On Tue, 3 Jun 2003, John wrote: > > Yes, you're probably right. Leave it to Red Hat to get rid of a perfectly > > good numbering scheme. I've been using RHL since 3.0.3 though so I guess > > Ditto. I've decided RH has left me, and I'm moving to Debian. hmmm -- you guys would skip watching a movie, simply because the leading lady tinted her hair, as well, I'd guess <grin> It's just a number. Respect RH, instead, for making it clear that binary compatability is _not_ clean and RHL 9 is a big step from RHL 8 with the NPTL changes. > Install with --nodeps;-) Anyone who does so without being able to predict the effects, has the pleasure of figuring out why something seems hosed later. It turns out not to be necessary -- but as John has been saying for six month's on Red Hat sponsored lists about Red Hat releases and tools, that he is moving to Debian, I guess it doesn't matter <gentle smile>. > Exactly. Besides, you get to run the wrong version of rpm. Can rpm from > Valhalla install RHL 9 packages? if rebuilt locally, certainly. You can push rpm-4.1 back, presently, back to RHL 4 series -- but it is ugly. > > if Red Hat would fix their makefiles to make them more relocatable, > > although I guess it's probably not in their best interest to do so. A filed Bugzilla talks, Reproduceable bugs against RawHide are louder, and patches loudest of all, for it allows other developer people to get to a workaround -- contrarywise, carping walks. Cite where you have filed a reproduceable erroneous makefile, improvidently closed, and _then_ re-open the issue. I can and have I few I think are really braindead, but by and large do not find this kind of error. > It's also possible to use Anaconda (that's how RH folk test it) on a > running system, but I'm not at all sure about running Anaconda 9 on > Valhalla. It is a tricky backport, scarcely worth the effort. It is far easier to backport needed pieces into the 'proper' anaconda for a given point release. -- Russ Herrold
