1999-10-31 Akim Demaille <akim@epita.fr>

* acspecific.m4 (AC_PROG_GNU_M4): New macro.
	* configure.in: Use it.

ChangeLog

@@ -1,3 +1,8 @@
+1999-10-31  Akim Demaille  <akim@epita.fr>
+
+	* acspecific.m4 (AC_PROG_GNU_M4): New macro.
+	* configure.in: Use it.
+
 1999-10-31  Akim Demaille  <akim@epita.fr>
 
 	Provide a means to specify commands to run before config.status is

TODO

@@ -13,11 +13,17 @@ These are things mandatory to fulfill before releasing 2.15.  There
 are also suggestions we should either satisfy right now (they're
 easy), or remove (obsoleted since then).
 
+** AC_PROG_GNU_M4
+Document.
+
+** Produce a static list of all the AC_SUBST.
+This would greatly help Automake and the independence of Automake and
+Autoconf.  Currently Automake needs inner details of Autoconf.
+
 ** autoconf.texi: config.status in node names.
-The info mode of Emacs seems to have problems with this.  If the info
-reader catches that properly, we should not care.  If no info reader
-can handle this, we should change the affected node names.  Bug
-triggered by looking for `CONFIG_FILES' in the index.
+The info mode of Emacs seems to have problems with this.  The info
+reader catches this properly, so we should report to the X?Emacs teams.
+Bug triggered by looking for `CONFIG_FILES' in the index.
 
 ** AC_CONFIG_FILES, AC_CONFIG_COMMANDS.
 Will we need both ICMD and CMD as is the case for AC_OUTPUT_COMMANDS?
@@ -133,6 +139,13 @@ From: Roland McGrath <roland@gnu.ai.mit.edu>
 
 * Autoconf 3
 
+** Split all those &#@! files.
+acgeneral and acspecific is a bad design.  The split should be done
+according to the families.  For instance AC_CHECK_FUNCS,
+AC_CHECK_FUNC, AC_FUNC_FNMATCH etc. *and* AH_CHECK_FUNCS should be in,
+acfunc.m4.  I suppose there should be small exceptions, e.g., the
+family AC_REPLACE_FUNCS should still be in acfunc.m4.
+
 ** Use m4 lists?
 I think one sad decision in Autoconf was to use white space separated
 lists for some arguments.  For instance AC_CHECK_FUNCS(foo bar).  I

acspecific.m4

@@ -564,6 +564,19 @@ esac
 ])dnl AC_PROG_CC_STDC
 
 
+dnl AC_PROG_GNU_M4
+dnl --------------
+dnl Check for GNU m4.
+AC_DEFUN(AC_PROG_GNU_M4,
+[AC_PATH_PROGS(M4, gm4 gnum4 m4, m4)
+AC_CACHE_CHECK(whether we are using GNU m4, ac_cv_prog_gnu_m4,
+[ac_cv_prog_gnu_m4=no
+if test x"$M4" != x; then
+  case `$M4 --help < /dev/null 2>&1` in
+    *traditional*) ac_cv_prog_gnu_m4=yes ;;
+  esac
+fi])])
+
 dnl AC_PROG_MAKE_SET
 dnl ----------------
 dnl Define SET_MAKE to set ${MAKE} if make doesn't.

configure.in

@@ -6,17 +6,13 @@ dnl We use a path for GNU m4 so even if users have another m4 first in
 dnl their path, the installer can configure with a path that has GNU m4
 dnl first and get that path embedded in the installed autoconf and
 dnl autoheader scripts.
-AC_PATH_PROGS(M4, gm4 gnum4 m4, m4)
-AC_CACHE_CHECK(whether we are using GNU m4, acac_cv_gnu_m4,
-[acac_cv_gnu_m4=no
-if test x"$M4" != x; then
-  case `$M4 --help < /dev/null 2>&1` in
-    *traditional*) acac_cv_gnu_m4=yes ;;
-  esac
-fi])
-if test x"$acac_cv_gnu_m4" != xyes; then
+AC_PROG_GNU_M4
+if test x"$ac_cv_prog_gnu_m4" != xyes; then
   AC_ERROR([GNU m4 is required])
 fi
+dnl This is needed because Automake does not seem to realize there is
+dnl an AC_SUBST inside AC_PROG_GNU_M4.  Grmph!
+AC_SUBST(M4)
 AC_PROG_AWK
 
 dnl We use a path for perl so the #! line in autoscan will work.

lib/autoconf/specific.m4

@@ -564,6 +564,19 @@ esac
 ])dnl AC_PROG_CC_STDC
 
 
+dnl AC_PROG_GNU_M4
+dnl --------------
+dnl Check for GNU m4.
+AC_DEFUN(AC_PROG_GNU_M4,
+[AC_PATH_PROGS(M4, gm4 gnum4 m4, m4)
+AC_CACHE_CHECK(whether we are using GNU m4, ac_cv_prog_gnu_m4,
+[ac_cv_prog_gnu_m4=no
+if test x"$M4" != x; then
+  case `$M4 --help < /dev/null 2>&1` in
+    *traditional*) ac_cv_prog_gnu_m4=yes ;;
+  esac
+fi])])
+
 dnl AC_PROG_MAKE_SET
 dnl ----------------
 dnl Define SET_MAKE to set ${MAKE} if make doesn't.