pugscode.org <<Overview Journal Logs>> | You have safely opened the door to many Perl 6 hackers. | pugs.kwiki.org | smoke: xrl.us/fmw2 | Mac OS X (195/3716) | Win2k r1534 (210/3779)
Set by kbrooks on 4 April 2005.
putter is there a favored nopaste for #perl6 haskell code? i've a three line untodo which I'd like stylistic feedback on... 00:07
stevan perlbot nopaste 00:08
perlbot Paste your code here and #<channel> will be able to view it: sial.org/pbot/<channel>
pasteling "putter" at 66.30.119.55 pasted "new code for Prim.hs" (4 lines, 180B) at sial.org/pbot/8942 00:10
stevan pugscode.org <<Overview Journal Logs>> | You have safely opened the door to many Perl 6 hackers. | pugs.kwiki.org | smoke: xrl.us/fmw2 | Mac OS X r1537 (212/3783) | Win2k r1534 (210/3779) 00:10
putter Caveats: it handles ["a"].join("x") but not ("a").join("x"), so its not a complete join.t untodo. my attempts at () were excessively slurpy. 00:11
feedback of any flavor welcome. 00:13
any flavor at all... 00:16
Schwern Pork ripple 00:18
Holy crap, that's all the code? 00:19
<--- non Haskell programmer
putter is there any interest in having tests consolidated. eg, there are eq tests in operators/cond.t and in operators/eq.t. with the former also doing halfhearted string<->number conversion. any thoughts? 01:07
er, thats test _files_ consolidated. 01:08
Khisanth S10 is rather brief! O_o 01:19
metaperl wow Juerd actually agreed with me for once! 01:25
Khisanth must be a ruse 01:31
stevan metaperl: is it a Full Moon tonight in Holland? 01:32
metaperl crosses himself with garlic
it must be a full moon for Juerd to agree with me 01:33
stevan loads his gun with silver bullets
putter would it be plausible to drop the implementation of additional operators in Prim.hs? they probably need a better home, but I dont immediately see one... 02:06
err, additional _filetest_ operators.
shapr will vCast work for every type? 02:09
Is the precision of log part of the specification?
I have this from log.t, should I just change the expected and check it in? # Expected: 1.6094379124341 \n # Got: 1.6094379124341003 02:11
I search, for a clue. 02:13
putter thinks things are quiet... tooo quiet...
shapr putter: is your nickname from perl golf?
putter ah, that's the question. just _which_ pun is being made. ;) 02:17
shapr All of them? 02:18
putter not telling.
shapr I like mystery.
Anyway, do you know if the precision of log is specified?
putter hmm, exploring... 02:19
shapr It's getting light outside, I should go to sleep. 02:24
btw, what's the more polite approach? should I knock over more builtins or leave them for camelfolk who are learning Haskell? 02:25
putter ah, I'm losing. why not commit. possible outcomes are need to back out, or that it demonstrates need to truncate as comment in file hypothesizes. 02:28
shapr works for me. 02:29
putter re knocking off builtins... given the range of tasks available, perhaps addressing things that novices can't will have a higher collective payoff. 02:30
shapr Yeah, true. 02:31
putter oh, and perhaps leave a comment in log.t about the change, just in case thought did go into the precision. 02:34
shapr How do you create complex numbers in perl? 02:36
Alias_ I'm sure there's a module 02:37
putter Math::Complex? in p6, checking...
Alias_ Math::Complex perhaps?
shapr I just want to know for the unit test. log should work on complex numbers just the same, I think.
putter s09: my complex $n; 02:40
oh, but that doesnt help you with a constructor... checking... 02:41
shapr I'll just check in what I have and go to sleep. 02:45
putter i dont see it in the assorted docs. Math::Complex does 4 - 3*i. perhaps do that, and add a question to... wherever we are accumulating questions for p6l.
sleep is good.
shapr I think someone new to Haskell should be able to implement exp.t and sign.t without too much trouble. 02:47
shapr falls over asleep 02:48
putter sounds good. i think i'll leave some of the filetest operators, after correcting the return type of the current ones.
good night shapr 02:49
does anyone know what the pugs "bootstrap philosophy" is? just enough coverage of p6 to bootstrap a rules engine? full coverage? something in between? if full, how is "i'm an object pretending to be a string" handled? (probably not the right time of day to ask...) 03:21
Alias_ Isn't the Rules engine Haskell? 03:24
1) Load Haskell 2) Parse Perl
:)
putter my understanding is p6 will eventually be parsed by p6 rules. 03:28
putter putters off for the night.
Alias_ sure. But not pugs
AFAIK
putter Alias_: tnx
Alias_ Pugs is P6 in Haskell... P6 is P6 in P6 03:29
There will be a limited bootstrapping parser of some sort though
autrijus pugs will eventually be translated from haskell to p6. 03:31
that's our strategy. see roadmap
Alias_ So Haskell will be the bootstrapping parser, with P6 as the real parser?
autrijus yes. 03:32
putter: the philosophy is that we do full coverage until the day that p6 macros is supported. 03:33
because without them there's little chance to translate pugs to p6
putter the "bootstrap philosophy" question came to mind as i went to fix a filetest operator, and it struck me that strings are being used in situations where an object-vs-its-stringification distinction may need to be drawn. so i was wondering how much dynamic pain pugs is signing up for... 03:35
autrijus oh. those will be rewritten using our OO core.
the OO core needs to be attached soon.
once we fix the remaining significant bugs and deem the next release (I think) 6.2.0. 03:36
again, please see the roadmap.
sdtr443w I've been seeing a lot of examples have something like *@xs in a subroutine prototype. Is that a reference to an array? 03:37
putter for the record, quoteth "6.2: Basic IO and control flow elements; mutable variables; assignment." 03:39
autrijus right. that means we /completed/ them.
then we start hacking in OO
once OO support is considered complete, we call that 6.28.0.
sdtr443w: search.cpan.org/dist/Perl6-Bible/li...urpy_array 03:40
Alias_ This versioning thing is gunna bite you in the ass everntually :)
sdtr443w I guess referencing and dereferencing has changed
autrijus Alias_: nope. there is a finite number of milestones :)
Alias_: after the last one, we'll be called "Perl 6.0.0."
sdtr443w Oh wait I guess I picked up somebody else's dialog first and got confused 03:41
Alias_ autrijus: For all you know pugs could end up becoming the Haskell -> Parrot doohicky permanently
sdtr443w I see it in a scalar context too: *$x
autrijus Alias_: I will not let that happen :) 03:42
of course if I get hit by a bus, oh well.
sdtr443w I have a short bus. Where you at?
autrijus ;)
ninereasons autrijus, I wrote a test that assumes that 'my (@a) = 1 .. 3' is the same as in perl5. was I correct? 03:43
autrijus sdtr443w: search.cpan.org/dist/Perl6-Bible/li...parameters 03:44
03:44 knome is now known as rblackwe
autrijus ninereasons: correct. 03:44
lunch. bbiab! & :)
sdtr443w autrijus: I scrolled to the top of the first link you sent me and I'm just reading it sequentially
putter still puzzled, but too tired to think. 03:45
ingy autrijus: ping 03:46
sdtr443w Hrm I see things in the grammar about explicitly declaring the return value of a sub, but is there a way to explicitly declare the parameters? 03:52
Oh finally... somewhere past a third down it gets into it 03:54
Is there a problem with this subroutine declaration in pugs: sub void_run_strarray(str @test_array is r) { } 04:00
I put a print in there and get nothing; it's giving me the silent treatment
ninereasons I don't think that "is" works yet, sdtr443w 04:01
and I think that 'str' will work if written as 'Str'; 04:02
sdtr443w Yeah it works with caps 04:37
I mean, capitalized 'S'
That's odd -- doesn't that mean I'm casting it as an object? I thought those don't work yet 04:38
gaal morning all 06:37
06:39 castaway_ is now known as castaway
gaal are builtins meant to really be built in? after all in S29 they have Perl signatures 06:42
so theoretically they can be implemented in a core library?
is the ordering in the list at the bottom of Prim.hs important? 06:48
Juerd gaal: Perl 6 will eventually be written in Perl 6 06:58
gaal: Fortunately, Perl 6 is fast and thus so will Perl 6 06:59
gaal yay! i added sign! :) 07:08
ingy hola 08:46
gaal aloh
ingy hi gaal 08:47
gaal hello. do you know where we stand in terms of embedding c?
inlining from perl, that is. 08:48
ingy well we can almost inline haskell
but there are outstanding dynaloaderish issues
I assume that once we resolve those issues, inlining C will be rather trivial 08:49
gaal neat!
in Prim's initSyms, what does ~ mean as a prefix to a type? eg.. \\n Scalar left && (Bool, ~Bool)\ 08:51
eval lazily?
Khisanth coerce into a bool? 08:53
gaal so why doesn't the first arg have it too?
it looks like it's there only in the short-circuits 08:54
any lambdafolk around? 09:23
theorbtwo Looking at Prim.hs line 859, gaal? 09:25
gaal more or less :) i'm doing exp now, which has an optional arg
theorbtwo I think you're right -- it looks to me like it's building a thunk, but I'm certianly not a lambdaite.
Just list it as defaulting to... whatever? 09:26
(Presumably e.)
gaal those aren't very common - so i was peeking at the implementation of rindex, which is simple enough
theorbtwo (Num, ?Num=2.7...)
gaal but i don't understand the syntax of 'let' there, which doesn't have an 'in' clause
no, i don't want to hardcode e
theorbtwo Or, IIRC, just ?Num, and check if it's undef.
gaal i want to use the best haskell function -- 09:27
either exp or ** (or maybe even ^ and ^^)
the p5 code i am thinking of is something like return defined $base ? $base ** $exponent : builtin_exp($exponent); 09:29
because haskell provides me with a presumably efficient and exact exp builtin.
so back to rindex:
about line 630 - it has this code to do a let based on the results of a vCast 09:30
which looks like what i want, but mimicing the syntax doesn't work
ah, i think i see what my problem was, i forgot the -> do 09:33
autrijus greetings 09:41
gaal heya!
autrijus the "in" is omitted in monads
theorbtwo Allo, autrijus.
gaal i changed it to an if, makes more sense anyway :)
but now i get errors saying Couldn't match `Eval Val' against `Exp'. is there an explanation about pugs types somewhere i can stare at? 09:42
autrijus yeah 09:43
pasteling "gaal" at 192.115.25.249 pasted "my naive attempt" (14 lines, 401B) at sial.org/pbot/8951 09:44
gaal doesn't work when i vCast, either. 09:45
autrijus a sec 09:46
theorbtwo I think you just need a "do" after that =, but could well be wrong. 09:47
gaal no, it's a,er, the find of function that doesn't do .. return 09:48
theorbtwo pure function.
But it's getting called impurely.
gaal is in the "programming in the dark" state
rgs beware of the dark side 09:49
gaal or, more like, programming in the twilight state.
jabbot pugs - 1544 - build fix for Eclipse 09:52
pugs - 1543 - sign builtin
pugs - 1542 - implemented log and log10.
pugs - 1541 - more hangman.p6 work; improved gameplay/
pugs - 1540 - some tests for return statement
pugs - 1539 - array .join added
pugs - 1538 - my(@a)=1..3 with parens should return th
theorbtwo Hm, I think I know what's going on, but not how to explain it.
jabbot pugs - 1537 - Correct the test count
pugs - 1536 - Add tests for list assignment bug, not d
pugs - 1535 - add code to demonstrate arg list flatten
gaal theorb2, i think my problem is one of types, not purity. 09:53
theorbtwo Except that Eval is a monad. 09:54
It's also a type, of course.
nothingmuch ginger.ucdavis.edu/paloverde/paloverde.html 09:59
autrijus back 10:03
<- just installed Mac Mini
using usb keypad no less
nothingmuch yay! 10:04
start a smoke server
i'll check in my "alternative" script
autrijus that sounds like an idea
please do
gaal haha. reminds me of the time i had to use a keyboard where the middle row didn't work
it was in DOS
nothingmuch no wait, i have an OSX one
gaal so ALT-numkeypad gave out chars by their ascii values
nothingmuch damnit, i need to get that real smoke server stuff done
gaal and thankfully, although many numbers were inaccessible -- anything with 4,5, or 6 in it -- 10:05
ALT accepted modulos!
castaway wonders if pasta is working again
gaal so if i needed eg 65, i could try 65+256
nothingmuch castaway: err, yes 10:06
what port do you want SEE on?
gaal which is 321
nothingmuch gaal++ # great heroism
gaal and that was how i typed "A" :)
theorbtwo Wow, gaal... I wouldn't have thought of that. 10:07
castaway pick one :)
gaal you get to think of all sorts of things when you're desperate :)
theorbtwo I would have thought of replacing the keyboard first... 10:08
castaway but then we have about 3 extra keyboards lying around
theorbtwo More, I think.
theorbtwo loves his IBM type M.
jabbot pugs - 1545 - * `\_ ->` is now spelled `const` 10:12
castaway does actual work, having spend all morning see hacking (ahem ,)
nothingmuch castaway: one minute, i need to carry up a monitor 10:14
castaway no hurry :) 10:15
nothingmuch see://woobling.org:6944 should work eventually, 10:16
theorbtwo WOOHOO!
