"Neil Parker" <nparker(at)microniche(dot)com> writes:
> This problem appears to be glibc's fault, not PostgreSQL's.
Indeed, so why are you complaining to us?
> Below is a patch that fixes the problem on the affected system.
I'm fairly unenthused about a kluge that hurts performance on
standard-conforming systems (by forcing them to traverse the slow code
path) in order to cater to a long-obsolete glibc whose behavior conforms
to neither the C standard nor its own documentation. If this bug is a
problem for you, update glibc.
regards, tom lane