I have found at least one bug. http://issues.apache.org/jira/browse/JXPATH-68
I created a quick fix for this, for my own use, but don't really think it is the correct solution. I've actually found a couple of other bugs surrounding the automatic creation of paths based on the AbstractFactory. Joseph Campolongo -----Original Message----- From: Paul Copeland [mailto:[EMAIL PROTECTED] Sent: Monday, July 10, 2006 2:00 PM To: Jakarta Commons Users List Subject: Re: <JXPath> RE: Is JXPath used and/or loved Good to see there are several JXPath users out there. I am not running into JXPath bugs (other than the javadoc left out of the distribution which breaks links in the manual - seems like such an obvious thing that it's strange it wasn't corrected). Two things I am wrestling with - (1) understanding how JXPath deals with cyclic references. I have my own solution to this with a jxpath DynamicPropertyHandler class. But now I see the documentation implies that JXPath handles this (seemed like my first testing found infinite loops in cyclic cases). If anybody has experience with that issue it would be interesting to hear about. (2) turning off access to java classes and functions within XPath expressions. This is a security hole in our application. If anybody knows how to do that it would be very helpful to hear how. - Paul Copeland Joseph Campolongo wrote: >I have been using JXPath for the last month or so, since it was built >specifically for my use case, even though it appears the project has been >wound down (I haven't seen any activity on it anywhere, by anybody). > >Joseph Campolongo > >-----Original Message----- >From: Jess Holle [mailto:[EMAIL PROTECTED] >Sent: Sunday, July 09, 2006 10:15 AM >To: Jakarta Commons Users List >Subject: Re: Is JXPath used and/or loved > >Hmmm.... > >It has been a while, but last time we checked it out it had >bugs-a-plenty when compared to something like Xalan for various XPath >expressions. > >Performance of recent Xalan versions (including that behind Java 5's >built-in XPath implementation) sucks for many XPath use cases as they >assume you're going to work a long time on each document without >changing it, i.e. the XSLT use case. We therefore have to stick with >Xalan 2.1.0 for such cases, but last I checked it still handled more >XPath expressions correctly than JXPath, Jaxen, etc. > >I'd love to see someone actually produce a highly standards compliant >XPath library for Java that worked well against a "live" DOM, i.e. >returned original DOM nodes and allowed changes to the DOM between XPath >executions without penalty. > >For all I know, Saxon may now provide this by implementing the new JAXP >XPath stuff. For the longest time it was all but impossible to get >"live", DOM nodes back from it's XPath API, but I assume that has >changed. That leaves questions about the penalty for changing the DOM, >of course. > >Torsten Curdt wrote: > > >>Over in cocoon land it is being used all over the place. >>I think it rocks ...yepp >> >>cheers >>-- >>Torsten >> >>--------------------------------------------------------------------- >>To unsubscribe, e-mail: [EMAIL PROTECTED] >>For additional commands, e-mail: [EMAIL PROTECTED] >> >> >> > >--------------------------------------------------------------------- >To unsubscribe, e-mail: [EMAIL PROTECTED] >For additional commands, e-mail: [EMAIL PROTECTED] > >--------------------------------------------------------------------- >To unsubscribe, e-mail: [EMAIL PROTECTED] >For additional commands, e-mail: [EMAIL PROTECTED] > > > > > > --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