It doesn't work yet, but I've gotten past the first hurdle.
gaal what does see:// mean?
castaway hurdle?
theorbtwo Thing that one has to jump over to continue the race. 10:17
castaway SubEthaEdit
castaway grins
theorbtwo emacswiki.org/wiki/SubEthaEdit
wolverian are there SEE-compatible, free software linux editors?
castaway adds a 'cw' in that url somewhere
gaal doesn't own a mac
theorbtwo Not yet, wolverian; castaway and I are working on one.
nothingmuch wolverian: they're working on it
castaway not yet, wolverian, but we're working on it
heh :)
wolverian I love you guys. :) 10:18
castaway blushes
wolverian what are you coding it in, and what are you using as components?
theorbtwo Thanks, wolverian.
We're writing in emacs lisp, and creating many of our own components.
castaway lets theorbtwo answer and does some work instead
wolverian oh, I hate you guys. emacs. :(
castaway well you can use our protocol deciphering to make your own 10:19
nothingmuch castaway: are there howl bindings for emacs?
castaway who/what?
wolverian mmhm. I wonder if simlpy using vim's perl bindings would be adequate.
autrijus gaal: what does exp do again?
wolverian simply, rather.
autrijus is it not the perl5 exp? 10:20
it should be a op1, no?
nothingmuch rendezvous
i think it should have two variants
one being op1, using haskell's exp
and vcasting it appropriately
and the other being an op2 alias to **
autrijus exp(2, 3) ? 10:21
nothingmuch uhuh
theorbtwo rendezvous forms only a very minor part of SEE's protocol stack; we're ignoring that bit.
Where is that specified, BTW?
autrijus is it defined in the bible?
theorbtwo wolverian: If you want to write your own SEE-compatable client, you'll need to start with bencode/bdecode and BEEP (RFC 3080)> 10:22
autrijus and isn't it serialized ObjC objects? 10:23
i.e. you have little chance to munge them without ObjC
theorbtwo No, autrijus, it isn't.
nothingmuch theorbtwo: it should be a nobrainer with some rendezvous browser
wolverian theorbtwo: thanks.
castaway nm, I doubt there is.. but one doesnt need that bit anyway, much
nothingmuch autrijus: in that case, won't GNUStep be able to provide good enough remote objects? 10:24
autrijus nothingmuch: no idea. I'm just respewing rumours
wolverian theorbtwo: what is bencode?
theorbtwo The encoding scheme used in torrent files; the spec is in www.bittorrent.com/protocol.html (search for "bencoding"). 10:25
wolverian right. there's a module for that.
thanks again.
theorbtwo Note that it uses a small extension to that specification: Strings can start with : or '.'.
Right; Convert::Bencode.
wolverian there's also BT::Bencode; I don't know how good that is or if it's even on CPAN.
theorbtwo I'm also working with a server in perl, but it's meant largely as a development tool. 10:26
wolverian (just googling around while reading the RFC.)
theorbtwo: are you going to implement BEEP in it, though?
nothingmuch theorbtwo: do you have some wikipage where these notes are scribbled?
wolverian: there's already BEEPCORE for perl
castaway emacswiki.org search for subetha
gaal sorry, was making coffee 10:27
wolverian nothingmuch: oh. so I wouldn't have to write any protocol cruft myself, just the application layer?
theorbtwo I know, nothingmuch, but it doesn't do what I want.
Quite possibly.
wolverian hmm. there is no POE BEEP component. :)
gaal 2arg exp is not **, it's ** with the order of args switched :)
castaway iek, POE :)
wolverian Net::BEEP::Lite looks nice.
gaal it's defined in S29 10:28
theorbtwo There isn't a callback for the creation of a channel in Net::BEEP::Lite, though.
wolverian yeah, the POD indicates it's _very_ lite
that is, no asynchronous channels
castaway wonders what its point is then
wolverian I don't see perl on beepcore.org :/ 10:30
autrijus gaal: got it.
wolverian (besides Net::BEEP::Lite, and it's probably not enough of an implementation.)
castaway DIY all the way, baby 10:31
(neither of them existed for elisp ,)
wolverian argh. :)
castaway which is why we're only just getting to the actual SEE stuff 10:32
nothingmuch autrijus.explain($_) for «vCast castV»
castaway iek
wolverian the python library seems to be more complete. augh. 10:33
castaway must remember to set utf-8 permanently
autrijus castV casts anything into a Val
vCast casts a Val into anything.
however vCast can't be used for a mutable variable
nothingmuch autrijus.elaborate
autrijus ok. the args for op are Val types 10:34
you can't directly perform operation on them
so you need to use vCast to read them 10:35
nothingmuch so in gaal's question
gaal autrijus, okay, so please look at the following:
nothingmuch op1 "exp" = \x -> exp(vCast x) appearantly doesn't work
pasteling "gaal" at 192.115.25.249 pasted "vCast?" (22 lines, 1K) at sial.org/pbot/8952
wolverian do you have any basic implementation notes somewhere on BEEP? if I decide to do this myself, I'm going to need some guidance. :)
nothingmuch :t exp says it wants to eat a Float 10:36
i assume you can make a VRat into a Floating
castaway BEEP has RFCs, and is fairly easy
rfc3080
theorbtwo Hm, now I segfault.
I'm not sure if I should define that as progress or not.
Also, 3081.
castaway has ignored 3081 so far
wolverian castaway: yeah, I'm reading that. I was just wondering if there are any specific issues not addressed in that, or anything else you'd like to share. :)
castaway I cant think of any offhand, its fairly simple 10:37
nothingmuch libetha should be C IMHO, if there are volunteers
castaway I did it from there, plus a tcpdump of a SEE session
nothingmuch that way everyone can eat it
castaway which you're welcome to have
C--
wolverian I'd very much like to have the dump.
nothingmuch C++ for being portable, is all
wolverian (that sounded wrong..)
nothingmuch ofcourse, when all our editors will be compiled for parrot... ;-) 10:38
castaway wishes other things also produced useful shared libs
theorbtwo wonders if we should start an SVN repo, and dump everything we've got in it.
castaway that is, if theorbtwo is in agreement
theorbtwo Hm? 10:39
In agreement on what?
castaway giving wolverian the ethereal dumps
theorbtwo Of course.
castaway is SVN safe?
theorbtwo Link to them from the wiki.
Safe?
wolverian you have a wiki? :)
autrijus gaal: you are missing a return.
gaal: you can't return things w/o a return.
theorbtwo We're abusing the emacswiki.
gaal when do i need to return things as opposed to just have a function evaluate to them? 10:40
theorbtwo wonders if we should start decorating the source code with L<...> markers.
castaway desert-island.dynodns.net/misc/dump.etherealdump (for now)
theorbtwo gzip that, love.
castaway safe, as in secure, if we run it locally 10:41
wolverian thanks!
autrijus gaal: when you are in a monad.
castaway done, add a .gz
autrijus > :t return
return :: (Monad m) => a -> m a
gaal theorbtwo was right :)
wolverian it's gzipped by the server automatically.
autrijus return takes a value and put it into a monad.
wolverian (if the client supports that, anyway. or so wget seems to claim.)
autrijus so I fixed your code and checked it in. 10:42
gaal so i need that in a do block as well, right?
autrijus thanks :)
theorbtwo ...and when I'll get the hubris to yell at ingy again, to tell him that Perl6::Bible should have a function that gives me the contents of the document I ask for (as raw POD).
autrijus well sure, because a do block is just a way to write monadic code.
gaal autrijus, my code? hah
autrijus gugod_: can we have jabbot announce freepan commits?
wolverian hmm, BEEP can run out of reply codes. :)
castaway hmm, if we want to get/chat about multiple implementations, we should probably add a section on the kwiki on perlmonk.org
wolverian I so suck at reading RFCs. 10:43
theorbtwo wolverian, it's supposed to wrap at that point, which is why the spec is written to say that you can reuse msgnos that have had both a MSG and a RPY on them.
gaal thanks, autrijus :)
wolverian theorbtwo: ah, thanks.
castaway right, but it has quite a few before it gets that far :) 10:44
gaal map (++) [ autrijus, nothingmuch, theorbtwo ]
theorbtwo Hm, any purticular reason that SHA1__0_0_1.hs gives extern__ = [['s', 'h', 'a', '1']], rather then extern__ = ["sha1"] ? 10:45
autrijus theorbtwo: because that's how TH prints out code.
"sha1" is a Data.PackedString converted
theorbtwo Ah, OK.
autrijus 's' etc is primitive to haskell
castaway wolverian: anyway, any questions, just ask.. I may even know the answers 10:46
theorbtwo reminds castaway that she's supposed to be working while she's at work. 10:47
wolverian castaway: thanks. I'll do that. :) 10:48
castaway oops :)
castaway fetches more tea instead 10:49
jabbot pugs - 1546 - * exp 10:52
theorbtwo Hm, how do I do a <- and explicitly give the type of the var I'm putting into?
extern :: [String] <- loadFunction mod "extern__" doesn't quite do it.
autrijus extern <- loadFunction mod "extern__" :: [String]
extern <- (loadFunction mod "extern__" :: [String]) 10:53
extern <- (loadFunction mod "extern__" :: Eval [String])
something like that
theorbtwo Thanks.
autrijus np :)
castaway sighs.. elisp/see hacking is so much more interesting than SQL project updating.. 10:54
somebody motivate me ,) 10:55
Alias_ pulls out the whip
Do some work and if you're good I'll use it! 10:56
theorbtwo You just told your boss that you don't have time; if he thinks that you don't have time because you're playing at work, he probably won't look kindly upon it.
Hey, watch it, there.
castaway hacking isnt playing!
wolverian wow, reading an RFC really makes you tired. 11:02
castaway and thats actually quite a sane one, IMO 11:03
autrijus the UTF9 one?
wolverian BEEP. 3080.
autrijus ah.
wolverian where is the wiki on perlmonk.org? 11:04
castaway theorbtwo.perlmonk.org/kwiki 11:05
wolverian thanks.
castaway as yet no SEE pages
autrijus hrm, sign() should prolly return undef for undef
castaway I could stuff my local "notes" ones in there
Alias_ So any news on a truly common SEE'like application?
wolverian castaway: I'd be grateful. 11:06
castaway Alias_: been asleep the last hour ? :)
Alias_ yes
castaway anything to keep me from working :)
Alias_ castaway: Give me the 20c update 11:07
wolverian I wonder how hard it would be to wrap beepcore-c 11:08
it doesn't sound like fun.
castaway Alias_: we're writing an elisp implementation .. 2/3s done maybe.
Alias_ elisp...
So in other words a SEE client for emacs? 11:09
castaway correct
wolverian Alias_: I'm currently planning my own implementation of the protocol stack. I'm not sure yet about the application layer; I'd like to tie it into vim somehow.
Alias_ This is why I hate it when editors have their own language
castaway wolverian: in no particular order (and not linked anywhere): BEEPDecode, BencodeBdecode, SubEthaEditProtocol, SubEthaEmacs (wiki pages=
wolverian castaway: thanks! castaway++ 11:10
castaway the orginal (local) wiki is actually reachable, but seems to be terribly slow for most people
wolverian damn, I need to get home to read these.
castaway hmm?
theorbtwo He's trying to actually work at work. 11:11
wolverian I don't have a comfortable hacking environment currently.
castaway Alias_: well, it was that or perl.. I dont touch C/C++ in my free time
Alias_ So with vim and emacs caught up, and a mere mortal Win32 Ultraedit user like me find a program that the plugin will work with?
s/and/what is/
castaway hmm, you'd need a DLL
presumably
jabbot pugs - 1547 - * according to s29 draft from rod, sign 11:12
Alias_ Is there a Win32 emacs or vim spawn that works "normally" and isn't 90% obscure key-combination driven? :)
theorbtwo emacs has menus.
castaway win32 emacs should run it
Alias_ ok
Then all is good good good :)
castaway will test at some point
what was I.. ? oh yeah, work..
theorbtwo But not while she's at work.
wolverian the rendezvous part of SEE isn't connected to BEEP in any way, is it? 11:14
castaway no, just replaces the typing of URIs
wolverian right.
castaway bah.. 11:15
castaway wonders if SQL can bitshift
wolverian that sounds nasty. 11:16
shapr I usually call that filesystem corruption.
castaway hmm? 11:17
shapr Sorry, standard smartass comment from me. When all my SQL is bitshifted, I call that filesystem corruption.
shapr boings cheerfully 11:19
I think I'm going to give up on Rob Nagler.
theorbtwo Rob Nagler?
shapr extremeperl
theorbtwo Doesn't ring any bells. 11:20
Oh, a book?
shapr groups.yahoo.com/group/extremeperl/messages
jabbot pugs - 1548 - * according to S29 atan2 is now spelled 11:22
Alias_ Will be interesting to see the combination of SEE and Test::Inline 11:23
Multiple people on one file writing the docs, tests and code simultaneously
castaway ick :) 11:26
umm, iek :)
wolverian shapr: how are you giving up on him? 11:28
theorbtwo nopaste? 11:29
jabbot, nopaste?
jabbot theorbtwo: nopaste is Use sial.org/pbot
theorbtwo jabbot, thank you.
jabbot theorbtwo: You're not really talking about me -- are you?
shapr wolverian: I'm tired of talking to him. It seems to me he's not looking for a better way, he's looking to stay inside the way he's got at all costs. 11:30
autrijus stevan: I'm hacking your hangman
sub cls returns Void {
system( ($?OS eq any(<MSWin32 mingw cygwin>)) ?? 'cls' :: 'clear');
}
that should work for now
mm any<MSWin32 mingw cygwin> 11:31
wolverian shapr: oh. :/
autrijus discovers a nice idiom for p6 :) 11:32
shapr As much as I enjoy using Haskell, Epigram, Joy, or whatever, I'm sure there will be better solutions available in fifty years. I'd like to steal parts of that better future and use them today.
theorbtwo Shapr: We hope to take less then 50 years to get perl6 usable. 11:33
shapr haha
I'm just saying that everything we use and know has beauty and quality, but is necessarily imperfect. So it's best to look for lots of grains of goodness and keep them all in mind. 11:34
pjcj perl6 is for the next 20 years, no? But five of them have already gone ;-)
theorbtwo shapr++ 11:35
You are indeed quite eloquent for what I am assuming is not your native language.
shapr Heh, I'm from Alabama. Redneck is my native language! 11:36
theorbtwo Oh, damn.
