>> Suppose a database D ...
A tablespace is created, backed up, deleted, and the file name of
the tablespace is reused to hold a tablespace with a different
> IMHO there is no problem. At least nothing special to tablespaces.
> You can construct like conflicts with most object types. The problem is
> with the user who reuses names for different things between backup and
Saying that the problem is the user makes sense in some cases, but
restoring over an existing database is a very destructive operation
and if it fails, leaves almost nothing. That's a high price for
the system to pay for the carelessness of some privileged user.
> I don't think we can ever guarantee that restore will work regardless of
> what the users changes since backup.
> However, if you want to say that this is one of the cases that we should
> be able to deal with, then it's something different.
I'd like to see that change managed - perhaps by recreating the
tablespace in a system-generated uniquely named file.