Run Perl6 Now! | pugscode.org ('Overview', 'Journal') | pugs.kwiki.org | chan now logged: xrl.us/e98m
Set by theorbtwo on 28 February 2005.
wolverian hah. nice bug with shift() 01:17
ingy back 02:01
stevan ingy: do you know if open() and readline() was working? 02:05
ingy stevan: I don't thinkso 02:06
I didn't see any tests or other code using file input
I need it for my Kwid test harness 02:07
wolverian creating a filehandle seems to work to some extent, but it is always closed 02:08
that is, I can't hGetLine or hPutStr.
ingy I want to port IO::All to pugs
stevan wolverian: I can write to a file, but not read it
thats enough to do a test though 02:09
I can just eval the read tests
wolverian stevan: how? 02:10
oh, > works.
stevan yeah
wolverian < is a unknown mode. :)
stevan yup
wolverian s/a/an/
stevan actually from what I can gather form the Haskell, read mode is "" 02:11
so open("file") defaults to read
wolverian but it doesn't work.
stevan yeah it doesnt :)
but broken stuff has not stopped me from writing tests yet :P
wolverian hehe.
stevan write will also create the file too 02:12
wolverian it's neat how irrational I can be about programming languages. I just don't feel comfortable coding python, even though I consider it a very good language. (just as good as perl, or ruby, and in many cases better.) 02:28
also, ruby would be my favourite language if not for the Proc/block separation. 02:31
stevan r404 and we are up to 926 test 04:14
my goal is 1000+ tests by the end of the week ;-)
nnunley ;)
No problem.
2 days. 80 tests. 04:15
stevan shouldnt be a problem, thats about the average lately anyway
nice! rt.openfoundry.org/Foundry/Project/.../pugs/log/ 04:17
I only just saw this :P
nnunley stevan: I realized there was another path towards testing the pretty printing... We could export a function into pugs that pretty prints an evaled statement. 04:19
Into a string.
And then use 'is' testing.
stevan nnunley: that would work
although I was considering learning Haskell to write that test 04:20
but realistically if I can do it in perl, it will get done much quicker
nnunley stevan: If at least 1 of us learns enough of QuickCheck to start testing the haskell core, then we can ensure some really nice coverage.
Preferably 2.
stevan nnunley: isn't QuickCheck the specification based testing tool? 04:21
nnunley stevan: Exactly.
stevan tmoertel's LectroTest is based on that
hmmmmmm 04:22
nnunley nods.
And shapr has a standing offer to help us with groking its usage.
stevan looks for that YaHT PDF he downloaded yesterday
I need file test operators too 04:24
to round out the tests t/op/io.t
*sigh*
wolverian does an infix:<foo> declaration work yet? :) 04:25
Steve_p shapr's mentioned he's done qite a bit with QuickCheck
buu wolverian:
wolverian: I seriously doubt it=]
nnunley Steve_p: Yeap. He's got a TDD version of QuickCheck.
stevan wolverian: doubt it, i was just reading Prim.hs and didnt see anything that looks remotely like it
wolverian stevan: right. it doesn't seem to work. 04:26
buu infix:<Z> { zip(@_) }
wolverian: What is the full syntax for those declarations anyways?
wolverian I'd like to infix is()
stevan drools at the idea of creating my own infix operators
nnunley nods at wolverian.
wolverian buu: sub infix:<foo> (...) { ... } 04:27
buu Oh
nnunley foo is bla because "I said so"
buu Simple enough.
I thought there already was an infix 'is' 04:28
nnunley nods
buu for @x -> $_ is rw { }
nnunley Actually, it's adverbial. 04:29
IIRC.
buu Perl6, why are you so scary
nnunley sub infix:<foo> is parsed /(.*)foo(.*)/ {...}
buu ! 04:30
Oh god
#writing sub languages in perl is bad idea
stevan uh oh I think buu's head is going to explode again
nnunley grins. 04:31
buu Too late =[
nnunley buu: And there will be rabbits, too.
buu nnunley: Oh goody. 04:32
Do I get my own rabbit?
wolverian nnunley: doesn't that need to be a macro?
(since it needs to parse the code itself, so it needs to do it at compiletime.)
(I think.) 04:33
buu: heh, I forgot that. :)
mauke sub infix:<IsNot> is patented {...}
tmoertel stevan: did you have questions about QuickCheck? 05:09
stevan tmoertel: no, not yet, still working on Haskell itself
nnunley: r405 and up to 969 tests 05:31
:)
nnunley Slow down, man... Save some for the rest of us. ;) 05:32
stevan++
stevan nnunley: no doubt there will be plenty more needed
we dont even have classes yet
and forget about reg-exps
nnunley nods 05:33
And rules
stevan yup
nnunley I'm guessing that rules might come next... With them and PGE, we very close to bootstrap. 05:34
stevan nice 05:35
although I have to admit, I want to try out the classes more
nnunley nods.
stevan ok,.. off to sleep for me, later all 05:37
nnunley waves. "Sleep well."
tmoertel 'night
ingy hola 07:06
autrijus: ping 07:07
buu HALLO 07:12
ingy hi buu 07:13
buu Hi ingy! I recruited someone else to make fun of kwid 07:14
ingy what is your deal?
buu um 07:15
3 queens and a two
I pass
ingy later
viirya_ yyyy 07:34
ingy viirya_: do you know haskell? 07:35
Khisanth the syntax looks scary
viirya_ ingy: a bbit 07:40
weridly, making pugs causes very high loading now. 07:47
what's wrong
buu Fine for me 07:49
ingy autrijus: hihi 07:50
autrijus_ greetings lambdacamels
ingy autrijus_: too
autrijus_ @ AOSS "Standards" working group
ingy I am transcribing YAHT to Kwid
autrijus_ oooh. 07:51
ingy while waiting for you to show up
autrijus_ I fixed the open"<" bug; working on open some mor
e
ingy I have a simple haskell question
autrijus_ sees many new tests committed
sure?
ingy you++
autrijus_ (also, #haskell is a good place to have simple haskell questions answered)
ingy I have an array of strings
and I want to listify it 07:52
autrijus_ you mean, a list of strings
but go ahead
hmm?
ingy well I want to use the list where strings are expected
btw, I am sitting on a very nice patch 07:53
autrijus_ you can use "concat"
ingy just need this to finish it
actually it is done
autrijus_ concat ["1","2","34"] -- "1234"
ingy no
autrijus_ write code, not words :)
ingy I want "1", "2", "3", "4" 07:54
autrijus_ understands code better than words
you want to join them with comma?
ingy ok...
Khisanth heh just got up to that part in YAHT :)
ingy hold on
I am using unlines 07:55
which takes a list of strings
autrijus_ and join them with "\n"
that's correct
ingy let me try something 07:56
viirya_ why new pulled code hard to build?
making high loading, and in slow scramble.
ingy foo = ["1", "2"] 07:57
unlines [foo, "3", "4"]
I want foo to list its contents 07:58
because the above is invalid
viirya_ just unlines $ foo ++ ["3","4"] 08:00
guess what you need is that. 08:01
ingy that works
I tried it without the $ 08:02
why is the $ needed?
viirya_ unlines (foo ++ ["3","4"]) works too.
ingy oh I see now 08:03
viirya_ unlines don't know what's foo ++ ["3","4"]
ingy I would think ++ binds tighter
thanks, it works 08:04
pugs now has -V
autrijus_ woot
ingy which comes from the new Config.hs
autrijus_ no, function application always binds tightest
can't have any operator that binds tighter than function application in haskell
clkao autrijus_: yo, come around #svk 08:05
autrijus_ however useful it may be
ingy cool 08:06
it is a pitiful -V
but it's a start
buu ingy: I can parse text, verbatim and .verbatim! 08:08
autrijus_ ooh!
buu Yeah =/
But I think I finally have a clean design to expand to the rest of kwid.
ingy buu: cool
buu yeah 08:09
cls_bsd ok, make nightly build again
svk up -s ; perl Makefile.PL; make
autrijus_ cls_bsd: you are holding nightly builds
?'
buu ingy: =header is invalid, right? It has to be: = header; ?
viirya_ compiling in turtle speed... 08:10
autrijus_ sorry, I'm involved in the conf right now, discussing revisions for the PO format
ingy buu: correct 08:11
(and documented)
buu Yeah yea.
