* teknomunk is away: Gone away for now.
% teknomunk is now known as teknomunk_afk
<Tene> Almost got it working, but now I'm getting:
<Tene> error:imcc:The opcode 'callmethodcc_p_i' (callmethodcc<2>) was not found. Check the type and number of the arguments
<Tene> on a line containing just ".end"
<Tene> Which means... somewhere in there I'm trying to call a method on an integer?
<nopaste> "tene" at 67.137.148.11 pasted "expression parsing for lolcode" (211 lines) at http://nopaste.snit.ch/12268
<Tene> Whoops, problem with that paste.
<nopaste> "tene" at 67.137.148.11 pasted "expression parsing for lolcode, try #2" (212 lines) at http://nopaste.snit.ch/12269
<Tene> That's better.
<Tene> One problem here... I'm trying to call the given sub with the given arguments, which I have in an array... I can pop them out and call the sub with an array of the arguments, but that's not the same...
<Tene> There's no $P0(|$P1) in pir...
<Tene> It goes away when I remove the calls to .arity()
<Tene> Compiles fine when I story the result of .arity() in a pmc instead of an int.
<Tene> now I'm getting a null pmc access, but if_null says that it isn't null... and it was working fine before...
<Tene> Okay, all errors fixed.  Now I'm down to the "How to call a function with args stored in an array" problem... time to grep docs/
<Tene> Looks like set_args...
% alvar has joined #parrot
% alvar has left alvar!~alvar@sub87-230-120-57.he-dsl2plus.de
% alvar has joined #parrot
% alvar has left alvar!~alvar@sub87-230-120-57.he-dsl2plus.de
% alvar has joined #parrot
% iblechbot has joined #parrot
* Tene finds .arg and .call
* Tene notices that he's being needlessly verbose in #parrot and stops.
<Tene> Okay, it works, except for return values in functions.  That can come later.  I'm done for tonight.
<Tene> purl: parrotbug
<purl> parrotbug is mailto:parrotbug@parrotcode.org or http://svn.perl.org/parrot/trunk/docs/submissions.pod
<Tene> Patch sent.  RT#50492
<Tene> Sorry for all the spam.
% IllvilJa has joined #parrot
% barney has joined #parrot
<spinclad> Tene: nothing wrong that i see with thinking out loud in a quiet channel
<svnbotlt> r25461 | bernhard++ | trunk:
<svnbotlt> : [Rakudo]
<svnbotlt> : Clean up perl6.o
<svnbotlt> diff: http://perlsix.org/svn/parrot/revision/?rev=25461
<Tene> Oops, looks like I made that diff from languages/lolcode, not from parrot root.
% HG` has joined #parrot
% ruoso has joined #parrot
% slightlyoff has left slightlyoff!~slightlyo@204.14.154.209
<Tene> Oh, looks like my patch causes test failures, too.
<Tene> Need to work on that. >.>
<Tene> Okay, looks like the big issue is with the special handling of VISIBLE
<Tene> Need to get varargs support in there too, so VISIBLE can be just another function.
% alvar has left alvar!~alvar@sub87-230-120-57.he-dsl2plus.de
<svnbotlt> r25462 | bernhard++ | trunk:
<svnbotlt> : [Rakudo]
<svnbotlt> : Fiddle with perl6/Makefile.
<svnbotlt> : - Use make variable $(EXE)
<svnbotlt> : - Remove setup of unused commands
<svnbotlt> : - Let 'make help' tell about the targets 'perl6' and 'xmas'.
<svnbotlt> : - Use $(RM_F) instead of $(RM_RF).
<svnbotlt> diff: http://perlsix.org/svn/parrot/revision/?rev=25462
% IllvilJa has left IllvilJa!~jilves@emea-netcache1.oracle.co.uk
% AndyA has left AndyA!~andy@82.152.157.85
% AndyA has joined #parrot
% cognominal_ has left cognominal_!~cognomina@82.67.232.89
% cognominal_ has joined #parrot
% alvar has joined #parrot
% turrepurre has left turrepurre!~chatzilla@90-229-128-103-no52.tbcn.telia.com
% kjs_ has joined #parrot
% contingencyplan has left contingencyplan!~contingen@cpe-76-186-27-146.tx.res.rr.com
% contingencyplan has joined #parrot
% DarkWolf84 has left DarkWolf84!~dwolf@89.215.234.147
% AndyA has left AndyA!~andy@82.152.157.85
% AndyA has joined #parrot
<svnbotlt> r25463 | bernhard++ | trunk:
<svnbotlt> : #39992: [TODO] t/op/01-parse_ops.t: figure out how to test parsing of pmc constant parameters
<svnbotlt> : Mention this, the original, ticket in comment.
<svnbotlt> diff: http://perlsix.org/svn/parrot/revision/?rev=25463
% DarkWolf84 has joined #parrot
% barney has left barney!~bernhard@p549A2A30.dip0.t-ipconnect.de
% jq has left jq!~jquelin@merlin.mongueurs.net
% jq has joined #parrot
<svnbotlt> r25464 | coke++ | trunk:
<svnbotlt> : [rakudo]
<svnbotlt> : Spelling fix; minor output consistentifcation.
