On Mon, Jul 14, 2008 at 6:56 PM, Martin Langhoff <[EMAIL PROTECTED]> wrote: > Version numbers are used to communicate API/ABI compat and degree/type > of changes to users. Later in this thread Eben suggests what everyone > else in the industry is using: major.minor - sounds good to me. Even > better - and more prevalent in OSS - is major.minor.bugfix .
If we were to make the change, I'd suggest supporting dotted integer sequences of arbitrary length. 1, 1.2, 1.2.3, and 1.2.3.4 are all reasonable version numbers. These are *not* floating point numbers; they sort based on integer comparison major component first, so 1.1 < 1.11. Once we've gone that far, we might as well go all the way and adopt the version number scheme fedora uses, with an epoch and everything. But this doesn't actually do anything to solve the problem Eben first posed, which is why I'm objecting: it seems needless complexity at the moment, to a format which was explicitly designed to Keep Things Simple. > In any case, this issue does not seem to deserve such a colourful > thread. Maybe we can save our time and effort for other stuff? After You think this is colorful? Geesh. I need to work harder on the other threads. > all, if an activity writer wants to use Klingon characters for > versioning, hey, let them go wild! I am compelled to strongly object. The updater needs to be able to compare version numbers. I refuse to implement Klingon numeral comparison. All this stuff is trivial until you are the one who has to implement the updater. --scott -- ( http://cscott.net/ ) _______________________________________________ Sugar mailing list [email protected] http://lists.laptop.org/listinfo/sugar
