On Mon, 16 Jun 2003, Dave Thomas wrote:
>
> [...] in an ideal world, the tests would be
> updated by the author of every functional change.
>

That seems like a very good idea.

> I view Rubicon as a totally shared resource. If folks have fixes or new
> tests to apply, I'd welcome them.
>

I understand.
I'll try to write/refine some tests ;)

In my original mail I asked som questions about the "versioning" of
the tests. Maybe I can ignore that for now. I assume that writing a
test without any "versioning" can be a first approximation ...

But I also assume that any new/changed features from 1.6 ought to
have a "version-guard" (if one knows that it is an intentionally
changed behaviour).

/Johan Holmberg