I'm not quite sure I follow you (I'm not familiar with that pattern). But the macromethod I imagine is the non-polymorphic one, and the one it expands to is the polymorphic one, if I'm guessing correctly. And you certianly could do that.
yes: http://patterndigest.com/patterns/TemplateMethod.html
But the actual I<parsing> could not be polymorphic. Thus, this wouldn't work:
my $db = new Database; $db.select ...
Because the compiler [may] not know that $db is a Database until runtime, after it's been parsed. However, this could work:
my Database $db = new Database; $db.select ...
To a large extent, this depends on how powerful the type-inference mechanisms are. In theory, if MyDatabase.new "returns Database", then the type of $db can be inferred. Even if the return type is not explicitly defined, we can look in the method-body, and see that its constructing/returning an object of base-type Database.
But you're right, there are situations where the (base) type might not be knowable: and these could result in syntax errors.
So, what I'm essentially saying is that C<macromethod> doesn't buy you much save syntactic sugar. Perhaps you can give a concrete counterexample?
I admit that I have a sweet-tooth: I like syntactic sugar ;-). And I'm one of those environmentalists who don't like (namespace) pollution.
The sugar I'm using here is to go from
$db.do_sql("select * from Foo");
to
$db.select * from Foo;I'm sure there is nice way to represent this with a leading keyword: but one of the things that Perl6 is doing (over perl5) is to make (almost) everything callable as methods. Thus if we allow (as a standard macro -- though a macromethod would also support it)
select $db: * from Foo;
then we should also allow the method-call syntax. I'll admit that this is one of those cases where the user already knows that there's some magic happenning -- that "select" is introducing a new syntax -- but it would be nice to keep the invocant part looking like Perl6 code.
Perhaps I should back off a bit: Perl6 *will* be powerful enough support the "Macros::macromethod" module, so perhaps I'll be able to write it in a year's time, or so. It'll be a good test of the extensability mechanisms.
Dave. -- http://dave.whipp.name
