peer to peer regression test proposal (was Re: QT controller on Vista: how to change volume?)

[Richard Gaskin]

Andre Garzia wrote:
> I have here a little proposal for the volunteers on this list. With
> every release of an OS or Revolution we have some things breaking.
> This is the nature of things, we can't change it but we can help fix
> it.
> 
> What about we build a community "regression test", some test stacks.
> Each volunteer or group of volunteer would build and mantain a little
> stack to test some subset of revolution. With each release and
> platform, they'd test it against the new thing and we could report
> back to runrev and web.
> 
> I can host this and I can build the report thing and I volunteer to
> build the regression test for socket routines and libURL (which never
> breaks!)
> 
> This way a small group can do some impact and help Bill at QC and
> RunRev team. Even if we can't test all of transcript and rev, if we
> test 40% if it, still we can guarantee that this 40% is working or
> broken which is nice anyway.
> 
> A group of people could take care of QT controler and routines testing
> and quickly catch things such as this volume issue very fast.
> 
> is this a good idea?

I think it's a great idea, but unfortunately some tests like what's 
needed to see this QT error require manual work.  So as long as there's 
a framework for both automated and manual tests it should be quite helpful.

-- 
  Richard Gaskin
  Managing Editor, revJournal
  _______________________________________________________
  Rev tips, tutorials and more: http://www.revJournal.com