new (some possible) testcases from Austin ML, and one fix

check.t

@@ -1,4 +1,4 @@
-# $MirOS: src/bin/mksh/check.t,v 1.432 2011/03/23 18:47:04 tg Exp $
+# $MirOS: src/bin/mksh/check.t,v 1.433 2011/03/26 15:32:35 tg Exp $
 # $OpenBSD: bksl-nl.t,v 1.2 2001/01/28 23:04:56 niklas Exp $
 # $OpenBSD: history.t,v 1.5 2001/01/28 23:04:56 niklas Exp $
 # $OpenBSD: read.t,v 1.3 2003/03/10 03:48:16 david Exp $
@@ -2080,6 +2080,156 @@ expected-stdout:
 	} vf={=f $x @=
 	} |
 ---
+name: heredoc-comsub-1
+description:
+	Tests for here documents in COMSUB, taken from Austin ML
+stdin:
+	text=$(cat <<EOF
+	here is the text
+	EOF)
+	echo = $text =
+expected-stdout:
+	= here is the text =
+---
+name: heredoc-comsub-2
+description:
+	Tests for here documents in COMSUB, taken from Austin ML
+stdin:
+	unbalanced=$(cat <<EOF
+	this paren ) is a problem
+	EOF)
+	echo = $unbalanced =
+expected-stdout:
+	= this paren ) is a problem =
+---
+name: heredoc-comsub-3
+description:
+	Tests for here documents in COMSUB, taken from Austin ML
+stdin:
+	balanced=$(cat <<EOF
+	these parens ( ) are not a problem
+	EOF)
+	echo = $balanced =
+expected-stdout:
+	= these parens ( ) are not a problem =
+---
+name: heredoc-comsub-4
+description:
+	Tests for here documents in COMSUB, taken from Austin ML
+stdin:
+	balanced=$(cat <<EOF
+	these parens \( ) are a problem
+	EOF)
+	echo = $balanced =
+expected-stdout:
+	= these parens \( ) are a problem =
+---
+name: heredoc-subshell-1
+description:
+	Tests for here documents in subshells, taken from Austin ML
+stdin:
+	(cat <<EOF
+	some text
+	EOF)
+	echo end
+expected-stdout:
+	some text
+	end
+---
+name: heredoc-subshell-2
+description:
+	Tests for here documents in subshells, taken from Austin ML
+stdin:
+	(cat <<EOF
+	some text
+	EOF
+	)
+	echo end
+expected-stdout:
+	some text
+	end
+---
+name: heredoc-subshell-3
+description:
+	Tests for here documents in subshells, taken from Austin ML
+stdin:
+	(cat <<EOF; )
+	some text
+	EOF
+	echo end
+expected-stdout:
+	some text
+	end
+---
+name: heredoc-weird-1
+description:
+	Tests for here documents, taken from Austin ML
+	Documents current state in mksh, *NOT* necessarily correct!
+stdin:
+	cat <<END
+	hello
+	END\
+	END
+	END
+	echo end
+expected-stdout:
+	hello
+	ENDEND
+	end
+---
+name: heredoc-weird-2
+description:
+	Tests for here documents, taken from Austin ML
+stdin:
+	cat <<'    END    '
+	hello
+	    END    
+	echo end
+expected-stdout:
+	hello
+	end
+---
+name: heredoc-weird-3
+description:
+	Tests for here documents, taken from Austin ML
+stdin:
+	cat <<x*x & touch 'x*x'
+	hello
+	x*x
+	echo end
+expected-stdout:
+	hello
+	end
+---
+name: heredoc-weird-4
+description:
+	Tests for here documents, taken from Austin ML
+	Documents current state in mksh, *NOT* necessarily correct!
+stdin:
+	cat <<END
+	hello\
+	END
+	END
+	echo end
+expected-stdout:
+	helloEND
+	end
+---
+name: heredoc-weird-5
+description:
+	Tests for here documents, taken from Austin ML
+	Documents current state in mksh, *NOT* necessarily correct!
+stdin:
+	cat <<END
+	hello
+	\END
+	END
+	echo end
+expected-stdout:
+	hello
+	\END
+	end
+---
 name: heredoc-quoting-unsubst
 description:
 	Check for correct handling of quoted characters in

lex.c

@@ -22,7 +22,7 @@
 
 #include "sh.h"
 
-__RCSID("$MirOS: src/bin/mksh/lex.c,v 1.142 2011/03/13 16:35:54 tg Exp $");
+__RCSID("$MirOS: src/bin/mksh/lex.c,v 1.143 2011/03/26 15:32:37 tg Exp $");
 
 /*
  * states while lexing word
@@ -108,7 +108,7 @@ void yyskiputf8bom(void);
 static int backslash_skip;
 static int ignore_backslash_newline;
 static struct sretrace_info *retrace_info = NULL;
-short comsub_nesting_level = 0;
+short subshell_nesting_level = 0;
 
 /* optimised getsc_bn() */
 #define _getsc()	(*source->str != '\0' && *source->str != '\\' && \
@@ -1198,10 +1198,13 @@ readhere(struct ioword *iop)
 		/* end of here document marker, what to do? */
 		switch (c) {
 		case /*(*/ ')':
-			if (!comsub_nesting_level)
+			if (!subshell_nesting_level)
-				/* not allowed outside $(...) => mismatch */
+				/*-
+				 * not allowed outside $(...) or (...)
+				 * => mismatch
+				 */
 				break;
-			/* Allow $(...) to close here */
+			/* allow $(...) or (...) to close here */
 			ungetsc(/*(*/ ')');
 			/* FALLTHROUGH */
 		case 0:

syn.c

@@ -22,7 +22,9 @@
 
 #include "sh.h"
 
-__RCSID("$MirOS: src/bin/mksh/syn.c,v 1.58 2011/03/21 21:57:35 tg Exp $");
+__RCSID("$MirOS: src/bin/mksh/syn.c,v 1.59 2011/03/26 15:32:37 tg Exp $");
+
+extern short subshell_nesting_level;
 
 struct nesting_state {
 	int start_token;	/* token than began nesting (eg, FOR) */
@@ -356,7 +358,9 @@ get_command(int cf)
 
 	case '(':
  Subshell:
+		++subshell_nesting_level;
 		t = nested(TPAREN, '(', ')');
+		--subshell_nesting_level;
 		break;
 
 	case '{': /*}*/
@@ -1098,10 +1102,9 @@ yyrecursive(void)
 	bool old_reject;
 	int old_symbol;
 	struct ioword **old_herep;
-	extern short comsub_nesting_level;
 
 	/* tell the lexer to accept a closing parenthesis as EOD */
-	++comsub_nesting_level;
+	++subshell_nesting_level;
 
 	/* push reject state, parse recursively, pop reject state */
 	old_reject = reject;
@@ -1118,6 +1121,6 @@ yyrecursive(void)
 	cp = snptreef(NULL, 0, "%T", t->left);
 	tfree(t, ATEMP);
 
-	--comsub_nesting_level;
+	--subshell_nesting_level;
 	return (cp);
 }