Re: rhel8 test failure confirmation?

[Karl Berry]

    Are you sure about that?

Yes. Well, as sure as I can be. I don't see any $(...) constructs in
Automake's *.m4 now, and this is certainly deliberate. We discussed
shell portability many times over the decades of Automake development.

I can't quickly find where it's documented, if anywhere, but as far as I
know it is the status quo and I certainly see no reason to break new
ground for this small patch.

    I got a fair bit of pushback on removing $(...) from config.guess

Yes, I know. I was one of the people who reported problems because of that.

    on the grounds that Autoconf locates a POSIX shell and uses it for
    the bulk of configure

My understanding is that Zack implemented that precisely to defend
against the config.guess snafu. So for all but the latest Autoconf
versions, it will not happen.

    Let's try:

Thanks much! --all the best, karl.