> "Jim.Hyslop" <[EMAIL PROTECTED]> wrote:
>
> Andy Jones [mailto:[EMAIL PROTECTED] wrote:
> > Allow me to suggest a simpler alternative.
> >
> > Developers do not book code into CVS until it is ready to be
> > promoted to QA.
> Pardon me for being blunt, but this is a Really Bad Idea. This encourages
> people to leave code out of the repository for weeks or even months on end.
> There are many nasty side effects to this practise, including the risk of
> losing code if your hard drive fails, and increasing the difficulty and
> amount of time it takes to integrate code once the code is ready to be
> checked in.
Why wouldn't the developers use a branch within the same CVS repository? We have a
formal "inspection" process before anything gets committed to the main branch (HEAD),
which is the one built on a nightly basis and what is released to QA.
> CVS works best when you check in frequently, and update your code from the
> repository frequently. By frequently, I mean at least once a day.
Agreed. I check-in to my branch very often, usually after I have written an
interesting piece of code.
> > Presumably you would not want code included
> > in the automatic build unless the developers had completed
> > it, so your "test" tag works fine as it is.
> There's a difference between completing code, and having the code ready to
> be reviewed by QA. It also depends on the purpose of your daily build - our
> daily build, for example, always uses the head of the trunk, and is
> primarily used as a sanity check to make sure nobody broke the build.
Our QAs don't see code, they see the compiled product, but the means to this end is
the same: QA sees the head of the trunk, developers use branches and there is a formal
gateway between those. Whenever developers release to QA, we formally tag the
repository so we know what QA got at that point in time.
Philippe
_______________________________________________
Info-cvs mailing list
[EMAIL PROTECTED]
http://mail.gnu.org/mailman/listinfo/info-cvs
