sebb wrote: > I think that adding extra processing to JMeter to deal with processing > client-side Javascript is counter-productive in general.
I agree for the "in general". Load testing is not producing a DoS on the server therefore I think that what really matters is the ability to reach the charge that should be tested. > Although it would be convenient for some tests, it's a waste to keep > reprocessing the Javascript in order to decide what URL to sample next > - or to keep scanning the HTML to find out how to submit a particular > form. Using the JMeter Proxy to record the session allows the work to > be done by the browser when creating the test plan; it does not have > to be repeated for every test. It's not that simple. With even more logic on the client side, captured request aren't necessary those that should occur each time. Just imagine an application that asks the server for the number of logged in visitors and depending on that can have different behavior. When you record your test you will have a single user and this is what you would simulate with a large number of users. This may be interesting but this would have nothing to do with the real load that could happen on the application. An additional point is the maintainability. Cheers, Marc. -- Blog: http://mguillem.wordpress.com --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
