes    for %h.pairs -> $_ { ... } broken
16:48 stevan fun
16:49 autrijus (my proposal on p6l (subj: Strange interaction between pairs and named binding) is that untyped Block params should be Any not Item)
16:49 autrijus supports welcome
16:51 stevan with Any it allows for Pair right?
16:51 stevan in that case
16:52 stevan so, I was trying to represent ::Method and ::Attribute objects in the metamodel
16:52 autrijus yes.
16:52 autrijus ok
16:53 stevan but this of course presents a metacircularity issue
16:53 autrijus right
16:53 stevan but I am thinking maybe that doesnt make sense, maybe those should be pseudo-primatives anyway
16:54 autrijus I think the question boils down to how much datatypes you want out of your host
16:54 autrijus s/much/many/
16:54 autrijus and what their (static, never extended) methods are
16:55 stevan ok
16:55 stevan because I have boiled Perl6::Object down to meta(), can(), isa() and AUTOLOAD
16:55 stevan meta to return the meta object
16:55 stevan can just dispatches to AUTOLOAD
16:55 stevan and so does isa
16:56 autrijus right.
16:56 stevan and AUTOLOAD just catches method invocation
16:56 autrijus you need host to provide function calls with invocant
16:56 autrijus for ::Method
16:56 autrijus yup
16:57 stevan ok, well I think I will hold off on any more bootstrapping until I read more of book
16:59 autrijus cool. nice to see you have timeslices for your work -- i.e. reading the book :)
16:59 stevan yes, but $work is getting a little CPU starved ;)
17:04 elmex has quit IRC ("leaving")
17:09 iblech Hm... Test::is's signature is: (Str $got, Str $expected, Str +$desc, +$todo, +$depends)
17:09 iblech But, according to S06, args which have a "+" in front of them (i.e. $desc) may *only* be passed by name
17:09 autrijus aye.
17:10 autrijus so drop the +
17:10 autrijus :)
17:10 iblech :) PIL2JS found a bug :)
17:10 autrijus good :)
17:11 lightstep has joined #perl6
17:12 stevan what is the current layout for sub types? Sub is Code is Block?
17:13 iblech Both Block and Sub are Code, I think
17:13 iblech But Sub is not a Block
17:13 stevan ok
17:13 stevan so Code is an executable chuck of expressions and statements
17:13 Limbic_Region has quit IRC ("CGI:IRC (EOF)")
17:14 stevan and sub is a named version of that
17:14 stevan and Block is unnamed
17:14 stevan does that sound right?
17:14 autrijus no.
17:14 iblech No, subs can be unnamed to (sub ($foo) {...})
17:14 stevan ah yes
17:14 autrijus you can't return from a Block.
17:14 autrijus that's the primary difference :)
17:14 * stevan slaps forhead
17:14 autrijus everything else is secondary.
17:14 stevan ok
17:14 iblech Well you can't, but you'll return from the surrounding Sub, right?
17:14 autrijus yup.
17:15 autrijus and the main program thing is Code
17:15 autrijus but not Block nor Sub
17:15 autrijus so you can't write
17:15 autrijus pugs -e '&?BLOCK(...)'
17:16 stevan now what about parameters, does Sub has-a Parameters?
17:16 autrijus no, Code has-a Params
17:17 autrijus actually I think it's called Signature
17:17 stevan ah yes, parameterized blocks
17:17 stevan ok
17:17 autrijus even bare can be parameterised.
17:17 autrijus $_ and $^x
17:18 stevan so the Signature is comprised of Params, Returns
17:18 stevan anything else?
17:19 autrijus actually I'll say that
17:19 autrijus Code has-a two slots
17:19 autrijus both of Signature type
17:19 autrijus one is the Signature it accepts, $.params
17:19 autrijus one is the Signature it returns, $.returns
17:20 autrijus but if you unify them into a single Singature type I'm fine with that too.
17:20 autrijus if so, then
17:21 autrijus class Signature { has ($.params, $.returns) is Array of Param;}
17:21 autrijus they are isomorphic so both is fine
17:22 stevan ok
17:22 stevan I am going to rough this stuff out
17:22 autrijus woot
17:22 stevan nothingmuch: and I might work on MMD this weekend
17:22 stevan I am a virtual participant in the Tel-Aviv Hackathon :)
17:23 stevan all this MMD arguing and nobodys got any code to show
17:23 wolverian you _can_ return from a Block, it's just not called returning, right?
