Re: UGLY_DEFS for darwin in configure.in

[John W. Eaton]

On  3-Oct-2007, Thomas Treichl wrote:

| My suggestion would be to reject because it really isn't a problem
| to install a very new GNU's sed on any platform and use this beside
| any other old sed that is installed

Does the following patch properly reject the default broken sed on
your system?

jwe

ChangeLog:

2007-10-03  John W. Eaton  <address@hidden>

        * aclocal.m4 (OCTAVE_PROG_SED): Also check for \(X\|Y\) style
        regular expression alternation.


Index: aclocal.m4
===================================================================
RCS file: /cvs/octave/aclocal.m4,v
retrieving revision 1.110
diff -u -u -r1.110 aclocal.m4
--- aclocal.m4  1 Oct 2007 18:58:29 -0000       1.110
+++ aclocal.m4  3 Oct 2007 19:02:24 -0000
@@ -806,9 +806,10 @@
 # OCTAVE_PROG_SED
 # --------------
 # Check for a fully-functional sed program, that truncates
-# as few characters as possible.  Prefer GNU sed if found.
+# as few characters as possible and that supports "\(X\|Y\)"
+# style regular expression alternation.  Prefer GNU sed if found.
 AC_DEFUN([OCTAVE_PROG_SED],
-[AC_MSG_CHECKING([for a sed that does not truncate output])
+[AC_MSG_CHECKING([for a usable sed])
 if test -z "$SED"; then
   AC_CACHE_VAL(ac_cv_path_sed, [
   # Loop through the user's path and test for sed and gsed.
@@ -837,6 +838,10 @@
        octave_cv_path_sed=${_sed}
        break;
       fi
+      # Reject if RE alternation is not handled.
+      if test "`echo 'this and that' | sed -n 's/\(this\|that\).*$/\1/p'`" != 
"this"; then
+        continue;
+      fi
       while true; do
        cat "$tmp/sed.in" "$tmp/sed.in" >"$tmp/sed.tmp"
        mv "$tmp/sed.tmp" "$tmp/sed.in"