[V3dot10] Re: 3.10

[Marcus Denker]

On 14.12.2006, at 11:57, Ralph Johnson wrote:

> It is important to keep SUnit for Squeak compatible with SUnit for
> VisualWorks and other Smalltalks.  It won't be identical, of course,
> but it should be fairly easy to write tests that run on any version of
> Smalltalk, once you get past the problem of actually reading them in.
>
> That is why messing with SUnit makes me nervous.  The version we have
> is pretty standard

I have to admit that I did add a little to SUnit...
	-> I refactored it to not call the sunit* methods everywhere (that just
             forwarded to the same method in squeak without the prefix)
         -> Added the logic for "expected failures"
	-> replaced the testrunner with Lukas' version.

but overall, I think it would be best to let SUnit be Sunit... and  
instead
put all the effort of new and likelly incompatible things into a fork  
that then
can take over at some point in the future.

SUnit has not seen much developent, so I think that we should not
restrict ourselves to much to be compatible with a quite dead project.
But on the other hand, a forked version should have another name, I  
think.
I should have done that with the changes I did, too.
SUnit's main goal is not to be a test-framework, but to run on all  
Smalltalks,
even at the expense of features or evolution of the framwork itself.

Lukas Renggli has some very nice first version of an improved Test
Framwork, too, that might be interesting to look at.


	Marcus