[Spread-users] Public CVS?

Jeremy Hylton jeremy at zope.com
Mon Sep 17 22:55:37 EDT 2001


I'm hesitant to comment on what seems like a sensitive issue, but I
thought I'd offer a few comments based on our experience using
SourceForge for Python development.  In my experience, there's every
topic of discussion is raised again by people who seem unaware of the
previous discussion or are (mistakenly) sure they have something new
to add.

I understand CVS access came up before, but I couldn't find a message
that offered a clear picture of the plan for CVS access.  If I
understand correctly, there is a CVS server available and accounts
will be issued open request.  Can anyone get an account?  If so, is
there some mechanism to restrict access so that someone doesn't
accidentally add or remove files or directories?

We've found it important for random users to have CVS access without
having any commit priveleges.  We've achieved that using anonymous
CVS.  It's important because random users, i.e. casual users rather
than active developers, submit bug reports and send patches.  We want
bug reports against the current CVS when possible and we want the
submitters to be well informed.  The CVS history often contains lots
of useful information about the code, which is useful for any
developer using the code.

CVS also provides a convenient way for a potential user to browse the
source and evaluate it.  I'm not so concerned about the need for
anonymity, but the requirement to apply for access seems an
unnecessary burden.

We've also found that having tools like the SF tracker are quite
helpful for managing bugs and patches.  An email list works when there
are only a few bugs, but it doesn't scale well.  The list traffic gets
to be too high.  You have to hunt for crucial information in your
mailbox.  You can't assign a patch to a person for review.

So that's the modest plug for anonymous access.  I'd like to have CVS
access myself; I'll be happy regardless of the mechanism, but I think
SF-style services are valuable.

Jeremy








More information about the Spread-users mailing list