doc issue (because no regression… even if it works in ksh93):

19:09 < nDuff> I'm reading the following behavior as contrary to the documentation. Could someone clarify?
19:09 < nDuff> $ words_pat='*@(cat|mice)*'; [[ 'there is a {cat}' = $words_pat ]]; echo $?
19:09 < nDuff> 1
19:09 < nDuff> $ [[ 'there is a {cat}' = *@(cat|mice)* ]]; echo $?
19:09 < nDuff> 0

mksh.1

@@ -1,4 +1,4 @@
-.\" $MirOS: src/bin/mksh/mksh.1,v 1.332 2014/03/07 22:47:49 tg Exp $
+.\" $MirOS: src/bin/mksh/mksh.1,v 1.333 2014/06/06 23:50:22 tg Exp $
 .\" $OpenBSD: ksh.1,v 1.151 2014/01/28 14:16:59 jmc Exp $
 .\"-
 .\" Copyright © 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009,
@@ -74,7 +74,7 @@
 .\" with -mandoc, it might implement .Mx itself, but we want to
 .\" use our own definition. And .Dd must come *first*, always.
 .\"
-.Dd $Mdocdate: March 7 2014 $
+.Dd $Mdocdate: June 6 2014 $
 .\"
 .\" Check which macro package we use, and do other -mdoc setup.
 .\"
@@ -964,7 +964,7 @@ The second operand of the
 .Sq !=
 and
 .Sq =
-expressions are patterns (e.g. the comparison
+expressions are a subset of patterns (e.g. the comparison
 .Ic \&[[ foobar = f*r ]]
 succeeds).
 This even works indirectly:
@@ -976,6 +976,7 @@ $ [[ $bar = \&"$baz" ]]; echo $?
 .Pp
 Perhaps surprisingly, the first comparison succeeds,
 whereas the second doesn't.
+This does not apply to all extglob metacharacters, currently.
 .El
 .El
 .Ss Quoting