17:23 stevan I figure a prototype is in order
17:23 autrijus wolverian: right it's called leaving.
17:23 iblech wolverian: Yes. &leave returns from the innermost block
17:24 wolverian yay, for once I remembered something right :)
17:24 wolverian oh, I might have tuits today to code URI more
17:24 putter hey stevan :)
17:25 stevan ok, I am going to take my metamodel break from $work now thanks autrijus, iblech :)
17:25 stevan hey putter
17:26 autrijus :)
17:28 iblech What should happen if I return() inside a try block? -- sub foo { try { return }; say "still there" }; foo() -- Should "still there" be printed or not?
17:28 autrijus nope, a return is a return.
17:28 wolverian no, that's a bare block.
17:29 lightstep some languages model returns as exceptions
17:29 iblech Ok. Pugs has that wrong then
17:29 iblech lightstep: For example PIL2JS :)
17:29 wolverian is try { } a statement?
17:29 autrijus and PIL2PIR :)
17:30 autrijus no, try is ordinary sub
17:30 wolverian oh. hm. why? since we have the neat do syntax now :)
17:30 autrijus because eval{} was expr.
17:30 svnbot6 r5717, iblech++ | * Test: Changed various +$foos in sub signatures to ?$foos. Tests now have an
17:30 svnbot6 r5717, iblech++ |   actual chance of passing.
17:30 svnbot6 r5717, iblech++ | * PIL2JS: Beginning of context handling:
17:30 svnbot6 r5717, iblech++ |   * All subs take a PIL2JS.Context object as first parameter.
17:30 svnbot6 r5717, iblech++ |   * When calling native functions, the context object obviously has to be
17:30 svnbot6 r5717, iblech++ |   * excluded.
17:30 svnbot6 r5717, iblech++ | * PIL2JS: Better error messages (thanks to new variable $CUR_SUBNAME),
17:30 svnbot6 r5717, iblech++ |   beginning of callchain support.
17:31 autrijus iblech: the thing is try{} should catch error exception not control exceptions
17:31 autrijus but pugs's try{} currently catches both.
17:31 wolverian autrijus, and I always hated it for being an expr. :) am I really the only one?
17:31 iblech autrijus: Right.
17:31 rindolf has joined #perl6
17:31 rindolf Hi all!
17:31 autrijus greetings rindolf.
17:31 PerlJam hello rindolf
17:31 rindolf autrijus: hi
17:31 autrijus wolverian: I don't hate it
17:31 rindolf perlbot: hi
17:31 perlbot hi rindolf
17:31 rindolf nothingmuch: here?
17:31 rindolf PerlJam: hi.
17:32 wolverian autrijus, it just feels inconsistent, both in perl5 and 6.
17:32 wolverian I know you're not who I should be complaining to :)
17:32 * autrijus nods and points to the implementation monkey hat
17:33 wolverian has this been specced and put in stone yet?
17:33 pupilzeng has quit IRC ("See you!")
17:33 autrijus yes and yes :)
17:33 * putter tries to picture the implementation monkey hat...
17:33 autrijus but stones have half life of 7 days on p6l.
17:34 Khisanth putter: it's very ugly :)
17:34 putter :)
17:34 wolverian autrijus, was it on p6l? I just want to read the thread :)
17:34 autrijus wolverian: no, i mean in syn
17:34 wolverian ah. thanks. reading now.
17:34 autrijus not sure about p6l.
17:40 * autrijus posts a strange patch on p6c ;)
17:41 PerlJam indeed.
17:41 autrijus it is, however, I think neccessary.
17:41 PerlJam don't want to answer that question?
17:41 PerlJam or denying that it's frequently asked?
17:41 autrijus I don't think it's answerable by me
17:42 autrijus if @Larry has an updated answer, that's cool
17:42 wolverian hmm. I decided it's probably not an issue, since all blocks can have CATCH, and it's just consistent with do { }
17:42 autrijus but the current answer is bad :)
17:42 PerlJam Well, what was there was a pretty good guess (maybe).  But someone should post the real answer.
