On Fri, Dec 07, 2012 at 09:56:38PM +0100, Christian Weisgerber wrote:
> Currently, infrastructure/db/config.site holds these openpty(3)-related
> values:
>
> ac_cv_func_openpty=${ac_cv_func_openpty=no}
> ac_cv_lib_util_openpty=${ac_cv_lib_util_openpty=yes}
>
> At first glance, this makes sense. We have openpty() in libutil,
> but not in libc. However, I'm already looking at the second port
> today that does something like this:
>
> AC_CHECK_LIB(util, openpty)
> AC_CHECK_FUNCS(openpty)
>
> This is a quite reasonable use of the available autoconf macros.
> Check that libutil holds openpty, then check for openpty itself
> with LIBS=-lutil. Unfortunately, our preloaded cache values cause
> this to fail:
>
> checking for openpty in -lutil... (cached) yes
> checking for openpty... (cached) no
>
> I move that we delete ac_cv_func_openpty from config.site.
>
> Anybody have a suggestions for a pithy comment we can put in there
> so ac_cv_func_openpty won't accidentally be included again?
I think this is worth a defect report to the autoconf people at the FSF.
So, let me see, the value cached does not mean anything, because it's
actually dependent on the list of libraries that are currently tested ?
Great way to go. What a pile of dung.
Pithy comment:
# XXX autoconf cache values are so much crap it's laughable
# don't ever cache "raw" openpty, autoconf is as likely to pick its nose
# in public as to change the LIBS list and fuck up whatever you do.
