You are viewing a plain text version of this content. The canonical link for it is here.
Posted to c-users@xalan.apache.org by Donald Leslie <dl...@apache.org> on 2002/09/22 20:09:36 UTC
Xalan-C++ 1.4 now available
We have just posted the Xalan-C++ version 1.4 distribution files and
updated the Xalan-C++ project site to reflect the new release.
The new release includes bug fixes, a 64-bit build, use of a new Linux
platform and compiler, alpha support for a number of EXSLT functions, and
a new sample application that illustrates how to perform transformations
with input in the form of a pre-built XalanDOM or XalanSourceTree.
For more information about what is new in this release, see
http://xml.apache.org/xalan-c/readme.html#done.
Xalan-C++ 1.4 works with Xerces-C++ 2.1. For information about downloading
Xalan-C++ 1.3 and Xerces-C++ 2.1, see
http://xml.apache.org/xalan-c/getstarted.html#download.
Schema question
Posted by Brian Quinlan <br...@sweetapp.com>.
It looks like Xalan 1.4 (really Xerces 2.1) is trying to load a schema
file when I turn validation on (using XMLParserLiason.setUseValidation).
My XML file looks like:
<!DOCTYPE message SYSTEM "test.dtd">
<messages><message>bye bye!</message></messages>
Is it intended that enabling validation triggers the search for a schema
file when the XML file already contains a DTD?
Cheers,
Brian
SystemID question
Posted by Brian Quinlan <br...@sweetapp.com>.
If I am doing a transformation from an in-memory XSLTInputSource, can I
set the system id of the input source to anything that I want?
Cheers,
Brian
Xalan-C 1.4 seems to report style nodes to problem listeners less
faithfully
Posted by Brian Quinlan <br...@sweetapp.com>.
In Xalan-C 1.3, usage of an undefined variable causes the
ProblemListener to be called with a valid style node. In Xalan-C 1.4,
the style node passed to the ProblemListener is NULL.
The same is true of messages generated using xsl:message.
Is this a bug or a feature that has been deliberately removed?
OTOH, the URI and locator information passed to the ProblemListener are
now more accurate :-)
Cheers,
Brian