<svnbotlt> diff: http://perlsix.org/svn/parrot/revision/?rev=25464
% jq has left jq!~jquelin@merlin.mongueurs.net
% jq has joined #parrot
% Coke has left #parrot
<svnbotlt> r25465 | kjs++ | trunk:
<svnbotlt> : [cardinal] fix grammar a bit based on ruby's parse.y file. Major todo is fixing the lexer part; ruby's lexer is cheating.
<svnbotlt> diff: http://perlsix.org/svn/parrot/revision/?rev=25465
% Coke has joined #parrot
% jq has left jq!~jquelin@merlin.mongueurs.net
% jq has joined #parrot
% alvar has left alvar!~alvar@sub87-230-120-57.he-dsl2plus.de
% Andy has left Andy!~Andy@64.81.227.163
% Nicol has joined #parrot
<svnbotlt> r25466 | kjs++ | trunk:
<svnbotlt> : [cardinal] make end-of-line backslash work
<svnbotlt> : + add some tests.
<svnbotlt> diff: http://perlsix.org/svn/parrot/revision/?rev=25466
% Nicol has left #parrot
% guru has joined #parrot
<svnbotlt> r25467 | kjs++ | trunk:
<svnbotlt> : [cardinal] 
<svnbotlt> : * add a few operators
<svnbotlt> : * distinguish between locals, globals and instance vars.
<svnbotlt> : * distinguish between integers and floats.
% Andy has joined #parrot
% Maddingue has left Maddingue!~Maddingue@profane.mongueurs.net
% Maddingue has joined #parrot
% jq has left jq!~jquelin@merlin.mongueurs.net
% marmic has left marmic!~chatzilla@89-253-66-101.customers.ownit.se
<svnbotlt> r25468 | kjs++ | trunk:
<svnbotlt> : [cardinal] 
<svnbotlt> : * add grammar support for arrays
<svnbotlt> : * add a "new" method to "array" class; this should be completed.
<svnbotlt> r25469 | jonathan++ | trunk:
% marmic has joined #parrot
% GeJ has left GeJ!~geraud@edna.nealab.net
% dalek has left dalek!dalek@feather.perl6.nl
% GeJ has joined #parrot
% dalek has joined #parrot
<svnbotlt> r25470 | petdance++ | trunk:
<svnbotlt> : localizing vars
<svnbotlt> diff: http://perlsix.org/svn/parrot/revision/?rev=25470
% Andy has left Andy!~Andy@64.81.227.163
<svnbotlt> r25471 | petdance++ | trunk:
<svnbotlt> : localizing & consting
<svnbotlt> diff: http://perlsix.org/svn/parrot/revision/?rev=25471
% alvar has joined #parrot
<svnbotlt> r25472 | jonathan++ | trunk:
<svnbotlt> : [rakudo] Allow classes whose last declaration is an attribute to compile.
<svnbotlt> diff: http://perlsix.org/svn/parrot/revision/?rev=25472
% lili20 has joined #parrot
% lili20 has left lili20!~lili20@ANantes-257-1-65-24.w90-25.abo.wanadoo.fr
% jq has joined #parrot
<zev> are nqp tests still useful right now?
<Tene> zev: eh?
<peepsalot> is there any good documentation for perl 6?  should I be reading the synopsis or what?  i'm coming from NOT knowing perl 5
<zev> when I last asked what needed doing, I was told nqp tests.  Several of them are now done.  I would do the method ops tests, but I'm not quite sure what that means.
<kjs_> peepsalot: there's a set of articles on perlbuzz.com
% barney has joined #parrot
<kjs_> micro articles they're called i think; theyre' nice intros
<mdxi> perlbuzz++
<peepsalot> kjs_, you mean this? http://perlbuzz.com/2008/01/take-a-nibble-of-perl-6-with-a-microarticle.html 
<shorten> peepsalot's url is at http://xrl.us/bfmqc
<peepsalot> oh, didn't see you just said micro article. ok
<peepsalot> they're really at perlfoundation though http://www.perlfoundation.org/perl6/index.cgi?perl_6_microarticles
<shorten> peepsalot's url is at http://xrl.us/bfmqe
<jonathan> zev: NQP is being used heavily, so of course texts for it are a good thing.
<jonathan> *tests
<jonathan> As for method ops tests - I guess that's method calling.
<svnbotlt> r25473 | jonathan++ | trunk:
<svnbotlt> : [rakudo] Have the new method on objects initialize the attributes to Undef or the supplied value. Note we aren't passing named parameters yet, so can't really use that bit. It does, however, let us mutate attributes through the accessor (we're always is rw by default at the moment, though that will change).
<svnbotlt> diff: http://perlsix.org/svn/parrot/revision/?rev=25473
<zev> jonathan: just testing that methods can be called on objects (ie. $foo.bar)?
<jonathan> zev: Yes, I think that's pretty much the only method operation supported in NQP.
