https://bugzilla.redhat.com/show_bug.cgi?id=1510565 --- Comment #8 from Pavel Zhukov <pzhukov@xxxxxxxxxx> --- (In reply to Miroslav Lichvar from comment #7) > (In reply to Pavel Zhukov from comment #6) > > (In reply to Miroslav Lichvar from comment #5) > > > What if a third package was added to fedora, which provided this service? > > > ntpstat can get the data from ntpd (using ntpq) and chronyd (using chronyc), > > > but that's all. > > Well it's theoretical question which applies for mysql example as well. > > There're few options: > > 1) Add support of the service to ntpstat > > 2) Add "Conflicts:" to either the service or ntpstat spec. > > In any case As far as ntpstat suggests chrony It's safe. In other words user > > has to explicitly install ntpstat and "the service". > > Wouldn't it better to just explicitly list the supported services in the > ntpstat spec using a rich dependency, e.g. "Requires: (ntp or chrony)" ? looks good for me. It's maintainer decision at the end :) Are you going to update spec or it'll be done after formal review? > > > > If old ntp is installed, I think an upgrade should result in both new ntp > > > and ntpstat installed. > > Something like Enhances: ntpstat in ntp.spec? > > That does not seem to work. I created a testing repo with ntp which > "enhances" ntpstat and also contains the new ntpstat package, but dnf seems > to ignore it: > > # dnf update --disablerepo=* --enablerepo=local > ... > Upgrading: > ntp x86_64 > 4.2.8p10-4.nontpstat.fc28 local > 670 k > ntpdate x86_64 > 4.2.8p10-4.nontpstat.fc28 local > 85 k > > I tried adding "Obsoletes: ntp < 4.2.8p10-4" to ntpstat.spec, expecting ntp > would be dropped without replacement as chrony is installed, but the result > was the same as before. weird. maybe bug in libsolv? Anyway I'll post formal review soon... -- You are receiving this mail because: You are on the CC list for the bug. You are always notified about changes to this product and component _______________________________________________ package-review mailing list -- package-review@xxxxxxxxxxxxxxxxxxxxxxx To unsubscribe send an email to package-review-leave@xxxxxxxxxxxxxxxxxxxxxxx
