This and other RFCs are available on the web at
  http://dev.perl.org/rfc/

=head1 TITLE

prototype-based method overloading

=head1 VERSION

  Maintainer: David Nicol <[EMAIL PROTECTED]>
  Date: 13 Aug 2000
  Last Modified: 29 Sep 2000
  Mailing List: [EMAIL PROTECTED]
  Number: 97 
  Version: 2
  Status: Frozen

=head1 freeze notes (29 sep)

I believe several prototyping systems have been proposed,
concerned with the exact nature of the parameter lists.  Since
this RFC does not specify the mechanism for the declaration, beyond
that it is a compatible superset of the current syntax, I believe
we have compatability.

=head1 ABSTRACT


When I read the chapter on OO in the second edition camel book I
was saddened that C++ style method overloading was not explicitly
described.  Ever hopeful, I wrote some quick code that I hoped would
do what I meant and discovered that it didn't.  This document is
an attempt to remedy the situation.


=head1 SUMMARY

        $frog_t = qs(frog);
        sub listargs($){ print "One arg, $_[0]"}
        sub listargs($$){ print "Two args, $_[0] and $[1]"}
        sub listargs($$frog_t){ print "$_[0] and a frog $[1]"}
        sub listargs { throw argsyntax, "odd arguments to listargs" }

        my $frog_t $k = new frog(type=>tree);
        listargs("franz","tree");       # prints "Two args..."
        listargs("franz",$k);           # prints "franz and ..."
        listargs($k,"franz");           # throws an argsyntax error


=head1 DESCRIPTION

It it now possible to define two subroutines with the same
name but different interfaces without error.  Perl will puzzle
out which one to call in a given situation based on lexical
information available in the program text.

Defining two subroutines with both the same name and the same calling
interface is undefined and may be an error.  This may change.

No coercion protocol is defined at this time.
This document will be updated as a protocol for coercing 
function calls with arguments that don't explicitly match is
developed, such as having a method name (which is now
distinct from a method, but still a strong grouping characteristic)
have a "coercion" method associated with it which would 
indicate what to do if no prototypes matched. 

For now, for the greatest ease
for implementors, calls to method C<foo> that exactly match none of the
prototypes of defined subroutines named C<foo>  
will fall through to a C<foo> with no prototype, should one exist.

Programmers using this feature are advised to include a full coercion
system into their unprototyped methods, when writing in a strongly typed
environment.


=head1 IMPLEMENTATION

At compile time, the keys in the big hash (be it global or per-package
or per-class) that holds the mapping from the names of the classes
to their coderefs is extended to include the prototype as part of the
name of each method.  The nature of this extension is beyond the
scope of this document.

Perhaps the method is first looked up by name, and then further
dispatch is done if needed.  This would exempt methods using "relaxed
perl5-ish semantics" from any additional processing overhead.



=head1 REFERENCES

RFC 57: Subroutine prototypes and parameters

RFC 61: Interfaces for linking C objects into perlsubs

RFC 75: structures and interface definitions

camel book, 2nd ed.


Reply via email to