On Fri, Aug 20, 2010 at 2:12 PM, David E. Wheeler <da...@kineticode.com> wrote: > Would it be possible to *always* use three integers? So the next release > would be "9.0.0beta5" or "9.0.0rc1"? In addition to being more consistent, it > also means that PostgreSQL would be adhering to Semantic Versioning > (http://semver.org/), which is a very simple format that's internally > consistent. I'm planning to require semantic versioning for PGXN, and it'd be > nice if the core could do the same thing (it will make it nicer for > specifying dependencies on core contrib modules, for example).
One thing that may be worth noting here is that even if we implemented this policy (and the consensus seems to be against it at the moment), we wouldn't be in compliance with semantic versioning, because our use of the first two components of the version number does not match that specification, and we aren't likely to make them match in the future. What that would mean is that certain kinds of changes would FORCE us to bump the major revision, and by historical precedent, pretty much every release cycle would have some. I've occasionally thought that it would be interesting to have something in between point releases and major releases, where, perhaps, we would implement changes that are more than what we'd allow for a minor version bump but nothing too invasive; and then use major releases for the real big stuff. But the problem with this is that it would greatly complicate development and testing and I think in the end we'd end up with a less reliable product and a lot more arguing about which branches things went into. I think the semantic versioning approach makes sense for libraries, but it is not too clear to me that it makes sense for other kinds of applications. YMMV, of course. -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise Postgres Company -- Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-hackers
