On Sun, Mar 28, 2021 at 01:26:10AM +0000, brian m. carlson wrote:
> On 2021-03-28 at 01:04:32, Jeffrey Walton wrote:
> > I see what happened... I stopped Autoreconfing so the configure.ac
> > changes were not picked up. The original configure is being used.
> >
> > I think Git is getting itself into that state.
> >
> > Without the Autoreconf, I was able to sidestep the issue with:
> >
> > if [[ "${IS_SOLARIS}" -eq 1 ]]; then
> > CONFIG_OPTS+=("ac_cv_func_inet_ntop=yes")
> > CONFIG_OPTS+=("ac_cv_func_inet_pton=yes")
> > fi
>
> Most developers don't use the autoconf stuff and just use the makefile.
> config.mak.uname doesn't have NO_INET_NTOP set for SunOS.
>
> If the autoconf scripts or config.mak.uname need fixing, which is
> possible, a patch would definitely be welcome.
This all seemed vaguely familiar, and indeed it came up about a year
ago:
https://lore.kernel.org/git/CAH8yC8m3JFvEcfFF3z1rrRnEPK-adHGObmkOhNZiph7QJKUWqA@xxxxxxxxxxxxxx/
Maybe worth resurrecting this patch (at the very least it needed a
signoff):
https://lore.kernel.org/git/CAH8yC8kaWXbN+RYMJnM9em7KKW54+N07JtyS1MZk0qppD=m2BA@xxxxxxxxxxxxxx/
-Peff
