Akim,
I assume the problem is for AM_FUNC_ERROR_AT_LINE, not AM_FUNC_ERROR.
As far as I can see, Jim doesn't define AC_FUNC_ERROR at all. The is
an AC_REQUIRE([AM_FUNC_ERROR_AT_LINE]) in jm-macros.m4. The actual
definition for AM_FUNC_ERROR_AT_LINE appears to come from aclocal/error.m4
in automake-1.4a.
>From your email, it appears that you are going to back out the change
that obsoleted these macros until an Automake with quoted macros is
released?
Dave
>
> >>>>> "John" == John David Anglin <[EMAIL PROTECTED]> writes:
>
> John> On a different subject, in trying to use autoconf-2.14a to
> John> regenerate configure for fileutils-4.0x, I get the error
> John> _LC_LANG is not defined in AC_LANG_SOURCE. Is this a bug or has
> John> there been a change in how this is initialized?
>
> OK, I have it. This is due to the fact that now Autoconf declares
> that AM_FUNC_ERROR and some other Automake macros are obsolete (since
> you are supposed to use the macros from Autoconf itself).
>
> To declare them obsolete, Autoconf defines AM_FUNC_ERROR. But the
> Automake used by Jim defines AC_FUNC_ERROR without quoting the name,
> hence, when autoconf see
>
> AC_DEFUN(AM_FUNC_ERROR, [...])
>
> an expansion is performed on AM_FUNC_ERROR leading to yuck or gross
> effects (YMMV).
>
> To bad. I think this definitely means Autoconf 2.50 must not obsolete
> the symbols from Automake until an Automake with quoted macro names is
> released. This problem is a good means to have Mike shot in Joe's
> foot.
>
> Akim
>
> PS/ To solve your issue with the fileutils, track the `AC_DEFUN(A' in
> aclocal.m4 and quote the name, ie. `AC_DEFUN([AM_FUNC_ERROR], ...)'.
>
--
J. David Anglin [EMAIL PROTECTED]
National Research Council of Canada (613) 990-0752 (FAX: 952-6605)
