[PATCH v2 0/5] fix some "dwz -m" regressions

[Tom Tromey <tromey at redhat dot com>]

This is a refresh of a patch series from February:

    http://sourceware.org/ml/gdb-patches/2013-02/msg00395.html

I've rebased this series and reordered it.  When re-reading my
original submission I saw that the first patch introduced some
regressions that subsequent patches then fixed.  I think it is better
not to introduce regressions, even temporarily, so this series
reorders the patches so the fixes are first.

Even after this series there are some "dwz -m" regressions relatively
to trunk.  I've noted the details in PR symtab/15028.  I haven't
investigated these yet.

I regression tested each patch in this series both normally and using
dwz -m.

I am checking this in.  The series has been available for review since
February and has only undergone minor changes, despite the reordering.

Tom