[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: ftp-master support fro Debian pre-versions xx~yy-1 [katie@ftp-master.debian.org: Processing of libpri_1.0~RC1-1_i386.changes]

On Tue, Jul 20, 2004 at 12:30:19PM +1000, Hamish Moffatt wrote:
> If it's just a case of interpreting correctly, I don't think it's a
> problem. Any package in sarge will have a version number greater than
> its version number in woody. woody's dpkg would interpret 1.0~pre-1 > 1.0
> while sarge's will behave the opposite way. However, none of the version
> numbers are going to decrease anyway.

Possible misbehaviours: 

	(a) you've got X~Y installed; X is available and should be
	    installed, but won't be

	(b) you've got X installed; X~Y is available and shouldn't
	    be installed but will be

(a) could happen via partial upgrades -- someone points apt at unstable
to get an updated gnome set, but doesn't upgrade dpkg, eg. It's fixable
if we make sure every .deb that uses a ~ version Depends:/Pre-Depends:
on an appropriate version of apt/dpkg though.

(b) could happen if you have bother testing and unstable in your
sources.list, and haven't updated apt yet.

The real misbehaviours are things like:

	(c) bar Depends: foo (>= 1.0), foo 1.0~pre1 is installed/available
	(d) bar Conflicts: foo (<< 1.0)

though. It seems fairly straightforward for those sort of things to arise,
and I'm not sure how you'd make sure you knew all the cases that could
make them arise so as to be confident none of them would.

We've lived with the existing situation for ages, it's not urgent enough
to risk screwing up our dependency handling; so unless someone can come
up with a way of making sure there aren't any problems, it'll just have
to wait 'til we're sure everyone's upgraded their apt & dpkg.


Anthony Towns <aj@humbug.org.au> <http://azure.humbug.org.au/~aj/>
Don't assume I speak for anyone but myself. GPG signed mail preferred.

``Like the ski resort of girls looking for husbands and husbands looking
  for girls, the situation is not as symmetrical as it might seem.''

Attachment: signature.asc
Description: Digital signature

Reply to: