A completely different approach to EH runtime

[Joe Buck]

Graham Murray writes:

> >From the discussion I have seen so far, it would seem as though the
> results are going to be the opposite. That is - in most cases I
> suspect that the user-installed gcc will be a later version (ie
> upgrade) in $prefix than the distribution-supplied one in
> /usr. Therefore, I think it is far more likely that
> $prefix/lib/libgcc_s.so will be backwards compatible with (the older)
> /lib/libgcc_s.so than the other way round. 

Exactly!  In almost all cases, a user will be adding a newer gcc to
a distribution that provides an older one, or will be assisting with
testing gcc snapshots.  It will be much rarer for a user to install
an older gcc on a system that has a newer one.  It is only in this
rare case where there might be a problem with having missing functionality
in the "private" libgcc_s.so.  The solution can be as simple as just
not installing the older libgcc_s.so at all and let all code (both
that compiled by the old and by the new compiler) use the newer
libgcc_so.so that will live (presumably) in /lib.