----- Original Message ----- From: "Noel J. Bergman" <[EMAIL PROTECTED]> > Harmeet, > > I have to say that I am unsure. Yes, I think that there should be a testing > sub-project. But is this something that needs to be dealt with now, or > afterwards as a PMC?
Either way seems good. It maybe a new proposal idea, but testing has been talked about for a looong time. A separate project may be the best way to pull together, focus testing effort and invite wider participation for testing. > > > Objective would be to have Protocol tests for > > - RFC Compliance > > - Load Testing > > - Performance Tesing. > > Yes in concept. And tools for doing those things. One thing I don't know > is whether or not the testing sub-project ought to work with the JMeter > folks to get our protocols into that code, or whatever project might replace > it. There are lot of Mail Servers, open source and otherwise. It would be useful and help harness energy and provide a common testing project for Mail Protocol Testing. There is good code and ideas already. I think this combination can yield a good structure to build tests on. - Ant tasks (already there.) - Protocol Simulation ability (in 2 places in James.) - A scripting/template facility to tie things together (Written locally but not committed. Requires library addition, long emails etc) - Possible Junit tie in - (already there) The main problem would be writing the large number of tests. For this we could interest developers from other opensource projects. There seems to be a lot of common value that can be realized. Having a general testing project would have these benefits. - Interest developers from other open source or commercial projects. There seems to be a lot of protocol servers and testing could provide common value to all at the same time help James. - Testing is a big and important task. It would be good to have a separate validation project. Regd: JMeter and other testing projects. I think we should investigate this. load/performance analysis would be esp. useful to reuse. I think it is good to leverage where possible, but we should have rfc testing for mail protocol as the primary goal. I don't think JMeter would care much about testing n variants of SMTP protocol. > > Testing could become a pretty consuming project > Exactly. If we can have a James neutral sub-project defination we may be able to solicit help from other mail server projects and reduce our testing burden. At the same time build a community that benefits from this. > requiring someone to focus > on it every bit as much as a release manager has to focus on a release. I have no problem with doing this. Darrell has done a lot of work. You have a lot of ideas on this. There may be others passionate about this too. Harmeet -- To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]> For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>
