[mythtv] Start using Mythtv/cvs

Matt Zimmerman mythtv-dev@snowman.net
Sun, 29 Dec 2002 13:51:40 -0500


On Sun, Dec 29, 2002 at 01:35:10PM -0500, Lego Andy wrote:

> Well in projects like CMake and VTK, we perform lots of tests every
> night to make sure things still work. When developers commit their
> changes, they make sure that things at least seem to work before they
> commit them in and so people actually use CVS versions for their
> work. Though this is not recommended, we do take that into an account
> when committing things into CVS. That said, people do break things
> every once in a while. 

If you would like to write a nice self-test suite for mythtv, I think that
would be great.  However, I suspect you'll find that it's a few orders of
magnitude more difficult than regression testing something like cmake or
vtk.

None of the problems you have reported have represented any loss of data or
even impaired your ability to use MythTV.  You seem to be implying that the
MythTV CVS code does not "at least seem to work", and that the developers do
not test significant changes before committing them.  This is insulting, and
seems to have no basis in fact.

> Sort of like Debian unstable. It does not really have to be completely
> stable, but there are several non debian developers who use it. It is not
> recommended, but I suspect things have to kind of work before they go in.

Debian unstable seems like a reasonable analogue.  The basic criteria for
unstable are things like "works for me" and "passes self-tests".  More
subtle problems than that are found through user testing.  Users volunteer
for testing by running unstable.

This seems to be more or less how MythTV CVS works as well.

-- 
 - mdz