RFC 12

[Steve Lime]

I agree with Sean's points here. Mandating won't work well, at least not initially. Perhaps that will change in the future. Ease of use and integration into the build process (a 'make' forces the tests to be run) will really help. I'm certainly willing to give it a try. 

Anything that leads to less bug chasing and more stable would be most welcome.

Steve

>>> Sean Gillies <sgillies at FRII.COM> 02/18/06 3:34 PM >>>
Umberto,

I think you will have to remove the language about test driven  
development from RFC 12. It's a distraction from your main  
proposition: that C unit tests will be good for MapServer.

I myself am game to help you out with C unit tests. However, it's no  
use trying to mandate that people use them. MapServer just doesn't  
have a culture of testing, and that's all there is to it.
All we can do is write them, use them, encourage others to do so, and  
apply a little (friendly) shaming on people who write code that  
breaks them :) If the tests are easy to run, and useful, there's a  
good chance that developers will grab onto them, but that's the best  
you're going to get.

So, trim the RFC, and let's see a first good example of using cunit.  
I'd like to see if it could replace much of the Python mapscript test  
suite.

cheers,
Sean