http://bugzilla.spamassassin.org/show_bug.cgi?id=3554
------- Additional Comments From [EMAIL PROTECTED] 2004-06-29 13:39 ------- > > as far as I know, RPMs are supposed to deal with that automatically > > where possible -- especially when the parameter was one that had > > been set up *as a default* by a previous release of the RPM. > > well, perhaps that'd be the expected thing (I personally wouldn't want > a package install modifying my configs), but we're not talking about > upgrading from our RPM to our RPM. we're talking about upgrading from > someone else's RPM to our RPM. who knows how "someone else" does things, > so we really can't deal with all of the possibilities of how they do it. Good point; sorry, I was mistakenly thinking our RPM did the same thing in 2.60. > what we could do, potentially, is remove the postun condrestart, and > leave it for the post section. that way, in this case, the old default > sysconfig file would be updated with our sysconfig file, and the restart > will work. I'm not sure why there's 2 condrestarts in there in the > first place. Will it overwrite user-modified sysconfig files? If not, then that sounds good. > as for the -a issue, I like having it as an error. this means people > will have to pay attention that something has changed. I still prefer a warning, since (if I recall correctly) we've never actually deprecated it -- just one release it works, the next release it's an error. That's extremely user-hostile behaviour, IMO. > I think we > can have a more useful message spit out than "unknown option" though. > But that's another ticket. well, agreed on that. --j. ------- You are receiving this mail because: ------- You are the assignee for the bug, or are watching the assignee.
