This is the mail archive of the automake@gnu.org mailing list for the automake project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: ac_config_libobj_dir broken


> From: Akim Demaille <akim@epita.fr>
> Date: 06 Sep 2002 13:50:20 +0200
> 
> >>>>> "Andreas" == Andreas Schwab <schwab@suse.de> writes:
> 
> Andreas> The default value of ac_config_libobj_dir is wrong, it should
> Andreas> be $srcdir.  Similarily, AC_CONFIG_LIBOBJ_DIR should prepend
> Andreas> $srcdir to the directory.
> 
> Arg, you are right.  But I'm not sure sticking srcdir in
> ac_config_libobj_dir is the right thing to do.  For instance, in the
> replacement of fnmatch, you have:
> 
> AC_CONFIG_LINKS([$ac_config_libobj_dir/fnmatch.h:$ac_config_libobj_dir/fnmatch_.h])
> 
> This idiom is expected to spread, for getopt and others, and it should
> not try to write in the $srcdir tree.  I suppose there will be more
> such examples where some replacement file will actually be in the
> build hierarchy :(
> 
> So I'm more inclined to paste $srcdir in from of $ac_config_libobj_dir
> where needed.
> 
> What do you (and other people :) think?

Yes, I tend to agree.  $ac_config_libobj_dir should be a relative
path, and it should be a subdirectory of both $srcdir and of
$builddir.  It sounds like Autoconf should reject attempts to use an
absolute directory as an argument to AC_CONFIG_LIBOBJ_DIR.



Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]