[stuff deleted]
> all you know is that the change has happened. With a good source code
> control system, you would know why. That is what makes the CVS trees
> from FreeBSD, at least, so valuable. People tend to document why they
> made the changes, so you'd know that the MUMBLEFOO bit was added to
> the initialization sequence to help BRAIN-DEAD-MONKEY keyboards
> operate properly....
[I'm joining this thread a little late. Forgive me if I repeat stuff
which has possibly been said before].
This certainly is a *very* good justification for using CVS. I am not a
CVS user myself (I plan on changing to CVS from RCS real soon now), but I
suspect that it would be possible to set up a 'read-only' tree accessible
by Joe Average but allow write access for important developers. If I
remember correctly, Linus' original objection was that he didn't want just
anyone hacking away at the tree ...
Linus could thus have the main branch and merge in code from the other
trusted developer branches (Alan etc.) as he saw fit. Official releases
would be done from the main branch. Joe Average would get to access any
part of the tree he wanted in read-only mode and of course be able to
access the log information. The current sporadic log information tacked
into files could be dispensed with. Patches from Joe Average could be
sent to the maintainer of whatever branch ...
Any CVS guru's out there want to set up a test-system like this that could
be demo'ed to Linus? Reading the CVS FAQ implies that restricted access to
the tree might not be altogether trivial to set up. Perhaps 2.1
development could be done on CVS if a convincing enough demo can be given
!!
--Craig