You were closer.
ingy you should carefully read perlkwid at least once
buu I did. 08:12
ingy maybe twice :p
buu Well yeah
I keep re-reading secitons 08:13
But while I'm asking you questions, should: =foo; be an error or just plain text?
ingy plain text 08:14
cls_bsd sudo pkg_delete /var/db/pkg/bsdpan-Perl6-Pugs-6.0.9/ ; sudo make install clean
hmm
always doing this
cls_bsd considering about make pugs in ports even when pugs is still alpha
viirya_ give up compilation 08:16
buu ingy: Is there any particular reason you made *foo*,*bar*,*baz not bold? 08:17
ingy it is too problematic imo 08:18
buu Well, I can see stuff like foo* bar* baz;
ingy *foo*, *bar*, *baz* is bold
buu Good point.
ingy and you can always {*foo*},{*bar*},{*baz*} 08:19
buu Well yes
ingy: line 190 "If you start ith bullets or numbers, stick with them, as formatters use the first "=item" type to decide how to format the list" 08:20
Should that be there?
theorbtwo Question: What is $foo{`clear`} ?
I'd insert a "may". 08:21
ingy buu: I'll fix that
buu theorbtwo: Er, kwid doesn't have =item tags =]
theorbtwo Oh. Right.
buu ingy: And inside a .list region, is there any difference between - and * thingies? 08:22
ingy yes
- foo() 08:23
blah blah
it is just like the diff between
=item foo()
and
=item * 08:24
buu Er, sorry, I'm not following
oh
"- for things that look nothing like bullets or numbers"
ingy :)
buu I see.
ingy pod has 3 list types
autrijus_ ingy: ooh xliff-tools.freedesktop.org/wiki/Pr...iffPoTools
buu So if I was rendering - foo; as html 08:25
autrijus_ ingy: you may or may not like that, but it looks sensible
buu Would I use <li>, despite the bullet?
ingy bulleted, numbered and worded
buu: html has nothing to do with parsing
buu ingy: Right.
I was just curious
ingy buu: you can use whatever you want 08:27
buu heh
ingy you can have multiple html formatters
buu Well yeah
ingy look at what various pod2html formatters do 08:28
theorbtwo I'd use li, certianly. If you really want a certian sort of bullet, use CSS or the type= parameter. 08:30
buu shrugs
Just an idle curiosity.
ingy autrijus_: Config stuff checked in 08:31
you should take a look
theorbtwo Semantic markup is the most important sort, IMnsHO; use CSS if your semantic markup doesn't make the browser display what you want.
autrijus_ ok, updating 08:32
ingy Next step is to rework the installation of modules into the perl6 paths
autrijus_ woot.
mmm bin/
can I move it to script/ ? 08:33
ingy I did some clever heuristics to make perl6 module directories
autrijus_ to agree with makemaker convention
buu theorbtwo: I need some way to use kwid as xml =[
autrijus_ and to avoid clashing with the haskell/eclipse conventino of putting executables produced to bin/
ingy um
it's not an installed script
it's part of the build process 08:34
autrijus_ then util/
ingy sure
autrijus_ is a better idea
please do so if you can :)
ingy ok
autrijus_ is entangled is weird ISO stuff
hi nothingmuch! 08:37
nothingmuch++ # donor
nothingmuch =)
autrijus_ nothingmuch++ # paypal'ing me for the cause
nothingmuch: I've synced darcs repo
nothingmuch autrijus_: grazie
say 08:38
yesterday i tried my first hours or so with pugs and stuff
theorbtwo Allo, nothingmuch!
autrijus_ cool.
nothingmuch hola, orb!
stevan said I should try testing the open op
autrijus_ the fatal error in eval"open()" is fixed
will commit soon
nothingmuch so i did
ingy autrijus_: bin->util done
nothingmuch but i didn't get far at all 08:39
autrijus_ ingy: danke
nothingmuch i'd like to get a sort of handle on what is available, language wise
autrijus_ nothingmuch: the bottom of Prim.hs 08:40
ingy night all...
autrijus_ nothingmuch: and grep for todo_ in t/
nothingmuch for example, my $handle = eval { open "non_existing_file" }; didn't work well for me
autrijus_ should give a pretty good idea
there is no eval {}
nothingmuch right
autrijus_ yet
there's only eval ""
because in p6 it's try{} now.
nothingmuch but it "worked", in that it didn't complain much
autrijus_ gone was the punning
nothingmuch i tried try first
but that didn't work at all
autrijus_ yeah, and try{} is not there :)
you want it?
I can code it up :)
nothingmuch so buu suggest eval {}, and it gave the impression that it was working
i'll make deal 08:41
if you give me try and catch, with throwable stuff
i'll give you tests
;-)
autrijus_ uh... the norm here is the other way around :) 08:42
tests first etc
nothingmuch anywho, i shouldn't get carried away
test & code simultaneously
autrijus_ but sure, considering your generosity, I can try to get try working
nothingmuch i'll do it this afternoon
well, no, bass lesson. This evening
autrijus_ cool!
nothingmuch don't rush it
buu nothingmuch: I have no idea what's actually working.
nothingmuch it will take me long
buu I was just randomly guessing.
nothingmuch spreading disinformation, huh?
anywho, i just dropped in to give buu a link (www.findinglisp.com/blog/2005/03/da...ited.html) 08:43
so now i have to go back to real work
ciao! 08:44
autrijus_ Can't locate Spiffy.pm in @INC (@INC contains: C:/perl/lib C:/perl/site/lib .) a
wilx I just updated pugs.
Chasing modules from: Main.hs
Main.hs:
Can't find module `Config'
autrijus_ ingy: thou shalt not pollute the tree
wilx Oh, now I get it.
Geez, do I really need to configure?
:/
ingy ?
wilx: rerun perl Makefile.PL
autrijus_: ? 08:45
autrijus_ ingy: nvm, fixed
wilx Yeah, but that would mean installing Win32 Perl :/
Oh well...
ingy ?
wilx Never mind.
autrijus_ and you did not escape stuff correctly
also fixing
ingy ?
explain all this 08:46
sorry about the Spiffy thing 08:47
did you fix that?>
autrijus_ yeah
I also fixed escapgin
ingy what escaping?
autrijus_ see r410
$value =~ s{\\}{\\\\}g;
ingy nice, thanks 08:49
how do single quotes work in haskell? 08:50
are they for single chars?
autrijus_ yup, just like a certain C language 08:51
but haskell has no interpolation
ingy I C
autrijus_ so just escaping \\ and \n is enough
and I truse Config.hs will have no \n
ingy nope
finished YAHT intro -> Kwid 09:24
ok zzzzzzz for real &
buu 75 tests
Matches inline listy items. 09:25
beth enlighten me, buu 09:29
buu Pugs is the new religion.
svn.perl.org/perl6/pugs/trunk 09:30
beth web.mit.edu/dyezone/www/images/Bolt...20Pugs.JPG ?
buu I doubt it, but I can't see that link
svn.perl.org/perl6/pugs/trunk/docs...rview.html
beth showing me where to get it is not the same as saying what it is!
buu ^
beth I am SO not enlightened
ok
buu Read!
beth re4dx0rz 09:32
damn you buu, I'm going to have to look at this now 09:33
didn't you know I'm already busy and stuff?!
buu hahaha 09:44
Yo're welcome.
theorbtwo beth, autrijus has cleared out many of his paying customers to work on this. It's kind of addictive. 09:47
Even I'm getting into it.
buu hrm 09:48
I think this is the first time I've seen you on irc.