That makes that reference somewhat less funny.
shapr My swedish isn't nearly as eloquent. 11:37
theorbtwo sial.org's pastebot seems to be down; what's a good alternative?
theorbtwo sighs... I'm getting closer, I think. 11:40
autrijus theorbtwo++ 11:41
theorbtwo: nopaste.snit.ch ?
jabbot pugs - 1549 - * more p6ification for hangman 11:42
stevan autrijus++ # nice hangman/clear screen fix :) 11:47
theorbtwo nopaste.snit.ch:8001/2256
autrijus any<foo bar baz> 11:48
is such a nice idiom.
stevan I like that too
theorbtwo Tis indeed.
I think it'd like it better with a ~~, though.
stevan is very much in love with junctions :)
theorbtwo $?OS ~~ any<MSWin32 mingw cygwin> 11:49
stevan theorbtwo: with the mutiple operators you have greater control
autrijus theorbtwo: you have this committer right thing
:)
stevan $?OS ne any<MSWin32 mingw cygwin>; $?OS eq any<MSWin32 mingw cygwin>; etc...
stevan has to unplug and drive to his weekly work meeting 11:50
bye all
theorbtwo True, but I'm not so sure if the code I have is actually better then the code that is already there. 11:51
autrijus oh, that works?
the bit you posted
theorbtwo Well, it works up to the (`mapM`) bit. That segfaults. 11:52
Alias_ autrijus: You're involved in various Module::(Build|Install) stuff right?
autrijus can you print out extern?
Alias_: not M::B
theorbtwo How?
autrijus M::I is my work yes.
theorbtwo: print extern
Alias_ autrijus: Is there any "standard" way to install a data file?
autrijus Alias_: no. I think we've had this discussion before :)
many moons ago
Alias_ I _am_ having a bit of deja-vu...
But them I'm working on Data::Package again, so that might be why 11:53
theorbtwo I don't see any output.
autrijus theorbtwo: hrm, try using trace 11:54
trace (show extern) return () 11:55
but, maybe you can commit that bit
so I can svk pull and go to dinner where there's no network :)
Alias_ autrijus: OK, so in that case, I'd like to _propose_ a standard way :)
autrijus: If you think it will integrate with M:I
autrijus I'd be happy to integ it w/ M::I. 11:56
theorbtwo Still no output, but ci'd anyway. 11:57
autrijus k.
bbiab. :) &
theorbtwo You'll have to change the hardcoded path, of course.
autrijus sure.
jabbot pugs - 1552 - r6230@speights: samv | 2005-04-05 19:5 12:02
pugs - 1551 - r6229@speights: samv | 2005-04-05 19:2
pugs - 1550 - inline haskell support -- I think this i
shapr hops cheerfully 12:13
theorbtwo You seem very cheerful, shapr.
shapr I am! I had food! I have sugary drinks! w00! 12:14
and I committed code to an open source project in the last 24 hours! w00h00!
theorbtwo grins. 12:15
Alias_ Question: Is a "lambda" basically the technical term for an anonymous subroutine? 12:26
shapr How much detail do you want? 12:27
wolverian I'd say the technical name is 'anonymous subroutine'. :)
Alias_ shapr: The level 2 version
shapr \x -> x + 1 is a nameless Haskell function that adds 1 to a number. 12:29
Technically it's the lambda from lambda calculus. 12:30
theorbtwo en.wikipedia.org/wiki/Lambda_calculus
shapr In short, the turing machine and the lambda calculus are two approaches to computational theory. They can both do anything a general purpose computer can do, but one or the other is easier to reason about for varying purposes. 12:31
Alias_ damn the informal description on wikipedia is dense 12:32
"In lambda calculus, every expression stands for a function with a single argument; the argument of the function is in turn a function with a single argument, and the value of the function is another function with a single argument. A function is anonymously defined by a lambda expression which expresses the function's action on its argument."
shapr The first example I saw of that dense description was these crazy things called 'Church numerals' (Church numbers?). 12:33
elmex shapr: it just popped into my mind, aren't u a regular #haskell idler?
shapr elmex: yeah, I show up there sometimes.
shapr grins 12:34
Alias_ so yes, a "lambda" is a function, just without a name
... and apparently only one argument
shapr Technically, Haskell functions only take one argument, even when they take several.
theorbtwo All functions in haskell only have one argument.
They just fake it really well.
elmex shapr: i only remember you, back from the days when i idled there and tried to learn haskell ;) now i barely remember how to code in haskell ;/ 12:35
Alias_ hmm... and it explains a lot about why they teach Haskell at university
shapr Alias_: it does?
gaal Gordon Freeman was a lisp hacker.
theorbtwo OTOH, all functions in perl5 only take one argument, and that argument is always named @_.
Alias_ Because is Haskell is a lambda-focused language, mathematicians probably love it
shapr I dunno. I've never had any CS classes, or any math classes. 12:36
theorbtwo Mathematicians proveably love it, you mean. ;)
Alias_ and thus in the "early" days of computer when it was mathematician heavy, haskell got used a lot
shapr theorbtwo: haha!
theorbtwo Alias, Haskell isn't that old.
shapr Alias_: actually, lambda calculus is older than computers, and Haskell is only fifteen years old.
castaway returns, out of breath
Alias_ And when those people got old and wanted to teach and introductory language they already knew, Haskell got pushed onto another generation of uni students
Anyways...
It spans comp and math
So I see why the unis like it 12:37
pjcj but there was a plethora of other functional languages
castaway falls on theorbtwo.
shapr I'm just looking for better ways of doing things, that's how I found Haskell.
Alias_ But Haskell was the coolest one of the most recent generation?
shapr Depends on what you mean by cool.
Alias_ The one all the cool kids were using 12:38
being the people that dictate the courses :)
pjcj haskell was designed to replace all the others, IIRC.
shapr Haskell was designed because the most popular lazy functional language was a commercial closed source product.
pjcj or at least the others in its class
shapr And a bunch of guys wanted something they could hack on. 12:39
castaway hmm, where'd he go?
shapr castaway: I'm sorry, you squished him.
Remember, don't jump on people from six stories up!
pjcj My college used Hope, but no one else did.
shapr I always feel like a total geek when I want to say "Oh, Hope had some really cool features!" 12:40
castaway hmpf, and I just climbed 4 floors, how did you know?
shapr laughs
castaway: are you really German?
castaway No
shapr whew.
castaway is a brit!
Odin- Poor thing. 12:41
theorbtwo kisses castaway. 12:42
(I was reading CNN.) 12:43
castaway Aha!
shapr I'm amazed that Haskell is so close to lambda calculus. The more I find powerful patterns in software, the more I think they depend on being powerful patterns from mathematics.
Odin- theorbtwo shares a host with castaway, so I feel I must ask: In real life, or?
:p 12:44
castaway correct
Odin- curious.
castaway snuggles theorbtwo. (mines!)
theorbtwo snuggles back.
shapr I'm tryin to code here! geez!
castaway sowwy shapr 12:45
,)
shapr grins
castaway attempts more SQL, while waiting for the next interruption
shapr select * from irc where username >= 5; 12:46
theorbtwo shapr, any idea about the segfault autrijus and I were discussing a bit ago?
Limbic_Region autrijus - www.perlmonks.org/index.pl?node_id=444853 # Perl6 and Ruby on Rails 12:47
shapr I used to work on this 65,000 line visual basic trust management application. It had SQL queries that were several pages long, with no line breaks. Debugging those was fun.
castaway ick
theorbtwo Limbic, autrijus is at dinner.
castaway BEEP! username is not numeric
shapr doh
Limbic_Region theorbtwo - he usually catches back up on the log, but I /msg'd him anyway 12:50
wolverian Limbic_Region: I like how he hasn't apparently read any apos or even synopses. (re: perlmonks) 12:52
I guess perlmonks isn't the right place to say RTFM.
theorbtwo We generally prefer that you say something else in addition to RTFM. 12:53
...like how to find the FM, and what FM to R.
wolverian I try to do that on #perlhelp and such, yes. 12:54
Limbic_Region wolverian - I haven't read the article yet. I make a daily scan of use.perl journal entries and posts at the Monastery and mention anything parrot related to chip and anything p6/pugs related to autrijus 12:55
shapr theorbtwo: have you considered the hightech option of throwing putStrLn in there? 12:56
castaway ooh, hightec
theorbtwo Hm, no, good idea, shapr.
wolverian Limbic_Region: oh, right. thanks, too, it's always good to read such things, even if they're by IGNORAMUSES. 12:57
I think I need to sleep. :/
shapr I get tired of reading stuff written by Ignoramuses. I'd rather read stuff by Oleg Kiselyov. Then I feel like an Ignoramuses who can improve!
I feel that way when reading of autrijus' code too. Like "wow, I can really learn something here." 12:59
castaway you can learn a fair bit attempting to fix other peoples crap code, too :)
Limbic_Region agrees with castaway 13:00
castaway But I agree that one gets tired of it ,)
shapr Alias_: btw, church numerals and all that lambda calculus stuff is tremendously fun. 13:01
Alias_ shapr: So you say :)
shapr c2.com/cgi/wiki?ChurchNumeral
Alias_ shapr: I don't program for fun. Just to be ! bored
shapr: Fortunately for CPAN, I get bored a lot
shapr I can intellectually realize that some people do not program for fun. But I do not think I'll ever really get it. 13:02
Alias_ shapr: I've done practically everything fun I can think of with programming 13:03
shapr My current email signature is "Programming is the Magic Executable Fridge Poetry, It is machines made of thought, fueled by ideas."
Alias_ shapr: And now I'm just out to make it not suck
shapr Alias_: if you don't know about lambda calculus, there's a whole new world to be found.
Alias_ meh... but what's the point, really 13:04
shapr What's the point of programming at all?
Alias_ I've probably discovered it myself independantly
mugwump laughs
Alias_ mugwump: It's a bad habit :)
shapr Alias_: c'est votre choix
castaway he speaks french too!
shapr Peut etre. Parlez vous français Mademoiselle? 13:05
Alias_ yes, but unfortunately my spelling issues are far worse in French, so I don't dare type it
castaway Non ,)
shapr quelle dommage.
castaway wot? :)
rgs "quel"
shapr rgs: man, I did that again! ouch!
rgs: thanks!
castaway got a B about 15 years ago (GCSE)
rgs :) 13:06
Alias_ I wonder...
Take a machine learning algorithm for recognising languages, add babelfish, and mix with an IRC client
shapr Last time I took french classes was 1992.
Alias_ The mis-tranlations would be hilarious... but it might actually be functional
Implementation is left as an exercise for the reader 13:07
castaway those things are bad enough when you give them a specific language
shapr But I spent a few hours in some bars in Lille last christmas speaking only French. That was fun. No one really understood me, but I sure felt cool.
Alias_ Masque: About the only knife I had I really like is my swiss army knife
shapr Alias_: oh, try a leatherman! 13:08
Alias_ Masque: milspec... from Switzerland
shapr Alias_: EMISCHAN
castaway looks confused.
