Thanks for collecting input from us. I have been using xalan in my applications for nearly 6 years, in various web containers (jboss/tomcat4 for dev, SunOne AS 8.1 in production).
My app assembles Oracle database records into XML trees, then transforms the tree with an XSL stylesheet that produces (X)HTML output. If RTF format is desired, a second transformer is used to transform the HTML into RTF. These stylesheets use xsl:message with the terminate='yes' attribute to abort the transformation if critical pieces of the input tree or incoming parameters between XSL templates are missing or invalid. The output is caught and logged in the webserver, but only a generic error is shown to the user. I have just read the JIRA record at https://issues.apache.org/jira/browse/XALANJ-2340, and I would not have a problem with the addition of a command-line switch to suppress it, as long as the location information remains intact in the other contexts. Timothy Jones -----Original Message----- From: Brian Minchau [mailto:[EMAIL PROTECTED] Sent: Monday, December 11, 2006 7:36 PM To: [email protected]; [email protected] Subject: A question on how users are using <xsl:message> Way back in history it seems that <xsl:message> produced just the message, but then some users asked for location information in the stylesheet of where the message was coming from. Recently XALANJ-2340 asks us to remove that location information, perhaps with some option. So the questions for users are: How are you using <xsl:message>? Is it being run on a server, or web-server, and the output of xsl:message going to stderr and caught in a log and processed by other tools, or what? What would happen for you if we removed the location information? - Brian
