Mark Knecht <markkne...@gmail.com> posted 5bdc1c8b0902051228j789c31a1i282d60a0328c2...@mail.gmail.com, excerpted below, on Thu, 05 Feb 2009 12:28:21 -0800:
> Duncan has scared me off of this path for now. That wasn't really my intention, but I did think you needed to know what you might be getting yourself into... But as I said, others have done it. It's initially a LOT to deal with, to get your head around thinking in slightly different ways (but it's still the Gentoo we know and love!) especially in network config and RAID/LVM etc if you run them. That's the BIG thing. But in the end it's better. But there is a reason it's not stable yet, and by the time it is, the transition will certainly have a nice upgrade document available and should be very smooth for those that follow it. It's just not there yet, and while upgrade is absolutely possible, it's not yet "smooth". > I'll continue down the > .critical file method until it succeeds or fails. All things considered, that's what I'd probably do if I were on stable in general. Of course I'm not and I actually enjoy the bit of extra challenge, pitting my mind against some bug in a package obviously not designed for my sometimes "unique" customization choices, but if I was to be on stable, for a production machine I was making lots of money with, or something... > I don't run anything ~amd64 in @system. I can wait for that. This isn't > a 'failure'. It's an annoyance only. I can live with it until it's fixed > in stable. > > I'm just surprised that I didn't have this problem until maybe the last > 5-10 weeks. Seems it's been there for a long time... That's actually a good point; something I can't quite answer myself. The best I could figure was that it was some sort of interaction between components that I (and apparently others) never did track down. For me, it was easier to work around it using CRITICAL_SERVICES, and then later, upgrading to baselayout-2/openrc. I suspect that's what most folks have found that stumbled on the problem -- it was easier to work around than to trace down to its root cause and fix. Thus, it remains somewhat of a mystery. OTOH, it's possible other folks have traced and fixed it on their systems, but whatever triggers it in the first place, perhaps a misplaced quote in an ebuild command (hmm, seems I know something about misplaced quotes right about now!) having an effect entirely other than intended and then it looks like, remains free to blow up in the faces of others. It might even be a common enough error that when it ultimately gets fixed in one place, it eventually shows up in another. But it was easy enough to work around, once I got a hint as to how. Hopefully CRITICAL_SERVICES, once we get it straightened out, works as well for you. -- Duncan - List replies preferred. No HTML msgs. "Every nonfree program has a lord, a master -- and if you use the program, he is your master." Richard Stallman