Alias_ shapr: Nah... I have a "real" swiss army knife, and that is enough for me
ack, channel
shapr I have a stick. It is enough for me. It is turing complete. 13:09
shapr beats himself with his stick. 13:10
xerox You need brainpower for fullify the turing completeness of your stick :-)
castaway stopped carrying a knife when people started complaining about them on various transport mediums
xerox ^__^
castaway (that was abt 15 years ago too)
Alias_ castaway: Now you just pack a sword in your carry-on? 13:11
castaway heh, no, theorbtwo has the sword ,)
Alias_ I like the idea of getting a sword
Even one of the knock off dodgy cheap ones 13:12
But instead of mounting it or some shit, just leave it lying around in the loungeroom
castaway heh
Alias_ "Oh, that's for roach killing"
theorbtwo Mine is indeed lying around the living-room somewhere.
Well, the small one; the big one is still in the US.
BTW, check them, not carry-on. 13:13
shapr Yeah, I have to check my leatherman too :-(
Alias_ "Crap a spider, $guest can you grab the sword? I think it fell down the back of the lounge"
castaway assumes Alias_ isnt quite that daft
elmex ?
castaway lost a pair of pliers that way.. (coming back that is, going they didnt notice/complain... sigh)
Alias_ castaway: Well, if I were to do it, it would be the biggest craziest looking over the top sword I can find 13:14
castaway yup, but putting it in the carry-on is asking for trouble
Alias_ 5-6 feet long, spikey bits, dragon's head on the pommel
The whole 9 yards
castaway :)
shapr that's not 5-6 feet
castaway 9 yards is 27 feet
castaway high-5s shapr
shapr Isn't that like the whole two yards?
Alias_ heh
shapr snickers 13:15
castaway dammit folks, this isnt SQL
shapr and this isn't $work
theorbtwo threatens castaway with the shortsword.
castaway iek! 13:16
jabbot pugs - 1553 - * fix parsefail bug on false unaries 13:22
Alias_ ok, I've read the description of lambda calculus on wikipedia and it makes practically no sense 13:23
Which probably means Haskell is not for me :)
autrijus the old parsebug fixed
Alias_ sorry lambas :(
autrijus next; last; return;
now all just work.
theorbtwo Alias, the description there seems very dense for programmers. 13:24
shapr autrijus: can you suggest a task that is one step up from implementing builtins? do you think the want.{Scalar,...} fits in there?
theorbtwo autrijus, any dice on my problem?
Alias_ I'm liking "lamba calculus as birds" a bit better though 13:25
autrijus theorbtwo: I had not looked; sorry
Alias_ uq.net.au/~zzdkeena/Lambda/index.htm
autrijus shapr: want.Scalar requires OO subsystem
Alias_ of course, I'm still not seeing the point
shapr oh
autrijus and we currently don't have one
maybe implement kill()? :D
shapr Alias_: whoa, that rocks.
autrijus that's still a builtin, though
shapr maybe a builtin that requires more than just filling in initSyms? 13:26
autrijus not sure... you grok GADTs?
shapr vaguely
autrijus kill() requires a bit more thought
shapr not grok :-)
ok, I'll check that out.
autrijus or maybe help theorbtwo with DynamicLoader :)
shapr ok
theorbtwo: is that paste all that you need to get the segfault? 13:27
Alias_ I don't like the use of . in lamba calc, because it makes it really bloody hard to read 13:28
castaway ieks, COBOL
autrijus shapr: you need just to check out the trunk of pugs
Alias_ I'm too used to . as method or concat
autrijus install that, then checkout tpe.freepan.org/repos/ingy/SHA1/ then install that (perl Makefile.PL ; make install)
then you can witness the segfault with
make ghci ; loadHaskell 13:29
theorbtwo Alias_: That's probably why Haskell spells it ->
(And \ instead of lambda.)
shapr I get an error building with the latest svn, anyone else? 13:33
shapr tries make clean && make
autrijus is it an impossible error? 13:34
shapr yup
fromJust: Nothing
autrijus I get that too
GHC bug. a clean fixes it
shapr ok
Alias_: this lambda calculus is birds is a great find, thanks! 13:36
Alias_ shapr: I'm still confused as hell, but I think I at least have an idea of the point of it, from the math side of things
in that sense the bird thing is handy 13:37
castaway bird thing?
Alias_ uq.net.au/~zzdkeena/Lambda/index.htm
"To shred a Mockingbird"
Lambda calculus animated and described in bird names
shapr snickers wildly, "In the case of the Idiot song being heard by an Idiot bird we can perform another beta reduction and obtain a result which is just the Idiot bird/song." 13:38
xerox "In the case of the Idiot song being heard by an Idiot bird we can perform another beta reduction and obtain a result which 13:39
Whoops.
castaway cute
mj t/pugsbugs/die_and_end.t 1 256 1 2 200.00% 1 :-) 13:41
jabbot pugs - 1554 - * there is no nullary eval form. 13:42
Alias_ castaway: Of course, reading any haskell file in pugs still makes no sense... :)
castaway It doesnt to me either, Alias_
shapr I can help!
Alias_ heh
castaway you can translate to perl? 13:43
shapr oh, probably not.
castaway elis?
elisp?
Alias_ I'm guess it's going to be easier to put a lamba on a camel than to put a camel on a lambda
shapr maybe to elisp.
castaway hmm 13:44
Alias_ shapr: Or maybe, "where the fuck do you start"
xerox I do like the effort going on in putting a camel on a lambda, as in pugs :-)
shapr I wrote a Baskin-Robbins introduction to Haskell for those who value methylphenidate - www.haskell.org/hawiki/HaskellDemo
Alias_ Haskell seems like the sort of language that really needs a video introduction
shapr Hey, that's a good idea! 13:45
theorbtwo vnc2swf ?
Alias_ Hi I'm Troy McClure.
shapr And I'm Shae Erisson
And today we'll introduce you to the new LambdaCamel TV series!
Alias_ You may have seem me in such educational videos as "Sex. Know you know how, DON'T DO IT!" and "Forks and Toaster. A bad thing?" 13:46
vnc2swf?
So THAT's how he did it
Alias_ saw a swf video demo some some program the other day
theorbtwo You may remember me from such educational films as Lambda does Latin Dancing and Beta Reduction -- the SEXY Way.
shapr And you may have seen me in such educational films as "Ivory Towers, more fun than tilting at windmills?" and "Will the university of Lyon lie down with the lambda?" 13:47
Alias_ So tell me, do you have to define "add" in every single program?
shapr no
Alias_ I also worry about anything where "true" is defined as a set of functions :)
shapr Oh, you mean the temperature equality typeclass? 13:48
Alias_ shapr: Are you taking suggestions for that demo page?
shapr I am.
I encourage them even.
Maybe there's no outright begging, but it is implied. 13:49
Alias_ somenumber = 1 (good)
somestring = "somestring" (good)
shapr you think monads shouldn't go right after that?
Alias_ add1 :: Integer -> Integer and add1 x = x + 1
slowing down a bit... but ok
theorbtwo The first line tells you the signature of add1, the second tells you the defintion. 13:50
Alias_ yeah, I managed to get the idea after 10 seconds
castaway just wait til you get to multiple arguments :)
Alias_ My next question, before you got any more complex, was "and how do you use it"
How does one "do" anything
theorbtwo Well, the sort answer is "just like perl, except for the parens". 13:51
Alias_ I need a 1 + 1 = 2 example, using your add1 function
castaway good idea, Alias_
Alias_ immediately after you defined ad1
xerox Missing function application?
shapr Alias_: excellent point.
Alias_ here is how to define the most basic function, and how you use it
shapr edits
Alias_ here is how you define a lambda, and how you use it
etc etc
slowly slowly... with lots of parallels to normal functional programming 13:52
castaway "a lambda"
?
jabbot pugs - 1557 - r6235@speights: samv | 2005-04-05 21:4
pugs - 1556 - Implemented sum.
pugs - 1555 - Add nick for shapr
Alias_ anonymous function ?
castaway ah
Alias_ Am I misusing the term?
-- anonymous function (we give this one a name anyway)
add2 = (\x -> x + 2)
castaway I dont know.. I know it from somewhere else :)
autrijus anonymous function is correct. 13:53
castaway the word "lambda" seems to get thrown about a lot and used for various things
shapr nick?
castaway nickname?
autrijus notes that 13:54
data People = Person Name Age
castaway digs out the LPC lambda stuff
autrijus reads worse than
data People where Person :: Name -> Age -> People
shapr truly
theorbtwo isn't sure what either of them are supposed to mean. 13:55
autrijus so I guess I'll convert more pugs code to use that form :D
shapr That's an interesting point. GADTs are more readable than boring datatypes?
autrijus theorbtwo: it defaults a new type "People"
to construct them, use "Person" function
13:55 Aankh|Clone is now known as Aankhen``
autrijus which takes two arguments of "Name" and "Age" type 13:55
and returns something that belongs to "People" type
s/defaults/defines/
mj WinXP ... 209/3795 subtests failed, 94.49% okay. After s/todo_//g ... 1087/3795 subtests failed, 71.36% okay ... $Pugs ~~ '71% percent Perl 6' ? 13:56
theorbtwo Something like that, mj.
shapr I really like that idea. 13:57
Alias_: what do you think about the add1 example I used?
castaway wait, the 2nd reads easier than the first, autrijus ?
Alias_ shapr: /me looks 13:58
castaway notes parallels between LPC closures and LISP (hmm, what fun) 13:59
theorbtwo Unsurprising -- doesn't the L in LPC stand for LISP? 14:00
shapr autrijus: GADTs for readability is a great idea, thanks!
castaway No
"Lars"
Lars Penji C (sp?)
shapr yay LPMuds! 14:01
castaway confusingly, its docs call "anonymous functions written in a weird mini language" - "closures" :)
precisely
shapr Closures have some other picky details that make them interesting. The word comes from the fact that you 'close over' the values.
Alias_ shapr: aaaah... so fortytwo = add1 41 is basically like a normaly function call but without the braces that add so much annoying readability
shapr: That is a Big Deal(TM)
theorbtwo Alias: Yep. 14:02
castaway yeah, but I'm not actually sure that LPCs do
Alias_ makes understanding at least somewhat possible, maybe
shapr Alias_: well, braces add readability for some folk, I assume.
castaway Alias_++
theorbtwo I think Haskell would be much more readability if they added some parens and commas.
castaway too
theorbtwo You can use parens, but they go around expressions, not argument lists.
Alias_ shapr: Well, for camels, the fact that add1 41 is add1(41) is gunna matter a LOT
theorbtwo (And around tupples.)
Alias_ shapr: Since we are used to full parens on things 14:03
castaway closure max;
max = lambda( ({ 'x, 'y }),
({ #'? ,({ #'>, 'x, 'y }), 'x, 'y }) );
return funcall(max,7,3);
(even more silly parentheses :)
shapr That is clearly closing over the values.
shapr eats a value for lunch.
theorbtwo Alias, recall that parens are optional in perl. However, they are /disallowed/ in haskell.
castaway which vals? I seem to be blind 14:04
shapr the 'x and 'y values.
well, I think so.
castaway umm, they're passed in, not external?
shapr ok, they are..
but it could be closing over values!
Alias_ theorbtwo: True, but most people don't use implicit braces any more, since most coding now is OO, and OO MUST have braces
shapr pouts
castaway I've never seen any that do
Alias_ There is remarkebly little use of function signatures to allow implicit braces
castaway tickles shapr.
shapr unpouts 14:05
castaway: can you get partial application by filling in one part of a lambda?
castaway Alias_: who are these "most people" ? :)
theorbtwo That's because p5 prototypes aren't very good at that.
Alias_ castaway: Most camels
castaway Alias_: I wouldnt agree..
shapr umm, I dont think so
Alias_ castaway: As a percentage of the code you've written in the last year, how much uses implicit braces 14:06
castaway: Hell... how much is non-OO
theorbtwo I take advantage of implicit braces and non-OO quite a bit, Alias.
I think OO is rather overrused by many.
Alias_ I've got maybe 2000 lines of functional stuff, and 150,000 lines of OO stuff
castaway umm, you can do it with unbound closures tho 14:07
shapr hm, neat.
castaway Alias_: ALL of minie does.. I just know/have seen a fair few people who dont, per default
theorbtwo Alias: You almost certainly mean imperative, not OO.
Er, imperative, not functional.
Alias_ sorry, procedural?
castaway And lots of it is non OO
Alias_ imperative...
theorbtwo shapr, you can always use closures to create currying.
shapr yay 14:08
Is a functor the opposite of an imperator?
theorbtwo I don't know what either of those are.
Alias_ (\x = x + 2) is equivalent to sub { return $_[0] + 2 } ?
theorbtwo Yes. 14:09
The return is optional, BTW.
pjcj perl -wle 'sub add1 { $_[0] + 1 } $forty_two = add1 41; print $forty_two' # I like missing parens
Alias_ right
just trying to be clearer
theorbtwo Oh, right... /me forgot which language was your native, and which you were uncertian about. 14:10
Alias_ shapr: And you use pooled idenfitiers?
theorbtwo (\x = x + 2) is exactly equvlent to sub { $_[0] + 2 }, down to the closing.
shapr Sorry, my identifiers can't swim. No really, what does that mean? 14:11
castaway grins
Alias_ is C< add1 x = x + 1 > equivalent to C< add1 = (\x -> x + 1) >
castaway lambda (x) (+ x 2) :)
(lambda (x) (+ x 2)) (of course) 14:12
shapr Alias_: yes it is.