Well, today and all.
theorbtwo I tend to lurk on here more then I talk. 09:50
buu I noticed 09:52
I see you a lot on the chatter box though
But this is the first time I've really spoken to you
If we can all this speaking
theorbtwo Heh. 09:54
pjcj Do your lips move as you type? 10:14
theorbtwo Rarely. 10:16
I also often don't get jokes over IRC/AIM/etc that "work" because the punchline sounds like something else. 10:17
metaperl does anyone understand the arguments to the value constructor ApplyArg? 11:16
in src/Junc.hs
theorbtwo Hm, hat doesn't like pugs, and it looks like ghc's inbuilt profiling stuff isn't as detailed as I want. 12:56
cls_bsd ok, time to translate synopsis 12:57
autrijus_ metaperl: you were looking for me? 13:02
metaperl well, what is the collapse field of ApplyArg used for? 13:03
data ApplyArg = ApplyArg
{ argName :: String
, argValue :: Val
, argCollapsed :: Bool
}
deriving (Show, Eq, Ord)
in Junc.hs 13:04
cls_bsd oh, will pugs become perl6? or just a aim to archive?
autrijus_ argCollapsed controls whether that arg avoids junctive autothreading 13:05
metaperl autrijus, that was my question?
autrijus_ if argCollapsed is true, it means it's not a junction, or it was a junction used in a nonautothreading context (typically Bool)
cls_bsd: it is not my place to answer that question :) 13:06
if the C-based team came forward with a working compiler before Pugs does, then that will become perl6
if however, by the time we completed the bootstrap, we are the only team around
then there is the possibility that pugs will be the perl 6.
metaperl and which synopsis covers junctions? 13:07
autrijus_ S09?
metaperl yes it does, thanks
autrijus_ :) 13:08
theorbtwo Allo, Joshua. 13:10
Limbic_Region morning James 13:13
guess it is early afternoon over there 13:14
theorbtwo Aye; 2.
cls_bsd hmmm 13:17
so there exists an perl6 contest :p
autrijus_ not really a contest :) 13:24
a context, maybe, not a contest
cls_bsd good point 13:25
autrijus_ pugs ported to GHC 6.4. 13:40
mmm much faster code. 13:41
(but ghc 6.4 had not been really released yet.)
Limbic_Region buu - you about yet? Ready and willing to help. 13:45
evening autrijus_
autrijus_ hi Limbic_Region 13:46
mm BOF finished 13:52
autrijus scw: thanks a lot on implementing pick() 14:30
I rescinded dump() though; in perl6, dump() is known as perl()
and works on all data types
theorbtwo Nifty!
Does it produce p6 code, or p5 code? 14:31
rgs ub perl5, dump() is known as CORE::dump()
s/^ub/in/
autrijus it produces p6 code. 14:32
rgs: uh. the context is Data::Dumper::Dump ;^H:)
Coke_ hey, autrijus. 14:47
Limbic_Region grrr - I hate it when this client flakes as I don't know it has 14:55
buu - you around? 14:56
Juerd 15:34 < rgs> ub perl5, dump() is known as CORE::dump() 15:06
Which I think has no place in P6
integral serialisable continuations otoh... 15:07
autrijus yeah, serialisable continuation is just a fancy term for dump/undump ;) 15:08
Juerd ehh :)
I hope the dumps will be a bit fancier than core dumps :) 15:09
autrijus it will be core dumps... it's not j^Wthe C core ;)
theorbtwo If dump works on all data types, then $?CURRENT_CONTINUATION must be one of those data types.
integral hmm, it'd be very neat for post-mortem debugging
Juerd theorbtwo: Isn't that just $?CONTINUATION?
theorbtwo Quite possibly.
Juerd As with $?SUB, which isn't $?CURRENT_SUB (unless I have not been paying attention well enough)
Juerd expects continuation to be abbreviated as cont, as subroutine is sub 15:10
theorbtwo I think it was supposed to have a long name so people weren't tempted to use it friviously.
Juerd Why not use it much?
autrijus uh. &?CALLER_CONTINUATION makes sense 15:11
&?CURRENT_CONTINUATION makes less sense.
Juerd autrijus: Why not? Serialize it, store it, exit()
autrijus which dynamic scope is "current"?
4 + &?CURRENT_CONTINUATION + 10;
basically continuations only makes sense on dynamic l^Hscope boundaries 15:12
s/lscope/scope/
Juerd Oh
integral it's a bit like a return instruction you can save a copy of to call in another scope
autrijus yeah. it's just \&return; 15:14
except more magical
nnunley Hrm. Problems with the ext/Kwid tests. The harness is attempting to run the tests with perl5, rather than perl6 15:26
autrijus yeah. 15:34
weird, HARNESS_PERL should be set.
malaire where can I read about &?CURRENT_CONTINUATION ? synopsis, etc..? 15:37
PerlJam malaire: It's a pugism 15:38
autrijus malaire: escape continuation support is I think mentioned by larry informally
but not part of synopses 15:39
malaire ok
autrijus &?CALLER_CONTINUATION is my interpretation on that
but it's a pugism, as perl said
PerlJam, even.
malaire so, what is it? 15:40
autrijus it is a function that, when invoked, returns on behalf of your caller.
you can also pass that function around and, when anywhere, returns on behalf of the original caller. 15:41
nnunley: Pretty for Value types must be eval-roundtrippable 15:51
because that is what ".perl" uses
I've fixed it
r418
nnunley autrijus: Didn't realize there was a fixpoint contract there. :) 16:01
autrijus feel free to add it as a comment
sorry; I should've write it down myself
but journal up; I need to sleep now :)
clkao autrijus: take a look at svk/win32 before sleep? 16:02
autrijus: or put the instructions on wiki and i can ask matthewd to do that 16:03
ingy hola 16:23
Steve_p bonjour
ingy are there any tests for junctions
nnunley ingy: One in 03operators.t that I know about. And the borken sendmoremoney.t
ingy yow! 16:35
my $osnames = any( 'darwin' | 'linux'); 16:36
is($?OSNAME, $osnames, "Test \$?OSNAME == $?OSNAME");
runs 2 test!!
that is freaky
surreal
Steve_p heh
ingy I'm on darwin
ones test passes, one fails
This works fine though: 16:37
my $osnames = any( 'darwin' | 'linux');
ok($?OSNAME eq $osnames, "Test \$?OSNAME == $?OSNAME");
nnunley ingy: Unless you force a binary collapse in the type signature, it will run both branches. 16:38
Err... boolean.
integral welcome to the evil world of auto threading. junctions thread as early as possible
ingy nnunley: yes, it appears so. but it still blows my mind
integral: interesting 16:39
nnunley Test::ok used to run each of the junctive branches before I changed its signature from $cond (implicit Any) to Boolean $cond
integral ingy: yep. personally I support the other side ;-)
ingy the cool thing is that the code actually ran and the test harness was fine with the output etc 16:40
nnunley I really, really want junctive narrowing. But that's because I'd like to use it for non-deterministic solving.
