DOC HOME SITE MAP MAN PAGES GNU INFO SEARCH
 

Apache::AxKit::Plugin::Fragment



NAME

Apache::AxKit::Plugin::Fragment - Fragment plugin


SYNOPSIS

  AxAddPlugin Apache::AxKit::Plugin::Fragment


DESCRIPTION

This module provides direct web access to XML fragments, using an XPath syntax. By simply providing a querystring containing an XPath query, this module will set the XML to be parsed to be the XML nodes returned by the query. The nodes will be wrapped in either <resultset>...</resultset> or you can specify the outer tag using:

        PerlSetVar XPathFragmentElement foo

to wrap it in <foo>...</foo>.


USAGE

Simply add this module to the plugin list before StyleFinder:

        PerlHandler Apache::AxKit::Plugins::Fragment \
                        AxKit

Then request a URL as follows:

        http://server/myfile.xml?/some/xpath/query

Queries that match the regular expression: ^\w+= are ignored, as are any invalid XPath queries.

Note that it's important to write your stylesheet to make use of this capability! If you intend to use this Fragment plugin, you can't assume that your stylesheet will just magically work. It will have to not make assumptions about the XML being passed into it. The apply_templates() method of XPathScript is extremely useful here, as is the xpath query 'name(/child::node())' which identifies the top level element's name. Here's how I got around this with my first experiments with this:

        <!-- Main document body -->
        <% if (findvalue('name(/child::node())') eq 'page') { %>
                <%= apply_templates('/page/body/section') %>
        <% } else { %>
                <%= apply_templates('/') %>
        <% } %>

Which checks that the top level element is called 'page', otherwise it simply does apply_templates() on all the nodes.