Re: configure problem

[Kenichi Handa <handa@m17n.org>]

In article <b9yvfu4z3q4.fsf@jpl.org>, Katsumi Yamaoka <yamaoka@jpl.org> writes:
> configure.in:50: error: possibly undefined macro: AC_DEFINE
>       If this token and others are legitimate, please use m4_pattern_allow.
>       See the Autoconf documentation.

> Here is a patch which has been confirmed by the NetBSD user.
> I don't know why the AC_DEFINE macro is not expanded there,
> though. :{

The reason is described in the comment above this jambo
"case" statement as this:

dnl This used to use changequote, but, apart from `changequote is evil'
dnl per the autoconf manual, we can speed up autoconf somewhat by quoting
dnl the great gob of text.  Thus it's not processed for possible expansion.
dnl Just make sure the brackets remain balanced.

I think it's not good to handle confdefs.h directly in
configure.in.  It's the task of autoconf.  So, have about
this alternative change?

---
Ken'ichi HANDA
handa@m17n.org


*** configure.in.~1.343.~	Tue Jul 15 13:11:42 2003
--- configure.in	Tue Jul 15 15:45:56 2003
***************
*** 222,231 ****
    ## NetBSD ports
    *-*-netbsd* )
      opsys=netbsd
-     if test -f /usr/lib/crti.o; then
-       AC_DEFINE(HAVE_CRTIN)
-     fi
- 
      case "${canonical}" in
        alpha*-*-netbsd*)	machine=alpha ;;
        i[3456]86-*-netbsd*) machine=intel386 ;;
--- 222,227 ----
***************
*** 1168,1173 ****
--- 1164,1175 ----
  fi
  
  ]
+ 
+ if test "${opsys}" = "netbsd"; then
+   if test -f /usr/lib/crti.o; then
+     AC_DEFINE(HAVE_CRTIN)
+   fi
+ fi
  
  if test $unported = yes; then
    AC_MSG_ERROR([Emacs hasn't been ported to `${canonical}' systems.