On Mon, May 01, 2006 at 09:39:04PM -0500, Bob Friesenhahn wrote:
> On Mon, 1 May 2006, Bob Rossi wrote:
> >
> >Would you mind quicly explaining how autoconf works? Meaning what are
> >the steps taken that generate a configure.sh script?
>
> Don't be needlessly confused by the m4 "gibberish" which is often
> posted here. It is not necessary to understand such "gibberish" in
> order to properly use autoconf.
>
> The 'aclocal' program is used to pull in any m4 macros defined in the
> autoconf/automake/libtool system install directory, as well as any
> project-specific macros from acinclude.m4, and put them in an
> aclocal.m4 file. The 'aclocal' program is smart and only pulls in the
> macros that configure.ac actually needs. When autoconf is executed,
> it uses m4 to create a shell script ('configure') based on expanding
> the macros from aclocal.m4 as they are encountered in configure.ac.
> Recently it has become popular to include macros in additional .m4
> files so that the macros may be bundled in the package without
> cluttering up the project-specific acinclude.m4 file or to make it
> more clear who maintains them. These may be placed in the same
> directory as configure.ac, or in a subdirectory, as long as autoconf
> is told where to find them.
Thanks! This greatly improves my understanding. I actually have a very
little bit of experience working with m4, as I'm working with Akim to
get bison acting as a push parser.
> >I understand that m4 is used. Is m4 the only tool that is run on the
> >configure.in file to generate the configure script? I'm trying to better
> >understand how macro's in an sh condition effect the outcome.
>
> Long ago I wrote a magazine article on Autoconf. Remarkably most of
> it is still correct, which shows that the GNU autoconf approach is
> quite stable. See "http://www.byte.com/art/9711/sec4/art2.htm";.
Thanks, I'll read this today.
Bob Rossi
_______________________________________________
Autoconf mailing list
Autoconf@xxxxxxx
http://lists.gnu.org/mailman/listinfo/autoconf
