Pugs t-shirts www.cafepress.com/pugscode | Pugs 6.2.9 released | pugscode.org | pugs.kwiki.org | paste: sial.org/pbot/perl6 | www.geeksunite.net
Set by stevan on 15 August 2005.
dudley_ mmm, I do like the union type over super/subclasses here. 00:10
but I have to go quickly, so I'll have to finish later... 00:11
svnbot r6995 | stevan++ | Perl6::MetaModel 2.0 - 00:43
r6995 | stevan++ | * fixed class attribute handling in Perl6::MetaModel
r6995 | stevan++ | * added a test from 1.0 which uses class attributes
02:09 sleepster_ is now known as typester
panicst8 how long has perl6 been around there are guys in th js room telling me it hasn't been around for 2 years and i say it has 04:37
luqui depends on how you define "around" 04:38
it's been in design for more than three years
but the implementation, pugs, only started this february
panicst8 around as in builds of it i know my brother cmilfo who hangs out in here has been monkeying around with it if not actually using it plus perl 6 essentials was put out by oreilly in 2003
so could he have been using it for 2 04:39
luqui oh, parrot!
Khisanth heh large part of the book is probably no longer valid
luqui and P6C in parrot (which was a miserable failure)
panicst8 but it has been around and would a true perl dork ;) my brother hehehe been playing it it if he eats breaths nad shits perl which he does
i do to but not to his extent. 04:40
luqui okay, well, again, depends on how you define "around"
the *current implementation* has not been around for two years. how about that? 04:41
panicst8 ;) done i just hate being told i am wrong. hehe even when i most probably am ;) 04:42
Supaplex wrong! 04:46
buu Seconded.
panicst8 buu at no time did i say current implimentation i said he has been using/working with it for 2 years and tomorrow during the day pop in and chat with him he hangs out in here for some reasaon cmilfo i beleive. 04:52
oh he is in the room now but most probably asleep
Supaplex not if you keep this racket up
luqui Khisanth, do you read p6l? 04:54
panicst8 ;) 04:55
Khisanth nope, I figure I'll wait till p6 is more or less spec out before doing any serious reading on it
luqui ahh. well, I proposed that @a >>+ 3 be equivalent to fmap (+3) @a...
if that makes sense (paskell :-)
Khisanth fmap? 04:56
luqui so folks can implement their own hyper functors
uh, from Haskell
Khisanth, you are from haskell background aren't you? 04:57
Khisanth nods 04:59
luqui so I basically just proposed that hypers are a short form for fmap. "@a >>+ 3" is [email@hidden.address] $x { $x + 3 })" 05:00
aanyway, so, my work with >><< is done
my next goal is to get the rest of the design team to accept it so I can turn junctions into a functor and begone with those horrible things that don't obey any laws 05:01
Khisanth hrm how is fmap different from map? 05:02
Supaplex the f part
Khisanth :)
luqui map works on lists, fmap works on all Functors (so for [a], fmap = map). But trees are also functors, and also all monads 05:03
Supaplex o.O
luqui but that's haskell's way of looking at it. In Perl, I'll say that lists, hashes, and junctions are all functors
so you'll say if (1|2|3) >>== 3 {...} 05:04
Khisanth ... 05:05
as if there weren't enough operators already :) 05:06
luqui hyper operators already exist
except in the current state of perl, they only work on lists
I'm just trying to get them to work on junctions (and other stuff) too
because, as it currently stands, treating junctions as normal values allows you to prove absurdities (like every object is equal to every other object) 05:07
so junctions must not be normal values (we knew that), but I want them to be. This is just a way to keep the functionality and allow junctions to be normal. 05:08
anyway, I should really get to my homework
?eval if 1|2|3 == 2 { say "2 is equal to 1 or 2 or 3" }
evalbot_6995 2 is equal to 1 or 2 or 3 bool::true
luqui &
nothingmuch morning 07:09
NSK hi 07:10
smith100 suc 07:15
dduncan very early morning 07:32
gaal why are pragma and module load declarations bundled together in the syntax, anyway? 08:08
wolverian can where { } blocks have side effects? 08:09
gaal when i say "use strict", i indeed want to load module strict if it wasn't loaded yet, but what i really mean is something like
pragma strict on
where "on" is some default activation
wolverian: yes
nothingmuch well, i guess you could say 08:10
gaal wolverian: you can put whatever you like in them. and you don't HAVE to have any where blocks in a given; you can do given $foo { bare_code(); when $x { smth() } more_bare_code() }
nothingmuch %*PRAGMAS<strict>.enable or something 08:11
strict::import is probably just that
wait, thats %?PRAGMAS since it's lexical
gaal so re: pragmas: i don't like how configuration of modules in general is conflated with importation, and i guess pragmas are a special case
wolverian gaal, I mean where { } as in 'subtype Foo of Bar where { ... }' etc 08:12
gaal ahh, THAT where :) dunno, sorry
wolverian I'm not aware of any other where :)
gaal oh right, i was thinking of when {}.
