Fix gsub losing white space when working on fields.

12 files changed, 388 insertions, 263 deletions

diff --git a/ChangeLog b/ChangeLog
index 0f01099a..2acb0567 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,21 @@
+2011-07-15         John Haque      <j.eh@mchsi.com>
+ 
+	* awk.h (Op_sub_builtin): New opcode.
+	(GSUB, GENSUB, AFTER_ASSIGN, LITERAL): New flags for 
+	Op_sub_builtin.
+	* awkgram.y (struct tokentab): Change opcode to	Op_sub_builtin
+	for sub, gsub and gensub.
+	(snode): Update processing of sub, gsub and gensub.
+	* builtin.c (do_sub, do_gsub, do_gensub): Nuke.
+	(sub_common): Renamed to do_sub. Relocate gensub argument
+	handling code from do_gensub to here; Simplify the code a
+	little bit. 
+	* eval.c (r_interpret): Handle Op_sub_builtin. Avoid field
+	re-splitting or $0 rebuilding if (g)sub target string is
+	a field and no substitutions were done.
+	* pprint (profile.c): Add case for the new opcode.
+	* print_instruction (debug.c): Ditto.
+
 2011-07-15         Arnold D. Robbins     <arnold@skeeve.com>
 
 	* awk.h: Typo fix: "loner" --> longer. Thanks to Nelson Beebe.
diff --git a/awk.h b/awk.h
index 2d13a213..d9da3b98 100644
--- a/awk.h
+++ b/awk.h
@@ -521,6 +521,7 @@ typedef enum opcodeval {
 	Op_K_nextfile,
 
 	Op_builtin,
+	Op_sub_builtin,		/* sub, gsub and gensub */
 	Op_in_array,		/* boolean test of membership in array */
 
 	/* function call instruction */
@@ -626,6 +627,16 @@ typedef struct exp_instruction {
 #define target_jmp      d.di
 #define target_break    x.xi
 
+/* Op_sub_builtin */
+#define sub_flags       d.dl
+#define GSUB            0x01	/* builtin is gsub */
+#define GENSUB          0x02	/* builtin is gensub */
+#define AFTER_ASSIGN    0x04	/* (g)sub target is a field or a special var with
+                            	 * set_XX routine.
+                            	 */
+#define LITERAL         0x08	/* target is a literal string */
+
+
 /* Op_K_exit */
 #define target_end      d.di
 #define target_atexit   x.xi	
@@ -1181,9 +1192,7 @@ extern NODE *do_cos(int nargs);
 extern NODE *do_rand(int nargs);
 extern NODE *do_srand(int nargs);
 extern NODE *do_match(int nargs);
-extern NODE *do_gsub(int nargs);
-extern NODE *do_sub(int nargs);
-extern NODE *do_gensub(int nargs);
+extern NODE *do_sub(int nargs, unsigned int flags, int *num_matches);
 extern NODE *format_tree(const char *, size_t, NODE **, long);
 extern NODE *do_lshift(int nargs);
 extern NODE *do_rshift(int nargs);
diff --git a/awkgram.c b/awkgram.c
index 4edec572..876821d5 100644
--- a/awkgram.c
+++ b/awkgram.c
@@ -4485,6 +4485,7 @@ struct token {
 #	define	RESX		0x0800	/* Bell Labs Research extension */
 #	define	BREAK		0x1000	/* break allowed inside */
 #	define	CONTINUE	0x2000	/* continue allowed inside */
+	
 	NODE *(*ptr)(int);	/* function that implements this keyword */
 };
 