<jonathan> If you want to clear up after me, the =:= operator I put in yesterday could use a test too. ;-)
<zev> heh, ok
* jonathan is messy.
<peepsalot> are there any docs for nqp?
<kjs_> peepsalot: nqp is Perl 6 (subset)
<peepsalot> great but how do i know what that subset is
<kjs_> heh. good point
<kjs_> i guess look into the grammar
<kjs_> nqp/src/parser/grammar.pg or so
<jonathan> Also there is README.pod which gives a vague idea.
<kjs_> and study existing nqp scripts: languages/{perl6,ecmascript,lua,pynie,punie,cardinal}
* jonathan has written quite a bit in NQP this weekend and likes it.
<zev> kjs_: the features supported are those features that are useful for implementing Rakudo
<kjs_> yes
<kjs_> in fact, the features of nqp are selected based on that, more or less
<kjs_> not especially for rakudo
<kjs_> but for using nqp in building compilers
<zev> oh, that was meant for peepsalot
<kjs_> ooh it's not a question :-)
<kjs_> parse error ;-)
<zev> well, it was my logic error :)
<kjs_> parse error ;-)
<kjs_> woops
<kjs_> too many windows...
<peepsalot> yeah, i'm interested in helping writing a compiler, but I'm coming from not knowing perl at all
<kjs_> neither am I :-)
<kjs_> i think I'm the biggest perl noob around here
<peepsalot> so i've been scrounging around the internet trying to frantically get up to speed
<kjs_> just use it seems to be the only way to get up to speed
<peepsalot> but there's a lot of incomplete and out of date perl6 documentation out there
<kjs_> well, the docs that are up to date are the synopses
<peepsalot> ok
<zev> and the latest synopsis can be found at perlcabal.org
<kjs_> but I'm not sure to what extent they're useful to a beginner
<kjs_> dunno, I always find them through perl.org -> perl 6 ->synopses
<peepsalot> well, i tried just reading through some .pgs first, but it wasn't making much sense.  maybe now that I know a couple basic operators it will mean a little more
<kjs_> you can start out your own little language to experiment; using the script in tools\dev\mk_language_shell or something
<kjs_> it 'll generate a complete working compiler including 1 test :-)
<Tene> peepsalot: This is an interesting problem we've never had before.  :)
<peepsalot> Tene, which problem?
<Tene> peepsalot: a willing contributor who doesn't know perl.
<peepsalot> hehe
<Tene> peepsalot: what sort of documentation are you looking for to learn from?
<peepsalot> i just like the idea of the parrot vm supporting a variety of languages, and want to help with ecmascript if I can
<kjs_> i did some work lately on ecmascript
<peepsalot> Tene, just whatever i need to understand the process of writing compilers for parrot
<kjs_> i never programmed in it , or any javascript variant
<Tene> peepsalot: personally, I'd read through compilers/nqp/t/*.t if you're looking to get a feel for nqp, and ask about any parts that look odd.
<peepsalot> ok
<Tene> Is probably the fastest way to get familiar with nqp, at least it was for me.
% jq has left jq!~jquelin@merlin.mongueurs.net
% jq has joined #parrot
<peepsalot> hmm, i'm a little confused, i thought the /languages dir contained the definitions for compilers 
<kjs_> languages/ contain all languages targeting parrot
<Tene> peepsalot: /compilers is for internal parrot compilers, /languages is for... yeah, that.
% schmalbe has joined #parrot
<peepsalot> and an internal compiler compiles what into what?
<kjs_> depending on the compiler ;=)
% silug has left silug!~steve@38.104.162.2
% silug has joined #parrot
<kjs_> the most important one is imcc
<kjs_> which compiles pir into parrot bytecode
<kjs_> there's nqp, who does the nqp subset of perl6 into PIR
<Tene> peepsalot: nqp was originally in /languages, but was moved to /compilers because it's mainly for building things for parrot, not for general use.
<peepsalot> so there is not really any difference in the way they are implemented?  just a sort of categorization?
<jonathan> It maybe is fair to say that compilers/ is stuff that will be shipped with a core Parrot without support for any "real" languages, and what's in languages may live in their own repositories one day and are implementations of languages for everyday people to use.
<peepsalot> ok
% barney has left barney!~bernhard@p549A25EB.dip0.t-ipconnect.de
<jonathan> So we'd not really expect "end users" to be writing PIR or NQP generally, but we would expect them to be writing in Ruby, Perl 6, etc.
% alvar has left alvar!~alvar@sub87-230-120-57.he-dsl2plus.de
% guru has left guru!~guru@74.12.142.176
<svnbotlt> r25474 | kjs++ | trunk:
<svnbotlt> : [cardinal]
<svnbotlt> : * add unless statement
<svnbotlt> : * add while statement
<svnbotlt> : * there's something weird going on with 'then' and 'do' rules. can't figure it out.
% schmalbe has left schmalbe!~bernhard@p549A1F0A.dip0.t-ipconnect.de
* Tene debates between "Prepare for trip" and "hack on cardinal"
* jonathan votes for cardinal
<jonathan> Preparation can be sqeezed into as short a time as is available.
<jonathan> As witnessed at various conferences. :-)
<kjs_> Tene: you're a ruby programmer?
<Tene> kjs_: nope, never touched it.
<kjs_> neither have I 
<kjs_> yet I'm writing a parser for it :-P
<kjs_> I thought Perl's parser was known for its complexity, but Ruby's not much better
<Tene> kjs_: shouldn't instance variables be :scope('package') ?
<kjs_> instance variables are object attributes, no?
<Tene> right.
<jonathan> :scope('attribute')
<kjs_> jonathan: yeah that's what I was proposing
<kjs_> or does it exist already
<kjs_> ?
<jonathan> kjs_: Yeah, I implemented it.
<Tene> jonathan++
<kjs_> oh cool
<kjs_> jonathan++
<jonathan> And various other bits of OO support in languages/perl6/
<kjs_> I just sent an email earlier today
<svnbotlt> r25475 | kjs++ | trunk:
<svnbotlt> : [cardinal] fix a bit
<svnbotlt> diff: http://perlsix.org/svn/parrot/revision/?rev=25475
<jonathan> I will go document it in the PAST PDD now, sorry for forgetting.
<kjs_> np; did you see my email with the proposal?
<jonathan> Just been madly hacking on getting the basics of Perl 6 OO stuff in place (got classes, attributes, accessors, roles, inheritance and role composition, though the second two have an ordering issue)
<Tene> Although, for that matter, /me also debates between cardinal and fixing the rest of the lolcode expression support patch.
<jonathan> kjs_: Let me take a look...
% alvar has joined #parrot
<kjs_> Tene: if you do start hacking cardinal, please gimme a ping, otherwise we'll have conflicts really soon
<jonathan> kjs_: Was there an attributes one, or do you mean the "register/result node referencing (or something)" one?
<kjs_> maybe the attr one didn't come through yet
<kjs_> sent it 2 hours ago or so
<jonathan> Yeah, don't see it in my mail.
<kjs_> or 1 hour? I don't keep track of time
<kjs_> ok; well if it's implemented already, great :-)
<jonathan> Appears to kinda work too, even.
<kjs_> jonathan: the reg/result thing is related to the problem you had the other day
<kjs_> if you remember
<kjs_> earlier we had a short discussion about evaluating some expression once, and re-using the result multiple times
<jonathan> kjs_: Yeah, though I was taking the wrong approach to that particular problem so won't need this in the end for that. *But* I agree that we need something like what you're suggesting.
<jonathan> As there will surely be a range of use cases for it.
<kjs_> yeah.
<jonathan> As you point out.
<jonathan> :-)
<kjs_> did you also encounter the use for handling self?
<kjs_> or "this"
<kjs_> (in other languages such as java)
<kjs_> for instance. self.foo = 42
<jonathan> Hmmm, hadn't thought about implementing that yet.
<kjs_> not sure, but i was thinking about an :inline('%r = self'), as self is already known in PIR
<jonathan> Was hoping we could just somehow emit a "self" at a PIR level.
<kjs_> but it might not always work
<kjs_> right, that'd be better/nicer
<kjs_> as most languages support OO one way or the other
<jonathan> Grrr. Can't see in STD.pm where "self" gets parsed.
<kjs_> does perl6 have a self?
<jonathan> Yeah.
% avar has left avar!avar@hlagh.mtfnpy
<jonathan> I'm not sure what it is classed as syntactically, though.
<kjs_> often it is called a pseudo-variable
<jonathan> "You may always access the current invocant using the keyword self."
<jonathan> Grr. Later on it refers to it as a function.
<kjs_> jonathan: :scope('attribute') assumes 'self' as the object if you don't specify a PAST node as the object?
% avar has joined #parrot
<TimToady> jonathan: I will add self as term:sym<self>, though it can also be thought of as simply a 0-ary function
<jonathan> TimToady: OK, thanks. Will have a look to see if I can get it to parse/work.
<jonathan> kjs_: Let me check... ;-)
<kjs_> (updating the new document about past nodes usage)
<jonathan> kjs_: Erm, it always operates on self at the moment.
<kjs_> ooh ok
<kjs_> well, it's a good default :-)
<jonathan> Sure, but it should probably take another optional node to supply self.
<kjs_> yeah
<jonathan> kjs_: If you have time, please pop a ticket in for that. Feel free to assign it to me.
<kjs_> you reckon it's a lot of work?
<kjs_> i mean, implementing the child evaluation
<jonathan> I don't think so.
<kjs_> I'll have a look myself now
<jonathan> But I need to go eat, and have other things to sort out, so maybe can't do much more tonight, and don't want to forget it. ;-)
<jonathan> Sure, if you can hack it yourself, even better! :-)
<jonathan> compilers/ptc/past/Compiler.pir
<kjs_> ok; i'll have a short look, but I don't expect I can fix this. we'll see. otherwise I'll file a ticket
<jonathan> search for 'attribute'
<jonathan> With the quotes
<kjs_> great, thnx
<svnbotlt> r25476 | kjs++ | trunk:
<svnbotlt> : [docs] update past_building_blocks.pod with :scope('attribute')
<svnbotlt> diff: http://perlsix.org/svn/parrot/revision/?rev=25476
<jonathan> It was my first time hacking on the PAST compiler.