wolverian ah, right.
nothingmuch gaal: did you see the compilation mish mash?
it was lots of fun
wolverian: yes they can
wolverian: there's no real way to limit them 08:13
gaal re: pragmas again - you can't say %?PRAGMA = x because the assignment has to happen at compile time
wolverian nothingmuch, I thought pugs can do that?
nothingmuch it's probably not a good pragma
gaal so it should get a keyword of its own, i think
nothingmuch ugh
nothingmuch confused himself
gaal nothingmuch, yes; and it made my head hurt
nothingmuch it's probably a good practice to make side-effectful where clauses
gaal as you no doubt intended
nothingmuch but there's no where to tell which side effects are validddd
wolverian hm. 08:14
nothingmuch actually i tried to make complex things slightly simpler
i reduced my headache at least ;-)
gaal heh
wolverian I thought where clauses are strictly for making stricter types, so it'd make sense for perl to be able to run them as many times as it likes
(without, uh, weird side effects..)
nothingmuch wolverian: in my mmd proposal I proposed that where clauses are executed in order of declaration, after all other constraints on the candidate list have been verified, and the first candidate that passes all the where clauses is taken 08:15
gaal nothingmuch: anyway, it deserves more attention than i could afford, so i haven't really attempted to grok it yet.
nothingmuch gaal: basically &infix:<+> is not a builtin
it's a foreign function
gaal i know, i read that part.
nothingmuch ah, so not much more to say 08:16
integral nothingmuch++ 08:43
nothingmuch integral: huh? 08:52
f0rth ` 10:08
nothingmuch dum dee dum 10:13
13:59 slvls is now known as Sylvain-Loiseau
dudley_ stevan: Some sucker started working on Scala's unit testing, so maybe the Scala mm is back on :-) 15:30
eric256 hello 15:50
dudley_ hello 15:54
wolverian where is everyone? :( 16:06
rafl is over here 16:07
stevan is hiding behind the type checker
dudley_: re: Scala MM -> cool,.. Scala has a lot of potential I think 16:10
dudley_ I agree. 16:11
It just needs a little help becoming more practical
stevan yup 16:14
dudley_ but it'll have to wait a little while so I can catch up on $work-- 16:16
stevan dudley_: I need to finish the Role stuff anyway 16:19
and I have a small $work project coming up,.. but it shouldnt interfere too much ;) 16:20
dudley_ Heh, famous last words...
stevan nah, I am pretty sure this will be painless
16:28 typester is now known as typeout 17:07 typeout is now known as typester
nothingmuch is bored 18:36
autrijus: how's work treating you? 18:37
18:38 typester is now known as sleepster
luqui wtf! why is gentoo's scala so screwed? 18:39
nothingmuch hiya luqui 18:40
dudley_ luqui: How do you mean?
