Re: Aclocal bug?

[Ralf Wildenhues]

Hello Gary,

* Gary V. Vaughan wrote on Wed, Nov 08, 2006 at 06:34:45AM CET:
> Bad news, I appear to have hit a bug in aclocal.  From a fresh checkout
> of m4 CVS HEAD:
[...]
> $ bootstrap
> ...
> bootstrap: running: autopoint --force
> bootstrap: running: gnulib-tool --update
> ...
> bootstrap: running: AUTOPOINT=true autoreconf --force --verbose -- 
> install --no-r
> ecursive
> autoreconf: Entering directory `.'
> autoreconf: running: true --force
> autoreconf: running: aclocal --force -I ltdl/m4
> autoreconf: configure.ac: tracing
> configure.ac:147: warning: AM_INTL_SUBDIR is m4_require'd but not  
> m4_defun'd
> ltdl/m4/gettext.m4:360: AM_GNU_GETTEXT is expanded from...

I'm pretty sure AM_INTL_SUBDIR isn't needed in CVS HEAD M4 (see the
logic in gettext.m4 for 'external').  My bootstrap passes, aclocal.m4
doesn't include ltdl/m4/intl.m4, but also doesn't warn.  I have perl
5.8.7, but otherwise CVS versions of Autoconf, Automake, Gettext, and M4
1.4.7a installed.

Please post the output of the aclocal step --verbose'ly, and report
the m4 version used by autom4te.  Also, do you by chance have Gettext
installed below a different $prefix than Automake, and if yes, do you
have $automake_prefix/share/aclocal/dirlist pointing to gettext.m4's
directory?  Likewise for Libtool.

FWIW, one way to ensure that CVS HEAD libtoolize was called could be to
test for non-existing directory 'libltdl' after calling autoreconf in
the bootstrap script.

Cheers, and I'll try following the riding blog... ;-)
Ralf