testing of CVS Autoconf

[Ralf Wildenhues]

FWIW, this is how current CVS (during the last few weeks) plus Paul's
pending gnulib additions fares on these systems:
(sorted by OS then CPU; no comment means no failures; using the native
compiler if present, GCC on the w32-based ones; GNU m4 everywhere, but
otherwise native tools; no special CFLAGS etc.)

powerpc-ibm-aix4.3.3.0
powerpc-ibm-aix5.1.0.0
powerpc-ibm-aix5.2.0.0
powerpc-ibm-aix5.3.0.0
i686-pc-cygwin
        OK with a (completely) binmounted cygwin.
        With text mode mounts I got several failures.
        Somebody should test this, maybe it was just my cygwin
        setup being broken.  How could one change this quickly
        without a full reinstall and regedit?
powerpc-apple-darwin8.2.0
i386-unknown-freebsd5.4
hppa2.0-hp-hpux10.20
hppa2.0w-hp-hpux11.00
hppa2.0w-hp-hpux11.11
hppa2.0w-hp-hpux11.23
ia64-hp-hpux11.23
mips-sgi-irix6.5
i686-pc-linux-gnu
        OK for several different distributions
hppa2.0-unknown-linux-gnu
x86_64-unknown-linux-gnu
        OK with GCC, Portland (5.2, 6.0, 6.1), Intel 9.0.
x86_64-unknown-linux-gnu
        with Pathscale compiler has two failures:
        110: GNU Fortran 77
        111: GNU Fortran
        because the compiler tricks configure into thinking it's the GNU
        compiler, but actually has a different flag for free-form.
        Probably safest to fix the test rather than the macro?
i686-pc-mingw32
i386-unknown-openbsd3.8
alphaev5-dec-osf4.0d
alphaev67-dec-osf5.1
sparc-sun-solaris2.6
sparc-sun-solaris2.7
sparc-sun-solaris2.8
sparc-sun-solaris2.9
sparc-sun-solaris2.10

It also passes maintainer-check and distcheck on i686-pc-linux-gnu.

Cheers,
Ralf