Gentoo's got an old version 18:43
luqui it points me to a wrapper script that is empty!
it's okay, I installed manually 18:44
nothingmuch grunt 18:47
ddd is not doing anything with perl 18:48
masak stevan++, iblech++, fglock++, rafl++ # for docs/journal/2005-09-12.kwid, thx 19:18
rafl still wonders if it will be journaled some day.. 19:23
Aankhen`` goes to sleep. 19:44
G'night.
Limbic_Region anyone about?
fglock hi
Limbic_Region I was wondering if anyone/everyone had received an email from an apparent editor interested in finding someone to write a Perl book 19:45
I don't get very much spam at all at that particular address so I was half thinking it was legit
I find it incredibly odd that it would come while I have a proposal into a different publisher 19:46
ok - so either no one is listening or no one cares or no one has received such an email 19:47
fglock I was checking my spam filter for something like that - no, no mail 19:48
Limbic_Region actually - it came with a phone number - I am calling it now
wow - it was legit 19:50
fglock are you going to write two books? 19:51
Limbic_Region I told them that unless the first book is turned down (no contract at this point just in the feeling it out process) or the topic was very different I was interested 19:52
s/was/wasn't/
writing a book is HARD work btw 19:53
All I have done is put together a detailed outline and put my research notes in related piles and it has consumed far more time then I would like to admit
fglock and little money - but it is nice work
Limbic_Region well - I am not interested in the money 19:54
if I were - I wouldn't work on OS projects or devote my timing helping people at the Monastery and IRC 19:55
so what's new in the world of Pugs? 19:56
Limbic_Region has been out of the loop for aforementioned reasons
fglock the perl5 backend is working better than expected 19:57
Limbic_Region it seems from my limited exposure that focus has been on backends and not expanding the p6 feature set - sound accurate? 19:58
nothingmuch sort of
fglock yes
nothingmuch autrijus is slightly swamped recently
Limbic_Region I have seen his HW problems 19:59
very sorry to hear that some of the stuff lost has no up to date backups
that happened to me in the great HD crash of '03
fglock re feature set - putter is working on rules, and stevan is working on the object model 20:01
ods15 hya nothingmuch !
fglock I'm working in Infinite/lazy lists 20:03
Limbic_Region fglock - do you understand how the object model fits into p6 objects?
or does it not currently
fglock the javascript and perl5 backends already use the metamodel 20:04
basically, it means that class is an object
nothingmuch hi ods15 20:05
Limbic_Region hmmm - I guess I am not asking my question very effectively
and my brain is dead so I won't bother trying 20:06
fglock no, maybe I didn't understand 20:07
Limbic_Region my understanding was this (which could very well be wrong) 20:08
before the metamodel, pugs had some of the OO that was outlined in AES12 20:09
it wasn't quite right and it wasn't quite complete but it was a start
nothingmuch wishes someone would comment on his threads
Limbic_Region then the metamodel came along whose original goal was to completely define the OO behavior in p5 20:10
which was to flesh things out and figure out what was really going on 20:11
nothingmuch the goal was to completely define stuff period
perl 5 was just the most accessible platform
Limbic_Region at which point it would be incorporated into Pugs replacing the existing OO
but as far as I know, that last step has been derailed to be incorporated into backends 20:12
not the Pugs core
so my question was about how much, if any, was the pugs core using the metamodel
fglock the "Haskell backend" should incorporate the matamodel 20:13
Limbic_Region nothingmuch - my understanding and your statements don't disagree though I may have not said them as succinctly
exactly fglock 20:14
fglock there is a lack of people hacking Haskell these days
Limbic_Region the "Haskell backend" is currently using a different model for its OO then the other supported backends 20:15
and my question was WRT to that backend - which I consider base/core functionality
fglock the Haskell backend is a bit stalled. most of the developments have been in small fixes to PIL 20:18
Limbic_Region ok - my intermittent exposure hasn't lead me astray 20:19
GeJ fglock: the plan was for stevan to finish the MM so autrijus was implementing it, wasn't it? 20:20
fglock yes, but autrijus has had little available time 20:21
GeJ yes, I read that.