theorbtwo Junctive narrowing?
ingy I mean... if they are threads, there could have been interspersed output, no?
nnunley theorbtwo: Chaining conditions such that I've narrowed a large junction into a much narrower set of states.
theorbtwo Oh, throw away the false ones, keep the true ones? 16:41
nnunley Right.
ingy anyway, I got my cute test using junctions working :) 16:42
nnunley ingy++
Limbic_Region ingy - I had volunteered to help buu with his kwid parser last night with work commencing today - haven't seen him - do you know where his work in progress is?
ingy and pugs now has $?OSNAME
Limbic_Region: no
Limbic_Region ok - to be honest it didn't seem like a fun thing to do anyway 16:43
ingy hehe
Limbic_Region: kwid parsing or working with buu? ;)
Limbic_Region little of both
ingy :)
Limbic_Region buu and I don't always see eye to eye
ingy aye and aye 16:44
Limbic_Region but he is writing a custom parser and not going with a grammar/lexer/parser approach
PerlJam Limbic_Region: Why don't you track down the bug in examples/hanoi.p6 for me :-)
ingy Limbic_Region: that seems sad
PerlJam er, the bug is in pugs, the exposition is examples/hanoi.p6
integral Limbic_Region: would you have used yapp?
Limbic_Region after having read through the kwid definition - likely 16:45
that's easy to say without getting your hands dirty though
jaap hi; I'm having some trouble compiling parrot; specifically, it complains "blib/include/unicode/ucnv.h" is missing, and I do indeed not see anything called blib in the source tree
Limbic_Region once your knee deep in mud with shrapnel overhead - you do what works even if it means sucking mud
integral jaap: you'd probably get an answer out of [email@hidden.address] if you asked there 16:46
jaap integral: thanks. i s'pose i'll also try the tarball, maybe the CVS happens to be in a non-working state 16:47
abstraction jaap, use latest cvs
Limbic_Region PerlJam - if there is something I can do in p5 without needed to know/learn Haskell/p6 - let me know
jaap abstraction: that's I've been doing, got it about 20 minutes ago 16:48
Coke_ hurm? what steps did you go through to compile parrot, and what's your OS?
(anything in "blib" should be generated, IIRC) 16:49
jaap Linux stein 2.6.8-1-686 #1 Thu Nov 25 04:34:30 UTC 2004 i686 GNU/Linux
get cvs; then "perl Configure.pl"; then "make"
Coke_ how old is your cvs distro?
PerlJam Limbic_Region: somehow I think you're in the wrong place if you don't want to know/learn haskell/perl6 ;-)
Limbic_Region jaap - my recommendation would be to get ICU directly and build that before attempting Parrot 16:50
Parrot is now mostly smart enough to detect a system ICU and use that
jaap coke_: updated to the latest debian/testing a few hours ago; claims to be cvs 1.12.9
Coke_ I thought blib was a perl thing, not an ICU thing, lr.
jaap o wait
that's obviously not what you're asking
Limbic_Region PerlJam - didn't say I didn't want to learn - just not offering up my time during paid work hours to do it
Coke_ (just make sure if you did an update, you do a -dP)
Limbic_Region Coke_ blib from a Parrot sense is a Parrot thing is it not? That's just where Parrot puts things before linking IIRC - and ICU is one of those things 16:53
of course, I haven't worked on Parrot since late November
pjcj has anyone? </troll> 16:54
abstraction what does ICU stand for?
jaap Limbic_region: i did apt-get icu, it installed fine, i ran configure.pl again, as well as cvs update, it doesnt complain, still get same error on build
something with unicode, i take it
integral jaap: wouldn't you want something like icu-dev?
Limbic_Region jaap - Dan's unicode patch is probably in by now - so I am likely talking out my @$$ 16:55
theorbtwo International Components for Unicode
abstraction his string pathc is in
theorbtwo www-306.ibm.com/software/globalizat.../index.jsp
abstraction that sounds painful
Limbic_Region icu.sourceforge.net/ 16:56
abstraction more pain
theorbtwo It is, but it's mostly "somebody needs to do it" sort of things.
integral haven't read anything about the icu dep being dropped though
abstraction better now than later...
Limbic_Region well, the plan was not to drop it from CVS but to make it completely optional when building
jaap integral: i do, indeed; it's called icu-data; i installed it; doesnt help; thanks though
theorbtwo icu-data includes the shared library? 16:57
Limbic_Region Dan originally said Parrot will ship with the ability to do Unicode - period. The installer can choose not to if they want
jaap it seems like it wants a header file that has to be within the parrot source tree, though, since it says make: *** No rule to make target `blib/include/unicode/ucnv.h', needed by `src/string_primitives.o'. Stop.
[weel@stein] /home/weel
Limbic_Region He has been known to change his mind - and I haven't followed the list since late November either
nnunley Limbic_Region: Did they decide to do their own unicode implementation? Last time I bothered paying attention, that seemed to be the jist of the conversation. 16:58
theorbtwo I think the idea is to avoid unicode more often, and consider it just another encoding. 16:59
nnunley nods.
Limbic_Region nnunley - my understanding was that Dan's big patch was to allow a standard interface for unicode libraries so ANY unicode library that played nice with the interface could be used 17:00
nnunley Good.
Limbic_Region that ICU was "it" for now, but that we should consider our own
and that until an alternative was available, Parrot would ship with ICU source
theorbtwo ...so if your input is big5, and your output is shift-jis, you may never go through Unicode in the middle.
Limbic_Region though using it would be optional
and that's where I leave the story 17:01
theorbtwo Of course, Larry has said that Unicode is a hard requirement for p6, so a parrot compiled without unicode support may well not be able to run p6, which rather spoils the point.
jaap the tarball seems to somehow magically not have the same problem, so I think I'll stop bugging you guys ;-) 17:02
Limbic_Region not really theorbtwo
integral just ignore Larry ;-)
Limbic_Region if I want the VM for a reason other than p6 where Unicode isn't needed, why should I be forced to use it
my point being that I think it is a "good thing" that Parrot and P6 are divorced
PerlJam indeed 17:03
theorbtwo I agree.
PerlJam theorbtwo: Were you saying that parrot without unicode spoils the point of parrot-as-perl6-VM ? 17:04
theorbtwo Yes.
PerlJam but it's perfectly fine for parrot-as-XXX-vm where XXX is almost anything else :)
theorbtwo Ja.
But while parrot is designed to be a general purpose VM, or at least general purpose for highly dynamic languages, the only large project really considering using it is p6. 17:06
jaap I think there's quite some interested from more academic circles for doing e.g. Scheme in Parrot 17:07
theorbtwo That'd be very nifty.
Limbic_Region and nothing about what I described as my interpretation of Dan's plan for Parrot/Unicode spoils that
the hooks are there - let the installer choose to use/not use them
IIRC, nothingmuch was looking at doing Scheme in Parrot as an academic exercise 17:08
PerlJam theorbtwo: So you don't think that efforts to BASIC, PHP, Python, etc. with parrot are important or valid or something?
s/to/to use/
integral hmm, I thought the python one could be important if it actually ran faster
jaap well, the thing about Scheme is that it has been implemented about 30 times by now; it is not at all like the major scripting languages that are dominated by 1 or two implementations
but I'm off to breakfast now; cheers all 17:09
theorbtwo I think that while using it for PHP, Python, etc, have been much discussed, they don't look purticularly like they're going to happen. BASIC has happened, but only a very old dialect. 17:10
Limbic_Region theorbtwo I think Python will happen if Dan returns to overule Leo - or someone else takes over - there's a pretty smart Python guy working that problem though I can't remember the handle ATM 17:11
theorbtwo Oh! That's great news. 17:12
Steve_p Sam Ruby?
theorbtwo "Dan returns"? He left?
