>> We had the problem that a directory was deleted,
>> and a new directory having the same name was added.
>> Because of this, the update failed; the new directory
>> could not be updated to the wc.
>
>
> Except that this is a very different use-case than what you originally
> presented. Adding a new directory with the same name (but not the same
> functionality) is an even worse workflow error, only on User A's part
> this time.
>
> User A should have deleted the directory and informed all users with
> WC's to update, then add the new directory and again alerted the other
> programmers. Reusing the directory name is bad enough, but expecting the
> VCS program to somehow know this is not the same code is even worse!
Sorry, but this is absolute nonsense.
First, this means that the change is not atomic anymore,
and there will be a revision which does not compile.
Having non-compiling revisions is a very bad philosophy.
Second, in our company, typically there is no time
where all users are at work and could do an update
at the same time.
Subversion should NEVER enforce all uses to do something
at some particular time.
Did you ever work in a larger team?
Cheers,
Folker
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Received on Fri Dec 19 08:47:00 2003