On 6 Feb 2008, at 8:39 am, Adrian Bridgett wrote:
On Wed, Feb 6, 2008 at 08:33:24 +0000 (+0000), Tim Cutts wrote:
Hi - to add to this one, as is often the case, version 4.1.4 of
tkdiff is required by tkcvs 8.1. Since tkdiff comes from the same
upstream tarball as tkcvs, and they're written by the same person,
is it perhaps time we made this a multiple binary package and built
both tkdiff and tkcvs from the same upstream source package? It
would solve these dependency problems every time tkcvs gets updated
upstream. Not that I want to tread on your toes, of course, but it
might be simpler that way.
That may well make sense.
There seem to be three possible approaches I could take, and I don't
know what the right thing to do is.
1) Have the tkcvs binary package provide the tkdiff binary, and then
conflict with tkdiff (in other words kill off the tkdiff package
altogether, and have the tkcvs package 'Provides' tkdiff.
2) Do the proper multiple binary package thing. I'm slightly
concerned about how the package versioning would work in this case,
because I suspect that the new tkdiff packages would always have the
same version as the parent tkcvs package, even though the actual
version of tkdiff is independent. I don't like that...
3) Just have tkcvs ship an internal version of tkdiff for its own
use, and allow people to use your package as the one they get on their
PATH.
None of these seem to be ideal. 3 is the easiest, and least
disruptive to the status quo. I've done 1) as an experiment, but it
seems ugly, and likely to cause users confusion when the package
disappears. It also seems unnecessary to have to install tkcvs if all
the user wants is tkdiff (not that it's a very big package, but it's
the principle of the thing!) 2 is less ugly, but the version number
thing is bound to confuse people.
Any thoughts?
One day the keyring maintainers might
actually get around to acknowledging that exist. ATM I can't do any
uploads (and havn't been able to for about 18 months now IIRC).
You have my sympathy. :-(
Tim
--
The Wellcome Trust Sanger Institute is operated by Genome Research
Limited, a charity registered in England with number 1021457 and a
company registered in England with number 2742969, whose registered
office is 215 Euston Road, London, NW1 2BE.
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