<kjs_> yeah', it's a bit scary :-)
<jonathan> TimToady: Not sure if it works as a term.
<jonathan> <noun> <postfix>*
<jonathan> That's what parses method calls.
<jonathan> Inside term
% jq has left jq!~jquelin@merlin.mongueurs.net
<jonathan> Oh, wait. A term is considered a noun.
<jonathan> In STD.pm.
% jq has joined #parrot
<jonathan> Wonder what happens if I do this in the Rakudo grammar... :-)
<jonathan> ...infinite recursion. :-)
% avar has left avar!avar@hlagh.mtfnpy
<kjs_> jonathan: step 1 is completed: the object is retrieved, now use it instead of 'self'
<kjs_> holy cow, I think it works already :-P
<jonathan> Nice!
<kjs_> I'll send a patch, then pmichaud or you can have a look 
<jonathan> kjs: I've got self working...
<jonathan> class Foo { method bar { say self.baz; }; method baz { "hi"; }; };
<jonathan> my $x = Foo.new(); $x.bar();
<jonathan> hi
<jonathan> But it's using a call to self every time at the moment, and %r
<kjs_> cool :-)
<jonathan> It's a start though.
<kjs_> right
<kjs_> like $P10 = self
<kjs_> and then use $p10 
<jonathan> Yeah, that's pretty much what it emits.
<jonathan> Yeah.
<jonathan> I tried a couple of other things.
<kjs_> yeah. That's also what I implemented in ecmascript i think
<kjs_> (or pynie? I forgot )
<kjs_> I created a patch just now, will send it to parrotbug for review
<jonathan> Bwaha. I got it to emit self.methodname() now
<jonathan> But it's REALLY sick.
<jonathan> $past := PAST::Val.new( :value('self'), :returns('Integer'), :node( $/ ) );
<kjs_> eeeuhw :-P
<particle> why Integer?
<jonathan> Erm, because it makes it not quote it...hang on...what if...
<Tene> 33
<kjs_> if you leave out :returns, it'll be quoted i think
<jonathan> Yeah. Anything other than Integer or Float will quote it.
<jonathan> Or do something odd.
<particle> that's a missing feature, then
<Tene> ECATONKEYBOARD
<particle> pct should recognize self, and treat it special
<particle> it is a parrot primitive
<jonathan> Yeah, but what if you actually have the string "self".
<jonathan> Need to disambiguate it somehow.
<kjs_> tthen :returns("string") ;-)
<particle> ok, how about :returns('self')
<jonathan> What if someone creates a type called self? ;-)
<particle> i think we should modify pct to deal with self
<kjs_> particle: returns('self') would be a bit inconsistent
<jonathan> Agree; just need to find the Right Way.
<particle> yeah,
<kjs_> :returns attribute takes a type
<particle> so maybe a new flag?
<particle> :self(1)
<kjs_> in what cases do we need to refer to 'self'?
<particle> is self better represented as a var or a val?
<jonathan> Since PAST::Val is so close to doing the right thing anyway...
<kjs_> good q.
<jonathan> Plus I can't think of a situation where assigning to self makes much sese....
<jonathan> "Oh hai, I can change my invocant!"
<kjs_> oh I think that can be done in some languages?
<jonathan> Oh?
<particle> yep, i think so too
<kjs_> never tried...
<jonathan> Wow.
<kjs_> java: this  = new Integer(42); ?
<kjs_> (should be in class Integer of course)
<jonathan> I've never tried that...if it works, wow, that's sick. :-)
<kjs_> never tried myself
<kjs_> don't havve java here
<kjs_> no javac, that is
<jonathan> I haven't used Java since I had to. At uni.
% DarkWolf84 has left DarkWolf84!~dwolf@89.215.234.147
<particle> what about perl 5's "my $invocant = shift;"
<kjs_> I think it's legal pir to write "self = new "Integer" "
* jonathan is going to check in the non-sick version of self for now.
<jonathan> particle: Do you know what the status of getting named parameters and pairs into Rakudo is? I seem to have some vague recollection of you doing them for NQP...
<kjs_> yeah it works in PIR :-)
<jonathan> Named parameters in NQP anyway, not pairs...
<particle> i think i've gotten named params in to rakudo
<particle> if not, i can do
<jonathan> Passing by them too?
<particle> i think so, but i don't think they're creating Pairs
<jonathan> class Foo { has $x; }
<jonathan> my $x = Foo.new(:x("hi"));
<jonathan> Statement not terminated properly at line 1, near "(:x(\"hi\"))"
<particle> ok, perhaps not.
<particle> what about :$x('hi') ?
