Re: INSTALL should tell about particular systems, HP-UX cc

[Bruno Haible]

Ralf Wildenhues wrote:
> > I can agree to dropping the recommendation to try CC="cc -Ae" if
> Is there an "or" or an "and" connecting (1) and (2)?

That was an implicit "and" :-)

Let me separate the two issues:

-----------------------------------------------------------------------------
1) AC_USE_SYSTEM_EXTENSIONS

> >   1) AC_USE_SYSTEM_EXTENSIONS is modified to add -Ae on HP-UX with cc, if
> >      it is not already present in $CC.

The platforms where this has an effect are those where the expensive HP-UX
add-on compiler is installed.

-----------------------------------------------------------------------------
2) AC_PROG_CC / AC_PROG_CC_C89

> >   - On a system where the expensive HP C compiler is not installed, the
> >     autoconf test certainly tried "cc -Ae" and this failed, so -Ae was
> >     not added to CC or the CFLAGS.
> 
> OK, IIUC then we don't need to care about this case at all, right?

Two things still to do:
  - The AC_PROG_CC should reject this compiler if it does not support ANSI C.
  - The documentation should mention that GCC binaries exist as a workaround.

> Judging from the manual, I'd expect AC_PROG_CC_STDC to achieve this
> (so if it doesn't, it should be fixed).

But AC_PROG_CC_STDC is not used by many packages. AC_PROG_CC_STDC tests for
C99, which many packages don't need. Amanda (from which this report comes
from) does not use it. Gnulib does not use it (except in module 'stdarg').
What these programs would need, according to the current autoconf, is:

  AC_PROG_CC_C89
  if test "$ac_cv_prog_cc_c89" = no; then
    echo "C compiler does not support ANSI C" 1>&2
    exit 1
  fi

But no one does this...

Bruno