svn switch --relocate svn.openfoundry.org/pugs svn.pugscode.org/pugs/ | run.pugscode.org | spec.pugscode.org | paste: sial.org/pbot/perl6 | pugs.blogs.com
Set by avar on 16 November 2006.
meppl guten morgen 02:21
svnbot6 r14893 | audreyt++ | * obra++ pointed out the repo links 404s on pugscode.org' overview section 05:11
svnbot6 r14894 | audreyt++ | * unTODO passing tests related to binding with array/hash elems as LHS. 05:35
gaal moose 07:19
ludan re 11:10
ofer0 re 11:16
andara hi all 14:53
anyone going to the POPL07 conference? 14:55
nwc10 I'm not, and I'm not sure if anyone is awake currently 14:59
andara nwc10: seems very quite indeed 15:01
fridim i've made : cpan -i Perl6::Pugs . the pugs version is : Version: 6.2.13. Is it really the last ? 17:14
andara fridim: that's the latest release, yes. You can get a more recent version from svn.pugscode.org/pugs 17:16
lambdabot Title: Revision 14894: /
fridim great :) 17:28
cj TimToady: we were talking about you 21:11
wolverian uh oh..
TimToady we were? 21:29
jdv79 will it be possible to build p5esque OO code under p6? 21:41
gaal just how esque? 21:42
TimToady yes, different kinds of objects can have different representation schemes and p5hash is one of them. It's just the default is p6opaque. 21:43
gaal I think the plan is to make the dispatcher pluggable too, so you can have a p5 @ISAy method resolution order if you like 21:44
TimToady obviously if two different representations are related by inheritance but have different schemes, you have to emulate that with some kind of delegation or composition.
jdv79 cause ai showed some p6 OO code to a colleague and he had a good point that its starting to look like java;) i think less structured/lighter-weight classes still have their uses.
gaal jdv79: are you looking for a class to which you can add fields on the fly? 21:44
jdv79 that's a large part of it, yeah. 21:45
gaal what else does the weight dimension relate to? perf?
nothingmuch and the other moose here have some ideas on making "everything is an object" not suck that way 21:46
TimToady as for the "this is starting to look like langage XXX" argument, that says nothing about whether we're stealing the right ideas or the wrong ideas from language XXX.
gaal * including attributes
jdv79 i guess what we mean is the way its very loosely coupled in p5 - methods can be called as subs, representation is dynamic, stuff like that.
TimToady you can still call a method as a sub--you just have to work at it a little more. 21:47
TimToady but there are really good reasons to keep subs and methods straight from each other by default. 21:48
gaal bbiab &
wolverian ..and that would probably be documented in apo12. I certainly like at least the visual documentation :) 22:07
gaal TimToady: is there a context-safe _value_ in Perl 6 for nothing? That is, one that if returned, looks like the callee had said "return", but without involving a control exception in the callee? 22:26
gaal I've wanted this in p5 in the past. perlmonks.org/?node_id=590133 is a p6 case 22:28
lambdabot Title: Re^2: 99 Problems in Perl6
diotalevi Say, this is a completely random question but is there likely to be a way to manipulate the call stack in p6? I just found myself in p5 wanting to either modify my caller's @_ and force it to restart or return a value for my caller. 22:30
gaal that's a scary thought, but I can see where it'd be useful 22:33
i wonder how far you could get with control exceptions in the general case 22:34
you want this w/o the cooperation of the caller, right? 22:35
diotalevi Incidentally, this was because Data::Dumper allows me to have it call into an object's method which is allowed to modify the reference/object but doesn't have a return value and it can't be substituted for just an unblessed value. :-(
Yes, absolutely without any cooperation on the part of the caller.
Data::Dumper kind of sucks when feeding it overloaded objects and tied variables. 22:38
So yeah, I just hadn't recalled hearing anyone talk about p6's call stack and didn't know if that was part of the spec or not. Maybe I just heard different terms than I expected. 22:39
gaal (sorry, back) yeah, ACK about the "if you need this you need it bad" nature of this feature 22:41
I don't recall moosing it either.
things like caller have very strong selectors, but not the kind of crazy mutations you're scheming. 22:42
(eg you can ask caller for the most recent block that had a certain label) 22:43
diotalevi On p5 to *really* solve that I'd write something to walk my data prior to DD seeing it and collapse my objects or I'd use PadWalker or such to reach directly into the C. 22:44
I'm not going to bother, fwiw.
But see... when it comes to solutions like "reach into the C and tweak something" then it's murky about what happens on p6 because "the C" doesn't exist anymore. :-/ 22:45

Powered by Cro and the Raku® Programming Language.

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

Thank you!