Alias_ so (\x = x + 2) differs in that while Perl would create a ref to the func, Haskell's IS the func?
wilx Hmm, OT question.
Does SVK allow background syncing of reps while checking out something on foreground? 14:13
shapr I don't quite understand the difference between a ref and the function itself.
Alias_ well... for example...
autrijus shapr: easy
Alias_ $reference = sub { $_[0] + 1 };
autrijus in perl, everything is IORef.
shapr oh!
Alias_ You have to escape the ref to use it
shapr that's wild! 14:14
autrijus and all programs are in the IO monad.
(with builtin ContT and ReaderT)
that's all :)
Alias_ autrijus: I think you two need to create a pair of primers
shapr appoints autrijus dual-citizenship lambdacamel
Or at least fluent translator
Alias_ autrijus: Lamba for Camels, and Camel for Lambdas
autrijus Alias_: I have some URLs in PA01... 14:15
Alias_ If the two communities are going to be mixing for a while, it may well be needed a lot
autrijus more suggestions welcome
theorbtwo shapr: Autrijus already is a fluent translator.
shapr It's sort of braintwisty for me to figure out how you would do stuff with everything wrapped in an IORef.
Alias_ theordtwo: Beware the English-only speaker who also knows only PHP
autrijus easy; you "deref" it.
shapr theorbtwo: yes, but he has now added Haskell <-> Perl to his lists.
Alias_ &$ref 14:16
The & derefs
shapr yeah, I dig it.
&IORef
autrijus and by deref, it actually means "cast the thing in IORef to one of <Code Scalar Hash Array>."
because you see, perl5 is a strongly typed language.
with no subtypes.
theorbtwo It's sort of braintwisty for me to figure out how you do stuff when you keep having to enter and escape monads.
shapr Dare I ask what those types signify?
theorbtwo It just only has a very short list of types, and not the ones you'd think.
Alias_ Code is an anonymous function
shapr I'll gues that, Code is a sub, Array is an array, and... 14:17
theorbtwo shapr, latin plural class, believe it or not.
autrijus Hash is a Data.Map
Alias_ Scalar is a string/number/reference
shapr oh wow
Limbic_Region autrijus - you get the link I /msg'd you?
theorbtwo Scalar is string/number/reference/undef (a bit catch-all).
shapr Geez, Larry Wall should hook up with Aarne Ranat.
um "Aarne Ranta"
autrijus Scalar is Either Ref Value
shapr huh
autrijus and Value would be the union of String, Num, Int 14:18
pjcj Also Regexp, IO
autrijus s/Num/Float/
Alias_ autrijus: We'll ignore Ref and Glob for now? :)
shapr Why nested refs?
autrijus shapr: because somebody like it.
shapr fair enough
autrijus shapr: because that's how perl does multidim arrays.
Alias_ shapr: It's partly the C heritage
autrijus perl arrays can't contain arrays
only scalars
theorbtwo Shapr: Arrays and hashes only hold scalars.
autrijus so those are references to arrays
Alias_ @inner = ( 1, 2, 3 );
@outer = ( \@inner ) 14:19
shapr Hm, interesting.
Alias_ bad example
autrijus shapr: e.g. perl allows for
[ [1,2], 3 ]
Alias_ autrijus: Or specifically, how perl DOESN'T to multidim arrays
autrijus which is ill typed in haskell.
the trick is to make it
Alias_ Perl has no proper multidimentional array, just arrays containing pointers to other arrays 14:20
autrijus [ (Ref Array [1,2]) (Int 3) ]
Alias_ So the length of each dimention is not enforced
theorbtwo s/pointers/references/
autrijus (Int) and (Ref a) being both constructors to Scalar.
castaway looks slightly lost.
autrijus so that has type [Scalar]
theorbtwo Indeed, the number of dimensions is not enforced.
autrijus and hence well typed.
theorbtwo castaway: A haskell array is limited to containing any number of elements, all of which have the same type. 14:21
shapr That kind of makes sense, an array of pointers.
castaway hmm, so are LISP ones, thats why it also has lists :)
autrijus the weirdo thing is that Scalar is a union of pointer or pure values.
castaway No pointers!
Alias_ shapr: Thus, to be camel-friendly, you may need to point out that add2 = (\x -> x + 2) is exactly equivalent to add2 x = x + 2
shapr Right, I'm beginning to see that. 14:22
Alias_ Big difference between "reference to" and "is" for us
The sigil is what defines the IS
castaway does haskell have anything that does allow a dynamic number of elements?
autrijus castaway: a list? :)
theorbtwo Sadly, perl 6 is trying to hide the difference, without getting rid of it.
castaway oh sorry, misread.. any number of els, same type 14:23
what about multi types then?
shapr You can do multi-types with HList, but it's not for the faint of heart.
Alias_ shapr: In fact, I'd do naked (\x -> x + 2) first, then show assigning it, THEN show it's the same
autrijus castaway: that's called a HList.
shapr Alias_: show me?
autrijus but usually in haskell, you either fix the number of els
or the type.
Alias_ -- anonymous function
autrijus if you fix the els, it's called a n-tuple
Alias_ (\x -> x + 2) 14:24
autrijus written as (a,b,c)
theorbtwo Tupples allow elements of any type, but you have to declare how many elements, and of what types, the purticular tupple will have.
autrijus right.
Alias_ -- but you can assign these anyway
autrijus if you want to keep both variadic
you want HLists.
castaway yeah, I gave up reading the tutorial somewhere there
Alias_ add2 = ...
-- assigning in this way is exactly the same as the basic function def
add2 x = x + 2 14:25
autrijus oh also, Alias_, "=" in haskell reads "binds to"
castaway hands around tea and biscuits.
autrijus it's roughly the same as :=
in perl6
chip autrijus: If you'll indulge me, what in your Rule modules should I look for as the key difference from stock Parsec? Interface? New concept(s)?
Alias_ autrijus: I'm sort of getting that
It's a type-less pooled identifier space
autrijus chip: uhm, from the outside, there's no difference really 14:26
theorbtwo Oooh, tea and biscuts!
autrijus chip: I added the "List" and "Chain" assocs
chip Alias_: but with packages! foo is Main.foo
autrijus chip: and there's some code cleanups
shapr Alias_: cool, I'll add that. thanks for the contribution!
autrijus uh it's not type-less
chip autrijus: I haven't read the code yet, thus the "obvious" question; I'm asking for what to look _for_
autrijus each identifier has one type
Alias_ sorry, non-type-split?
autrijus chip: ok. assume they are identical.
"sigilless" is the term. 14:27
Alias_ in the sense that perl has $foo and @foo and etc
theorbtwo D'oh!
autrijus Alias_: do you know that in perl6 you can say
theorbtwo is foo <- bar in a monad lazy?
That is, will bar never be executed if I never use foo?
autrijus my %foo is Array; %foo[0] = 1;
theorbtwo: that is mostly the case.
there are exceptions.
theorbtwo Blast.
Alias_ autrijus: Yeah, I've been getting some funny smells from certain p6l posts, so it doesn't surprise me too much 14:28
autrijus hence my suggestion in using trace()
theorbtwo That explains it; I'm segfaulting earlier then I thought.
autrijus which forces strictness.
chip autrijus: Maybe you could pass it as an argument to a function that declares its parameter with "!" for forced evaluation. I forget what that's called.
autrijus strictness annotation.
chip ah
autrijus the most usual form is $!
Alias_ autrijus: I'm presuming some form of use strict will stop that from being legal?
autrijus which evals the rhs and hands it to lhs for apply, just like $
Alias_: I have no idea. I think that may be intentional. 14:29
intentionally legal anyway.
shapr Alias_: better?
Alias_ looks
shapr It's exactly what you wrote, so...
Alias_ whoa... hang on
you added something
shapr I did?
Alias_ add2' <--? WTF you didn't mention that before
shapr huh?
oh
Alias_ the quote
theorbtwo ' is a perfectly valid character in identifiers in haskell. 14:30
(I also find this confusing.)
xerox autrijus: I didn't understand, is it $! different from $ ?
shapr right, and it's a commonly used math thingy, a and a' means a and a prime
theorbtwo xerox: A bare $ is meaningless in perl.
Alias_ theorbtwo: kinda
xerox theorbtwo: whoops, perl's $! equals Haskell's $ ?
Alias_ theorbtwo: The sigil doesn't strictly have to touch the idenfier 14:31
$ foo is the same as $foo
xerox I tought you were talking about Haskell's $!.
autrijus I'm talking about haskell's $!
Prelude> :t ($)
($) :: (a -> b) -> a -> b
Prelude> :t ($!)
($!) :: (a -> b) -> a -> b
see, same type
shapr Alias_: if you have any other suggestions, I'd like to hear 'em.
xerox autrijus: yep, different semantics? 14:32
Alias_ shapr: So you added ' to make it a different name, but using ' was a convention of familiarity to math'heads?
autrijus the only difference being that ($!) evaluates the right hand side strictly.
castaway shapr: Is there an infinite arg function? eg: (defun addall (&rest args) .. ?
autrijus while ($) does not evaluate right hand side until needed.
Alias_ shapr: In that case, you could put the add2 x = x + 2 in the comment
shapr Alias_: I changed to add3 x = x + 3
Alias_ better to make it truly identical
autrijus castaway: yes there is :)
Alias_ just put it in the comment
shapr ok
castaway care to add that, while you're there?
xerox autrijus: I see, thank you.
autrijus castaway: see Text.Printf
it needs tricks. 14:33
i.e. not "elementary" haskell.
castaway is trying to improve the demo too :)
Alias_ oh!
shapr castaway: you can also do sum [1,2,3,4,5]
Alias_ just noticed this page is a wiki
castaway Alias_: heh, didnt notice?
shapr Yeah, I'm maintainer of the HaskellWiki too.
It's funny how admin/maintainer so often means "hi-tek janitor" 14:34
Alias_ Can I make minor changes?
shapr sure
Alias_ just to presentation really
shapr it's a wiki
PerlJam hands shapr a broom
Get to work!
;-)
shapr upgrades to broom 2.0
castaway as yet doesnt even remember what [..] means in haskell
PerlJam castaway: just like perl IIRC 14:35
shapr castaway: in what context?
theorbtwo Creates a list.
More or less like [] in perl.
castaway your sum example
shapr ok, now we have the big guns.
castaway hugs lambdabot
shapr @plugs [1..5] 14:36
The plugs command runs a full Haskell subprocess (with IO limitations, of course) and gives the results.
theorbtwo Odd, shapr, lambdabot just gave the answer to that to me in a /msg.
castaway PerlJam: that would mean it creates a reference to an anon-array :)
Alias_ shapr: Damn, it's getting denser :)
shapr theorbtwo: it's actually, a notice
theorbtwo Oh. GAIM doesn't handle those well, then. 14:37
castaway gets it in the channel, in purple
shapr theorbtwo: The IRC RFC specifies that automated processes should only send via NOTICE because otherwise those same processes might get into a loop.
list via PRIVMSG, send via NOTICE
theorbtwo Makes some degree of sense.
castaway so it wont read its own output?
@plus sum [1,2,3,4,5] 14:38
shapr Right, it might get into a loop.
theorbtwo Most bots I've seen ignore that, though.
castaway @plugs sum [1,2,3,4,5]
ooh, I can do haskell, me :)
shapr yay!
castaway @plugs sum [1..2]
theorbtwo Is there a way to use lambdabot to do the equiv of :t from ghci? 14:39
castaway pats lambdabot on the head.
xerox @type foldr
Alias_ shapr: You've got a bit of a circular description there, re partial application
theorbtwo Ah, nifty.
shapr Alias_: how so?
ninereasons @plugs product [1,2,3,4]
cool
xerox @plugs intersperse 'Z' "zzzz" 14:40
Alias_ You describe the multi-arg is invoking partials, which I haven't seen, then shows partial using the multi :)
xerox loves intersperse.
Alias_ s/is/by/
autrijus notes that lambdabot is much safer than Safe.pm :)
shapr Alias_: It's difficult to explain one without the other. Can you think of something better?
castaway what a useful function, xerox ,)
xerox castaway: it is! :-) 14:41
Alias_ shapr: I'll see what I can do
chip autrijus: sorry to bother you again, but I can't find any definition for ($). Where to look?
shapr I guess I can just start with multi-arg functions.
autrijus Alias_: do you have pugs installed? :)
Alias_ shapr: But I find this idea of partial evaluation built into the language to be quite amazing
autrijus: I have a checkout... but no
shapr Alias_: yeah, I agree. It r0xx0rs.
Alias_: And it's totally based on lambda calculus.
