[Groff] Re: patch for configure.ac and aclocal.m4

[Ralph Corderoy]

Hi Gaius,

There's one portability problem with the patch, and a suggestion whilst
I'm at it.

> @@ -82,6 +84,29 @@
>  AC_SUBST(make_html)
>  AC_SUBST(make_install_html)])dnl
>  dnl
> +dnl check to see whether pnmtops can handle the -nosetpage option
> +dnl
> +AC_DEFUN(PNMTOPS_NOSETPAGE,
> +[AC_MSG_CHECKING([whether pnmtops can handle the -nosetpage option])
> +cat <<EOF | pstopnm > conftest.pnm 2>/dev/null
> +%!PS-Adobe-3.0
> +/Times-Bold findfont 16 scalefont setfont
> +/inch {72 mul} def
> +/Times-Roman findfont 10 scalefont setfont
> +4.5 inch 0.812500 1 add inch moveto ( testing ) show
> +showpage
> +EOF

If there's no particular reason why the test image has to be generated
from PostScript then pbmmake or pbmtext could be used instead in the
pipeline on the line below.  See `Image Generators' in netpbm(1).

> +if pnmtops -nosetpage conftest.pnm >& /dev/null ; then

`>&' isn't portable.  I'd suggest

    if pnmtops -nosetpage conftest.pnm >/dev/null 2>&1; then

> +    AC_MSG_RESULT(yes)
> +    pnm_to_ps_nosetpage="pnmtops -nosetpage"
> +else
> +    AC_MSG_RESULT(no)
> +    pnm_to_ps_nosetpage="pnmtops"
> +fi
> +rm -f conftest.pnm

The temporary file wouldn't be needed then.

> +AC_SUBST(pnm_to_ps_nosetpage)
> +])dnl
> +dnl
>  dnl
>  dnl GROFF_CSH_HACK(if hack present, if not present)
>  dnl

Cheers,


Ralph.