This is a followup on something that was discovered while reviewing
a stabs patch:

   http://www.sourceware.org/ml/gdb-patches/2007-02/msg00139.html

In particular, this is where we discovered this little typo:

   http://www.sourceware.org/ml/gdb-patches/2007-02/msg00368.html

I confirm that I misanalyzed the second problem I mentioned in my
original message, and that this is the source of the issue I was seeing:

   Breakpoint 1, pck.lock.set (x=Unrecognized 0-bit floating-point type.
   ) at pck.adb:11

Basically, the effect of the typo is that we forgot to update the
type length of the other types in the same CV ring. Fixed thusly.

2007-02-28  Joel Brobecker  <brobecker@adacore.com>

        * gdbtypes.c (replace_type): Fix typo that caused us to not update
        length of the types referenced by the new type CV ring.

Tested on x86-linux, no regression.  OK to apply?

Thanks,
-- 
Joel