¼,
½ and ¾ was broken.
² and
³ (superscript two and superscript three)
was ⊃ and has been corrected to ²
and ³
-help option to print out
the command-line help.
-model to set different XPath model
implementations. This is a convenient way to integrate
jd.bigxml.
"/ | *" and
pattern2 = "item".use attribute of a key was evaluated then
the current node was not defined. It is now equal to the context node.
-validate option had no effect when used for
the identity transformation (i.e. the predefined stylesheet url
urn:jdxslt:identity).
-trace option to log
evaluated expressions, patterns and instantiated templates and template rules.
IndexOutOfBoundsException occurred when large
text strings (with length > 4096 characters) were written to the output.
IndexOutOfBoundsException occurred when the
length argument passed to the substring function was negative.
<xsl:strip-space elements="*"/> did not
work when applied to elements with a non null namespace uri.
checkDocumentRead(String uri)
to the XsltSecurityManager API, to control the security risk of the
document function.
href).
xsl:excluded-result-prefix
atttribute had value '#default' and the default namespace was not declared. Now the
exclusion is ignored.
-msglistener to
set a listener for xsl:message outputs.
-repeat option is set then in verbose mode
only a summary of the transformation time is printed.
jd.xml.xpath.model.XPathNode (e.g XPathNode.getXpName()
is now XPathNode.getName()).
com.sarvega.JdXsltDriver
for the Sarvega XSLT benchmark.
System.exit() with an suitable
exit code.
build.xml to compile the source,
create the jar file and build the javadocs.
jd.jar to jdxslt.jar and
the distribution file jd.zip to jdxslt.zip.
xsl:script instruction for script languages you will have to upgrade your
third party libraries accordingly.
-verbose option contains now
the peek memory usage during the transformation.
urn:jdxslt:identity, since the old one was not syntactically correct.
empty which does not produce
any output. Especially it can be used with the verbose option
to measure the transformation time without the output time.
xsl:message texts are now per default output to System.err instead of System.out.
doc\api now contains an Ant buildfile if you want to build the
documentation for yourself.
<xsl:value-of select="document('book.xml#chapter1')">
urn:jd.xslt:identity. The old uri {identity} still
works but is no longer documented.
-out:method value which allowed
to set the output method was changed to allow to set every property of
the xsl:output instruction: Use -out:property value where
property is the name of any xsl:output property.
-validate to instruct the xml parser
to validate documents. (Default is not to validate).
jd.xml.xslt.Stylesheet: Instead of callingjava -cp jd.jar jd.xml.xslt.Stylesheet [args], you can now writejava -jar jd.jar [args].
java jd.xml.xslt.Stylesheet test.xml test.xsl {identity}).
canonical-xml which produces
a XML document in canonical form.
{identity} as stylesheet URI to denote the
identity transformation.
-out:method which lets you override
the output method specified by the stylesheet. In conjunction with the identity tranformation
this provides an easy way to convert a XML document to its canonical form: java jd.xml.xslt.Stylesheet -out:method canonical-xml test.xml {identity}
-w0, -w1, -w2
to specify how to treat a recoverable error.
-ignoredocerr whose functionality
is now handled in a more general way by the -wx options.
-uriresolver options is
now also used to resolve stylesheet uris in xsl:import and xsl:include instructions.
-entitiyresolver to set a SAX EntityResolver.
-repeat, -ignoredocerr
and -proxy.
Detailed description of this options can be found here.
This version contains a major rewrite of the internal tree model. Until the last version, DOM 2
was used as representation of the input XML document. Because of the differences between
the XPath tree model and DOM a new internal tree model was created which fulfills the needs
of a clear and performant implementation of the XPath tree model.
As an effect this version gains a performance increase of 40-50% on the XSLTMark benchmark
compared to the last version.
Great care was taken that this major change will not introduce any unintentional bugs (and of course
no intended bugs too). Besides my own ever increasing test suite, this version was tested
against the XSLT tests of NIST, the public available tests of other popular processors and
with Docbook example documents.
-parser, -parserxml,
-parserxsl, -uriresolver and -errorstack.
Details can be found on the new Command-line page.