> But does it handle the case where there are a bunch of packages
> installed, and foo changes from math/foo to science/foo, and the user
> doesn't notice, and issues the 'update everything' command?

If math/foo was explicitly selected as "wanted" by the user, the user will 
have to change the want entry to science/foo. If math/foo was brought in as a 
dependency by some other package, it will be handled automatically (but 
math/foo will be deleted and then science/foo will be installed at the 
appropriate time later on).

I don't remember off hand whether I actually implemented warnings on 
non-existent want list entries on start-up; I might have. Even if not, I will 
in the re-write, sooner or later...

Although a warning would be nice, this is not *that* fatal IMO since for a 
package the user actually wanted explicitly the user tends to know what's 
going on (even if it's a pain finding where the package went for each such 
package). But things like the new package conflicting with the old version of 
the package is always avoided as a result of the old package being deleted 
first.

> I don't see how it could possibly do this without parsing the renamed
> lines in CHANGES (which I suppose are there, so this is a mere matter
> of programming).

Right. I'll have to have another look at that; my recollection was that it was 
not obviously parseable, but perhaps it is. If so that would definitely be 
something to support.

-- 
/ Peter Schuller, InfiDyne Technologies HB

PGP userID: 0xE9758B7D or 'Peter Schuller <peter.schuller@infidyne.com>'
Key retrieval: Send an E-Mail to getpgpkey@scode.org
E-Mail: peter.schuller@infidyne.com Web: http://www.scode.org