What about 'shutdown -h now'?

Unless you've already tried it, you should check out revdep-rebuild from app-portage/gentoolkit.
Alternativly you could try qdepends from app-portage/portage-utils, which allows you to do something like this:

emerge --oneshot $(qdepends -NC <pkg> | cut -f2 -d ':')

This will (hopefully) rebuild ...

The segmentation fault suggests to me that one of your memory modules is defect, however that might not be the case if you're only able to reproduce this when the test USE flag is disabled. Have you tried memtest, or something similar, to verify that your hardware is working like it should?

yoshi314 wrote:that would mean that eselect depends on portage. is that true?

I don't think eselect explicitly depends on portage.

what about just using 'src_unpack () { true }' or smth?

I think xorg-x11 is just a meta package, you probably need to rebuild the xorg-server package.

Well, in case you didn't know, it is, as far as I know, possible to undo prelinking on a per-file basis (prelink -u <path/to/file>, or something to that effect).
Btw., the prelink mask thingy doesn't do anything by itself, it just tells prelink to skip certain files, i.e. it doesn't affect files ...

Have you considered undoing the prelink operation?
If you really must have prelink, you could try adding the paths to the apps that won't work to PRELINK_PATH_MASK in /etc/env.d/60prelink.

I've discovered that when installing certain haskell packages (cabal, f.ex), paludis seems to mix up 'merge' from /usr/libexec/paludis/utils with /usr/bin/merge (from app-text/rcs), which leads to failure, for obvious reasons.
My guess would be that this happens due to some funky things in the ...

tried commenting out/removing 'PORTDIR_OVERLAY' in /etc/make.conf ?

Woops.

tried binding to some key/mouse press?

It's supposed to give you a list of packages to rebuild, it doesn't really fix anything by itself.

not very pretty, but it might be a starting point.

Can we have your emerge info?
Sure.

Portage 2.1.1_pre4-r4 (default-linux/x86/2006.0, gcc-4.1.1/vanilla, glibc-2.4-r4
, 2.6.17-ck1-r2 i686)
=================================================================
System uname: 2.6.17-ck1-r2 i686 AMD Athlon(tm) XP 2600+
Gentoo Base System version 1.12.4 ...

Works for me.

You could use gzip to store the pickles.
Turning on the new bin format (cPickle.dump(data, fp, True)) might also be beneficial, but I'm not sure whether or not it affects the size of the pickle, but it should make pickling more efficient.
Other than that I don't really know what you could do, unless ...

Did you try compiling xmms without '-Wl,--as-needed'?

No error messages at all?

I don't think so. AFAIK gcc does not take into account what architectures the kernel is supposed to support..

AFAIK the main downside of FORTIFY_SOURCE is a minor decrease in performance because of the extra run time checks. Also, I believe some broken applications might refuse to work/compile, but I'm not certain, as I've never used it.

-Wl,-z,combreloc is enabled by default, I think, but you might be ...

I believe these are called reverse dependencies.
You could use revdep-rebuild (from the gentoolkit package). It tries to identify broken libraries/binaries and re-emerge them.

I think tabs allow grouping of windows, which seems to be what you want.
AFAIK fluxbox has this feature, and fvwm supports it through a module.