view release on metacpan or  search on metacpan

lib/Acme/Pythonic.pm  view on Meta::CPAN

package Acme::Pythonic;

# Please, if you tested it in some earlier version of Perl and works let
# me know! The versions of Filter::Simple, Text::Tabs, and Test::More
# would be useful as well.
use 5.006_001;
use strict;
use warnings;

our ($VERSION, $DEBUG, $CALLER);
$VERSION = '0.47';

use Text::Tabs;

sub import {
    my ($package, %cfg) = @_;
    $DEBUG = $cfg{debug};
    $CALLER = caller() # to be able to check sub prototypes
}


use Filter::Simple;
FILTER_ONLY code => sub {
    unpythonize();
    cuddle_elses_and_friends();
    if ($DEBUG) {
        s/$Filter::Simple::placeholder/BLANKED_OUT/g;
        print;
        $_ = '1;';
    }
};


# This regexp matches a 7-bit ASCII identifier. We use atomic grouping
# because an identifier cannot be backtracked.
my $id = qr/(?>[_a-zA-Z](?:[_a-zA-Z0-9']|::)*)/;

# Shorthand to put an eventual trailing comment in some regexps.
my $tc = qr/(?<!\$)#.*/;


# Tries its best at converting Pythonic code to Perl.
sub unpythonize {
    # Sometimes Filter::Simple adds newlines blanking out stuff, which
    # interferes with Pythonic conventions.
    my %bos = (); # BlanketOutS
    my $count = 0;
    s<$Filter::Simple::placeholder>
     <my $bo = "$;BLANKED_OUT_".$count++."$;";
      $bos{$bo} = $&;
      $bo>geo;

    # In addition, we can now normalize newlines without breaking
    # Filter::Simple's identifiers.
    normalize_newlines();
    my @lines = split /\n/;
    return unless @lines;

    # If unsure about the ending indentation level, add an extra
    # non-indented line to ensure the stack gets emptied.
    push @lines, '1; # added by Acme::Pythonic' if $lines[-1] =~ /^(?:\s|\s*#)/;

    my ($comment,             # comment in the current line, if any
        $indent,              # indentation of the current logical line
        $id_at_sob,           # identifier at StartOfBlock, for instance "else", or "eval"
        $prev_line_with_code, # previous line with code
        $might_be_modifier,   # flag: current logical line might be a modifier
        $line_with_modifier,  # physical line which started the current modifier
        $joining,             # flag: are we joining lines?
        $unbalanced_paren,    # flag: we opened a paren that remains to be closed
        @stack,               # keeps track of indentation stuff
    );

    @stack = ();
    foreach my $line (@lines) {
        # We remove any trailing comment so that we can assert stuff
        # easily about the end of the code in this line. It is later

lib/Acme/Pythonic.pm  view on Meta::CPAN

        2

If a line ends in a comma or arrow (C<< => >>) it is conceptually joined
with the following as well:

    my %authors = (Perl   => "Larry Wall",
                   Python => "Guido van Rossum")

As in Python, comments can be intermixed there:

    my %hello = (Catalan => 'Hola',   # my mother tongue
                 English => 'Hello',)


Acme::Pythonic munges a source that has already been processed by L<Filter::Simple>. In particular, L<Filter::Simple> blanks out quotelikes whose content is not even seen by Acme::Pythonic so backslashes in C<qw//> and friends won't be removed:

    # Do not put backslashes here because qw// is bypassed
    my @colors = qw(Red
                    Blue
                    Green)


=head1 CAVEATS

Although this module makes possible some Python-like syntax in Perl,
there are some remarkable limitations in the current implementation:

=over 4

=item * Compound statement bodies are not recognized in header
lines. This would be valid according to Python syntax:

    if $n % 2: $n = 3*$n + 1
    else: $n /= 2

but it does not work in Acme::Pythonic. The reason for this is that it
would be hard to identify the colon that closes the expression without
parsing Perl, consider for instance:

    if keys %foo::bar ? keys %main:: : keys %foo::: print "foo\n"

=item * In Python statements may span lines if they're enclosed in
C<()>, C<{}>, or C<[]> pairs. Acme::Pythonic does not support this rule,
however, though it understands the common case where you break the line
in a comma in list literals, subroutine calls, etc.

=back

Remember that source filters do not work if they are called at runtime,
for instance via C<require> or C<eval EXPR>. The source code was already
consumed in the compilation phase by then.


=head1 DEBUG

L<Filter::ExtractSource> can be used to inspect the source code
generated by Acme::Pythonic:

    perl -c -MFilter::ExtractSource pythonic_script.pl

Acme::Pythonic itself has a C<debug> flag though:

    use Acme::Pythonic debug => 1;

In debug mode the module prints to standard output the code it has
generated, and passes just a dummy C<1;> to L<Filter::Simple>.

This happens I<before> L<Filter::Simple> undoes the blanking out of
PODs, strings, and regexps. Those parts are marked with the label
C<BLANKED_OUT> for easy identification.

Acme::Pythonic generates human readable Perl following L<perlstyle>, and
tries meticulously to be respectful with the original source code.
Blank lines and comments are preserved.


=head1 BUGS

This module uses a regexp approach and the superb help of
Filter::Simple. The regexp part of this means it is broken from the
start, though I've tried hard to make it as robust as I could. Bug
reports will be very welcome, just drop me a line!


=head1 THANKS

Damian Conway gave his full blessing if I wanted to write a module like
this based on his unpublished Language::Pythonesque. The code that
handles indentation is inspired by his.

Also, Dr. Conway is the author of L<Filter::Simple>, which aids a lot
blanking out PODs, strings, etc. so you can munge the source with
certain confidence. Without Filter::Simple this module would be
infinitely more broken.

Esteve Fernandez helped testing the module under 5.6.1 and contributed
a Sieve of Eratosthenes for F<t/algorithms.t>. Thank you dude!

=head1 SEE ALSO

L<perlfilter>, L<Filter::Simple>, L<Filter::ExtractSource>, L<SuperPython>, L<Acme::Dot>.


=head1 AUTHOR

Xavier Noria (FXN), E<lt>fxn@cpan.orgE<gt>


=head1 COPYRIGHT AND LICENSE

Copyright (C) 2004-2012 by Xavier Noria

This library is free software; you can redistribute it and/or modify
it under the same terms as Perl itself, either Perl version 5.8.2 or,
at your option, any later version of Perl 5 you may have available.

=cut