Home > Articles

The XSLT Transformation Environment

  • Print
  • + Share This
Learn the basics of XSLT by going through the aspects of the environment and where the XSLT processor is used. Topics include stylesheet basics and communicating with the processor.
This chapter is from the book

This chapter is from the book

Some aspects of our stylesheets are of a global nature that we must consider before we delve into the details of template instruction execution and behaviors. This chapter overviews the aspects of our environment in which we use our XSLT processor from the perspectives of the stylesheet contents, the serialized output (if any), and the operator invoking the processor.

All explicitly declared stylesheets follow a required shape of container and top-level (container children) elements. Methods are also specified for including arbitrary information in a stylesheet file, useful for supplemental information for processing, or as documentation of the stylesheet content.

The stylesheet can declare its desire for values of certain parameters of the output serialization that influence the contents of the reified result node tree.

In addition, there are a number of ways available to communicate with an XSLT processor that is interpreting a stylesheet resource against a source resource. Communication to the processor can be engaged at invocation as well as from the processor during execution.

Finally, this chapter reviews aspects of the transformation environment that cannot be controlled by the operator or the stylesheet. It is important to understand the limitations of what can be asked for or even supported by the XSLT processor.

This chapter includes discussion of the following XSLT instructions regarding the transformation environment in which a stylesheet is used.

Instructions for wrapping the content of a stylesheet are as follows.

  • <xsl:stylesheet>

    • encapsulates a stylesheet specification.

  • <xsl:transform>

    • encapsulates a stylesheet specification.

Instructions for serializing the result tree are as follows.

  • <xsl:namespace-alias>

    • specifies a result tree namespace translation.

  • <xsl:output>

    • specifies the desired serialization of the result tree.

Instructions for communicating with the operator are as follows.

  • <xsl:message>

    • reports a stylesheet condition to the operator.

  • <xsl:param>

    • supplies a parameterized value from the operator.

5.1 Stylesheet basics

5.1.1 The stylesheet document/container element

Two identical and interchangeable choices for the document element are:

  • <xsl:stylesheet>

  • <xsl:transform>

They can also be used as a container element for a stylesheet embedded in another context.

  • They may use id="unique identifier".

    • It identifies the stylesheet when there are multiple ones from which to choose.

    • The use of this XML IDattribute is outside the scope of the Recommendation.

    • It could be used as a fragment identifier by the stylesheet association processing instruction or by other techniques to identify a given stylesheet among many.

Controls available on container element or any literal result element are:

  • exclude-result-prefixes="whitespace-separated-prefixes"  

    • scope of influence is all descendent elements in stylesheet;  

    • this declaration indicates which stylesheet namespace prefixes are not expected in the result, thus are not to be included in the stylesheet tree;  

      • a list of whitespace-separated namespace prefixes specifies prefixes that are to be explicitly excluded from the stylesheet tree (using #default as the name to reference the default namespace, which is sometimes unofficiallycalled the null namespace);  

    • user-specified prefixes and associated namespace declarations are often used in XSLT stylesheets (but not desired in the result) for various purposes such as:

      • top-level documentation,

      • embedded structured data,

      • named XSLT constructs;

    • recall that copying an element node from the stylesheet to the result will copy all attached namespace nodes, thus stylesheet namespace declarations can easily end up in the result tree;

      • a stylesheet wrapper-element namespace declaration is typically used for top-level namespace usage, thus the document element of the result will typically end up with the same declarations;

    • this exclusion declaration tells the XSLT processor to not include the specified namespace nodes on descendent nodes of the stylesheet tree;

    • this exclusion declaration has no effect on namespace nodes of the source tree,

  • extension-element-prefixes="whitespace-separated-prefixes"

    • scope of influence is all descendent elements in stylesheet;

    • this declaration indicates which stylesheet namespace prefixes are instruction prefixes;

      • a list of whitespace-separated namespace prefixes specifying prefixes that are extension namespaces to be recognized by the XSLT processor (using #default as the name to reference the default namespace);

    • recall that everything that is not an instruction is considered to be a literal result element;

    • elements prefixed with the namespace prefix associated with the XSLT URI are interpreted as instructions;

    • this declaration tells the XSLT processor what other prefixes are to be interpreted as instructions because they are extension elements required by the stylesheet;

    • the processor need not implement the extension elements (detailed in Chapter 6).

Child elements of the document or container element are referred to as "top-level" elements.

  • If present, the following must occur before all other top-level elements:

    • xsl:import

      • see Chapter 6.

  • If present, the following (listed alphabetically) may occur in any order as top-level elements:

    • xsl:attribute-set

      • see Chapter 7,

    • xsl:include

      • see Chapter 6,

    • xsl:key

      • see Chapter 8,

    • xsl:decimal-format

      • see Chapter 8,

    • xsl:namespace-alias

      • see this chapter,

    • xsl:output

      • see this chapter,

    • xsl:preserve-space

      • see Chapter 3,

    • xsl:strip-space

      • see Chapter 3,

    • xsl:template

      • see Chapter 4.

  • The following are used not only as top-level elements, while all others listed above are only used as top-level elements:

    • xsl:param

      • see Chapter 6,

    • xsl:variable

      • 6 see Chapter 6.

