Re: Possible explanation for readline configuration problems

I wrote:
> ... how could we see reports like this:

> tab-complete.c:734: `filename_completion_function' undeclared (first use in this function)

> when the code makes a point of providing a declaration for
> filename_completion_function if configure didn't see it in the headers?

Never mind ...

I pulled down readline 4.2, and the answer is depressingly clear: the
Readline boys have decided to rename filename_completion_function to
rl_filename_completion_function. This graphically illustrates the
fundamental bogosity of AC_EGREP_HEADER: it still finds a match,
blithely ignoring the fact that it matched only part of an identifier.

Most of the other compiler warnings that we've been hearing about arise
because the readline headers have been const-ified. Suppressing these
warnings across both old and new readlines will be a pain in the neck :-(

regards, tom lane