Alias_ shapr: Since I've been debating how to do partial evaluation in Perl 14:42
autrijus chip: sure. haskell.org/onlinereport/standard-prelude.html
shapr lambda calculus has a bunch of cool tricks like that.
chip ah, of course
autrijus f $ x = f x
Alias_ shapr: Mainly be using PPI to "read" perl and manipulate the source manually, then eval it :)
shapr: Which would be absolutely insanely evil
s/be/by/
autrijus Alias_: there is search.cpan.org/~dconway/Perl6-Curr...urrying.pm
Alias_ You'll have to escuse my little spelling/grammarisms. I type quickly and have got a little mild ADD/dyslexia 14:43
autrijus Alias_: also if you have pugs installed, you can try it out easily :)
pugs> my $add = sub ($x, $y) { $x + $y }
theorbtwo From the top of my head: sub partial {my ($coderef, @args)=@_; return sub {$coderef->(@args, @_) }}
autrijus pugs> my $add1 = $add.assuming(1)
pugs> $add1(2)
3
that's partial application for you. 14:44
theorbtwo This is what I said earlier about building currying off of closures.
shapr Alias_: It's okay, I have heavy ADHD. I can dig it.
Speaking of which, it's time for more DRUGS!
autrijus and the perl6 ways is better because you can partially apply $z first for a function that takes $x $y $z
Alias_ Well it's no wonder Haskell makes for such nice optimising compilers then
wolverian autrijus: does &add.assuming(1) work? (I don't have a recent pugs due to the new GHC requirement.) 14:45
autrijus to do that in haskell using point free style is... an exercise in madness.
shapr Alias_: just wait till you see monads. They make compilers soooo much easier.
autrijus wolverian: uhm, what's your platform that prevents you having GHC 6.4?
wolverian: yes it works
Alias_ You can take things like if ( $os eq 'Win32' ) { and remove it from the functions for the current platform
wolverian autrijus: it doesn't prevent, I'm just lazy and haven't compiled it myself yet
autrijus Gruber: welcome aboard!
theorbtwo wonders who Gruber is.
castaway .. and where 14:46
Gruber bows
shapr Anton Berezin?
shapr knots
theorbtwo Ah, hello, Gruber.
castaway wonders how it knows which param is being assumed
autrijus is proud to have tobez, the BSDPAN czar, on board :)
castaway: always the first. but you can also do
$add.assuming( :y(3) )
shapr waits for flip
castaway ah, sneaky 14:47
Alias_ shapr: I think the problem with invoking partials in the multi-variable comment is that you insert $difficult topic in the middle of a description of another $difficult concept
autrijus castaway: it's absolutely crazily nice.
Alias_ shapr: My head was fully occupied with Integer -> Integer -> Integer and you threw in a wrench
autrijus nothingmuch++ # implemented it in 3 days with _no_ prior haskell knowledge
castaway Yeah, that'll get rid of most of the stuff I'd keep global :9
wow
assuming($dbh) - yay 14:48
autrijus and largely without my help
Alias_ Haskell will rock for doing platform-adaptive coding
theorbtwo Note that it just replaces global variables with global functions.
shapr Alias_: what about moving the partial application explanation down under the multi-arg function, to just before the first demonstration of partial application?
autrijus theorbtwo: strangely, we find global functions easier to reason about :)
Alias_ shapr: Yeah, split it in half or something
castaway you cant have everything.. (anyway whats wrong with global funcs, are we going java-like now?)
autrijus theorbtwo: partly because you can't assign into a function.
wolverian hmm. aren't statement modifiers (postfix 'if' and such) macros in perl6? but can they affect the parsing of the RHS? hmm. 14:49
Alias_ shapr: If I read the first three lines of the comment 2 or 3 times I "get" it
castaway assuming("sql is already written") -> false, damn
Alias_ shapr: It was the last two that screwed me up mid-trying-to-understand
shapr Alias_: better? 14:50
autrijus wolverian: yes. now, try to get macros specced :-(
wolverian: I can fake/improvise lots of missing synopses
but macro I cannot
wolverian autrijus: eugh. they need the whole parse tree interface.
autrijus because there's nothing equiv in perl5.
wolverian: we have that.
shapr y0 metaperl
wolverian in perl6?
autrijus oh, you mean we bind Exp as a perl6 structure? 14:51
Alias_ shapr: No... I'll show you
autrijus sure, I can do that if you want
wolverian that's what I assume larry wants to do
it's also the only way I can see this will work
shapr Alias_: you want the type sig part up with add, and the two lines of partial app down with partial app itself?
wolverian but I don't know anything really.
autrijus sure, but that's the semantics
I also need the syntax.
Alias_ shapr: Editing it now...
wolverian yes. that's what I can't provide. :)
autrijus yeah :) 14:52
jabbot pugs - 1558 - This test reads its own source; this be
theorbtwo Hm, what syntax do you need that's not in the subs Apoc?
ninereasons autrijus, I still get a seg fault in pugs-i, when I say '?' or ':r' -- well known?
wolverian hmm, isn't the parse tree just a match object as returned by rules?
autrijus theorbtwo: how "is parsed" works; also I need a definition for if/elsif/else. 14:53
wolverian I would implement if/elsif/else as macros as well
but that needs the former.
autrijus wolverian: sure. try implementing it, assuming that the Pugs AST in AST.hs (Exp) is the object :) 14:54
Alias_ shapr: Done
wolverian autrijus: hmm. :)
autrijus I tried and failed to come up with anythiing
but I didn't give it much effort
note to test/example writers: if you want to read from the source file itself, do not use hardcoded pathname 14:55
my $fh = open $*PROGRAM_NAME;
# that's the way to go
thanks to Gruber for catching it
wolverian I really don't have a clue how to do this
autrijus yeah. but it's the thing that prevents pugs from fully bootstrapping.
fortunately we still has lots of things to do before that. 14:56
wolverian we just need the design team to call in with the macro stuff
autrijus checks for design team presence here
chip? obra? :D
theorbtwo Hm, L<A06/" is parsed( <rule> )"> looks to have enough information to get started, I think...
Gruber autrijus: ah, even better! 14:57
theorbtwo pokes chip on #parrot.
autrijus theorbtwo: you want to help? :)
the [Update:] that begins with
"The default parse rule probably turns out to be dependent "
theorbtwo OTOH, the bit spec'd there isn't enough to create non-prefix macros. 14:58
autrijus is vague enough that I can't get nontrivial macros to work, let alone the entire Syn category.
Syn here being "special forms" that Pugs currently use for macros.
theorbtwo spre, you mean?
autrijus no, Syn 14:59
see Eval.hs
Syn "if"
Syn "given"
etc.
Alias_ ew... type stuff bogs me down a bit
autrijus Alias_: quick, when you see type, think class :)
and when you see class, think roles
wolverian we need the match object interface, I guess 15:00
Alias_ What amazes me the most is that you can make ANYTHING defined in these sorts of terms run on a processor, let alone run fast
wolverian if we want to be able to modify what has been already parsed
autrijus wolverian: write tests :)
wolverian (postfix macros need this, I think)
autrijus: I don't know what the interface should look like.
theorbtwo We need that anyway, as the sub half of the macro needs to get passed the results of the is parsed half of the macro. 15:01
autrijus read up and improvise? post to p6l?
Alias_ shapr: I think the concept "type" needs to be introduced earlier
wolverian good point, theorbtwo
Alias_ shapr: Give me a sec...
I'll edit
wolverian autrijus: very well.
(post to p6l)
jabbot pugs - 1559 - * do not hard code filename 15:02
wolverian do we need anything else besides the match object? 15:03
reading S05. 15:04
theorbtwo Perhaps the is parsed half should start where the keywork was seen, as specced outside of the square brackets, and $_ should be the match object of the thus-far bit?
theorbtwo checks the S06.
castaway attempts to persuade gentoo to emerge ghc 6.4, and fails - any ideas?
theorbtwo Fails how? 15:05
castaway it insists on giving me 6.2.2 even tho I set the ~x86 keyword 15:06
gaal by the way:
find t -name \*.t | xargs wc -l -> 14174 total
find src -name \*.hs | xargs wc -l -> 10321 total
autrijus nice.
shapr Poo, $work calls. Nay, it screams. 15:09
castaway Oh.. it has a ghc 6.4 but not a ghc-bin 6.4 ebuild (whatever the difference may be, the first depends upon the second for something)
castaway syncs and crosses her fingers
shapr takes the fun out of work with a defun
PerlJam shapr: that's always been a problem with LISP-like languages ;) 15:10
wolverian I just can't figure out what the parse tree would look like
Alias_ autrijus: type Name = String 15:11
just equivalent, or more like transparent subclass?
shapr equivalent
Alias_ shapr: Check out the page now
You needed to introduce types a bit earlier, when I was more receptive 15:12
castaway no gentoo users in here?
jabbot pugs - 1561 - Added tests for pi(), sin(), cos(), tan(
pugs - 1560 - Implemented tan. (Well, I had to add two
Alias_ I completely skipped over that initial comment about types, because I saw the shiny C< something = 1 >
shapr Alias_: oh, I like that.
I'd say what we *could* have written. 15:13
Alias_ Very important to introduce ideas one at a time and fully explore each :)
shapr yeah, I agree.
For the most part, you don't need type signatures. 15:14
Alias_ which was the bit of what you wrote I kept
castaway blinks cos the partial stuff on shapr's page now makes sense :)
Alias_ But it read as "don't need type signature... huh? don't need them for what?"
Since I hadn't even seen something = 1 yet
shapr oh I see 15:15
wolverian hmm, what primitives can a 'if' macro use? :)
metaperl hi shapr... on my way to the store and then work... cya in 45mins
shapr ok
I'll be working in a few minutes, cya tomorra
castaway later shapr 15:16
shapr I really like the way autrijus unified all the other type and function understanding with GADTs in the Person type
I'll add that in a moment of work frustration.
autrijus :)
Alias_ shapr: You skip from C< data Temp = Cold | Hot > to Show too quickly
autrijus GADTs is da way of da future
shapr Alias_: I appreciate your help on HaskellDemo, and hope you find more to clarify :-)
autrijus: truly, but I had not yet noticed.
Alias_ shapr: Happy to help as I go 15:17
shapr right, I'll be back later
theorbtwo Hm, from L<A6/Macros>, we've got more then I thought...
iblech castaway: ghc 6.4 is masked in /usr/portage/profiles/package.mask
castaway hmm, I cant unmask it in /etc/portage/package.keywords? 15:18
(cant check now, emerge sync is running)
Alias_ autrijus: Damn this is amazing dense stuff... cool, but oh so nastily dense 15:19
even the basic demo 15:20
iblech castaway: Dunno about portage.keywords. ghc-bin is mostly precompiled (took 25min to install), ghc is not (took 2h40min) 15:21
castaway package.keywords
ninereasons ghc-bin must surely be the bootstrap binary?
iblech right, typo
castaway presumably
ninereasons I wonder why the package would require ghc-bin if ghc is already installed? 15:22
castaway It isnt
ninereasons ah -
chip yo!
autrijus Alias_: it just require you to newfs your brain a bit 15:23
theorbtwo Yo, chip.
autrijus try find some more unused partition in it -- I hear the fs is 90% unallocated
iblech ninereasons: ghc requires virtual/ghc, and both ghc and ghc-bin provide virtual/ghc. Now you first need ghc-bin, and then ghc can get compiled. Afterwards, you can unmerge ghc-bin
castaway ah, sneaky :)
Alias_ autrijus: I've yet to see anything particularly new... it's just packed into with a high degree of Huffman optimisation
theorbtwo You guys need to design macros.
autrijus Alias_: wait till you see monads :) 15:24
chip I wish to understand why PIR should have macros. That's what HLLs are for.
castaway drums fingers..
theorbtwo I was using "you guys" in a rather large sense.
Alias_ autirjus: As I said earlier, I have a bad habit of independantly discovering things, only to find out later they have names :)
autrijus Alias_: MJD said to me that in his Higher Order Perl, the more advanced things are, the more penalty perl5 gives
theorbtwo IE the p6 cabal.
castaway Alias_: yeah, sucks dont it :)
autrijus Alias_: and he could not do monads in p5 reasonably anymore
so monads are not in HOP.
(not to mention arrows or functors) 15:25
Alias_ autrijus: Hell, partial evaluation itself is a bitch in Perl
autrijus Alias_: true, but it's just verbose
theorbtwo autrijus: Doesn't that mean that the hufmanization is right? More advanced things should be harder.
castaway not really, you just wrap a subref around
autrijus and linearly verbose at that
theorbtwo: sadly, the advanced things are also common things.
Alias_ autrijus: I wanted to use PPI to find things like $0 eq 'Win32' and convert them into a constant so that the perl tree-pruner would optimise it
theorbtwo What? I just implemented partial application in less then 128 characters...
autrijus theorbtwo: huffmanization uses that "uncommon" things needs to be long. 15:26
Alias_ probably uses the wrong magic var
autrijus theorbtwo: however, programmers uses recursion etc all the time.
