pugscode.org/ | nopaste: sial.org/pbot/perl6 | pugs: [~] <m oo se> (or rakudo:, kp6:, smop: etc.) || We do Haskell, too | > reverse . show $ foldl1 (*) [1..4] | irclog: irc.pugscode.org/
Set by TimToady on 25 January 2008.
weirdo hi 09:23
what is the rationale for the presence of sigils in perl6?
masak weirdo: same as in perl5 I guess 09:24
they form little namespaces in themselves
marking "scalar", "array" and "hash" because these properties are deemed worthy of marking
weirdo i found typing "push @{$foo->{bar}}" particularly tiring. don't such cases overweight the benefit of having syntactic sugar for symbol cells? 09:25
xinming weirdo: In perl 6, an [] auto deferences depending on the context, so, IMO, $foo.bar.push() is enough in perl 6. ;-) 09:28
wolverian $foo<bar>.push 09:29
xinming weirdo: every thing in perl 6 is an object, and for array and hash, they are automatically deferences.
wolverian: ;-) there should be an accessor. :-D 09:30
wolverian hashes are not illegal... even in perl 6 :)
weirdo any progress on macro specification?
wolverian (but hash is!)
xinming weirdo: wolverian is right.
wolverian weirdo, the spec has some things on macros. 09:31
weirdo can an equivalent of a PROG2 lisp macro be easily defined?
wolverian perlcabal.org/syn/S06.html#Macros
lambdabot Title: S06
xinming weirdo: read the synopsis, and you will see the what language features after 20 years we still use. :-) 09:32
weirdo do i see correctly, the spec defined macros as text to be parsed with regexes? 09:33
wolverian macros may return strings as code or syntax trees. 09:34
weirdo can macros accept syntax trees as argument?
s/arguments
wolverian the spec seems to indicate so. 09:36
I'm very unfamiliar with lisp macros, so I can't really say any better, sorry. 09:37
I suppose I'm equally unfamiliar with perl 6 macros. :)
but, yes, the arguments are ASTs, afaik. 09:38
weirdo IMO perl could win really big with adequate macros 09:39
wolverian well, if you don't think the macros in the spec are adequate enough, please suggest improvements (here or on perl6-language)
if, on the other hand, you think the spec is simply unclear, do the same. :)
weirdo i haven't yet researched the matter, but i don't understand why macros aren't simple nested arrayrefs 09:41
wolverian probably because perl 6's syntax isn't sexprs? 09:42
weirdo the main question is whether non-trivial cases can be expressed as perl macros, but i yet have to get a recent pugs trunk to check them out :-) 09:43
if a DESTRUCTURING-BIND common-lisp macro could be expressed in a not-so-hairy way, they would be really adequate :-) 09:44
wolverian I don't know if pugs has macros at all..
weirdo is there any implementation that does?
wolverian I don't think so
(if pugs doesn't have them)
pugs: macro foo { 'bar' }
exp_evalbot RESULT[\macro :(@_) "$_" := "Scalar" #<Scalar:0xb71c8b04>␤ "@_" := "Array" #<Array:0xb71c76fc>␤ "&?ROUTINE" := "Sub" #<Sub:0xb6604cec>␤ "&?BLOCK" := "Sub" #<Sub:0xb6604cec> {"&foo" := "Sub" #<Sub:0xb7a18830>, "$_" := "Scalar" #<Scalar:0xb7170ce8>␤ 09:45
.. ...
wolverian hmm, I guess pugs does have them in some form
rakudo: macro foo { 'bar' }
exp_evalbot OUTPUT[Null PMC access in invoke()␤current instr.: '_block10' pc 28 (EVAL_11:14)␤called from Sub 'parrot;PCT::HLLCompiler;eval' pc 785 (src/PCT/HLLCompiler.pir:458)␤called from Sub 'parrot;PCT::HLLCompiler;evalfiles' pc 1067 (src/PCT/HLLCompiler.pir:587)␤called from Sub
..'parrot;PCT::HLLCompiler;command_line' pc 1246 (src/PCT/HLLCompiler.pir:676)...
wolverian rakudo doesn't :)
I have no idea what destructuring-bind does, even after reading the lisp spec :) 09:47
weirdo destructuring-bind binds variables to contents of an expression returning a list
perl5 only allows shallow destructuring, i.e. my ($foo, $bar, @baz) = (1 .. 42) 09:48
d-bind allows recursive destructuring, i.e. (destructuring-bind (foo (bar baz &key test on-error)) ...) 09:49
wolverian I think you can do arbitrary destructuring (extraction / pattern matching in haskell parlance) in any signature context 09:50
see e.g. perlcabal.org/syn/S06.html#Unpackin...parameters and the sections following it in subroutine argument context
lambdabot Title: S06
weirdo i'm not interested in destructuring in particular :) i've only arbitrarily appointed destructuring as a test of a macro facility's capabilities 09:51
wolverian ah. 09:52
well, I don't see why not, as long as your argument grammar specification is loose enough..
weirdo well, that's the problem with algol-ish syntax :/ 09:53
wolverian the grammar is available to the user, so perhaps it's not a problem. 09:54
BinGOs a/win 33 12:26
bugger
cognominal_ screen misshap? 12:36
pmurias weirdo: the macros should be powerfull enough to compile passed code and load it through NCI 13:28
allowing a lispish destructuring bind 13:30
say "this is perl"; (destructuring-bind ((a &optional (b 'bee)) one two three)
`((alpha) ,@(iota 3))
;say "this is still perl" 13:31
should be possible also
weirdo don't mind the parens :) 13:41
i mean, it doesn't actually have lisp syntax
weirdo something along the lines of destructuring-bind ($a, $b, ($c, $d, @f)) { ... } perfectly demonstrates expressibility of a macro facility :) 13:43
pmurias weirdo: macros don't modifify the way stuff is parsed by default, unless you use 'is parsed' which lets you embed anything you wish 15:02

Powered by Cro and the Raku® Programming Language.

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

Thank you!