I wish sometimes The Man would have a clone() method implemented :) 20:22
PerlJam It would be good enough if TPF could fund autrijus for a year 20:23
fglock fortunately the MM have been evolving, so when it finally gets implemented, it would be "right" 20:24
GeJ does he have a paypal account or something?
Limbic_Region GeJ yes
I wrote a journal entry about his current state of funding and accepting donations 20:25
let me see if I can find it
GeJ Limbic_Region: oh yes, I remember that
Limbic_Region use.perl.org/~Limbic%20Region/journal/26062
GeJ Limbic_Region: mail only... ok. 20:30
geoffb is finally back with a new system 20:39
fglock welcome back geoffb, be welcome, system 20:42
geoffb :-) 20:44
rafl, ping
rafl geoffb: pong 20:45
geoffb rafl, where is pugs in the upload process? 20:46
I was going to take the opportunity with the new system to try an apt-get
rafl geoffb: ftp-master.debian.org/new.html 20:47
geoffb: If you would like I can provide you some debs built with pbuilder.
geoffb Nah, I was just going to test and then overwrite with a hand-built one anyway. 20:48
thanks though.
Is it just waiting for time to pass, or is something else blocking it?
rafl geoffb: No. It's only waiting for ftp-master approval. 20:49
geoffb nodnod
rafl Well, haskell-src-exts also is. If it's approved I'll also do an hs-plugins upload so we'll get haskell eval support for pugs. 20:50
Maybe I'll also do that upload before, but we'll need to wait for the ftp-masters anyway. 20:52
geoffb: But what you should be able to install right now is everything that pugs needs to build/run/smoke/whatever (except, hs-plugins and parrot) (if I didn't miss something obvious). 20:53
geoffb rafl, nodnod 20:54
right now I'm trying to find my old subversion password so I can be a committer again. :-)
rafl Uhm, I think debian can't help you with that. 20:55
geoffb rafl, heh
sleepster_ hi
geoffb hi sleepster 20:56
sleepster_ moo.
geoffb Oh wow, over 7000 revs. Go team! 20:58
fglock geoffb: rt.openfoundry.org/Foundry/Home/Gue...inder.html - password reminder 21:05
geoffb fglock, thanks for the info. Thankfully, I had not nuked my old machine -- of dubious merit, subversion stores passwords in plaintext
right now I'm just doing a massive sync, and seeing how many of pugs's build-depends I can actually install right now 21:06
rafl Yeeeeeehaaaaw! 1596 050914 Debian Installer ( 21) pugs_6.2.9-1_i386.changes ACCEPTED 22:25
rafl does the "pugs is in Debian"-dance
integral wow! that's fantastic! 22:26
rafl haskell-src-exts as well. Jippi. 22:30
Hm. I have no more packages waiting in the NEW queue now.. I think that should be changed. :-) 22:33
rafl considers to go to sleep after being awake about 34 hours.. 22:43
luqui rafl, probably a good idea 22:46
luqui can't go much more than 24
rafl Well, I still have half a beer to go. :-) 22:47
Grr.. I can't sleep anymore. Help! 23:34
luqui *whack* 23:35
rafl Hm. Something useful to to would be cool. Do you want some of your stuff packaged in Debian, luqui? 23:38
luqui some of my stuff? 23:39
rafl s/stuff/CPAN $&/
luqui huh. sure. 23:40
luqui wonders what he's proud of enough...
rafl Or did you write some other cool software that may be usefull in Debian?
luqui nope, I haven't :-)
rafl How about Perl6::Class? 23:44
svnbot seems to be dead again. :-( 23:49
No? OK, I'll play some blobwars in this case.. 23:53