<particle> look for <colonpair>
<kjs_> jonathan: did you update pdd26 with the "attribute" scope?
<kjs_> if not, I'll do it
<jonathan> colonpair not found in grammar.pg
<jonathan> kjs_: Not yet, I got distracted doing self!
<kjs_> ok; will do it now
<jonathan> Thanks!
<particle> ok, colonpair was from nqp. maybe i haven't gotten it into rakudo yet :(
<jonathan> particle: Seems not.
<particle> ah, token parameter
<svnbotlt> r25477 | jonathan++ | trunk:
<svnbotlt> : [rakudo] Implement self. Note that we can't make it a term yet, where it belongs, so it's a noun for now.
<svnbotlt> diff: http://perlsix.org/svn/parrot/revision/?rev=25477
<particle> $<named>=[':'?] ...
<particle> yep, there's named params
<jonathan> Sure, but that's declaring subs that take named arguments, not passing the arguments themselves.
<particle> ah, right.
<jonathan> Which isn't quite what I said, but what I meant. :-)
<particle> ok, then i ned to look at...
<jonathan> Sure, no hurry, would just be cool to have.
<jonathan> Then we can do Foo.new( :blah('worrevaminga' )
<jonathan> )
<jonathan> Once the past -> post ordering issue is sorted out (or if it's not going to, I shuffle things around some), I'll blog about inheritance and role composition and stuff.
<jonathan> But don't really want to get people trying it until that is resolved.
<jonathan> At least I'll be able to run much of the Puppy example now though. :-)
<particle> ok. great. i won't have a chance to look at past->post ordering until after the superbowl, and when the giants win... tomorrow :)
<kjs_> jonathan: committed; can you read the diff and see if'ts any good?
<jonathan> particle: I've looked, but I don't grok it enough to work out the fix.
<svnbotlt> r25478 | kjs++ | trunk:
<svnbotlt> : [pdd26] add description of :scope('attribute')
<svnbotlt> diff: http://perlsix.org/svn/parrot/revision/?rev=25478
<particle> everybody remember to submit your oscon talks today!
<jonathan> kjs_: Looks fine.
* jonathan can't afford to fly over to Portland this year.
<kjs_> great, thanks.
<jonathan> Flights to the US in the summer are non-cheap.
<svnbotlt> r25479 | kjs++ | trunk:
<svnbotlt> : [cardinal] change :scope of instance vars to 'attribute', which is newly available.
<svnbotlt> diff: http://perlsix.org/svn/parrot/revision/?rev=25479
<svnbotlt> r25480 | kjs++ | trunk:
<svnbotlt> : [cardinal] change array constructor; a :slurpy param already initializes an array with the passed in value, so just return that param.
<svnbotlt> diff: http://perlsix.org/svn/parrot/revision/?rev=25480
% marmic has left marmic!~chatzilla@89-253-66-101.customers.ownit.se
% marmic has joined #parrot
% avar has joined #parrot
% Limbic_Region has joined #parrot
% tcoppi has left tcoppi!~thisnukes@dhcp-west-76.resnet.nmt.edu
% tcoppi has joined #parrot
% mire has joined #parrot
* teknomunk_afk is back.
% teknomunk_afk is now known as teknomunk
% Khisanth has left Khisanth!~Khisanth@pool-151-204-136-50.ny325.east.verizon.net
% DarkWolf84 has joined #parrot
% grim_fandango has joined #parrot
% zaphod has joined #parrot
<zaphod> after working with PAST for a couple of days there are a couple of things that don't seem optimal to me.  It might just be that I'm not getting the reasoning behind it though.
<zaphod> PAST::Block having 'immediate' and 'declaration' doesn't seem quite right
<zaphod> 'immediate' is to create a new scope (and immediately enter it)
<zaphod> 'declaration' kinda creates a scope, but doesn't enter it.  And you can't enter it until you perform a newclosure on the result
<zaphod> it seems like there should be 3 things: create a scope and enter, create a sub, create a closure
% flora29 has joined #parrot
% silug has left silug!~steve@38.104.162.2
% flora29 has left flora29!~flora29@ANantes-257-1-65-24.w90-25.abo.wanadoo.fr
% Khisanth has joined #parrot
% Khisanth has left Khisanth!~Khisanth@pool-151-204-136-50.ny325.east.verizon.net
% Khisanth has joined #parrot
<kjs_> zaphod: immediate is for scopes that are executed immediately, for instance the <block> of a while loop
<kjs_> zaphod: declaration is for scopes that are declared, can be assigned/binded to some identifier, and invoked later
<kjs_> by looking up the symbol with that same identifier.
<zaphod> well, declaration is halfway there.  It creates the .sub and sets the outer, but doesn't do the newclosure so you can't really use it later
<zaphod> until you perform a newclosure on the result of the declaration
<kjs_> but it is used currently by several languages
<zaphod> what I was suggesting is leave immediate as is and change declaration to do a .sub without an :outer
<kjs_> maybe I'm not clear on the newclosure stuff
<zaphod> have a closure type for doing a .sub with an :outer and the newclosure op
% Andy has joined #parrot
<zaphod> a PAST::Block with blocktype of immediate will currently create .sub with an :outer
<zaphod> that is fine until you try to call the block that was created.
