XOM API for Processing XML with Java

[Steve Loughran <steve.loughran AT gmail.com>]

On 6/30/05, Andrew Stewart <andrew.stewart AT aspectcapital.com> wrote:
> 
> > > http://www.clarkware.com/software/JUnitPerf.html
> >
> > How does this handle varying system speeds? i.e. if the tests
> > were run on two machines, one twice as fast as the other,
> > could they pass on the faster system and fail on the slower?
> > Ideally I'd like the tests to be relative to some underlying
> > measure of speed rather than absolute numbers.
> 
> 
> As far as I know, it does not take system speed into consideration.  So
> tests could pass on a fast machine and fail on a slow machine.
> 
> 

yeah, its very, very brittle. 

I don't personally think you can/should encode response times in the
test suite itself. really that should be data calibrated on a single
box and saved, then future runs compared to detect significant
deviations from the existing stuff.

I am not aware of a junit extension that does that. 

-steve