RFC 199 (v2) Short-circuiting C<grep>, C<map>, and C<reduce> with C<last>

(or "Allowing built-in functions to use loop blocks")
Reply-To: [email protected]

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

=head1 TITLE

Short-circuiting C<grep>, C<map>, and C<reduce> with C<last>
(or "Allowing built-in functions to use loop blocks")

=head1 VERSION

  Maintainer: Garrett Goebel <[email protected]>
  Date: 6 Sep 2000
  Last Modified: 7 Sep 2000
  Mailing List: [email protected]
  Number: 199
  Version: 2
  Status: Developing

=head1 ABSTRACT

Allow built-in functions like C<grep>, C<map>, and C<reduce> which take
blocks as a parameter to be short-circuited with C<last>

=head1 DESCRIPTION

It'd be nice if one could use C<grep> to find out if a value is
held in a list without having to iterate through every element
of said list. It would follow that one should also be able to
label such built-in functions. 

Damian already has plans for C<reduce> that would
require this. I'm sure others can figure out their own uses for
short-circuiting C<grep>, C<map>, and C<reduce>. 

Adoption of this suggestion would create problems with any Perl 5 code
that goes against conventional wisdowm and uses C<last> within a C<grep>
or C<map> block. The workaround would require the Perl 5 to 6 translator
to generate labels for any unlabled C<next>, C<last>, and C<redo> 
statements and their associated C<while>, C<until>, C<for>, and
C<foreach> statements.

Neither Tom Christiansen nor Jarkko Hietaniemi profess to know why the
suggestion with regards to C<grep> hasn't already been implemented.
Perhaps it has to do with the fact that the blocks in C<grep> and C<map>
are only appear to be loops, where as C<while>, C<until>, C<for>, and 
C<foreach> are loops. I have practically no knowledge of the Perl
internals or possible history related to this issue. Please 
enlighten me, and I'll include it in a later revision of this RFC.

If a the built-in's block were terminated by a call to C<last>,
the built-in immediately returns the last block value (i.e.
C<undef> on the first block call, $_[0] otherwise). Or maybe that's
@_ otherwise... someone straighten me out please.

@a = grep( {$_ < 3 or last } (1, 2, 3, 2, 1);

results in:

@a = (1, 2);


@a = grep( { {$_ < 3 or last} } (1, 2, 3, 2, 1);

results in:

@a = (1, 2, 2, 1);

=head1 IMPLEMENTATION

I haven't a clue. Please feel free to suggest one for the next revision.

=head1 REFERENCES

RFC's that require this functionality:

RFC 76:  Builtin: reduce

Ideas that offer alternative routes to similar functionality:

RFC 22: Builtin switch statement

RFC ??: Damian Conway's forthcoming superpositions RFC