Because an XSL stylesheet is an XML document —

  • XML comments can be used to provide documentation about the stylesheet,

  • all XML comments and processing instructions found in an XSL stylesheet are ignored.

    • Note that some XML editing tools may leave processing instructions in files for remembering locations such as the last cursor position.

Adding richly marked up documentation to a stylesheet:

  • allows the stylesheet to be run through a documenting stylesheet to extract the documentation in any fashion desired,

  • is accomplished by including non-XSLT constructs as top-level elements (children of the stylesheet document element) provided that the default namespace is not used as the namespace for such constructs, as in the following example:

Example 5–1 Using non-XSLT constructs as top-level elements

Line 1  <?xml version="1.0"?>    <!--hellodoc.xsl-->
     2  <!--XSLT 1.0 - http://www.CraneSoftwrights.com/training -->
     4  <xsl:transform xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
     5  version="1.0" exclude-result-prefixes="mydoc"
     6  xmlns:mydoc="http://www.mycompany.com/mydoc">
     8  <xsl:output method="html"/>
    10  <mydoc:para>
    11  The following construct is the root template.
    12  </mydoc:para>
    14  <xsl:template match="/">	<!--root rule-->
    15  <b><i><u><xsl:value-of select="greeting"/></u></i></b>
    16  <?test a processing instruction here?>
    17  </xsl:template>
    19  </xsl:transform>

Note the use of exclude-result-prefixes= in the document element above to tell the XSLT processor to not emit a namespace declaration for the prefix of the documentation namespace —

  • if the stylesheet writer knows that namespace will never be needed in the result;

  • because the XSLT processor doesn't know when creating the document element node of the result tree whether the namespace will ever be needed in the instance, so by default the declaration is emitted.

5.1.3 Namespace protection

Some special concern regarding the use of namespaces are as follows.

  • The "transformation by example" paradigm utilizes literal result elements.

    • It represents result tree element nodes with associated attribute nodes.

      • An element is written with associated attributes in a template in the stylesheet and can use —  

        • the default namespace,

        • a namespace prefix and associated namespace URI.

      • An alternative described later is the use of XSLT instructions to synthesize result tree nodes.

  • Some namespaces can be sensitive in the document processing environment;

    • this includes automatically-triggered platform services;

    • for example: digital signature processing.

  • If the result tree requires a sensitive namespace, the stylesheet can't use the namespace in a literal result element —

    • to produce an XSLT script as the output of translation;

      • the XSLT processor would incorrectly interpret the result vocabulary as input,

    • to use a platform service for the output of translation;

      • the stylesheet use of the URI would incorrectly trigger the service.

  • The <xsl:namespace-alias attributes/> top-level element can be used, which:

    • is an instruction to translate a namespace prefix in the stylesheet into another namespace prefix when used in the result;

    • with attribute stylesheet-prefix="prefix":

      • specifies the prefix used in the stylesheet tree that is being added to the result tree by the stylesheet,

        • has no influence or recognition in the source tree,

    • with attribute result-prefix="prefix":

      • specifies the prefix of the stylesheet tree whose URI is to be used for the result tree prefix,

      • must have this attribute declared in the stylesheet even if no element in the stylesheet uses the prefix.

Note in the example below how the XSLT namespace URI cannot be used for the declaration for the xslo prefix, otherwise the xslo prefixed elements would be interpreted as XSLT instructions.

Example 5–2 A stylesheet that produces a stylesheet

Line 1  <?xml version="1.0"?>    <!--xsl.xsl-->
     2  <!--XSLT 1.0 - http://www.CraneSoftwrights.com/training -->
     4  <xsl:stylesheet xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
     5  xmlns:xslo="any-URI" version="1.0">
     7  <xsl:output indent="yes"/>
     9  <xsl:namespace-alias stylesheet-prefix="xslo"
    10  result-prefix="xsl"/>
    12  <xsl:template match="/">    <!--root rule-->
    13  <xslo:stylesheet version="1.0">
    14  <xslo:template match="/">
    15  <html>
    16  <p>Hello world</p>
    17  </html>
    18  </xslo:template>
    19  </xslo:stylesheet>
    20  </xsl:template>
    22  </xsl:stylesheet>
  • When this particular stylesheet is run with itself (or any XML file) as the source, the XSLT processor will assign the "xslo:" prefix's URI used in the result tree with the URI for the "xsl" prefix as indicated in the <xsl:namespace-alias> instruction, thus using the XSLT URI when the result tree is serialized as XML markup:

Example 5–3 A stylesheet produced by a stylesheet

Line 1  <xslo:stylesheet version="1.0"
     2  xmlns:xslo="http://www.w3.org/1999/XSL/Transform">
     3  <xslo:template match="/">
     4  <html>
     5  <p>Hello world</p>
     6  </html>
     7  </xslo:template>
     8  </xslo:stylesheet>
  • When run with itself (or any XML file) as the source, the output will be:

Example 5–4 The result produced by the produced stylesheet

Line 1  <!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN"> 
     2  <html> 
     3  <p>Hello world</p>
     4  </html>
  • + Share This
  • 🔖 Save To Your Account