[PATCHv8] nss_db: protect against empty mappings

[DJ Delorie]

Rafal Luzynski <digitalfreak@lingonborough.com> writes:
> Thank you for explanation and sorry for the noise.

As Carlos said, this isn't noise, this is feedback.  We knew "back then"
that this type of thing might happen, but that doesn't mean we can't
optimize, document, or otherwise tweak it to be less surprising to glibc
developers.

Input on your experience with the testroot system, and possible ways to
improve it, are always welcome :-)