[Simulavr-devel] regression tests

[address@hidden]

On Wed Sep 30  0:53 , "Weddington, Eric"  sent:

>I don't care how the project progresses, meaning that I don't have any say in
architecture or direction, or how a particular feature should be implemented. My
only personal concern is that it continues to build for, and run on, Windows
either Cygwin or MinGW (preferrably MinGW). I think that no one wants that the
project go "backwards", meaning that it continues to build on the same 
platforms,
that there aren't any serious regressions. Yes, I understand that there is very
little testing in place to ensure this, so I have to trust that the developers
keep this in mind when working on this project.

Portability issues shouldn't be too much of a problem.
The sockets and the guis are the most likely prospects.
Most of the rest is just a matter of doing the math.

>This is why I'm more than happy to add people to the project as developers or
admins, as long as these overarching goals are kept in mind. I don't think that
consensus would be that hard to reach in getting features added.

One of the features I would like to add is regression tests for IN and OUT.
They will need to be cpu-dependent.
They will access the port registers.
Do the savannah code and the Onno code use
the same regression test organization?

IIRC the Onno code uses a different port register
access syntax from the savannah code.
If so, I'd like a discussion about whether
to commit it to the savannah code.

--
Michael Hennebry
address@hidden
"War is only a hobby."
---- Msg sent via CableONE.net MyMail - http://www.cableone.net