Marcus Hüwe wrote:
But this maybe leads to some problems too. One possible problem could be that the user removes the pear-package via the pear commandline tool but the package still exists in the rpm database..?! I think this is a generic problem, it would be interesting to see how/whether is it solved in cpan or gems. Darix, Anicka, are you there? ;-) i have the similar problem with my rubygem-* rpms. i thought about hacking the cmdline tool to check if the module was installed via rpm and error out in this case. so far i didnt find time to implement it.
How do you want to find out if the module was installed via rpm or via the commandline tool?
A) extend pear's database to hold that information OR B) run exec("rpm -q " . escapeshellarg($rpmname), $out, $ret) in the commandline tool A) would IMO be more difficult to implement, B) would cost some time to query the rpm database (probably feasible), plus it would require a consistent, machine deducible naming of the rpm packages (desirable anyway ;)). Either way, we would need to find all possible code paths wher a packages are manipulated, that is not only direct install/update/erasure of the package, but also install/update via dependencies. So the whole think might turn out to be way too difficult, esp. when this would probably never be accepted upstream. But the %post %postun scriplets look like a goot idea to me. Michal --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-packaging+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-packaging+help@opensuse.org