The arguments we currently have in bin/jmeter.sh are about as bad as they could possibly be. I will now put some which will be better in 100% (or more) of cases, but which will still probably be sub-optimal in just as many occasions.
In order to establish reasonable values --that is, values which are close to optimal for at least one real-life[ish] situation, we would need a reasonably complex test script to run and a set of pages for it to hit.
Such pages and test script(s) could also serve as functional app-level tests. We currently have some unit tests (definitely not enough), but there are certain features which are extremely difficult to test at the unit level.
I propose that we start building such app-level tests. Ideas on how to harness them beyond a readme file with instructions?
I have some static pages + apache config files that can serve as a start for that -- I just need a suggestion for a path where I should place them in CVS.
-- Salut,
Jordi.
--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
