Dominic Hargreaves

NAME

Wiki::Toolkit::Plugin::Diff - format differences between two Wiki::Toolkit pages

SYNOPSIS

  use Wiki::Toolkit::Plugin::Diff;
  my $plugin = Wiki::Toolkit::Plugin::Diff->new;
  $wiki->register_plugin( plugin => $plugin );   # called before any node reads
  my %diff = $plugin->differences( node => 'Imperial College',
                                left_version => 3,
                                right_version => 5);

DESCRIPTION

A plug-in for Wiki::Toolkit sites, which provides a nice extract of differences between two versions of a node.

BASIC USAGE

differences

  my %diff_vars = $plugin->differences(
      node          => "Home Page",
      left_version  => 3,
      right_version => 5
  );

Takes a series of key/value pairs:

  • left_version

    The node version whose content we're considering canonical.

  • right_version

    The node version that we're showing the differences from.

  • meta_include

    Filter the list of metadata fields to only include a certain list in the diff output. The default is to include all metadata fields.

  • meta_exclude

    Filter the list of metadata fields to exclude certain fields from the diff output. The default is the following list, to match previous version (OpenGuides) behaviour: qw( username comment __categories__checksum __locales__checksum )

    Agreed this list is hopelessly inadequate, especially for OpenGuides. Hopefully, future wiki designers will use the meta_include parameter to specify exactly what metadata they want to appear on the diff.

The differences method returns a list of key/value pairs, which can be assigned to a hash:

left_version

The node version whose content we're considering canonical.

right_version

The node version that we're showing the differences from.

content

The (formatted) contents of the Left version of the node.

diff

An array of hashrefs of hunks of differences between the versions. It is assumed that the display will be rendered in HTML, and SPAN tags are inserted with a class of diff1 or diff2, to highlight which individual words have actually changed. Display the contents of diff using a <table>, with each member of the array corresponding to a row <TR>, and keys {left} and {right} being two columns <TD>.

Usually you will want to feed this through a templating system, such as Template Toolkit, which makes iterating the AoH very easy.

ADVANCED

Wiki::Toolkit::Plugin::Diff allows for a more flexible approach than HTML only rendering of pages. In particular, there are optional parameters to the constructor which control fine detail of the resultant output.

If this is not sufficient, the module is also subclassable, and the programmer can supply alternative methods.

METHODS

Most of these are called internally by the plugin, but provide hooks for alternative code if the module is subclassed.

new
  my $plugin = Wiki::Toolkit::Plugin::Diff->new( option => value, option => value...);

Here, option can be one of the following:

metadata_separator

A string which is inserted between each metadata field, and also between the contents and the metadata. This defaults to "<br />\n" so as to render each metadata field on a new line.

line_number_format

Used to lay out the head line number for a difference hunk. The string is eval'ed with $_ containing the line number. Default is "== Line \$_ ==" .

word_matcher

is a regular expression, used to tokenize the input string. This is the way of grouping together atomic sequences, so as to give a readable result. The default is the following:

            &.+?;                   #HTML special characters e.g. &lt;
            |<br\s*/>               #Line breaks
            |\w+\s*                 #Word with trailing spaces 
            |.                      #Any other single character
differences

see above.

line_number

This method is called to format a line number into a suitable string. The supplied routine performs the necessary substitution of $self->{line_number_format} using eval.

serialise_metadata

The purpose of this method is to turn a metadata hash into a string suitable for diffing.

content_escape

This method is used to apply the necessary quoting or escaping of characters that could appear in the content body, that could interfere with the rendering of the diff text.

intradiff

This method turns an array of hunks as returned by VCS::Lite::Delta->hunks into a side by side diff listing, with highlights indicating different words and punctuation.

Currently, this is hardcoded to present the differences with HTML tags.

This module is a prime candidate for migration into VCS::Lite, where this functionality really belongs.

get_token

This allows the "false positive" bug discovered by Earle Martin to be solved (rt.cpan.org #6284). Effectively, the input strings (diff1 and diff2) are tokenised using the word_matcher regexp (see above), and turned into arrays of tokens.

The default regexp absorbs trailing whitespace, to give a more readable result. However, if a word is followed by differing whitespace or no whitespace at all, this was throwing up a diff.

The get_token method supplied removes trailing whitespace from the key values before they are compared.

TODO

Move intradiff functionality into VCS::Lite.

BUGS AND ENHANCEMENTS

Please use rt.cpan.org to report any bugs in this module. If you have any ideas for how this module could be enhanced, please email the author, or post to the Wiki::Toolkit list (cgi (hyphen) wiki (hyphen) dev (at) earth (dot) li).

AUTHOR

I. P. Williams (ivorw_openguides [at] xemaps {dot} com) The Wiki::Toolkit team (http://www.wiki-toolkit.org/)

COPYRIGHT

     Copyright (C) 2003-2004 I. P. Williams (ivorw_openguides [at] xemaps {dot} com).
     Copyright (C) 2006 the Wiki::Toolkit team (http://www.wiki-toolkit.org/)
     All Rights Reserved.

This module is free software; you can redistribute it and/or modify it under the same terms as Perl itself.

SEE ALSO

VCS::Lite, Wiki::Toolkit, Wiki::Toolkit::Plugin, OpenGuides