* Michael Meskes (meskes@postgresql.org) wrote:
> > I think the current coding is extremely fragile (if it indeed works at
> > all) because of its assumption that <limits.h> has been included
>
> Well, this is the case in the code so far.
Right, the existing code is after limits.h is included, my suggestion to
put it in c.h would have lost limits.h and broken things. Sorry about
that. I didn't realize the dependency and make check didn't complain
(not that I'm sure there's even a way we could have a regression test
for this..). I didn't intend to imply the currently-committed code
didn't work (I figured it was probably fine :), was just trying to tidy
a bit.
Thanks!
Stephen