> As noted in a previous thread, I am hitting a wall trying to run thecactus tests in Ant. I want to commit the changes I've made to help Maven run cactus tests, and I've already had to reconcile a bunch of conflicts once...
The error seems to come with the first Cactus test case (org.apache.struts.action.TestActionServlet) and is "Error instantiating class" -- so I'm guessing maybe I haven't gotten all of the necessary jars defined correctly (the build.properties.sample is pretty confusing the way it repeats blocks of properties over and over).
For those hanging on the edge of your seats... I had compiled a local copy of the Cactus/Ant integration jar which has patches to work with Mac OS X, and I neglected to also compile a copy of the core Cactus framework; apparently (and unsurprisingly, really), there was some compiled incompatibility between the versions. So compiling a CVS copy of Cactus got me past that problem... and a lot farther once I turned off the halt on failure/error values.
How much farther? well, I'm passing 83 out of 103 test classes cleanly; however, I've run the tests against a fresh repository checkout and I get the same results as I do against my modifications (barring a few line number differences in stack traces). It's a bit of a puzzle that configuration problems could lead to a distribution like that, with so many tests passing, and yet a substantial number failing.
If anyone has all the cactus tests passing under Ant and would like to give my "maven changes" a shot, let me know. Otherwise, stay tuned for the next exciting installment...
Joe
--
Joe Germuska [EMAIL PROTECTED] http://blog.germuska.com "We want beef in dessert if we can get it there."
-- Betty Hogan, Director of New Product Development, National Cattlemen's Beef Association
--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
