Tutorial
Manual
Modules
Library
Tools
FAQ
Release
   
Template Toolkit | Modules | Template::Plugin::XML::DOM

Template::Plugin::XML::DOM

  Template::Plugin::Wrap
Template::Plugin::XML::RSS  
Plugin interface to XML::DOM

Table of Contents

SYNOPSIS Index  ] [ Modules  ] [ Top  ]

    # load plugin
    [% USE dom = XML.DOM %]
    # also provide XML::Parser options
    [% USE dom = XML.DOM(ProtocolEncoding =E<gt> 'ISO-8859-1') %]
    # parse an XML file
    [% doc = dom.parse(filename) %]
    [% doc = dom.parse(file => filename) %]
    # parse XML text
    [% doc = dom.parse(xmltext) %]
    [% doc = dom.parse(text => xmltext) %]
    # call any XML::DOM methods on document/element nodes
    [% FOREACH node = doc.getElementsByTagName('report') %]
       * [% node.getAttribute('title') %]     # or just '[% node.title %]'
    [% END %]
    # define VIEW to present node(s)
    [% VIEW report notfound='xmlstring' %]
       # handler block for a <report>...</report> element
       [% BLOCK report %]
          [% item.content(view) %]
       [% END %]
       # handler block for a <section title="...">...</section> element
       [% BLOCK section %]
       <h1>[% item.title %]</h1>
       [% item.content(view) %]
       [% END %]
       # default template block converts item to string representation
       [% BLOCK xmlstring; item.toString; END %]
       
       # block to generate simple text
       [% BLOCK text; item; END %]
    [% END %]
    # now present node (and children) via view
    [% report.print(node) %]
    # or print node content via view
    [% node.content(report) %]
    # following methods are soon to be deprecated in favour of views
    [% node.toTemplate %]
    [% node.childrenToTemplate %]
    [% node.allChildrenToTemplate %]

PRE-REQUISITES Index  ] [ Modules  ] [ Top  ]

This plugin requires that the XML::Parser (2.19 or later) and XML::DOM (1.27 or later) modules be installed. These are available from CPAN:

    http://www.cpan.org/modules/by-module/XML

Note that the XML::DOM module is now distributed as part of the 'libxml-enno' bundle.

DESCRIPTION Index  ] [ Modules  ] [ Top  ]

This is a Template Toolkit plugin interfacing to the XML::DOM module. The plugin loads the XML::DOM module and creates an XML::DOM::Parser object which is stored internally. The parse() method can then be called on the plugin to parse an XML stream into a DOM document.

    [% USE dom = XML.DOM %]
    [% doc = dom.parse('/tmp/myxmlfile') %]

NOTE: earlier versions of this XML::DOM plugin expected a filename to be passed as an argument to the constructor. This is no longer supported due to the fact that it caused a serious memory leak. We apologise for the inconvenience but must insist that you change your templates as shown:

    # OLD STYLE: now fails with a warning
    [% USE dom = XML.DOM('tmp/myxmlfile') %]
    # NEW STYLE: do this instead
    [% USE dom = XML.DOM %]
    [% doc = dom.parse('tmp/myxmlfile') %]

The root of the problem lies in XML::DOM creating massive circular references in the object models it constructs. The dispose() method must be called on each document to release the memory that it would otherwise hold indefinately. The XML::DOM plugin object (i.e. 'dom' in these examples) acts as a sentinel for the documents it creates ('doc' and any others). When the plugin object goes out of scope at the end of the current template, it will automatically call dispose() on any documents that it has created. Note that if you dispose of the the plugin object before the end of the block (i.e. by assigning a new value to the 'dom' variable) then the documents will also be disposed at that point and should not be used thereafter.

    [% USE dom = XML.DOM %]
    [% doc = dom.parse('/tmp/myfile') %]
    [% dom = 'new value' %]     # releases XML.DOM plugin and calls
                                # dispose() on 'doc', so don't use it!

Any template processing parameters (see toTemplate() method and friends, below) can be specified with the constructor and will be used to define defaults for the object.

    [% USE dom = XML.DOM(prefix => 'theme1/') %]

The plugin constructor will also accept configuration options destined for the XML::Parser object:

    [% USE dom = XML.DOM(ProtocolEncoding => 'ISO-8859-1') %]

METHODS Index  ] [ Modules  ] [ Top  ]

parse()

The parse() method accepts a positional parameter which contains a filename or XML string. It is assumed to be a filename unless it contains a < character.

    [% xmlfile = '/tmp/foo.xml' %]
    [% doc = dom.parse(xmlfile) %]
    [% xmltext = BLOCK %]
    <xml>
      <blah><etc/></blah>
      ...
    </xml>
    [% END %]
    [% doc = dom.parse(xmltext) %]

The named parameters 'file' (or 'filename') and 'text' (or 'xml') can also be used:

    [% doc = dom.parse(file = xmlfile) %]
    [% doc = dom.parse(text = xmltext) %]

The parse() method returns an instance of the XML::DOM::Document object representing the parsed document in DOM form. You can then call any XML::DOM methods on the document node and other nodes that its methods may return. See XML::DOM for full details.

    [% FOREACH node = doc.getElementsByTagName('CODEBASE') %]
       * [% node.getAttribute('href') %]
    [% END %]

This plugin also provides an AUTOLOAD method for XML::DOM::Node which calls getAttribute() for any undefined methods. Thus, you can use the short form of

    [% node.attrib %]

in place of

    [% node.getAttribute('attrib') %]

toTemplate()

NOTE: This method will soon be deprecated in favour of the VIEW based approach desribed below.

This method will process a template for the current node on which it is called. The template name is constructed from the node TagName with any optional 'prefix' and/or 'suffix' options applied. A 'default' template can be named to be used when the specific template cannot be found. The node object is available to the template as the 'node' variable.