17:42 autrijus as in, actively bad.
17:42 autrijus huh, Q3 2005 release a good guess, based on the incoherent 2004 synopses?
17:42 autrijus that's probably not a good guess :)
17:43 autrijus it might be, not anymore.
17:43 PerlJam autrijus: sure it was. p6 devel has come a long way making the Q3 2005 seem even more likely
17:43 PerlJam (the key word is "seem" ;-)
17:44 autrijus Q3 2005 is in 2 months away :)
17:44 sapper has joined #perl6
17:44 autrijus and pugs was never based on the 2004 synopses :)
17:44 PerlJam Look mister miracle worker, don't tell me you don't believe in miracles  ;)
17:44 osfameron has quit IRC (Read error: 113 (No route to host))
17:44 autrijus nah. it says the 6.0.0 production compiler
17:45 autrijus which is beyond my control anyway, and I don't think pugs can make it in 2 months.
17:45 PerlJam Pm has told me several times that he's still targetting a Q3 2005 release for the first alpha perl6 compiler.  (I keep telling him that he's crazy though)
17:45 autrijus (neither was that the goal.)
17:45 autrijus Pm's compiler can't yet parse nested expressions.
17:45 autrijus so, not likely :)
17:46 PerlJam Well, next time I ask him. I'll be sure to remind him of the insanity again  ;)
17:46 autrijus (I'm talking about the expression parser that is yet to be finished)
17:46 integral pm has a compiler?
17:46 PerlJam Pm's problem is another one of those N > 1 masters again.
17:47 autrijus integral: well, a plan for one, based on PGE and a shift/reduce bottom-up parser.
17:47 autrijus languages/punie/ is a prototype compiler based on that approach
17:47 integral ah, I thought you meant he was beyond that :-)
17:47 integral oh, punie uses that stuff?  neato :-)
17:47 PerlJam autrijus: he has an actual implementation that does ... stuff (I'm not sure as he hasn't shown it to me beyond the first baby steps a while ago)
17:47 autrijus yeah. except the s/r parser isn't available
17:48 autrijus so it can't handle 1+2+3 yet.
17:48 autrijus PerlJam: right, I saw that in hackathon, but at that time it can't handle parens.
17:48 autrijus so maybe 1+2+3 but not 1+(2+3).
17:48 PerlJam heh
17:48 autrijus in any case... to go from that to full production 6.0.0 compiler in 2 months time is. well. er.
17:48 PerlJam well, I'll continue to prod Pm and be optimistic anyway  :)
17:49 autrijus :)
17:49 autrijus anyway, I don't think it's fair to put the Q3 time pressure on pm.
17:49 PerlJam pm built a full fledged pascal compiler with optimizations in about 4 days when he was a grad student.
17:49 autrijus which is part of the reason why I submitted that patch.
17:50 autrijus not using parrot tools :)
17:50 PerlJam granted.
17:50 autrijus and my position was always that that was the critical bottleneck.
17:50 PerlJam If we could just nail down that parrot flux, we'd be good
17:50 autrijus but anyway... that's neither here or there.
17:51 * PerlJam hatches a plan to prod chip now ;-)
17:51 autrijus the thing is that promising the 2004 snapshot is actively harmful :)
17:51 autrijus and Q3 2005 is equally actively harmful.
17:51 PerlJam yeah, I agree with you.  I'm just being ornery
17:51 autrijus :)
17:52 lightstep why not always promise to deliver in the next quarter? like todo_ in pugs
17:53 autrijus todo_ in pugs never carries a time promise...
17:53 PerlJam lightstep: why not promise to deliver it by some future unspecified christmas?
17:53 autrijus right.
17:53 autrijus "Perl 6 will be announced by Christmas."
17:53 autrijus "we reserve the right of deciding which Christ."
17:54 Arathorn lol
17:57 PerlJam I like Simon's likening the "by christmas" release to the second coming  
17:57 rindolf has left "Client Exiting"
17:57 PerlJam (though a tad depressing when you think about it)
17:57 * autrijus suppresses an inappropriate multiorgasmic joke.
17:58 lightstep has quit IRC ("why do i have to run?")
