[Opensim-dev] Another idea for the OpenSimulator 0.7 roadmap - increased automated unit test / regression testing

[Justin Clark-Casey]

In addition to the ideas that MW (and others) have discussed, I'd also 
like to put forward the idea of increasing our level of automated unit 
and regression tests for the 0.7 release.

Previously I was more in favour of pushing forward quickly rather than 
spending too much time test writing (and I think this was how it was 
when I joined the project).  However. we're now approaching the point 
where a reasonable amount of functionality has been filled out in 
OpenSim (at least from the Second Life POV).  Greater reliability is 
going to be necessary if the platform is going to be used seriously for 
grid and enterprise level activity.   Naturally, there's a level of 
self/company-interest here ;-) but I think that spending time on this 
will benefit many if not all of us.

Of course, automated tests certainly won't be a perfect safety net but 
hopefully it would help to improve OpenSim by catching the revisions in 
which regressions are introduced.  This will also mean refactoring will 
be much less error prone.  I'm most probably preaching to the choir here.

I think that part of this will mean pushing the test results from builds 
into the IRC channel (as already discussed with Sean), and making test 
results more prominent (perhaps even on the opensimulator.org front page?)

Personally, I hope to start putting some of FRI's coding time where my 
mouth is and start doing some work towards this this later on next month.

-- 
justincc
Justin Clark-Casey
http://justincc.wordpress.com