Re: configure.ac: modernize check for hosts_ctl

On Tue, Mar 10, 2026 at 4:45=E2=80=AFAM Simon Josefsson <[email protected]=
g> wrote:
>
> Mike Gilbert <[email protected]> writes:
>
> > On Mon, Mar 9, 2026 at 3:50=E2=80=AFPM Simon Josefsson <simon@josefsson=
.org> wrote:
> >>
> >> Mike Gilbert <[email protected]> writes:
> >>
> >> > Get the proper declaration from tcpd.h and pass the appropriate numb=
er
> >> > of pointer arguments. This avoids relying on an implicit declaration=
.
> >>
> >> Thanks, installed.  I intended to make you the git author but
> >> accidentally pushed the change without that, and force push to savanna=
h
> >> seems unavailable so I can't fix it easily.  I added a NEWS entry for
> >> this with your name though.
> >
> > No problem!
> >
> > If you want to avoid this mistake in the future, I suggest making use
> > of the "git am" command to apply patches sent in this format.
>
> Right, I'm used to using 'git am' for attachments, but not entire
> e-mails.  I'll try to adapt!  Also
> https://codeberg.org/inetutils/inetutils is open for pull requests
> (although for larger contributions, copyright assignment will be
> required).
>
> Thanks again for noticing and fixing this.  Is your QA tool that found
> this easily usable from a GitLab CI pipeline?  It would be nice to have
> tools that caught regressions in this area.  Or for use in other
> projects, I suspect similar issues are common.

The tinderbox tool is pretty Gentoo-specific, and probably won't be
much use in a CI pipeline. It's basically a collection of scripts that
attempt to build every package in Gentoo with semi-randomized
configuration options. The scripts are available here if you are
curious.

https://github.com/toralf/tinderbox

The issue was actually flagged by a QA script provided by our package
manager Portage. You might find this script (and the others like it)
more useful.

https://github.com/gentoo/portage/blob/master/bin/install-qa-check.d/90conf=
ig-impl-decl