Experimental Physics and
Industrial Control System

[email protected] (Jeff Hill) · Tue, 16 Nov 1999 16:21:48 -0700

> I followed the thread so far, and all I can say is: personally I am fine
> with any solution that enables me to control the number of people with
> write permission on the CVS subtree in question. For example, I'd rather
> make sure that these people understand CVS enough to make a branch
> before they check-in anything. Cleaning up a rotten CVS tree is not fun!
>

Agree that this capability would be a nice addition, but with the
current modest sized group of developers I think that we can get by
on trust, and the mutual agreement that we will not modify someone
else's code without contacting them over E-mail first. Since all
persons with write access will be provided with a unique user name
it will be possible to determine, when changes are inappropriate,
who was responsible, and of course this person will be obliged to
buy beer at the next collaboration meeting.

Since well tested production versions will be released less frequently
and they will be assigned a CVS tag, then perhaps we can afford to
sometimes take an indirect path to the final goal if this allows more man
power to be applied to the problem.

> Another issue might be what my (and other people's) bosses allow. I fear
> that the people who are responsible for reliable uninterrupted service
> of a facility won't like the idea that 'their' software is maintained on
> some server half around the planet. Actually this is not a great
> problem, since everybody has the checked-out copy on their local
> machines, but you never know ...
>

Hopefully in production systems our application development environment
will only use a "shared" source code if it is CVS exported from a particular
CVS tag in a particular CVS repository. Since there would be a local copy of
this
exported code we can apply local emergency patches at any time, but in
my experience the CVS repository at APS is almost always available.

>
> P.S.: Why was this thread not posted on the tech-talk????

shoulda been

Jeff

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System