On Mon, 25 Mar 2002, David Talkington wrote: > >The "problem" with trying to install one of the > >packages was due to trying to use "rpm -i" instead of "rpm -U" ... > > Notice that the -i attempt followed a failed -F run, and was then > followed by a failed -U.
Yes, but that's because you didn't upgrade them together. Also you tried to install popt-something, instead of installing python-popt which was the one missing (the python popt module used to be in popt IIRC, and has been split out into a new package). I'm surprised they did that split in the updates, because the ideal is that you can indeed do a "rpm -Fvh *" in your updates directory. There are several packages that depend on the rpm libraries, so have to be updated at the same time. You do them all together and it works (so far ... they *could* break this by building rpm2html, rpmfind or gnorpm using the new rpm package and use a new rpm feature; *that* really would be dependency hell ;o) > > It's not rocket science. > > Neither is ./configure && make && make install, but I'd rather not have > to resort to that for basic system utilities. Something is clearly out > of whack with the latest rpms. Or with rpm itself. All I'm saying is, if rpm didn't work that way, people would just cheerfully ignore the dependencies, then complain that things didn't work ... it happens all the time when people compile stuff from the tarballs and break their systems. I'd rather have to deal with the occasional hiccough in upgrades than have to work out dependencies by hand. I think once they've tried both, people tend to understand the benefits. _______________________________________________ Redhat-list mailing list [EMAIL PROTECTED] https://listman.redhat.com/mailman/listinfo/redhat-list