On Wed, Mar 22, 2006 at 10:15:40PM -0800, Steve Langasek wrote: > Ok, this is pretty much an unsupported configuration in Debian. Debian > policy assumes, for instance, that any symlinks within a single top-level > directory can be made relative symlinks: e.g., /usr/include/X11 -> > /usr/X11R6/include/X11. If your /usr data doesn't fit on a single > partition, I strongly recommend using bind mounts instead of symlinks > (supported in 2.4 and 2.6 Linux kernels). Sounds doable. > Now that I understand the nature of the bug you're describing, I agree that > it's a bug and probably should be fixed (though that's left up to the > package maintainer, not me). It's just not one that anyone else is likely > to trip over. What constitutes a bug is certainly debatable but the way it seems the path is resolved appears to be pretty much the adopted policy at the GCC headquaters. It may not be that easy to change their mind about that ... ;) > > Hmm. Wasn't the ``-V'' option specifically introduced for purposes that > > require a specific compiler version (it's even documented here: > > http://gcc.gnu.org/onlinedocs/gcc-4.0.3/gcc/Target-Options.html) ? > > I have no idea; I'm just not surprised that the various bits of the compiler > aren't truly interchangeable. It's been a while since I delved into the internals of GCC but I'd very much expect things to be consistent with what is stated about them in the documentation. That linking problem, however, is indeed directly related to the location of /usr/lib/gcc/powerpc-linux-gnu/3.4.6. Same resolution policy here. Christian
Attachment:
signature.asc
Description: Digital signature