Re: AC_FUNC_LSTAT_FOLLOWS_SLASHED_SYMLINK

[Tim Van Holder]

> Tim> So I changed it to $LN_S, but forgot that was only substituted if
> Tim> AC_PROG_LN_S is invoked (does that make sense, btw?  we always
> Tim> compute it, so why don't we substitute it too?)
> 
> If we do compute it systematically, please, indeed, get rid of
> ac_ln_s.  If not, I would not be against computing it anyway.

So it's OK to move the

AC_SUBST(LN_S, $as_ln_s)

to _AS_PREPARE_LN_S
and make AC_PROG_LN_S a virtual no-op?