On Fri, 14 Jan 2005, Philip Hazel wrote:
> We are in close contact. :-) I had started issuing snapshots before
> Nigel decided that we should have an interim release. It is my fault for
> issuing snapshots with "real release" numbers. I should have used 4.491
> or something like that. Too late now. We need to make up some "rules"
> for next time.
>
> This is the first time we have done any parallel development on Exim. We
> are new at the game and it isn't as tidy as it could be. Perhaps we can
> do better in future.
>
> Anyway, if the interim release is 4.44 and the snapshots are 4.46, 4.47
> etc, the historical sequence is "correct" because 4.44 contains a subset
> of the changes for 4.46. (Later today there may be 4.47.)
One possibility is to regard the snapshots as "Release Candidates". So
the first snapshot for the 4.50 release would identify itself as
4.50RC1, the second would identify itself as 4.50RC2 etc, culminating
in the 4.50 release which would identify itself as 4.50 as normal.
The above scheme isn't original. The snort IDS project seems to operate
using something similar. However it does allow two parallel forks in
exim's development to be clearly identified as such and allow for a 4.44
release without any ambiguity. I'm sure there are other ways of
achieving this. So perhaps a more general discussion is required.
--
Dennis Davis, BUCS, University of Bath, Bath, BA2 7AY, UK
D.H.Davis@??? Phone: +44 1225 386101