Alias_ There's way too much OS-specific testing code that isn't compiled away
autrijus or reduce, or schwartzian transform
chip I was being paged?
theorbtwo Alias, $^O.
chip Was the macro question the reason?
Alias_ theorbtwo, right
theorbtwo Aye, chip. 15:27
autrijus Alias_: use constant IsWin32 => ($^O eq 'MSWin32')
like that?
Alias_ autrijus: exactly, but you'd be amazed how much core module don't do that
autrijus $^O lookup is cheap anyway
unless you are inside a real tight loop
Alias_ And there any many cases of such
chip Hm. My part of macros will probably be the AST, but I'm weeks away from being able to work on that.
Alias_ $^O reduntantly checking OZ millions of times 15:28
autrijus chip: let me know when you're able :)
Alias_ OS
chip autrijus: you'll be the 2nd to know :)
autrijus that is, of course, that you actually need that millions of times :)
autrijus raises the benchmark flag
Alias_ yeah
in any case, I've been thinking about some form of Const module 15:29
which just defines the 100 most common useful-as-constant things
castaway looks at the clock
theorbtwo Hm, I wouldn't think it'd be a horribly difficult thing to make the p5 constant folder aware that $^O is a constant.
Alias_ theorbtwo: I'm not sure it can fold the equality though 15:30
cognominal what is the meaning of "use v6" in pugs programs? Does pug intend to support perl5 with "use v5"?
theorbtwo Yeah, folding equality is easy.
Alias_ autrijus: Does it only fold plain booleans?
autrijus Alias_: it can.
[not|autrijus]~$ perl -MO=Deparse -e '1 eq 1'
'???';
there you go.
Alias_ hmm..
let me check something
theorbtwo perl -MO=Deparse,-p -e '"x" eq "x" and print "Yep!"'
perl -MO=Deparse,-p -e '"x" eq "y" and print "Yep!"' 15:31
autrijus, you can't actually tell in that context; it could have been optimizing that away as don't-care.
cognominal and why "use v6" and "require Test"; why not "use Test" as well?
theorbtwo Because 'use' is only implemented for use-version. 15:32
autrijus right.
cognominal can you elaborate?
ok, I got it
theorbtwo use Test isn't implemented, only use v6.
Alias_ in any case... there's quite a bit of code that would benefit from $^O being read-only and constant 15:33
theorbtwo wonders if he should get in the habit of writing "use 5;" at the top of his p5 scripts.
Alias_ Perl would most certainly get faster for many things 15:34
and have a lower overhead
rgs theorbtwo: no, but maybe "no 6" if I get this implemented :)
theorbtwo finds with some amazement that $^O isn't already read-only.
cognominal ho boy, this is good to read Perl6 code. /me reading Test.pm
Alias_ theorgtwo: Which is what worries me a little... if it's writable, there may well be a reason 15:35
theorbtwo The concept scares me.
Alias_ theorbtwo: Giving you enough rope, and all that
castaway wanders off home
theorbtwo Since setting $^O to 'MSWin32' does not cause my machine to reboot, I'd say it's a bug.
rgs $^O being writable is a feature, iirc
Alias_ theorbtwo: Use File::Find::PPI and see if anything in CPAN actually modifies it 15:36
theorbtwo Sounds like work. I'll let the pumpking handle that.
I just suggest ideas.
iblech cognominal: You might also want to read some ports of Perl 5 modules to Perl 6 on freepan.org/ 15:37
pjcj Windows builds from way back used to claim they were Irix, so setting $^O was very useful then. It's also handy for testing purposes.
Alias_ Anyways, I'm stuck at data Temp = Cold | Hot
I'll leave it till next time
theorbtwo Alias, that says that there is a data type named Temp, and that there are two functions that construct it, Cold and Hot.
chip And you can tell which function constructed it.
Alias_ So what does type Foo = Bar say? 15:38
autrijus type Foo = Bar says that Foo is a synonym to Bar.
"data" vs "type"
Alias_ wait a sec, Cold and Hot are functions?
autrijus they are constructor functions.
Alias_ Why use an uppercase name?
autrijus because they are not user-defined functions.
and also, because they can be pattern-matched.
Alias_ I really need to see what how you actually USE data Temp = Hot | Cold, before continuing 15:39
autrijus sure.
data Temp = Cold | Hot
mix Cold Hot = "nice and warm"
mix Hot Hot = "boiling, man"
Alias_ aaaah!
autrijus mix _ _ = "they do not mix"
Alias_ screams
my head!
autrijus mix :: Temp -> Temp -> String 15:40
cognominal autrijus, I will write in a French mag about pugs in a few month. Can you tell me why you crated #perl6 on freenode? Is the idea similar to Larry infecting shell related newgroup with perl posts?
Alias_ still explodey
autrijus cognominal: uh, yes, I think.
Alias_ More examples needed for basic data Temp in use 15:41
autrijus ok... try this
cognominal autrijus: thx
Alias_ Tell, See, Do
autrijus fromCelsius :: Int -> Temp
fromCelsius n | n > 30 = Hot
fromCelsius _ = Cold
Alias_ Medicine has taught using "Hear it, See it, Do it" forever
autrijus or, written less fancily 15:42
Alias_ So Hot is something like a constructor/constant?
autrijus fromCelsius n = if n > 30 then Hot else Cold
yes. it is a constructor, and a constant.
Alias_ That's the bit missing from the examples
autrijus it has the type
Hot :: Temp
Alias_ Need it to communicate that "Temp is one of two values" 15:43
autrijus actually in GHC 6.4 and above you can say exactly that.
data Temp where
Hot :: Temp
Cold :: Temp
which is crystal clear.
chip why would you want to?
autrijus want to say that?
or want to use that notation?
it reads more clear :) 15:44
chip why be able to say "::Foo" when it's redundant?
why add the grammar for it?
autrijus but more importantly, because that notation lets you say things like:
theorbtwo Oooh, that syntax is far more clear then the other ways you said it earlier.
autrijus data Term a where
TrueT :: Term Bool
ZeroT :: Term Int
which Cannot Be Done using the old notation. 15:45
Alias_ autrijus: Time for me to head home. I'll take up this whole demo thing another time
autrijus Alias_: sure :)
autrijus.org/tmp/old.hs # old, error-prone, impossible to reason about notation
Alias_ But it's no wonder nobody codes commercially in Haskell :)
autrijus autrijus.org/tmp/gadt.hs # new, guaranteed correct notation 15:46
Alias_ Only technical wizards in places nobody else gets to look
:)
autrijus Alias_: eh, I do :)
Alias_ I could never tech this crap to my LUG'mates
It's the sort of language that has a minimum IQ of 120
ninereasons I have a friend who programs ATMs and Mil crypto in Haskell
Alias_ Or you spend a year learning
autrijus notes that Java people say that that about Perl too.
Alias_ autrijus: Perl is just twisty.. there's nothing there that's particularly twisty 15:47
errr
hard
autrijus ah that's because you have learned it :)
Alias_ regex maybe
regular expressions were the only true "hard" thing, from memory
chip autrijus: you could do that with named fields in records, I think, but it would be uglier
autrijus chip: you can't typecheck the named fields. 15:48
chip: which is the whole point :)
Alias_ And the initial translation from functions to OO
Odin- autrijus: Nah. There's different types of people. There's java people, there's perl people, there's haskell people, there's lisp people... ;)
Alias_ Which just required "Collections of data that know where their functions are without you having to tell them"
autrijus chip: using the old notation you can write nonsensical things like "Succ TrueT" 15:49
Alias_ But regex and pattern'stuff was a very different concept
Alias_ &
autrijus i.e. "bool::true++"
Alias_ night
chip autrijus: named fields have types, though.
<- confused
autrijus chip: and they are all of the same type.
chip: the point of GADTs is allow each term to carry different type.
chip I think I see 15:50
autrijus the example:
data Term a where
TrueT :: Term Bool
ZeroT :: Term Int
cannot be rewritten using records
data Term
= TrueT { typeOfTerm :: Bool }
| ZeroT { typeOfTerm :: Int }
is invalid.
because all field selectors of the same name needs to have same type. 15:51
and that did not let you write
chip I see
autrijus eval :: Term a -> a
chip Is this new syntax standard or just ghc-author-fu?
autrijus this new syntax is shiny new and specific to ghc.
ninereasons it won't work in hugs? 15:52
autrijus pugs's symbol table type "Pad" and "Symbol" has just been converted to GADT.
ninereasons: not today. eventually in the future
converting them to GADT has raised some subtle bugs :) 15:53
I expect more bugs to be uncovered after we go all the way to GADTs.
mm 46 revs today. 15:55
autrijus journaling
chip: oh. the point of GADTs is that there is no need to write a "context propagator". 15:56
remember how hard it was in perl5? or in languages/perl6/ (P6C in parrot)?
GADTs makes them all _obsolete_ :) 15:57
chip neat
chip fixed a bug in Devel::Profile WRT contexts
autrijus cool 15:58
autrijus tries out the new ebug
ninereasons ebug is very slow, but what a concept!
theorbtwo ebug?
ninereasons a new cpan debugger maker 15:59
autrijus theorbtwo: search.cpan.org/dist/Devel-ebug/
one of the fruits generated from YAPC::Taipei hackathon
fueled by bad ideas from lots of people, including yours truly :) 16:00
Alias_ still impressed with it being a web front end
Alias_ leaves really this time
:)
chip autrijus: now I know I must use ghc 6.4 despite Debian slowness to upgrade, thank you 16:03
theorbtwo chip: it's in experimental.
chip theorbtwo: but trying to install it uninstalls some of the ghc-dependent libraries, whcih means I have to rebuild them myself 16:04
theorbtwo Ah.
chip theorbtwo: but now I know that's worth the trouble
ebug++ # autrijus is a hub of innovation 16:06
oh wait, Leon wrote it. leon++
autrijus acme++ 16:07
PerlJam chip: still acme had to rub up against autrijus for it to happen ;) 16:08
autrijus and obra. and beer. :)
beer being perhaps the most imporatnt.
PerlJam chip: go hang out with those guys. The productivity seems infectous!
ninereasons I don't really understand String-Koremutake 16:09
what does ebug use it for? 16:11
use it to do? that is
jabbot pugs - 1562 - Added todo tests for the delete builtin. 16:12
chip PerlJam: I'm here, aren't I? :-P
theorbtwo Hmm, what's the syntax at the top of Eval.hs's reduce: reduce env@Env{ envContext = cxt } exp@(Syn name exps) = case name of...
What's with the @s?
PerlJam chip: you might need physical contact to catch the disease 16:13
ninereasons perl -MString::Koremutake -le 'my $k = String::Koremutake->new(); my $s = $k->integer_to_koremutake(123456789); print $s;' # nugrohigo 16:14
PerlJam theorbtwo: I forget what it's called but that's a syntax to make a name that is the whole of the parts.
autrijus theorbtwo: what PerlJam said
fun arg@(x:xs) = ...
here x is head arg 16:15
and xs is tail arg
cognominal autrijus: what is the GADT ,entionned in your journal? 16:16
autrijus cognominal: see the log above :D 16:18
also compare autrijus.org/tmp/old.hs with autrijus.org/tmp/gadt.hs
ninereasons i see ... a way to keep secrets 16:19
PerlJam secrets? 16:20
It looks more like a way to eliminate "useless" code :-)
cognominal stands, for Glasgow Abstract Data Tree?
ninereasons sorry PerlJam, I'm OT 16:21
talking to myself
chip I wonder when autrijus will be visiting $Continent::NA
autrijus chip: YAPC::NA
cognominal: Generalised Algebraic Data Types
jabbot pugs - 1563 - removed 2 globals (kind of) and further 16:22
chip autrijus: Marvelous! 16:25
chip will be there too
autrijus arrive 4 days earlier for our hackathon?
www.nntp.perl.org/group/perl.perl6.compiler/487 16:26
crysflame "and of course, a lake to put them in" 16:27
chip autrijus: I think I cannot resist. The pull is too strong! 16:28
autrijus yay!
chip++
stevan oooh hackathon for YAPC::NA 16:31
but wait,.. wont we be finished by then ;) 16:32
jabbot pugs - 1564 - Added tests for pairs and exists on arra
autrijus stevan: no way ;) 16:33
stevan autrijus: should I consider AoH and AoP(airs) as todo or not todo? 16:37
autrijus I don't know. I think not todo 16:41
stevan ok
jabbot pugs - 1565 - Added tests for kv, keys, and values (mo 16:42
Limbic_Region autrijus - out of curiosity, how long before Apocryphon 2? 16:43
autrijus Limbic_Region: 1 week I think 16:44
now that the conf is over :D
autrijus kept getting drawn into coding 16:45
but, sleep first :) journal up.
&
theorbtwo G'night.
