@conrad - Sorry it took a few days to get back with you, the burden of
OSS I guess :)
To speak to our current implementation, the ant tasks make a command
line call the air test runner, it's the only abstraction the ant tasks
perform.  The airtestrunner has a dependency on the fluint.swc which
your tests are built against, calling the same methods as if you
created a FlexTestRunner.  So each component in fluint reuses the same
code you write your tests with, so I'm confused on how the ant task
which calls the air runner would provide different output than calling
the air runner from the command line directly.  Am I misunderstanding
your question?
Looking over your code, it looks correct.  Can you confirm that the
same module SWF is being run against your ant script and air test
runner?  This may be a bug with the XML output not showing the stack
trace from the setUp/tearDowns method however.   To confirm this, can
you run the airtestrunner without the headless argument and see if the
visual runner give you debugging output?  Let me know what you can
find out and we'll figure out what's going on.
-Brian
On May 1, 6:35 am, "
conrad.winches...@googlemail.com"