<kjs_> yeah, because there is only 1 top-level Block node, and all "child" Blocks are related to their parents through the :outer flag
<zaphod> then, since there was no closure created for the block (a binding to the current lexical environment), you get that strange error that I was getting before (wrong number of args)
<kjs_> you have to set the :name flag, or assign the resulting block to some id
<kjs_> ooh, you mean with the "wrong"outer value?
<kjs_> (changing :outer("block11") into :outer("block10") did the trick
<zaphod> but then the block would be in the wrong scope
<kjs_> ok. well, I don't know enough about this stuff I guess..
<kjs_> why not send it to list?
<zaphod> ok, it does get a little involved :)
<kjs_> heh yeah, it's a bit complicated
<zaphod> I played around with PCT some more and added another compilation stage to the toy language that I'm working with to perform tailcall optimization.  Turned out to be pretty easy
<zaphod> except for the floundering I did to figure out how to do the optimization itself correctly :P
<moritz> zaphod: do you have a repostory online where we could view your stuff?
<zaphod> nope, sorry
<zaphod> is there somewhere that I can put it?
% iblechbot has left iblechbot!~iblechbot@ppp-62-216-197-152.dynamic.mnet-online.de
% Andy has left Andy!~Andy@64.81.227.163
<moritz> sourceforge?
<purl> i guess sourceforge is making money
<moritz> d'oh
<zaphod> I thought about that, but it seems odd to make a sf project for it.  I'll take a look though
<moritz> google code offers free svn for open source projects as well
<zev> does anyone know what the following means from nqp/TODO.pod: "We should be able to load an external module (Pm suggests Test.pbc)
<zev> and exercise methods using that.
<zev> "
<moritz> implemet "use $module;"?
<zev> I can't tell whether it's supposed to be a bullet point or attached to method ops above it
<zaphod> moritz: you should be able to see my code at http://code.google.com/p/parrotfun/source/browse
<moritz> zaphod: thanks
<moritz> All tests successful.
<purl> pass(); ok(1); is(1,1);
% wknight8111 has joined #parrot
% silug has joined #parrot
<moritz> zaphod: does FUN have any IO?
<zaphod> nope....
<zaphod> it is about as simple of a language as can be 
<moritz> ok
<zaphod> it just evaluates and prints the result
<moritz> it's scary how much I understand (in grammar-actions.pl)
<zaphod> :) glad to know that it is understandable to someone else
<moritz> I'm not all that familiar with PCT, but I have some solid p6 understanding
<moritz> so I couldn't tell if what you wrote is right or not, but I can guess what you did, and why
<zaphod> most of the explanations for that $() and family mean are cribbed from the abc compiler (I just copied and gutted it for the start)
<moritz> $($index) is just shorthand for $/[$index], isn't it?
<moritz> at least in ordinary perl 6
<zaphod> In nqp $( $index ) would be the AST for $index and $/{$index} would be the parse(?) for $index
<zaphod> I was doing a bit of cargo-cult programming for that stuff :)
<moritz> from S05:2126: $() is a shorthand for $($/)
<moritz> that is $/ in item context
<zaphod> I'm not up on p6 yet...new contexts!
<moritz> just think "scalar context" ;)
<zaphod> ah
<moritz> which can be numeric, string or bool (iirc)
<zaphod> who woulda thunk it, reading through S05 gives some insight into what the nqp stuff is :)
<zaphod> I'm starting to understand all those things I think
<moritz> ;)
% Andy has joined #parrot
% Andy has left Andy!~Andy@64.81.227.163
% kjs_ has left kjs_!~IceChat7@ip565fd420.direct-adsl.nl
% Theory has left Theory!~Theory@dsl093-038-250.pdx1.dsl.speakeasy.net
% Andy has joined #parrot
% zaphod has left #parrot
% marmic has left marmic!~chatzilla@89-253-66-101.customers.ownit.se
% grim_fandango has left grim_fandango!~matt@bas2-kingston08-1096740991.dsl.bell.ca
% grim_fandango has joined #parrot
% jeffreykegler has joined #parrot
% marmic has joined #parrot
% HG` has left HG`!~wells@user-544731fa.wfd85a.dsl.pol.co.uk
% jeffreykegler has left jeffreykegler!~jeffreyke@cpe-69-204-164-42.nycap.res.rr.com
% itsme20F has joined #parrot
% contingencyplan has left contingencyplan!~contingen@76.186.27.146
% itsme20F has left itsme20F!~itsme20F@ANantes-257-1-65-24.w90-25.abo.wanadoo.fr
% jeffreykegler has joined #parrot
% contingencyplan has joined #parrot
<jonathan> afk, sleepz
% alvar has left alvar!~alvar@sub87-230-120-57.he-dsl2plus.de
% Theory has joined #parrot
<wknight8111> are there any win32 parrot hackers in here?
