Re: AC_DEFINE problem

[Ralf Wildenhues]

[ Your mailer has very amusing ways of messing with addresses. :) ]

Hi Keith,

* Keith MARSHALL wrote on Fri, Apr 28, 2006 at 05:48:31PM CEST:
> Ralf Wildenhues wrote, quoting Peter Breitenlohner:
> >> Isn't this (i.e., aclocal.m4 containing private macros[*] without
> >> their sources) something that should be VERY STRONGLY discouraged.
> >> After all, running "reautoconf" may (does?) invoke aclocal.
> >
> > Yes, I agree.
> 
> Eh?  Quoting autoconf-2.59 documentation:
> |To create a `configure' script with Autoconf, you need to write an
> |Autoconf input file `configure.ac' (or `configure.in') and run
> |`autoconf' on it.  If you write your own feature tests to supplement
> |those that come with Autoconf, you might also write files called
> |`aclocal.m4' and `acsite.m4'.
> 
> If I follow this advice, and write `aclocal.m4' myself, then *by
> definition*, `aclocal.m4' *is* the source!

Yes, you are right.  But then your aclocal.m4 file doesn't start with:

| # generated automatically by aclocal 1.7.9 -*- Autoconf -*-

and thus autoreconf won't cause aclocal to overwrite it.  Sorry for
being less precise than necessary here, and thanks for the reminder.

Cheers,
Ralf