@@ -4542,9 +4543,9 @@ static const struct token tokentab[] = {
 {"for",		Op_K_for,	 LEX_FOR,	BREAK|CONTINUE,	0},
 {"func",	Op_func, LEX_FUNCTION,	NOT_POSIX|NOT_OLD,	0},
 {"function",Op_func, LEX_FUNCTION,	NOT_OLD,	0},
-{"gensub",	Op_builtin,	 LEX_BUILTIN,	GAWKX|A(3)|A(4), do_gensub},
+{"gensub",	Op_sub_builtin,	 LEX_BUILTIN,	GAWKX|A(3)|A(4), 0},
 {"getline",	Op_K_getline_redir,	 LEX_GETLINE,	NOT_OLD,	0},
-{"gsub",	Op_builtin,	 LEX_BUILTIN,	NOT_OLD|A(2)|A(3), do_gsub},
+{"gsub",	Op_sub_builtin,	 LEX_BUILTIN,	NOT_OLD|A(2)|A(3), 0},
 {"if",		Op_K_if,	 LEX_IF,	0,		0},
 {"in",		Op_symbol,	 LEX_IN,	0,		0},
 {"include",  Op_symbol,	 LEX_INCLUDE,	GAWKX,	0},
@@ -4575,7 +4576,7 @@ static const struct token tokentab[] = {
 #endif
 {"strftime",	Op_builtin,	 LEX_BUILTIN,	GAWKX|A(0)|A(1)|A(2)|A(3), do_strftime},
 {"strtonum",	Op_builtin,    LEX_BUILTIN,	GAWKX|A(1),	do_strtonum},
-{"sub",		Op_builtin,	 LEX_BUILTIN,	NOT_OLD|A(2)|A(3), do_sub},
+{"sub",		Op_sub_builtin,	 LEX_BUILTIN,	NOT_OLD|A(2)|A(3), 0},
 {"substr",	Op_builtin,	 LEX_BUILTIN,	A(2)|A(3),	do_substr},
 {"switch",	Op_K_switch,	 LEX_SWITCH,	GAWKX|BREAK,	0},
 {"system",	Op_builtin,	 LEX_BUILTIN,	NOT_OLD|A(1),	do_system},
@@ -6286,8 +6287,6 @@ snode(INSTRUCTION *subn, INSTRUCTION *r)
 		assert(nexp > 0);
 	}		
 
-	r->builtin = tokentab[idx].ptr;
-
 	/* check against how many args. are allowed for this builtin */
 	args_allowed = tokentab[idx].flags & ARGS;
 	if (args_allowed && (args_allowed & A(nexp)) == 0) {
@@ -6296,7 +6295,86 @@ snode(INSTRUCTION *subn, INSTRUCTION *r)
 		return NULL;
 	}
 