17:58 PerlJam heh
17:58 PerlJam so *that's* what they were talking about wrt MMD!
17:58 PerlJam ;-)
17:59 chip well, I *do* experience a little death whenever I read the Perl 6 documents
18:00 * Arathorn rushes an answer to p6l about wretched namespace hierarchis and sprints to frisbee
18:00 Arathorn is now known as Aragone
18:00 nothingmuch has quit IRC (Success)
18:00 * stevan suppresses equally inappropriate crucifixion jokes
18:01 sili- has quit IRC ("Reconnecting")
18:01 sili- has joined #perl6
18:04 PerlJam has quit IRC ("Lost terminal")
18:32 autrijus sigh, need to wake up early for $work tomorrow too
18:32 mj41 has joined #perl6
18:32 vcv has quit IRC (Read error: 104 (Connection reset by peer))
18:33 * autrijus skips journal for tonight and sleeps
18:33 * autrijus waves &
18:33 putter &
18:36 elmex has joined #perl6
18:37 nothingmuch has joined #perl6
18:38 nothingmuch has quit IRC (Client Quit)
18:38 mj41 Synopsis 2 :  for *$handle :prompt('$ ') { say $_ + 1 } .. is it correct?   or  =$handle ?
18:46 awwaiid has joined #perl6
18:49 putter ?
18:50 stevan putter, iblech: care to take a look at my first draft of Code/Sub objects?
18:52 putter sure...
18:52 svnbot6 r5718, Stevan++ | Perl6::MetaModel - first rough cut at Sub/Code/Block, Signature objects
18:53 stevan perl5/Perl6-MetaModel/t/80_Code.t
18:53 * putter keeps forgetting how breathtakingly _large_ the regexp language has become...
18:53 osfameron has joined #perl6
18:56 cventers_ has quit IRC (Read error: 113 (No route to host))
18:56 cventers_ has joined #perl6
18:58 Nattfodd has quit IRC (No route to host)
19:05 putter stevan: it looks vaguely plausible to me.  I'm sure others could give better feedback - I haven't looked closely at p6 call semantics (except to worridly note that it seems to be one corner where p6 is not following the policy "be powerful enough to subsume everything anyone else is doing").
19:06 putter progress!
19:07 stevan putter: ok
19:07 stevan it does not need to be the "true" version, but just enough to prototype some MMD stuff
19:07 putter k
19:08 stevan so vaguely plausible is good :)
19:08 putter :)
19:09 putter it will be sooooo nice to have objects.
19:12 dada has quit IRC ("zu spaet um makellos zu sein")
19:12 putter mmm, mmd prototype... stevan++
19:14 Darren_Duncan has joined #perl6
19:17 iblech stevan: re. Looks sane :) Questions:
19:18 iblech In the introduction, you speak of $.params and $.returns, but in the actual code there's only one signature
19:19 iblech And you'll need an own &return later (i.e. not Perl 5's one), right?
19:21 iblech And BTW, the code is very readable :)
19:23 meppl has quit IRC (Remote closed the connection)
19:24 stevan iblech: $.returns is :todo :)
19:24 stevan not entirely sure how to procede with that
19:25 stevan my thought is to capture the return value, check it's container_type against wantarray maybe
19:26 stevan iblech: yes, I think in some ways I need to own return
19:26 stevan but I am blurring over those details for now ;)
19:28 kolibrie um, should $canidate be spelled $candidate?
19:29 stevan kolibrie: it is a special case of candidate :)
19:29 kolibrie :)
19:29 kolibrie just nit-picking, like iblech says, your code is very readable
19:30 stevan thanks
19:30 stevan yeah, my spelling is pretty bad, I blame it on Phonics :)
19:31 stevan is http://svn.openfoundry.org/​pugs/src/Pugs/Types/Code.hs relevant?
19:31 stevan or is something else used?