Thus, for this XML fragment:

    <page title="Hello World!">
       ...
    </page>

and this template definition:

    [% BLOCK page %]
    Page: [% node.title %]
    [% END %]

the output of calling toTemplate() on the <page> node would be:

    Page: Hello World!

childrenToTemplate()

NOTE: This method will soon be deprecated in favour of the VIEW based approach desribed below.

Effectively calls toTemplate() for the current node and then for each of the node's children. By default, the parent template is processed first, followed by each of the children. The 'children' closure can be called from within the parent template to have them processed and output at that point. This then suppresses the children from being processed after the parent template.

Thus, for this XML fragment:

    <foo>
      <bar id="1"/>
      <bar id="2"/>
    </foo>

and these template definitions:

    [% BLOCK foo %]
    start of foo
    end of foo 
    [% END %]
    [% BLOCK bar %]
    bar [% node.id %]
    [% END %]

the output of calling childrenToTemplate() on the parent <foo> node would be:

    start of foo
    end of foo
    bar 1
    bar 2

Adding a call to [% children %] in the 'foo' template:

    [% BLOCK foo %]
    start of foo
    [% children %]
    end of foo 
    [% END %]

then creates output as:

    start of foo
    bar 1 
    bar 2
    end of foo

The 'children' closure can also be called as a method of the node, if you prefer:

    [% BLOCK foo %]
    start of foo
    [% node.children %]
    end of foo 
    [% END %]

The 'prune' closure is also defined and can be called as [% prune %] or [% node.prune %]. It prunes the currrent node, preventing any descendants from being further processed.

    [% BLOCK anynode %]
    [% node.toString; node.prune %]
    [% END %]

allChildrenToTemplate()

NOTE: This method will soon be deprecated in favour of the VIEW based approach desribed below.

Similar to childrenToTemplate() but processing all descendants (i.e. children of children and so on) recursively. This is identical to calling the childrenToTemplate() method with the 'deep' flag set to any true value.

PRESENTING DOM NODES USING VIEWS Index  ] [ Modules  ] [ Top  ]

You can define a VIEW to present all or part of a DOM tree by automatically mapping elements onto templates. Consider a source document like the following:

    <report>
      <section title="Introduction">
        <p>
        Blah blah.
        <ul>
          <li>Item 1</li>
          <li>item 2</li>
        </ul>
        </p>
      </section>
      <section title="The Gory Details">
        ...
      </section>
    </report>

We can load it up via the XML::DOM plugin and fetch the node for the <report> element.

    [% USE dom = XML.DOM;
       doc = dom.parse(file => filename);
       report = doc.getElementsByTagName('report')
    %]

We can then define a VIEW as follows to present this document fragment in a particular way. The Template::Manual::Views documentation contains further details on the VIEW directive and various configuration options it supports.

    [% VIEW report_view notfound='xmlstring' %]
       # handler block for a <report>...</report> element
       [% BLOCK report %]
          [% item.content(view) %]
       [% END %]
       # handler block for a <section title="...">...</section> element
       [% BLOCK section %]
       <h1>[% item.title %]</h1>
       [% item.content(view) %]
       [% END %]
       # default template block converts item to string representation
       [% BLOCK xmlstring; item.toString; END %]
       
       # block to generate simple text
       [% BLOCK text; item; END %]
    [% END %]

Each BLOCK defined within the VIEW represents a presentation style for a particular element or elements. The current node is available via the 'item' variable. Elements that contain other content can generate it according to the current view by calling [% item.content(view) %]. Elements that don't have a specific template defined are mapped to the 'xmlstring' template via the 'notfound' parameter specified in the VIEW header. This replicates the node as an XML string, effectively allowing general XML/XHTML markup to be passed through unmodified.

To present the report node via the view, we simply call:

    [% report_view.print(report) %]

The output from the above example would look something like this:

    <h1>Introduction</h1>
    <p>
    Blah blah.
    <ul>
      <li>Item 1</li>
      <li>item 2</li>
    </ul>
    </p>
  
    <h1>The Gory Details</h1>
    ...

To print just the content of the report node (i.e. don't process the 'report' template for the report node), you can call:

    [% report.content(report_view) %]

AUTHORS Index  ] [ Modules  ] [ Top  ]

This plugin module was written by Andy Wardley <[email protected]> and Simon Matthews <[email protected]>.

The XML::DOM module is by Enno Derksen <[email protected]> and Clark Cooper <[email protected]>. It extends the the XML::Parser module, also by Clark Cooper which itself is built on James Clark's expat library.

VERSION Index  ] [ Modules  ] [ Top  ]

2.6, distributed as part of the Template Toolkit version 2.14, released on 04 October 2004.

HISTORY Index  ] [ Modules  ] [ Top  ]

Version 2.5 : updated for use with version 1.27 of the XML::DOM module.

  • XML::DOM 1.27 now uses array references as the underlying data type for DOM nodes instead of hash array references. User data is now bound to the _UserData node entry instead of being forced directly into the node hash.

BUGS Index  ] [ Modules  ] [ Top  ]

The childrenToTemplate() and allChildrenToTemplate() methods can easily slip into deep recursion.

The 'verbose' and 'nospace' options are not documented. They may change in the near future.

COPYRIGHT Index  ] [ Modules  ] [ Top  ]

Copyright (C) 2000-2001 Andy Wardley, Simon Matthews. 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 Index  ] [ Modules  ] [ Top  ]

Template::Plugin, XML::DOM, XML::Parser


  Template::Plugin::Wrap
Template::Plugin::XML::RSS  
Tutorial
Manual
Modules
Library
Tools
FAQ
Release
http://www.template-toolkit.org/