Patch to handle compressed sections

[Craig Silverstein]

} I see.  Well, you could use bfd_getb32 twice... I just felt that the
} eight copies of that line were excessive.

Then you'd do something like
   size = (a << 32) + b;
which likewise won't work well (well, will give a compiler warning) on
32-bit systems.

I agree the code ends up being verbose, but it's pretty
straightforward and robust.  Would you be ok with the code going in
the way it is now?

craig