Limbic_Region theorbtwo - Dan has/had been out of the loop for some time
one of the reasons why I stopped following Parrot
Khisanth Limbic_Region: I always thought we were more likely to get a python on parrot than perl6 on parrot :)
Limbic_Region Sam Ruby sounds right Steve_p - but it has been several months since I followed the list
theorbtwo That's horrible news. 17:13
PerlJam yes, it's sam ruby who's working on the snake eating bird.
Limbic_Region he (Dan) re-emerged for a short time but his rl job sucked him back in
Steve_p He has his fingers in *many* pies though
theorbtwo Sam Ruby was working on it last I payed attention. That was quite a while ago, though.
PerlJam And rubys is *actively* working on it though. (at least the last time I looked at #parrot (last week)) 17:14
theorbtwo (It always confused me -- Ruby/Python mental discouninuity.)
Limbic_Region my point being that he was making lots of headway and the stumbling blocks were issues he couldn't get resolved between just him and Leo
so I don't think Python on Parrot is a pipe-dream 17:15
theorbtwo Great!
Limbic_Region I should be able to dive in head deep in the 3 ps in June (we will have closed/moved into the new house) 17:16
where the 3 ps are p6, pugs, parrot
PerlJam Check groups-beta.google.com/group/perl.p....internals . There have been messages about python on parrot even today. 17:17
Limbic_Region: excellent
Limbic_Region: I'm treating pugs as the perl6 reference implementation and possible bootstrap. I talked with Pm yesterday and it appears that he's working more on PGE, so it's a race to see who implements the better bootstrap first :-) 17:19
Limbic_Region for now, I am working a full-time job, part-time job, in the process of buying a house, closely following his wife's paperwork to become permanent resident/citizen, and also making things happen so she can drive
Steve_p Yes, I thought the majority of Python had been implemented already 17:20
PerlJam DCCs some luck and congratulations and general warm fuzzies to Limbic_Region
Limbic_Region PerlJam - forgetting long term benefits, pugs has some very real tangible RIGHT-NOW benefits
it is getting the community excited about p6, it is getting other communities interested
PerlJam Limbic_Region: indeed. I can't tell you the thrill I get from writing and *running* actual perl6 code :)
Limbic_Region it is helping iron out the synopsises where things aren't exactly clear 17:21
it is letting people start seeing real-world working p6 code
if it ends up not being the p6 bootstrap of choice - I think no one would be overly disapointed
theorbtwo I think pugs can be not just a reference implementation or a bootstrap, but the real thing. 17:22
PerlJam I really think that pugs will make a better bootstrap too. PGE is focused on the "rules" syntax, pugs is focused on basically everything else. As soon as you add rules to pugs, you've got a perl6 suitable for implementing perl6.
theorbtwo: I don't think so.
Limbic_Region to me it doesn't matter - honestly - when is friendly competition not a good thing?
PerlJam yep!
theorbtwo: ghc is not as widely ported as it would need to be for pugs to be the real thing. 17:23
theorbtwo I think the best path is to do a straight port of pugs from haskell to p6.
...but perhaps I simply don't understand Hasell well enough to know why this is impossible. 17:24
Steve_p Hmmm...Haskell on Parrot?
PerlJam theorbtwo: it's not impossible, just highly unlikely :)
integral easier to port STG to parrot? 17:25
theorbtwo STG?
integral spineless tagless G-machine, GHC's virtual machine
theorbtwo Ah.
PerlJam integral: now *there's* an idea that has some potential
Limbic_Region notes the time and realizes he hasn't eaten yet 17:26
Limbic_Region wanders off
integral possibly not hard either. Just write a PMC for the G-machine closure, and GHC already does most of the rest
Steve_p Exactly. Let the G-Machine generate the bytecode or imcc. Very intoxicating :) 17:27
theorbtwo Virtual virtual machines...
Steve_p heh 17:28
theorbtwo Now run the whole thing inside of qemu, running inside of vmware!
Limbic_Region Virtual virtual machines - reminicent of "The 13th Floor" 17:58
www.imdb.com/title/tt0139809/
nnunley The fix-point virtual machine...
Turtles all the way down. 17:59
Limbic_Region perlbot karma for autrijus 18:07
perlbot Karma for autrijus: 1
Limbic_Region heh 18:08
rjbs perlbot: autrijus++
Limbic_Region no need for perlbot 18:11
just autrijus works fine
except my client eats plus symbols
pjcj autrijus += 2pi 18:12
theorbtwo perlbot, journal is use.perl.org/~autrijus/journal/ 18:13
perlbot added journal to the database
theorbtwo perlbot, journal?
perlbot use.perl.org/~autrijus/journal/
theorbtwo When did we get a perlbot? 18:15
Limbic_Region the other day 18:16
I asked why purl wasn't here
the response was that it wasn't easy to get purl to join another network
buu (in #perl) asked Chris62vw to invite perlbot here 18:17
perlbot forget journal
perlbot removed journal from the database
Limbic_Region perlbot learn autrijus journal as use.perl.org/~autrijus/journal/ 18:18
perlbot added autrijus journal to the database
Limbic_Region hmmm - should that have been possesive?
jdv79 maybe the reason i haven't been here yet is because parrot has my attention more than p6 lately 18:21
nnunley perlbot learn pugs revisions log as rt.openfoundry.org/Foundry/Project/.../pugs/log/ 18:22
perlbot added pugs revisions log to the database
Limbic_Region for those who don't know, perlbot was written by Chris62vw whose database spans all channels it is connected to (which currently also includes #perl) 18:23
perlbot help 18:24
perlbot Syntax: (fact) :: tell (who) about (what) :: (who) > (what) :: learn (what) as (info) :: relearn (fact) as (info) :: facts (search term) :: shorten (url) :: shorten it :: search (module) :: docs (module) :: perldoc -f (function) :: math :: fortune :: flip :: host (type) (record) :: rot13 :: roll (die) :: tempconv (temp) :: scramble (text) :: 8ball :: slap :: diss :: what time is it :: highest karma :: lowest karma
Limbic_Region perlbot .ca 18:25
perlbot .ca is Canada
Limbic_Region 8ball will pugs beat pge as the p6 bootstrapper? 18:26
perlbot 8ball will pugs beat pge as the p6 bootstrapper?
perlbot Magic 8ball says: Yes
theorbtwo roll 10d3.14159
Limbic_Region you need to say perlbot 18:27
theorbtwo perlbot roll 10d3.14159
Limbic_Region when all else fails - ask Chris62vw in #perl
Khisanth or
perlbot: source 18:28
perlbot check out my insides: chrisangell.com/incoming/chrisbot/v3
Khisanth :P
jdv79 haha, how did buu get into that article: www.perl.com/cs/user/print/a/908 18:31
theorbtwo Hm, I don't like the lack of prefix on chromatic's lines. 18:33
jdv79 it is written by him...
his name's at the top
nnunley Hrm. So, is grammar ~= module ? Other than stating that it's creating a grammatical module? 18:34
jdv79 "but thing quickly went out of control" - best quote so far 18:35
Khisanth jdv79: because the whole interview thing was done in here 18:37
theorbtwo There was some side stuff that chromatic clipped. 18:38
But it's a useful question.
jdv79 oh wow, i missed a good time!
