On Tue, 23 Apr 2002 13:01:28 -0700 "Joseph Carter" <knghtbrd@bluecherry.net> wrote: > On Tue, Apr 23, 2002 at 04:51:20PM +1000, Glenn McGrath wrote: > > > No, it shouldn't. Consider, eg, treating 2.2.9 and 2.2.10 as "2, > > > 0.2, 0.9" and "2, 0.2, 0.10". > > > > > Yes, i am obviously wrong, trying to compare 0.9 and 0.10 is > > incorrect, as is treating 0.1 as equal to 0.01 > > Should it have been 2.2.09 then? What do we do when we reach 2.2.99 and > have need of a 2.2.100? It happened with 2.3.x didn't it? > > There is a reason things are done as they are, even if it seems to not > be obvious at first. Frankly, some standard was needed, and the one > that was chosen was the one which seemed to make the most sense to the > people who were doing the work at the time. They chose to read . as a > delimiter rather than a decimal because that is how it is most commonly > used in version numbers, There are much more annoying things about how > dpkg handles versions to argue about and IMO the effort would be better > spent solving those problems - for example dpkg still lacks a method to > identify a pre-version. > I realise that the current system still works, and 0.1 being equal to 0.01 is a limitation that we can work around, but ideally the way we evaluate version numbers should correlate with the way most upstream authors evaluate version numbers. If we compare numbers that follow a decimal point as strings then i think that will fix the problem.... but hey, i might be wrong, im not perfect, so here is another chance for aj to be a !@#*#$%. Glenn
Attachment:
pgpC4yIxQoB6v.pgp
Description: PGP signature