19:36 shantanoo has joined #perl6
19:49 shantanoo is now known as dhoomketu
19:52 cventers_ ?eval class bob { };
19:52 evalbot6 undef
19:52 cventers_ odd
19:53 cventers_ pugs> class bob { };
19:53 cventers_ 'Perl6 User\'s Golfing System, version 6.2.8, July 13, 2005 (r5718)'
19:53 iwlx has joined #perl6
19:53 wilx has quit IRC (Read error: 104 (Connection reset by peer))
20:00 dhoomketu is now known as shantanoo
20:04 vcv has joined #perl6
20:06 jhorwitz has quit IRC ("Chatzilla 0.9.67 [Firefox 1.0.4/20050511]")
20:06 svnbot6 r5719, putter++ | Another cm of progress on the Perl6::Rx grammar.  Quite a way to go.  Failing PGE test submitted to parrot.
20:08 putter &
20:08 putter has quit IRC ("Leaving")
20:19 sapper has quit IRC ("Leaving")
20:23 jdv79 has joined #perl6
20:29 * Aankhen`` goes to sleep.
20:29 Aankhen`` G'night.
20:29 svnbot6 r5720, Stevan++ | Perl6::Sub - more tests and refactoring
20:29 Aankhen`` has quit IRC ("If it wasn't for C, we would be using BASI, PASAL and OBOL! [Time wasted online: 4hrs 41mins 31secs]")
20:29 chip Anybody know if the AES docs are state of the art on colons and MMD?
20:31 shantanoo has left "Leaving"
20:31 osfameron has quit IRC (Read error: 110 (Connection timed out))
20:32 geoffb chip, I don't know, but I find it highly unlikely they are, since there have been discussions here over the past days about both,
20:32 geoffb and I gather that has leaked to p6l (though I don't sub to the lists yet)
20:34 iblech stevan: re. Ok then :) /me looks forward of integrating Perl6-MetaModel into PIL2JS :)
20:35 svnbot6 r5721, iblech++ | * Usual svn properties.
20:35 svnbot6 r5721, iblech++ | * "sub foo { try { return }; say 42 }" should return from &foo and not print
20:35 svnbot6 r5721, iblech++ |   42, added a test for this bug to t/builtins/control_flow/try.t (FWIW, PIL2JS
20:35 svnbot6 r5721, iblech++ |   behaves correctly :)).
20:35 svnbot6 r5721, iblech++ | * Pugs.Compile -- Emit &prefix:<*> for Syn "*" so I can implement *(...) in PIL2JS.
20:35 svnbot6 r5721, iblech++ | * PIL2JS: Fixed &die (but it's actually a bug in Pugs' Parser, will write a
20:36 svnbot6 r5721, iblech++ |   test tomorrow).
20:36 svnbot6 r5721, iblech++ | * PIL2JS: New builtins: &Code::arity, &Array::push, &Array::shift, &try,
20:36 svnbot6 r5721, iblech++ |   &statement_control:<for> (works with n-ary subs)
20:36 svnbot6 r5721, iblech++ | * Updated PIL2JS/README and added PIL2JS to pugs::hack.
20:36 iblech Night all :)
20:37 iblech has quit IRC ("sleep &")
20:57 iblechbot has quit IRC (Connection timed out)
20:57 vcv has quit IRC (Read error: 104 (Connection reset by peer))
21:08 kolibrie has quit IRC ("leaving")
21:20 Amnesiac has joined #perl6
21:22 mj41 has quit IRC (Read error: 104 (Connection reset by peer))
21:36 Nattfodd has joined #perl6
21:43 osfameron has joined #perl6
21:44 osfameron has quit IRC (Read error: 104 (Connection reset by peer))
21:46 osfameron has joined #perl6
21:59 jdv79 has quit IRC ("Leaving")
22:25 Nattfodd has quit IRC ("Leaving")
22:36 vcv has joined #perl6
23:01 frodo72 has joined #perl6
23:01 vcv has quit IRC (Read error: 104 (Connection reset by peer))
23:02 frodo72 is now known as poletti
23:04 poletti has quit IRC ("Client exiting")
23:05 poletti has joined #perl6
23:15 mml_ has joined #perl6
23:15 mml_ has left
23:16 mml_ has joined #perl6
23:17 Nattfodd has joined #perl6
23:38 elmex has quit IRC ("bis morgen")
23:52 larsen has quit IRC ("night")
23:56 mml_ has quit IRC (Read error: 110 (Connection timed out))
23:58 jdv79 has joined #perl6

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

Perl 6 | Reference Documentation | Rakudo