theorbtwo I only read it via the archive. 18:39
Limbic_Region Apparently, M$ does Perl - www.microsoft.com/technet/scriptcen...fault.mspx 18:47
is there a way to browse pugs test suite on-line? 18:52
theorbtwo I wonder... is it possible to make the parsec bit automagically create examples? 18:56
stevan Limbic_Region: svn.perl.org/perl6/pugs/trunk/t/
Limbic_Region thanks stevan
stevan Limbic_Region: your welcome
rjbs Limbic_Region: haha. hungarian notation in Perl 18:57
Limbic_Region there have been some interesting problems in p5 at the Monastery recently - I was thinking about writing p6 solutions and seeing if Pugs was up to the challenge yet 18:58
jdv79 like what?
stevan Limbic_Region: remember that Pugs does not have classes yet, or multi-dimensional arrays (they all flatten)
AFAIK, pairs are the old nested data structure possible at the moment 18:59
Limbic_Region no need for classes, but complex data structures would be a "good thing"
stevan Limbic_Region: you can always write an example and eventually it will run :)
hanoi.p6 does not run yet 19:00
Limbic_Region that's the point stevan
stevan Limbic_Region: go for it them
s/them/then/
metaperl_ can anyone get this junction example to work? nomorepasting.com/paste.php?pasteID=33301 19:41
PerlJam Perl5's ?: operator is now spelled ?? :: 19:43
And I think you need the parens on the call to any 19:44
metaperl_ PerlJam, thanks
PerlJam oh, and you'll need to my @color, otherwise pugs will complain
metaperl_ PerlJam, the working example is committed as rev 422 19:47
in pugs/examples/junctions/1.p6
PerlJam cool. 19:48
More "real world" examples of junctions are needed however.
metaperl_ yes, I am moving slowly but surely 19:49
I want to present at our perl mongers meeting on junctions next wednesday
how do you read from STDIN in pugs?
my $x = <STDIN>; # assigns the string STDIN to $x 19:50
integral =$*IN 19:54
theorbtwo Is that working now?
integral somewhat
pugs -e'say =$*IN' behaves strangely 19:55
stevan reading and writing of files is working
t/op/io.t
theorbtwo Great! 19:56
stevan actually I take that back, the readline support is not all there yet
but creating, writing and appending files is 19:57
metaperl_: I usually check Prim.hs to see what is working and what is not 19:59
metaperl_: line 169 is the unary '=' and autrijus has a big block commented out there 20:00
metaperl_ oh ok
should this work? 20:02
my $x = $*IN;
print $x;
the print is failing not the reading from STDIN
stevan try say() instead of print() 20:03
metaperl_ my $x = $*IN;
say $x;
fails
Fail: <stdin>: hPutStr: illegal operation (handle is not open for writing)
PerlJam As it should.
stevan yeah I am getting the same thing
metaperl_ perhaps I need to specify where to write it
Coke_ it looks like you're trying to print something to STDIN... 20:04
PerlJam you need to use unary = to actually read from stdin
metaperl_ unary =
meaning what?
PerlJam my $x = =<$*IN>;
Coke_ (wow, that's ugly =-)
metaperl_ holy god jesus christ
theorbtwo Umm, both = and <...>?
metaperl_ i know, for such a common thing
PerlJam well, I'm not sure about the <>
theorbtwo I thought it was my $x = =$*IN
...which also looks like shut. 20:05
PerlJam I know that =<> is the replacement for magic <> as in perl5
theorbtwo shit.
stevan I think <...> stuff is not valid perl6
sorje Oh bummer, hyper ops aren't implemented yet
PerlJam (this week)
theorbtwo Er, s/s..t/junk/.
stevan although I can't remember where I saw that
Khisanth EWW I hope that isn't the only way, the double = looks rather ugly :p
PerlJam anyway $*IN.readline or somesuch should work too (once we get objects in pugs) 20:06
theorbtwo That looks /much/ better.
PerlJam Khisanth: you should know by now that there is no such thing as "the only way" in perl :)
theorbtwo Though .rl would be nice.
stevan PerlJam: that should work pre-object actually
Limbic_Region well, once objects...
PerlJam stevan: Are you sure?
Khisanth PerlJam: true for most things but ...
stevan PerlJam: the filehandle works as the invocant
Limbic_Region yeah - you can do fh.method once that's implmented it will look cleaner
stevan $*IN.readline == readline($*IN) 20:07
AFAIK that is
PerlJam stevan: yeah, yeah, you're right. But is it implemented?
stevan t/op/io.t shows it
PerlJam: readline is currently not complete
PerlJam: but $fh.print works 20:08
actually readline works for files 20:09
stevan wrote t/op/io.t late last night and his memory is failing
svn.perl.org/perl6/pugs/trunk/t/op/io.t 20:10
ingy Larry is so hardcore: www.nntp.perl.org/group/perl.perl6....uage/19468 20:12
metaperl_ how do you chomp your input read from $*IN
theorbtwo You get in chomped, unless you ask for it unchomped, IIRC. 20:15
Steve_p ingu, so you got your answer ;) 20:16
Limbic_Region and when you print you automatically get "\n" added on unless you ask for it not too, IIRC
ingy Steve_p: no :( 20:18
there is no support for objects at all yet, right?
and pugs is a practical project
metaperl_ theorbtwo, can anyone get revision 423? I am typing in something but it apparently is not chomped 20:19
ingy so I'll just have to wing it to accomplish my evil goals
metaperl_ pugs/examples/junctions/2.p6
theorbtwo Er, no, L~R, that's the difference between say and print. (Of course, it's not an accident that say is shorter...)
Limbic_Region theorbtwo - depends on how you parse what I said, I didn't say "use print", I said print - meaning what I was thinking was correct but I wasn't specific enough in what I typed 20:22
nothinmuch - correct me if I am wrong, but weren't you considered Scheme on Parrot a while back as an academic exercise? 20:27
nothingmuch no
i don't even know scheme, sadly
i'm learning lisp from the sicp lectures on line
they are wonderful
stevan Limbic_Region: I think Peirs had something to do with that
nothingmuch very fast paced
my other limitation is that i don't know parrot 20:28
qmole heh
nothingmuch and my third one is that i have time for neither =)
so,
L~R->correct
Limbic_Region hmmm - I know it was a monk, but don't remember which one - fairly certain the monk I am thinking of wasn't Piers
nothingmuch stevan - remember the patch from yesterday?
stevan nothingmuch: the Test::Mockobject 20:29
nothingmuch yup
i glued Test::Deep to it
stevan ??
nothingmuch for argument verification
stevan Test::Deep rules :)
nothingmuch (Test::MockObject also records arguments, so now you can declare them as part of the script and verify that)
it so does
especially for this
superset of is 100% correct 20:30
anywho
i can't decide on a reasonable way to make it optional, but unambiguous 20:31
right now, if you remember, steps in the script are like [ set_true => "foo" ]
stevan nothingmuch: the only way I can see to make it optional is to implement the validation yourself 20:32
nothingmuch i mean interface wise
stevan nothingmuch: or leave it out all together
nothingmuch i did all the logic
right now if the last element of that array is an array ref, or a bless object whose class =~ /^Test::Deep
then it is assumed it's an argument list 20:33
at that point it tries to require Test::Deep and dies violently if you don't have it
but if you don't want to validate args, it's not required
stevan nothingmuch: wait I am not sure I am understanding you
nothingmuch: can I see the updated patch?
nothingmuch my problem is what do i do when someone says something like [ set_always => "blah", [ ] ]; when they want blah to return an array ref? 20:34
sure
unified diff? or should I darcs record?
