On Tue, Jan 27, 2009 at 02:02:55PM -0600, Shawn Walker wrote: >> I would also argue that "3.2.4" doesn't match "3.2.4.*", which I think >> this implies. > > Actually, it's intentional that 3.2.4 matches 3.2.4.*, since I'm fairly > certain (based on questions I've asked others in #pkg-discuss and my own > understanding), that users will expect "pkg list [email protected].*" to list all > 2.2 versions of apache, including just "2.2".
Huh. Okay. Not intuitive to me, but I appear to have a completely different intuition than everyone else on the planet. >> - get_short_version(): Why is this copied from Version? With the >> exception of the test, it's identical, and if you need a specific test >> against None (do you?), couldn't that just move into Version? > > Because MatchingVersion does not inherit from Version because not all of > the methods on Version are applicable to MatchingVersion. I'm not sure I agree. Some of the methods may not be terribly useful, but then they probably won't be used. I would worry more about code reuse and maintainability than making it impossible to confuse a MatchingVersion object with a Version object. But perhaps you have specific reasons why you think this is a bad idea? Danek _______________________________________________ pkg-discuss mailing list [email protected] http://mail.opensolaris.org/mailman/listinfo/pkg-discuss
