Greg Hudson <ghudson@MIT.EDU> writes:
> > Now, it may be that the actual changes are scattered among entities
> > far down in the tree, leaving many intermediate directories with no
> > changes at all.
>
> Trying to optimize this with multiple replace_root() calls seems
> pretty squishy to me. I'm more comfortable with restricting ourselves
> to a single replace-root operation. If nothing else, it allows extra
> information to be passed in during the get_foo_editor call which can
> give information about the root.
Okay. I'm kind of ambivalent on the issue, and am ignoring it while
concentrating on filesystem work. For now, let's leave replace_root()
as it is.
-K
Received on Sat Oct 21 14:36:19 2006