Denis Leroy пишет:
If you catch new version of the package from CVS or SVN directly the version of it can't be still old (like x.y.z for last release) almost in all cases.I don't understand this one. The release is not the version number, and it's something that we completely control (unlike the upstream version number), so should it be anything other that a linearly increasing single number (just like a build number) ? (I can see an exception being made for things with kernel versions in them otoh).please don't, we need 0.x to mark release candidates and others pre-real-release versions
Look for example.
Last release: x.y.z-1%{?dist}
Preparing release: x.y.(z+n)-0.1.%{date}cvs%{?dist}
New release: x.y.(z+n)-1%{?dist} (attent to version tag).
And case just after release (i.e. in case when new cvs/svn version more
preferable than couple of patches): x.y.(z+n)-1.1.%{date}cvs%{?dist}
-- With best regards, Andy Shevchenko. mailto: andy smile org ua