package Perl::Critic::Policy::Objects::ProhibitIndirectSyntax;

use 5.006001;
use strict;
use warnings;

use Carp;
use English qw(-no_match_vars);
use Perl::Critic::Utils qw{ :severities :classification };
use Readonly;

use base 'Perl::Critic::Policy';

our $VERSION = '1.140';


Readonly::Hash my %COMMA => {
    q<,> => 1,
    q{=>} => 1,
Readonly::Scalar my $DOLLAR => q<$>;

Readonly::Scalar my $DESC => 'Subroutine "%s" called using indirect syntax';
Readonly::Scalar my $EXPL => [ 349 ];


sub supported_parameters {
    return (
            name                       => 'forbid',
            description                => 'Indirect method syntax is forbidden for these methods.',
            behavior                   => 'string list',
            list_always_present_values => [ qw{ new } ],

sub default_severity     { return $SEVERITY_HIGH             }
sub default_themes       { return qw( core pbp maintenance certrule ) }
sub applies_to           { return 'PPI::Token::Word'         }


sub violates {
    my ( $self, $elem, $doc ) = @_;

    # We are only interested in the functions we have been told to check.
    # Do this before calling is_function_call() because we want to weed
    # out as many candidate tokens as possible before calling it.
    return if not $self->{_forbid}->{$elem->content()};

    # Make sure it really is a function call.
    return if not is_function_call($elem);

    # Per perlobj, it is only an indirect object call if the next sibling
    # is a word, a scalar symbol, or a block.
    my $object = $elem->snext_sibling() or return;
    return if not (
            $object->isa( 'PPI::Token::Word' )
        or      $object->isa( 'PPI::Token::Symbol' )
            and $DOLLAR eq $object->raw_type()
        or  $object->isa( 'PPI::Structure::Block' )

    # Per perlobj, it is not an indirect object call if the operator after
    # the possible indirect object is a comma.
    if ( my $operator = $object->snext_sibling() ) {
        return if
                $operator->isa( 'PPI::Token::Operator' )
            and $COMMA{ $operator->content() };

    my $message = sprintf $DESC, $elem->content();

    return $self->violation( $message, $EXPL, $elem );





=head1 NAME

Perl::Critic::Policy::Objects::ProhibitIndirectSyntax - Prohibit indirect object call syntax.


This Policy is part of the core L<Perl::Critic|Perl::Critic>


Indirect object syntax is commonly used in other object-oriented languages for
instantiating objects. Perl allows this, but to say that it supports it may be
going too far. Instead of writing

    my $foo = new Foo;

it is preferable to write

    my $foo = Foo->new;

The problem is that Perl needs to make a number of assumptions at compile time
to disambiguate the first form, so it tends to be fragile and to produce
hard-to-track-down bugs.


Indirect object syntax is also hard for Perl::Critic to disambiguate, so this
policy only checks certain subroutine calls. The names of the subroutines can
be configured using the C<forbid> configuration option:

    forbid = create destroy

The C<new> subroutine is configured by default; any additional C<forbid>
values are in addition to C<new>.

=head1 CAVEATS

The general situation can not be handled via static analysis.

=head1 SEE ALSO

and L<indirect|indirect> both do a better job with this, but they require that you
compile/execute your code.

=head1 AUTHOR

Thomas R. Wyant, III F<wyant at cpan dot org>


Copyright (c) 2009-2011 Tom Wyant.

This program is free software; you can redistribute it and/or modify
it under the same terms as Perl itself.  The full text of this license
can be found in the LICENSE file included with this module.


# Local Variables:
#   mode: cperl
#   cperl-indent-level: 4
#   fill-column: 78
#   indent-tabs-mode: nil
#   c-indentation-style: bsd
# End:
# ex: set ts=8 sts=4 sw=4 tw=78 ft=perl expandtab shiftround :