bfd_size_type

[Alan Modra]

I've been looking at the huge number of warnings you get with -Wconversion.
Most of these are in fact errors if you're trying to compile binutils
with a K&R compiler from a 32 bit host to a 64 bit target, but I guess no
one cares too much about trying to do that these days.

Anyway, a large number of the warnings are about bfd_read args.  I'm
wondering why bfd_read has bfd_size_type args, as it's rather futile
to try to read more than 2^32 bytes on a 32 bit host.  Does anyone
have an objection to simply using size_t for bfd_read?

Alan