2007/8/8, B. Nice <[EMAIL PROTECTED]>: > > Peter, > I'm running a non-~amd64 box as well. My package.keywords file is > getting rather large, but it's still a "stable" system. I'd have to > guess that it is something with the new kernel, as I'm running > x11-drivers/nvidia-drivers-100.14.09 under the 2.6.20-gentoo-r8 kernel > and not having problems with the sandbox violations. Granted this may > be a faulty supposition as some of the major "unstable" packages I'm > running is gcc-4.2.0 and glibc-2.6. > > I am curious why you had to turn sandbox off for > nvidia-drivers. I've > never had to. If memory serves, the last time I had trouble with the > sandbox was firefox-1.5.xx.
i don't really why you have to do this. for what i know the sandbox violation come when some package tries to use the root in a strange way. this may be due to some ebuild bug. anyway i didn't experienced many problems of sandbox violation, but when i did, putting feature -sandbox when compiling it has given me no problem. Ain't Gentoo fun. Two systems running the same things and two > different experiences. I love it. > > > On Wed, 2007-08-08 at 08:51 +0100, Peter Humphrey wrote: > > On Tuesday 07 August 2007 20:06, Beso wrote: > > > have you switched to the no-multilib profile by chance?! > > > > If I'd tried anything of that magnitude I'd not have been puzzled by > such a > > small consequence as this. > > > > > or do you have an nvidia package?! > > > > Yes, I have nvidia-drivers. It's one of the packages that need sandbox > > switched off. How does that help? > > > > > try adding "-sandbox" to your features in the make.conf and see if it > > > works. > > > > Of course I tried that, and it did. Actually, I said "FEATURES=-sandbox > > emerge --resume", and the emerge finished properly. > > > > The problem seems to be in switching to this new kernel version. > > > > -- > > Rgds > > Peter. > > Linux Counter 5290, Aug 93 > > -- > [EMAIL PROTECTED] mailing list > > -- dott. ing. beso