Limbic_Region Sleep well
stevan iblech: are you around? 16:46
Limbic_Region is also wondering about huge jumps in versioning considering things in future versions (by roadmap) are already being implemented
iblech stevan: Yep
stevan hey
can we move all those delete, kv, pairs test into like a hashs_and_arrays/ folder? 16:47
or maybe I am being over-organized 16:48
iblech Of course
If we don't stay organized, we'll have 100 tests in one directory soon :)
stevan iblech: cool
stevan likes to use #perl6 as a sanity tester :) 16:49
iblech done as r1568 16:51
stevan iblech: thank you 16:52
jabbot pugs - 1568 - Moved some builtin tests concerning both
pugs - 1567 - (whoops forgot to save file before commi
pugs - 1566 - some tests for mixed type multi-dimensio
stevan iblech++ # for writing soooo many tests and soooo many modules in sooo short a time :)
Limbic_Region perlbot karma for iblech
perlbot Karma for iblech: 14
Limbic_Region perlbot karma for autrijus
perlbot Karma for autrijus: 60
iblech thanks :) 16:53
stevan perlbot highest karma
perlbot The top 5 karma entries: C: 65, autrijus: 60, nothingmuch: 46, stevan: 38, ~brad2901: 29
stevan I think iblech needs to knock brad2901 off the list :) 16:54
Limbic_Region and C plus plus too
rgs C cheats
iblech is writing some more tests then :)
nothingmuch how the heck do i have that much karma? 16:55
stevan C-- is great
nothingmuch++ # for humility
nothingmuch nono, seriusly
stevan C-- is my favorite language now
nothingmuch i was lurking for 2 weeks now
and i get ~10+, i think
stevan nothingmuch: it's still left over 16:56
Limbic_Region nothingmuch, I don't believe perlbot distinguishes between channels
so any karma from #perl applies here
nothingmuch Limbic_Region: i'm only occasionally on gentoo
and that's it
stevan perlbot karma for Limbic_Region
perlbot Karma for Limbic_Region: 1
stevan Limbic_Region++ # for slogging through perlmonks, use.perl etc 16:57
nothingmuch anyway, this motivates me to try and clear more time for pugs
otherwise i'll feel like a sellout ;-)
Limbic_Region stevan - I purposely keep my karma neutral, but thanks
Limbic_Region--
Limbic_Region-- 16:58
nothingmuch you can get the same effect by ++ yourself... much more fun
nothingmuch++
perlbot What kind of idiot karmas himself? Your kind of idiot!
Limbic_Region hmmm - those should have been in /msg's to perlbot - sorry
nothingmuch goes to dinner 16:59
castaway wonders if she can crash SEE
putter filetest operators -d -f currently return bool, rather than the filename, so they dont chain. shall i fix? 17:11
stevan pugscode.org <<Overview Journal Logs>> | You have safely opened the door to many Perl 6 hackers. | pugs.kwiki.org | smoke: xrl.us/fmw2 | Mac OS X r1567 (267/3922) | Win2k r1534 (210/3779) 17:11
stevan putter: yes please :) 17:11
rgs do -f etc accept filehandles ? 17:12
or only filenames ?
stevan rgs: good question
Limbic_Region out of curiosity - does the chaining of -f -d -r etc re-stat the file or use the equivalent of p5's _ ?
putter :)
stevan Limbic_Region: good question too
rgs Limbic_Region: in perl 5.9.2 this doesn't re-stat 17:13
stevan hmm, who wants to post to p6l? ( I don't this is addressed in the Synopsis )
rgs stevan, about the re-stating ?
stevan actually rgs's question is probably for p6l and Limbic_Region's is for p6c
rgs stevan, no, I remember having read that -f $filehandle is legal perl 6 in p6l 17:14
stevan rgs: even better (we already know the answer)
Limbic_Region the cached stat is supposed to be available per file (IIRC), my question was more a matter of does Pugs support that with chained file tests now 17:15
rgs and I think that -f -r $foo shouldn't re-stat, since I remember having already asked this question while implementing it in perl 5
ninereasons for <1 2 3 4> -> $_ {say $_ } # 1 3 : odd behavior ? (unintended pun) 17:16
for <1 2 3 4> -> $a {say $a } # 1 2 3 4 : compare 17:17
stevan pugs -e 'for (1, 2, 3, 4) -> $_ {say $_ }' # prints 1 3 as well 17:18
ninereasons: post a bug to perl6-compiler please
ninereasons yea. it's the explicit designation of $_ as a topic that does it, I think
Limbic_Region rgs - dev.perl.org/perl6/doc/design/apo/A03.html 17:19
it appears that the operators are supposed to be smart enough to return the right thing to avoid the re-stat. If I am reading it correctly 17:20
rgs niiice
Limbic_Region search for "RFC 290"
rgs I've seen
larry++
stevan putter: feeling adveturous I hope :) 17:21
putter Limbic_Region: currently pugs doesnt support chainging at all. or any filetest operators other than -d and -f. my intent is it will shortly support chaining, reflecting whatever the underlying haskell stat-ing is. excess stat avoidance i'll leave until the filetests are more fleshed out. i believe we will also need a filetest builtin.
rgs there's chaining (-d -r -w) and grouping (-drw)
$handle = open -r -w -x $file or die; 17:22
putter anyone like to create a bunch of tests for t/operators/filetest.t ?
Limbic_Region putter - thanks - I am an innocent bystander just watching the flow of progress (to be read as I am not involved with pugs development) 17:23
stevan iblech: wanna take on putter's request?
stevan has to go set up servers soon
putter it's _very_ easy to create tests... ok(-r -w 't/operators/file_test_ops.t','-r -w file works'); 17:24
iblech hm... such a test is a bit tricky, me thinks, because you've to create files with "unnormal" permissions, and sockets, and that 17:25
but as putter says, -r and -w should be simple -- will do
stevan iblech++ 17:26
putter++
stevan heads down to the server room
putter iblech: i check in a revised file_test_ops.t in a moment. (hmm, that should be renamed) 17:29
iblech: re -r -w, the p5 test file makes interesting reading. suggest using the recent mumble OS = any<mumble> idiom as a guard. 17:30
gaal shapr still around? 17:31
rgs iblech: be sure to take the t/op/filetest.t from perl 5.9.2. It tests stacked operators.
gaal shapr++; # www.haskell.org/hawiki/HaskellDemo
(and anyone else writing there)
jabbot pugs - 1569 - * Added tests for splice. 17:32
iblech rgs: Is there a online svn/cvs repo somewhere?
rgs iblech: search.cpan.org ?
iblech rgs: of course, thanks :) 17:33
brb 17:35
putter anyone object to my renaming t/operators/file_test_ops.t to t/operators/file_test.t? (it uses its own name in the file, so I might as well deal now if this seems worthwhile...)
gaal putter: go for it 17:39
putter its now operators/filetest.t. ill check in a revised version in a few minutes. 17:45
gaal cool
nothingmuch castaway|theorbtwo: is see on woobling at port+2 working? 17:50
theorbtwo will check.
nothingmuch anyway, i'm off to a friends to fix his TV 17:51
gaal pugs should talk like this, perhaps: www.effect.net.au/lukastan/humour/C...es-Mac.txt 17:52
theorbtwo G'luck.
jabbot pugs - 1570 - renamed file
gaal is there a technical name for the kind of structure that can contain an XML node? ie something whose content can be thought of as text; or as an array of elements; or sometimes as an ordered hash of elements 17:58
putter iblech: this might help... web.mit.edu/ghc/www/libraries/base/...ctory.html 18:00
castaway nope, not working nm
jabbot pugs - 1571 - Added more filetest tests. 18:12
putter iblech: beat me to it. looks like you did a superset of my changes (hmm, and nicer style too), so i'm done with filetest.t. 18:19
iblech will do some cleanup and, finally, a is_rw.t now
18:25 Aankh|Clone is now known as Aankhen``
iblech ~~ off, school tomorrow 18:26
jabbot pugs - 1575 - Some organizational cleanups. 18:32
pugs - 1574 - Added test for "is rw".
pugs - 1573 - build fix for Eclipse, again
pugs - 1572 - Added tests for stacking/cascading of fi
gaal anyone with Makefile-fu up? FileSpec isn't being make'd into blib 19:01
jabbot pugs - 1576 - restored some forgotten semicolons (on m 19:12
putter in haskell, is there a short variable name implying it contains a bool? eg, in scheme, "p". 19:35
wilx isSomething? 19:40
Maybe.
stevan nothingmuch: ping! 19:56
19:59 crysflame is now known as krysflame
putter tnx 20:01
Odin-LAP ) 20:22
Bah.
Corion Oooo. Segfaults have now reached Pugs/Win2k :)) 20:27
(segfault in operators/arith.t)
... backloggin
pugscode.org <<Overview Journal Logs>> | You have safely opened the door to many Perl 6 hackers. | pugs.kwiki.org | smoke: xrl.us/fmw2 | Mac OS X r1567 (267/3922) | Win2k r1576 (281/4129)+1Unex 20:30
Corion I've patched lots of tests to "if($?OS eq any<MSWin32 mingw cygwin>) {" where they previously used "if $?OS eq 'MSWin32'" - it might be that I broke some tests that way. Gaal - please test the msys stuff (I don't know how unixish it is) 20:40
... well, I haven't committed yet, and I'm running a local MSWin32 test to verify that the number remains constant before committing :)))
metaperl_ what is the colon after the parenthesis doing here:
multi sub abs (: Num ?$x = $CALLER::_ ) returns Num
nothingmuch theorbtwo|castaway: retry see://woobling.org:6944 20:49
Corion Ooo - you got the SEE decoding going ? I need to hook Zaphod up with that :)) 20:50
gaal Whoops, Corion, wasn't really here (and going to bed soon)
Corion gaal: np - I'm going to bed as well - I just wanted to tell everybody that I changed stuff so there is no frantic search in the code when just the tests broke :) 20:51
gaal go ahead and commit whenever you want, i'll test it in the morning (or, well, whoever uses msys other than myself)
:)
Corion Pugs porn was updated at datenzoo.de/pugs/win2k.html
gaal: Ah well - Perl6 should aim for as many platforms as Perl5 already sits on!
jabbot pugs - 1577 - $?OS eq 'MSWin32' => if($?OS eq any<MSWi 20:52
gaal if it works, it'd be a good idea to scan the whole tree for MSWin32 and judiciously apply that idiom
corion, sure.
Corion gaal: I did just that :)
re :) 20:53
I did just that, grep for MSWin32 and then after a quick code scan replaced the code with the junction
... so I might have broken some tests for the platforms that are more unixish 20:54
gaal including outisde t/ that is - though of course, not everywhere.
Corion gaal: Oh, yes! 20:55
Aaah - the one unexpected TODO is autrijus fixing "last()" to become "last" - I'll promote that test from TODO to real
HTML test viewer++
gaal Corion, consider doing the catalog thingie too so they become cross-referenced 20:57
Corion Hmmm. autrijus promised that I was writing a real HTTP proxy, instead of the fakey proxy I've been using to simulate my testing environment. Maybe I should port HTTP::Proxy :)
gaal: "catalog thingie" ?
gaal nothingmuch.woobling.org/pugs_test_status/ and click on a test 20:58
created with catalog_tests.jmm.pl 21:00
i'm off to sleep now -- see you later!
Corion Ah, cute! Yeah, I should find out how to do that, and do it as well :) Ah, like that. I'll put that into my smoking batch too.
good night
gaal zzzz 21:01
jabbot pugs - 1578 - Promoted TODO test for 'last;' to real t 21:02
Corion Gah. The second I decide to look at a port of HTTP::Proxy, BooK releases a new version with more features :)) 21:25
(But HTTP::Proxy might actually work without fudging under Win32 if ported to Pugs)
& # zZzzzzZzzz 21:32
21:42 krysflame is now known as crysflame
crysflame oo, neat 21:54
Corion: nice :)
putter ok, -r -w -x -e are in. sort of. and they chain. but they do multiple stat()s. we have to figure out how we are going to portablably do stat(). wonder what p5 does...? 22:07
jabbot pugs - 1579 - added more filetest operators, and they 22:12
rgs putter: it's awful 22:17
putter: see the filetest pragma for example
need to handle acls etc...
in my opinion this will be deferred to parrot at some point
ask chip about that
putter blech. what a zoo. 22:20
my fuzzy impression is darcs uses a fake posix to get windows support. perhaps we might snarf it... 22:21
rgs: some of the filetests (eg, -s ) probably cant be deferred to parrot, unfortunately. 22:39
darcs approach: abridgegame.org/cgi-bin/darcs.cgi/d...c=annotate 22:41
Limbic_Region not sure if autrijus still plays catch up with the log, but www.perlmonks.org/?node_id=445097 is a nice reply in the thread I pointed out earlier showing the impressiveness of Pugs 22:49
putter Limbic_Region: :) 22:54
it looks like pugs on non-posix systems already sheds functionality (eg, sleep)... 22:58
jabbot pugs - 1580 - fixed golf.t typo 23:02

Powered by Cro and the Raku® Programming Language.

Please report any issues / comments / feature requests as an issue on App::Raku::Log.

Thank you!