On Fri, Jul 29, 2005 at 10:24:30PM +0200, Adib Taraben wrote:
bulia byak schrieb:
... And here are my candidates for MUST FIX bugs for 0.42.1: ...
I am a little bit confused about version numbering
tar has 0.42 win32 0.42-1 rpm 0.42-0.1 rpm 64b 0.42-1
what will then a interims version will be ?
Yes, release numbering can be a bit confusing, but this is what a release number is made up of:
0.42.2-3
The part before the dash is the "version number":
0 - this part is the "major version number". This number is incremented very infrequently and is generally used to mark a major compatibility change, such as a switch to a different file or protocol format.
42 - this part is the "minor version number". This is the number that is normally incremented when we have a main release.
2 - this part is the "point version number". This is used to number small incremental releases off of the main code branch, usually for bug fixes and other non-feature enhancements.
These three numbers are used to tag the official source code release, which is done as a tarball. We always drop the point version number if it is zero, so you'd never see 0.42.0. Note that this shouldn't be considered a floating point number - 0.4 and 0.40 do not mean the same thing (the first means release number four, the second is release number forty.)
The part after the dash is used only by the packagers:
3 - this is the "package patch number". This is used by the packagers to denote new releases of the package. They might do this if an earlier version was packaged wrongly or if they wish to include some special patches that are not part of the version release. Note that package patch numbers always start from 1, rather than from 0 as the version numbers do. I don't know why this is, but it's convention.
Hopefully this explains the strange numbers that pop up with releases. I don't know what '0.42-0.1' is; it might be a mistake. However, sometimes some distros use somewhat different schemes for doing their package patch number.
Bryce