Saxon 10.3 PE and SqlExtension
Added by Sebastian Vogel almost 4 years ago
Hello dear Saxon Team!
I am trying to get saxon-10.3 PE running with SqlExtensions but I get this message on the command line when I try to run a transformation:
Warning on line 5 column 112 of saxonConfig.xml: Extension elements are not available in Saxon-HE Extension elements are not available in Saxon-HE
I run the command via:
java -cp /home/node/saxon-pe/saxon-pe-10.3.jar:/home/node/saxon-pe/saxon-sql-10.3.jar net.sf.saxon.Transform -s:/tmp/cXML.xml -xsl:/tmp/xslt.xslt -o:/tmp/saleOrder.xml -config:/tmp/saxonConfig.xml -t
I have a valid evaluation license in the same location as the saxon-pe-10.3.jar file
My saxonConfig.xml file looks like this
<?xml version="1.0" encoding="UTF-8"?>
<configuration xmlns="http://saxon.sf.net/ns/configuration" edition="PE">
<global allowExternalFunctions="true" versionOfXml="1.0"/>
<xslt>
<extensionElement namespace="http://saxon.sf.net/sql" factory="net.sf.saxon.option.sql.SQLElementFactory"/>
</xslt>
</configuration>
Why is the command line reporting that I am running HE but the binary and the config say PE? The command is failing and it's not outputting any data.
Regards Sebastian
Replies (2)
RE: Saxon 10.3 PE and SqlExtension - Added by Michael Kay almost 4 years ago
Curious. I've reproduced it from the command line, but having no luck so far trying to debug it in the IDE.
Internally, there's a ConfigurationReader class for Saxon-HE, and a ConfigurationReaderPE class which overrides it selectively for PE. The error message is coming from the HE version of a method ConfigurationReader.readExtensionElement(), which is overridden in ConfigurationReaderPE. Java loader tracing (-verbose:class) shows that a ProfessionalConfiguration and a ConfigurationReaderPE are being loaded. So I'm having trouble seeing why the HE version of the relevant method is being called.
RE: Saxon 10.3 PE and SqlExtension - Added by Michael Kay almost 4 years ago
This is Bug #4849 -- I thought it looked familiar.
But although it's marked resolved, as far as I can see the fix isn't committed on the 10 branch (let alone being released).
A bit of further investigation reveals what happened: we were in the middle of a move to a new repository structure at the time, and I committed the fix on the old repository.
Please register to reply