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

=head1 TITLE

lvalue subs: parameters, explicit assignment, and wantarray() changes

=head1 VERSION

  Maintainer: Nathan Torkington <[EMAIL PROTECTED]>
  Date: Aug 16, 2000
  Version: 1
  Mailing List: [EMAIL PROTECTED]
  Number: 118

=head1 ABSTRACT

RFC 107 proposed that lvalue subs should receive their rvalues as
subroutine arguments.  I counter-propose that lvalue subs should
receive their rvalues as lexical variables named in a prototype
associated with the :lvalue modifier.  The lvalue-ness of a given
subroutine call should be discoverable with the wantarray() function
(or its replacement), and the assignment behaviour should be written
in the subroutine rather than implicitly done by Perl.

=head1 DESCRIPTION

Subroutines may be called with multiple values.  Lvalue subroutines
may be assigned multiple values.  Simple passing on the argument
list conflates the two:

  foo(@args) = @rvalues;

would be identical to:

  foo(@args, @rvalues);

Perl's list flattening would prevent the subroutine from knowing where
one began and the other ended.  Better would be if the rvalue were
passed as a last or first argument, making it equivalent to:

  foo(\@rvalues, @args);

or

  foo(@args, \@rvalues);

I don't think this is appropriate, either.  Perl's OO tried implicit
arguments, and now we're rebelling against them.  Better would be to
have the values magically appear as lexicals in the subroutine.

=head2 Prototypes

To the language it would appear as a prototype associated with the
:lvalue modifier:

  sub foo (@) : lvalue ($first, $second, @rest) {
    print "My first rvalue is $first\n";
    print "My second rvalue is $second\n";
    print "The rest of my rvalues are: @rest\n";
    print "My arguments are: @_\n";
  }

=head2 Explicit assignment

This would change the perl5 meaning of lvalue subroutines.  Currently
you must have the lvalue as the last value before the return, and the
assignment is implicitly done by Perl.  I advocate making it explicit:

  # this is perl5
  sub foo :lvalue {
    $variable;          # am I rvalue or lvalue sub?  I don't know
  }
  foo = 5;              # implicitly $variable = 5

I would now have this as:

  # this is perl6
  sub foo :lvalue ($new) {
    $variable = $new;
  }

=head2 Lvalue context discoverable

Then the wantarray() replacement, whatever it may be, can be used
to say whether you're lvalue or rvalue and what to do.  That makes
an lvalue sub more like the STORE method of a tied variable, where
you can decide whether the storage is acceptable:

  # still perl6
  sub foo :lvalue ($new) {
    if (want("lvalue")) {
      if ($new >= 0) {
        $variable = $new;
      } else {
        croak "Attempt to store negative number in foo";
      }
    } else {
      return $variable;
    }
  }

=head1 IMPLEMENTATION

Straightforward.  Requires changes to:

=over 4

=item prototypes

Permit named function prototypes.  On the lvalue modifier.

=item wantarray

Identify lvalue context.

=back

=head1 REFERENCES

RFC 107: lvalue subs should receive the rvalue as an argument

RFC 57: Subroutine prototypes and parameters

RFC 21: Replace wantarray with a generic want function

Reply via email to