<wknight8111> the configure script keeps craping on the compiler test, and i can't figure out what options I need to make it work
<wknight8111> ...nevermind, i figured it out
% AndyA has left AndyA!~andy@82.152.157.85
<Coke> whee
% AndyA has joined #parrot
<wknight8111> cl.exe emits A LOT of errors when building parroty
<wknight8111> but it works fine
% Partizx has joined #parrot
% Theory has left Theory!~Theory@dsl093-038-250.pdx1.dsl.speakeasy.net
% Limbic_Region has left Limbic_Region!~Limbic_Re@c-68-49-236-220.hsd1.md.comcast.net
<svnbotlt> r25481 | petdance++ | trunk:
<svnbotlt> : consting a declaration in Parrot_new_INTVAL_hash
<svnbotlt> diff: http://perlsix.org/svn/parrot/revision/?rev=25481
% jeffreykegler has left #parrot
% Theory has joined #parrot
<wknight8111> what is the best source of information on porting a new language to parrot?
<svnbotlt> r25482 | rgrjr++ | trunk:
<svnbotlt> : * docs/running.pod:
<svnbotlt> :    + Add an "About runcores" section, based heavily on a post to p2 by
<svnbotlt> :      chromatic++ on Fri, 25 Jan 2008 11:03:12 -0800.
<svnbotlt> :    + Also, canonicalize wsp in "runcore" and "command line", change
% kid51 has joined #parrot
<svnbotlt> r25483 | petdance++ | trunk:
<svnbotlt> : Optimize two checks against 0 in list_splice
<svnbotlt> r25484 | petdance++ | trunk:
<svnbotlt> : consting, and fixed an instance of broken formatting
<svnbotlt> diff: http://perlsix.org/svn/parrot/revision/?rev=25484
% Andy has left Andy!~Andy@c-98-206-188-185.hsd1.il.comcast.net
<svnbotlt> r25485 | rgrjr++ | trunk:
<svnbotlt> :    + --debug is now --imcc-debug; make this more consistent with -D.
<svnbotlt> :    + Remove references to "IMCC" as a program.
<svnbotlt> :    + Add "-V" para.
<svnbotlt> :    + Explain that "Generated files" applies to JIT debugging.
<svnbotlt> r25486 | rgrjr++ | trunk:
<svnbotlt> : * compilers/imcc/cfg.c:
<svnbotlt> :    + (edge_count):  Fix a segfault in "-d 4" debugging output.
<svnbotlt> diff: http://perlsix.org/svn/parrot/revision/?rev=25486
<svnbotlt> r25487 | rgrjr++ | trunk:
<svnbotlt> : * src/pmc/sub.pmc:
<svnbotlt> :    + Oops; r19529 changed "nee" to "see" in flag descriptions when "n?\195?\169e"
<svnbotlt> :      was really what was meant.  Changed to "originally" for clarity.
<svnbotlt> diff: http://perlsix.org/svn/parrot/revision/?rev=25487
% kid51 has left kid51!~jkeen@70.107.13.232
% Andy has joined #parrot
% Ademan has joined #parrot
% wknight8111 has left wknight8111!~nobody@c-71-230-33-251.hsd1.pa.comcast.net
<svnbotlt> r25488 | chromatic++ | trunk:
<svnbotlt> : [IMCC] Update signature in yacc file to match signature in header.
<svnbotlt> diff: http://perlsix.org/svn/parrot/revision/?rev=25488
% DarkWolf84 has left DarkWolf84!~dwolf@89.215.234.147
% sjansen has joined #parrot
<svnbotlt> r25489 | chromatic++ | trunk:
<svnbotlt> : [IMCC] Minor cleanups and tidyings, mostly coding standards but some variable
<svnbotlt> : declaration relocations.
<svnbotlt> : This will be handy when trying to fix the memory leaks in IMCC, as well as to
<svnbotlt> : try to clean up the intent of the code as well.
% sjansen has left sjansen!~sjansen@75-169-87-14.slkc.qwest.net
% Ademan has left Ademan!~dan@h-68-164-171-226.snfccasy.dynamic.covad.net
% sjansen has joined #parrot
<Andy> oh, chromatic, why don't you show up?
<sjansen> Perhaps a memory leak has trapped him in an alternate dimension.
% grim_fandango has left grim_fandango!~matt@bas2-kingston08-1096740991.dsl.bell.ca
<svnbotlt> r25490 | petdance++ | trunk:
<svnbotlt> : Overhauled PIO_unix_write with localization, and renaming variables
<svnbotlt> diff: http://perlsix.org/svn/parrot/revision/?rev=25490
% uniejo has joined #parrot
% grim_fandango has joined #parrot
<svnbotlt> r25491 | petdance++ | trunk:
<svnbotlt> : The "unreachable code" warning is too noisy.  Demoted to a cage warning.
<svnbotlt> diff: http://perlsix.org/svn/parrot/revision/?rev=25491
% sjansen has left sjansen!~sjansen@75-169-87-14.slkc.qwest.net