stevan unified is fine, I speak diff :)
nothingmuch ok, the patch is on nothingmuch.woobling.org/Test-MockObject 20:36
and it depends on the script patch being applied too
the test is very self explanatory 20:37
it's the last part of the diff
stevan nothingmuch: what did chromatic think about adding Test::Deep ? 20:39
nothingmuch i don't know, i just got a reply from him about the original patch 20:40
he is considering whether it's better placed in a subclass or not. With Test::Deep and argument validation, i reckon it is
even though it's pretty tightly coupled to the internals
without it, I think it's a compact enough change, but it's up to him to decide 20:41
stevan nothingmuch: I think he might have a point actually
I did something similar to this in DBD::Mock, i called it DBD::Mock::Sessions
but DBD::Mock is highly aware of DBD::Mock::Session objects, so they are tightly coupled
nothingmuch i considered it at first 20:42
but i didn't want to release something this small without checking if he wants to integrate it beforehand 20:43
i'll be happy either way
(although happier if i'm not the maintainer ;-)
stevan nothingmuch: I actually think you could expand this more :)
nothingmuch how?
stevan Test::MockObject::Scripts
nothingmuch i tend to be spooked out of advanced features
the optional argument validation param is already starting to feel not right for me 20:44
even though it's a feature I need, the interface needs work
stevan nothingmuch: the first thought that comes to mind is reading a script from a config file
nothingmuch complicating it further is not something i think i want to do
ah, in that case
just store YAML in your __DATA__ and unserialize that
a script is an array of array refs, is all
stevan nothingmuch: yeah something like that
nothingmuch: I kind of see it as a way to automate Test::MockObject 20:45
r424 just commited, we now have 1003 tests in the suite :P 20:48
nothingmuch sorry, back 20:50
stevan ingy: you still around?? 20:51
I have a kwid question 20:52
Limbic_Region wonders how buu got on with his kwid parser 20:55
jdv79 he still didn't finish his search.cpan replacement
Limbic_Region he asked me to help him last night 20:59
I offered to help today
he hasn't pinged me
to be honest, I don't think a custom parser is the way to go given the grammar definition - but he asked 21:00
stevan kinda likes POD actually 21:01
stevan will give KWID a try though 21:02
jdv79 yeah, what's wrong with POD?
stevan jdv79: ask ingy :)
Limbic_Region blame Ingy
wolverian POD is not semantic enough. 21:04
ingy hola 21:05
jdv79 i don't think it was meant to be
ingy nothing is wrong with POD
Kwid is just much nicer
stevan ingy: whats the status of the kwid parser/converter stuff?
I was going to document Test.pm, and I figured it best to use kwid (if that is the official PUGS format that is) 21:06
ingy I'm still working on the systems to install it
stevan: go ahead
it is easy to convert between the two
or do it in POD 21:07
jdv79 lossy though i guess
stevan ingy: didnt you and autrijus hack a kwid -> HTML converter?
ingy not lossy
stevan: yes, but it doesn't really work yet ;)
patience perl6 programmer 21:08
stevan ingy: ok, I will just wing it then :)
ingy: rt.openfoundry.org/Foundry/Project/...n/?rev=426 21:09
ingy stevan: cool. I think you will really enjoy writing Kwid. I do. 21:10
did you see docs/yaht.kwid? 21:11
jdv79: read lib/Perl6/lib/perlkwid.kwid
it is a transcription of perlpod.pod 21:12
as such it really pushes the language
Limbic_Region . o O ( with little white lies )
ingy ;)
Limbic_Region I was going to ask buu if he had many examples to test against but he hasn't shown up - in the off chance I still agree to help him - where should I look for examples ingy? 21:14
ingy Limbic_Region: grep the pugs source for kwid files 21:17
metaperl the YAHT is writtein kwid
pugs/docs
metaperl returns to work
(which means leaving home, where he is now)
mugwump puts on a cheesy grin 21:18
"Try my new documentation formatting system. I think you might like it!"
mugwump drinks a glass of cranberry juice
stevan ingy: r428 has the Test.pm kwid doc in it, can you tell me if it looks okay? 21:37
ingy sure 21:39
jdv79 is he modeling this after his POD parser?
ingy you mean me? 21:40
:P
jdv79 buu i meant
ingy ah
jdv79 he wrote a POD parser a few weeks ago
stevan is he writing it in perl5?
jdv79 the POD one was 21:41
ingy created docs/src/Junc.kwid for metaperl 21:42
trivial
but so much nicer to read
metaperl_ I dont want to work in kwid 21:43
buu heh
ingy why not?
jdv79 yeah
is your Kwid parser gona be like the POD one you did?
ingy stevan: looks ok 21:45
metaperl_ jdv79, i dont follow your comments about the POD parser... what POD parser are you talking about?
ingy stevan: the authors section will word wrap though
gotta run& 21:46
metaperl_ ingy, if autrijus says all docs in kwid, then so be it
but we need one format, be it POD, kwid or whatever and then we all use it
jdv79 buu wrote a POD parser a montth or 2 ago. now i hear he is gonna buld a kwid parser. i'm wondering if they are going to be similar.
metaperl_ I like head1 head2 head3 better than = == ===
autrijus is not around today is he? where is he physically these days? 21:47
stevan metaperl: autrijus is sleeping now (i assume)
buu I return!
metaperl_ is he in Taiwan?
well anyway, I think we need to have a collaborative decision on kwid vs. pod 21:48
buu jdv79: Technically speaking, I didn't really write a parser, I just wrote a wrapper around Pod::Parser
metaperl_ with autrijus' vote being authoritiative
jdv79 buu, punk2
buu =[ 21:49
metaperl_ i'm gong to post the RFC on pugs docs to perl6.compiler... ok?
obra autrijus may still be in the PRC for a conference
clkao yes he is. 21:50
metaperl_ oh, that's why he was in a hotel, etc.
I see
clkao there's going to be a mini perl lunch today i belive
stevan metaperl: I agree that POD is not quite enough, but I am not sure of kwid either 21:53
mugwump this is because deep down, something about wikiformat makes your gut wrench :) 21:54
stevan my problem with wiki format is that every wiki seems to do it slightly differntly 21:55
much like every POD parser
buu stevan: Kwid obviously needs some kind of formatter guidelines.
mugwump I think that POD should look a bit like Perl
buu As for Kwid, how would you specify an item / definition list so it would be very easy to parse?
As in, perlfunc, or perlvar
metaperl_ what is wrong with POD? 21:56
we all know it, for one
buu Ask ingy.
I kind of like kwid better, just for formatting text.
stevan metaperl: it does have it's limits, and depending upon which parser you are using it can be very unforgiving
mugwump POD fills a gap fine, in that it is deliberately simple and so easy to make work with editors, parse, etc 21:57
kwid looks like a refinement of POD based on experience on Wikis
theorbtwo Mmm... I don't much care for kwikd, and like POD much more.
metaperl_ what did the perl6 community decide would be the doc format for perl6?
edgewalker I don't particularly like kwid 21:58
metaperl_ SGML? :)
theorbtwo I think POD just needs a bit of a cleanup with L<> semantics.
