NAME

       XML::Filter::Sort::Buffer - Implementation class used by XML::Filter::Sort

DESCRIPTION

       The documentation is targetted at developers wishing to extend or replace this class.  For
       user documentation, see XML::Filter::Sort.

       For an overview of the classes and methods used for buffering, see
       XML::Filter::Sort::BufferMgr.

BUFFER LIFE CYCLE

       A XML::Filter::Sort::Buffer object is created by a XML::Filter::Sort::BufferMgr object
       using the "new()" method.

       The XML::Filter::Sort object will then propagate any SAX events it receives, to the buffer
       object until the end of the record is reached.  As each element is added to the buffer,
       its contents are compared to the sort key paths and the sort key values are extracted.

       When the end of the record is reached, the "close()" method is called.  The return value
       from this method is the list of sort keys.

       The buffer manager will store the buffer until the end of the record sequence is reached.
       Then it will retrieve each buffer in order of the sort key values and call the buffer's
       "to_sax()" method to send all buffered events to the downstream handler.

       Following the call to "to_sax()", the buffer is discarded.  No destructor method is used -
       everything is handled by Perl's garbage collector.

DATA STRUCTURES

       The buffer contains a 'tree' of SAX events.  The tree is simply an array of 'nodes'.  Text
       nodes are represented as scalars.  Other nodes are represented as arrayrefs.  The first
       element of a node array is a single character identifying the node type:

         e - element
         c - comment
         p - processing instruction

       The second element is the node data (the hash from the original SAX event).  The child
       nodes of an element node are represented by the third element as an arrayref.

       For example, this XML:

         <person age="27">
           <lastname>smith</lastname>
         </person>

       Would be buffered as this data structure:

         [
           [
             'e',
             {
               'Name' => 'person'
               'Prefix' => '',
               'LocalName' => 'person',
               'NamespaceURI' => '',
               'Attributes' => {
                 '{}age' => {
                   'LocalName' => 'age',
                   'NamespaceURI' => '',
                   'Value' => '27',
                   'Prefix' => '',
                   'Name' => 'age'
                 }
               },
             },
             [
               "\n  ",
               [
                 'e',
                 {
                   'Name' => 'lastname'
                   'Prefix' => '',
                   'LocalName' => 'lastname',
                   'NamespaceURI' => '',
                   'Attributes' => {},
                 },
                 [
                   'smith'
                 ]
               ],
               "\n  ",
             ]
           ]
         ]

COPYRIGHT

       Copyright 2002 Grant McLean <grantm@cpan.org>

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