Robert Bresner wrote:
[... Drawing omitted...]
> This drawing works for me because it shows that when the V1.0 branch
> is made, all of the files are at the same point in time, ergo, are
> exactly the same. For some reason, people hear "branch" and imagined
> some magical transformation had taken place in all the files.
As a matter of fact, some (not-quite) magical transformation *has* taken place
in all of the files. Namely a branch tag has been added into all the RCS
files.
> And I
> would point out that MAINLINE is always MAINLINE (which is what we
> call it here), but when we make a V2.0 branch, it's not crazy to refer
> to MAINLINE as V3.0, which it will become eventually...
You might not want to use numbers like "1.0", "2.0", "3.0", because these
revision numbers a) do not correspond to the default revision numbers CVS will
most likely use, b) might encourage people to try to force CVS to use these
revision numbers (which is a very bad idea), c) even if you try to force CVS to
use particular revisions, somebody else is surely going to come along and
(unknowingly) use "cvs commit -r" to disrupt your "nice" scheme (anybody with
commit access can do this).
I would recommend modifying the explanation to omit revision numbers and just
use symbolic names. That's what CVS does anyway, the revision numbers for each
file on a particular branch may vary wildly. Or maybe explain that revision
numbers on individual files don't correspond to the numbers you're using in
your branch names. And BTW, "." is not allowed as part of a branch name,
something you might want to consider in your drawing.
Just my 2 cents.
-- steve
__________________________________________________
Do You Yahoo!?
Get Yahoo! Mail � Free email you can access from anywhere!
http://mail.yahoo.com/