But I also think it's my bedtime.
edgewalker what theorbtwo says
stevan metaperl: I dont think the perl6 community has decided
buu metaperl_: The perl community rarely decides these things, if you haven't noticed
edgewalker also, POD needs some cleanup with regards to list markup
metaperl_ POD needs a CPAN tag or something like that (like you were saying about L<>)
mugwump the perl community is not a sentient being
stevan metaperl: I think autrijus and ingy decided
buu metaperl_: What would cpan do?
mugwump it is a collective, so you can't talk about what "it" decides :) 21:59
buu stevan: I think ingy made it up and convinced autrijus =]
stevan buu: I think you are probably right
theorbtwo I think the perl community should just steal PM's bracket-link code.
buu theorbtwo: Dude, did you read kwid?
edgewalker I think the final decision for Perl6-the-Real-Thing is going to be with Larry and crew
stevan I think we should just all use XML :P
buu edgewalker: I doubt it. I suspect the final decision will be determined by how many use what.
mugwump p6l would be the place to discuss I guess
buu stevan: I was seriously considering it =[
stevan buu: if only to drive ingy more crazy right?
Coke_ pod is mentioned in one of the apocolypses that it will be treated in APOC 23 or some such, which is no where in site. 22:00
metaperl_ I think we should Microsoft Word
buu stevan: =]
stevan Wordstar !!!!!
metaperl_ that's my final offer
edgewalker LyX
metaperl_ Printshop Pro
buu brainfuck.
stevan and write the parser in BASIC
Coke_ I find kwid seems to be easier than POD for what I use POD for.
buu Coke_: That's because you don't have to parse it =[
stevan I just wrote a kwid doc for the Test.pm module, 22:01
its not bad, my hands wanted to type POD though
theorbtwo Everybody should just learn HTML already. ;)
Really, bedtime.
buu theorbtwo: Well, yeah.
rgs pod with kwid extensions ?
stevan Javadoc??
rgs python doc strings.
stevan I actually think POD could be re-worked and be fine 22:02
rgs iirc that was The Plan 22:03
stevan just clean up the ugly corners
I once tried to mix XML and POD, but that got really ugly
metaperl_ i have emailed autrijus and asked for a final decision 22:04
edgewalker well, what we really should be using is markdown, but that's another story
stevan metaperl: this might really be an issue for p6l
buu edgewalker: Mark down?
edgewalker buu: google it
stevan and then again, there is nothing to say PUGS can't support POD and KWID (it currently does)
buu edgewalker: I'm playing war3 =[ 22:05
metaperl_ I have used Text::Template to inline Perl code without manually pasting it into pod files.. Most of Parse::RecDescent::FAQ is done by text-template loading things in from outside files
edgewalker buu: then google it when you're done with war3. it's basically wiki "markup" on steroids, heavily influenced by the formatting of email, and reads as well in plaintext form as it does as HTML 22:06
buu Sounds like kwid.
Coke_ (don't have to parse it) ... so? It should be easy for the person WRITING it, not the person parsing it. =-) 22:07
buu Coke_: Well, yeah, I'm bitter.
edgewalker buu: except it's about 10x better than kwid
or any other wiki markup for that matter 22:08
anyway, the only complaints I have with POD: lists with many short items are awkward; L<> semantics are underspecified; and fixed width text can either have formatting *or* be prewrapped, but not both 22:11
other than those I like it better than most wiki markup; incl kwid
markdown is an order of magnitude more readable and writable, but markdown is also complex to parse 22:13
of course if Perl6 ships with a really good module for manipulating whatever documentation format is settled on, then the parsing complexity issue is moot
metaperl_ markdown, textile. there are a few out there 22:23
edgewalker textile is really nothing but a sparser way to write HTML 22:28
Limbic_Region kicks buu hard
buu limbic! 22:29
Limbic_Region I asked since 8AM EST for you/about you to help you with your kwid parser
edgewalker markdown is written with HTML in mind as well, but the markdown markup itself isn't HTML specific
Limbic_Region and you were no where around
metaperl_ in revision 430, does anyone have time to look at pugs/examples/junctions/3.p6? it isn't working as I expected 22:32
buu Limbic_Region: I was asleep =./ 22:36
Limbic_Region: Sorry.
Limbic_Region sleep is a "good thing" 22:45
buu heh
I have odd hours.
Limbic_Region in any acct, having read more of the spec - I think a proper grammar/lexer/parser is more appropriate
but - if you still want to roll your own - I am willing to help
buu Limbic_Region: Well, I kind of agree about the proper grammar thing, but I have my reservatios about it. Anyways, What I have so far: 22:46
jdv79 yaccin' it? 22:47
Limbic_Region I was thinking yapp 22:48
but yeah
buu erxz.com/pugdoc.zip
Note t/
jdv79 YAPP!
buu heh 22:49
jdv79 i've only heard of yacc
Limbic_Region is that link semi-static or should I forward myself a copy to work as well?
jdv79 isn't that your server buu?
buu It'll be there for a tleast a month
yes
until I forget what it is
and delete it
jdv79 someone has buu.com 22:50
ugh
buu YES
Limbic_Region jdv79 - see Parse::Yapp
buu they've fucking had it for like 7 years
and they did fuck all with it
jdv79 i hate that
buu It used to be this like one page about a health club
jdv79 took me 3 years and almost $500 to get my lastname domain
buu =[ 22:51
jdv79 PITA
buu No wait, that was buu.org
theorbtwo Somebody bought mastros.com about 3 hours before my father finally decided to buy it.
buu buu.com is being scalped
They've been wanting like 3 grand for it for 5 years now
jdv79 hahaha
i got off cheap
theorbtwo So he's mastros.biz, and printed off all sorts of busniess cards, stationary, etc.
buu heh
Reasonable 22:52
theorbtwo ...and of course, mastros.com becomes available again a year later.
buu Your tale is one of woe and tragedy
theorbtwo Actually, my tale is one of "eh, whatever". 22:53
buu Limbic_Region: If you really care I can set up a svn server
Limbic_Region no - just will likely do most of any work I contribute during the day and not in the evenings 22:55
I can't sleep if I am doing programming of any complexity after 7pm 22:56
buu heh
jdv79 Parse::Rec... isn't good Limbic_Region?
buu Maybe I should consider that
Limbic_Region jdv79 - there isn't anything wrong with P::RD 22:57
well, other than it is slow
but some grammars are more suited for different kinds of parsers
buu Limbic_Region: The basic structure of Parser.pm should be fairly obvious 22:58
Limbic_Region yeah - it is a lot cleaner then you lead me to believe
buu Well, I thought about it for a whilke
My next problem is how to deal with .list. regions
Limbic_Region well, like I said - not going to play with it until tomorrow 22:59
buu Eh, well, ok 23:00
Limbic_Region one thing I would like to do if possible is determine which regex to apply and apply it in only two steps
jdv79 is it really worth rolling it like that?
buu Yes!
Limbic_Region a la dispatch table - instead of for looping over all of them
buu Limbic_Region: Er, how? 23:01
Limbic_Region well - you need to join all those regexes together with capturing parens
buu hrm
Limbic_Region and then use whatever is captured as the key
but the trouble is it isn't fixed key still
buu I have to go do an errand, I'll be back in 10.
Limbic_Region won't 23:02
oh - 10 minutes
not 10pm
The dispatch table is still a possibility but probably shouldn't try and pursue it until more of the grammar is covered since it might not fit in as easily 23:04
buu Limbic_Region: Yeah. Maybe. 23:05
I see what you're getting at, but I'm not sure it's worth it. 23:06
Anyways, dropping sister off, we shall discuss in a bit.
23:26 edgewalk1r is now known as edgewalker
Limbic_Region notes that he will give buu another 15 minutes before wandering off himself 23:42
www.perlmonks.org/index.pl?node_id=431364 23:48
for instance
grrr - wrong channel - sorry 23:49
Limbic_Region wanders off 23:59

Powered by Cro and the Raku® Programming Language.

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

Thank you!