In the Apache Sling Scripting W3C XML Processing Support bundle, Juanjo Vázquez has implemented XSLT processing support for Sling as another new scripting engine, based on the Cocoon 3 pipeline engine.
An XML pipeline specifies a sequence of operations to be performed on zero or more XML documents. There are a number of advantages to using pipelines above all in terms of separation of concerns. On the other hand, we talk about XSLT processing if the operations in a pipeline are performed executing or applying XSLT stylesheets.
XSLT processing support is implemented in Sling as an scripting engine bundle named Apache Sling Scripting W3C XML Processing Support. This bundle is based on the Cocoon 3 pipeline engine and uses the W3C XProc language in order to specify pipelines to be processed.
For the time being, XProc is partially implemented and it´s not clear that Sling must support all W3C recomendation yet. This could depend of concrete user requirements or use cases.
The source code is found in the contrib/scripting/xproc module.
Install the org.apache.sling.scripting.xproc bundle in order to work with XProc. You can achieve this either building it from contrib/scripting/xproc folder in the Sling trunk or by downloading it from the Apache Snapshot repository here: org.apache.sling.scripting.xproc-2.0.0-incubator-20090403.114403-1.jar.
To deploy the bundle go to the bundles page of Apache Felix Web Console (http://localhost:8888/system/console/bundles), select the bundle file to upload, check the Start check box and click Install or Update button.
In order to check whether XProc scripting engine has been installed, go to the Script Engines page of the Apache Felix Web Console and see the entry for XProc there:
As explained above, the bundle is able to perform a sequence of XSLT transforms on an XML document just as is expressed in a pipeline definition. A pipeline definition is a file with an xpl extension that follows the W3C XProc grammar. Only p:xslt steps are supported at the moment.
For the XML input of pipeline, the processing uses a Cocoon generator named SlingGenerator that tries to resolve the requested resource as (in order of preference):
- an static XML file
- a dynamically generated XML
- the underlying node´s export document view
Let´s see some samples in order to understand the processing behaviour.
1. Create some content
2. Use WebDAV or curl to create a pipeline script at /apps/xproc/xproc.xpl :
3. Store the XSLT transforms in the repository:
4. Request foo.html to execute the pipeline:
In this case, the node´s document view has been the pipeline´s source.
5. Now, store an static XML in the repository:
6. Again, request foo.html to execute the pipeline:
This time the pipeline´s source has been an static XML file.
7. Store an script in the repository:
8. Delete previous static xml file (/foo.xml).
9. Request foo.html to execute the pipeline:
This time the pipeline´s source has been a dinamically generated XML.