Peter Antman wrote:
On 25 Sep, Marcel Ruff wrote:
Peter Antman wrote:
It is good idea to have an XPath content filter similar to the regex plugin.Hi, I have written an XPath based I_Access filter. No rocket science, but quite handy to have if you have many types of XML content beeing published and.
1. Do not want to place the XML content in the key 2. Do not want all XPath queries beeing placed against all published messages.
With it you may subsribe on an exact oid, and only messages with that OID will be matched against the XPathFilter.
Do you want it in the XmlBlaster source tree? Now I have it in org.xmlBlaster.engine.mime.xpath.
(Its based on the Jaxen XPath engine, so three more jars has to be added to XmlBlaster.
//Peter
The 3 additional jars - how big are they?
[pra at pra lib]$ ls -l xpath/
total 204
-rw-rw-r-- 1 pra pra 160967 Sep 25 10:10 jaxen-core.jar
-rw-rw-r-- 1 pra pra 13707 Sep 25 10:10 jaxen-dom.jar
-rw-rw-r-- 1 pra pra 23563 Sep 25 10:11 saxpath.jar
OK, not too big.
Should we stuff it into a "lib/mimeXpath" directory and specify them to the plugin when loaded like:
MimeAccessPlugin[XPathFilter][1.0]=org.xmlBlaster.engine.mime.xpath.XPathFileter,classpath=mimeXpath/jaxen.jar:mimeXpath/xy.jar
I tried that when running the testsuite with only the new TestXPathSubscribeFilter, but it did not work at all. Got a strange and basically untracable Linkade exception on DocumentBuilderFactory, wich is not even part of the new jar-files.
Ahh, the classloader seems to have problems ...
Marcel
Is it possible to reuse the lib/ant/xalan.jar ??
Nope.
//Peter
Marcel