+	/* special processing for sub, gsub and gensub */
+
+	if (tokentab[idx].value == Op_sub_builtin) {
+		const char *operator = tokentab[idx].operator;
+
+		r->sub_flags = 0;
+
+		arg = subn->nexti;		/* first arg list */
+		(void) mk_rexp(arg);
+
+		if (strcmp(operator, "gensub") != 0) {
+			/* sub and gsub */
+
+			if (strcmp(operator, "gsub") == 0)
+				r->sub_flags |= GSUB;
+
+			arg = arg->lasti->nexti;	/* 2nd arg list */
+			if (nexp == 2) {
+				INSTRUCTION *expr;
+
+				expr = list_create(instruction(Op_push_i));
+				expr->nexti->memory = mk_number((AWKNUM) 0.0, (PERM|NUMCUR|NUMBER));
+				(void) mk_expression_list(subn,
+						list_append(expr, instruction(Op_field_spec)));
+			}
+
+			arg = arg->lasti->nexti; 	/* third arg list */
+			ip = arg->lasti;
+			if (ip->opcode == Op_push_i) {
+				if (do_lint)
+					lintwarn(_("%s: string literal as last arg of substitute has no effect"),
+						operator);
+				r->sub_flags |=	LITERAL;
+			} else {
+				if (make_assignable(ip) == NULL)
+					yyerror(_("%s third parameter is not a changeable object"),
+						operator);
+				else
+					ip->do_reference = TRUE;
+			}
+
+			r->expr_count = count_expressions(&subn, FALSE);
+			ip = subn->lasti;
+
+			(void) list_append(subn, r);
+
+			/* add after_assign code */
+			if (ip->opcode == Op_push_lhs && ip->memory->type == Node_var && ip->memory->var_assign) {
+				(void) list_append(subn, instruction(Op_var_assign));
+				subn->lasti->memory = ip->memory;
+				subn->lasti->assign_var = ip->memory->var_assign;
+				r->sub_flags |= AFTER_ASSIGN;
+			} else if (ip->opcode == Op_field_spec_lhs) {
+				(void) list_append(subn, instruction(Op_field_assign));
+				subn->lasti->field_assign = (Func_ptr) 0;
+				ip->target_assign = subn->lasti;
+				r->sub_flags |= AFTER_ASSIGN;
+			}
+			return subn;	
+
+		} else {
+			/* gensub */
+
+			r->sub_flags |= GENSUB;
+			if (nexp == 3) {
+				ip = instruction(Op_push_i);
+				ip->memory = mk_number((AWKNUM) 0.0, (PERM|NUMCUR|NUMBER));
+				(void) mk_expression_list(subn,
+						list_append(list_create(ip), instruction(Op_field_spec)));
+			}
+
+			r->expr_count = count_expressions(&subn, FALSE);
+			return list_append(subn, r);
+		}
+	}
+
+	r->builtin = tokentab[idx].ptr;
+
 	/* special case processing for a few builtins */
+
 	if (r->builtin == do_length) {
 		if (nexp == 0) {		
 		    /* no args. Use $0 */
@@ -6338,71 +6416,6 @@ snode(INSTRUCTION *subn, INSTRUCTION *r)
 			if (/*ip == arg->nexti  && */ ip->opcode == Op_push)
 				ip->opcode = Op_push_array;
 		}
-	} else if (r->builtin == do_sub || r->builtin == do_gsub) {
-		int literal = FALSE;
-		
-		arg = subn->nexti;		/* first arg list */
-		(void) mk_rexp(arg);
-
-		arg = arg->lasti->nexti;	/* 2nd arg list */
-		if (nexp == 2) {
-			INSTRUCTION *expr;
-			expr = list_create(instruction(Op_push_i));
-			expr->nexti->memory = mk_number((AWKNUM) 0.0, (PERM|NUMCUR|NUMBER));
-			(void) mk_expression_list(subn,
-					list_append(expr, instruction(Op_field_spec)));
-		}
-
-		arg = arg->lasti->nexti; 	/* third arg list */
-		ip = arg->lasti;
-		if (ip->opcode == Op_push_i) {
-			if (do_lint)
-				lintwarn(_("%s: string literal as last arg of substitute has no effect"),
-					(r->builtin == do_sub) ? "sub" : "gsub");
-			literal = TRUE;
-		} else {
-			if (make_assignable(ip) == NULL)
-				yyerror(_("%s third parameter is not a changeable object"),
-					(r->builtin == do_sub) ? "sub" : "gsub");
-			else
-				ip->do_reference = TRUE;
-		}
-
-		/* kludge: This is one of the few cases
-		 * when we need to know the type of item on stack.
-		 * In case of string literal as the last argument,
-		 * pass 4 as # of args (See sub_common code in builtin.c).
-		 * Other cases like length(array or scalar) seem
-		 * to work out ok.
-		 */
-		
-		r->expr_count = count_expressions(&subn, FALSE) + !!literal;
-		ip = subn->lasti;
-
-		(void) list_append(subn, r);
-
-		/* add after_assign bytecode(s) */
-		if (ip->opcode == Op_push_lhs && ip->memory->type == Node_var && ip->memory->var_assign) {
-			(void) list_append(subn, instruction(Op_var_assign));
-			subn->lasti->memory = ip->memory;
-			subn->lasti->assign_var = ip->memory->var_assign;
-		} else if (ip->opcode == Op_field_spec_lhs) {
-			(void) list_append(subn, instruction(Op_field_assign));
-			subn->lasti->field_assign = (Func_ptr) 0;
-			ip->target_assign = subn->lasti;
-		}
-		return subn;	
-	} else if (r->builtin == do_gensub) {
-		if (nexp == 3) {
-			arg = subn->nexti->lasti->nexti->lasti->nexti;	/* 3rd arg list */
-			ip = instruction(Op_push_i);
-			ip->memory = mk_number((AWKNUM) 0.0, (PERM|NUMCUR|NUMBER));
-			(void) mk_expression_list(subn,
-					list_append(list_create(ip),
-						instruction(Op_field_spec)));
-		}
-		arg = subn->nexti;		/* first arg list */
-		(void) mk_rexp(arg);
 	} else if (r->builtin == do_split) {
 		arg = subn->nexti->lasti->nexti;	/* 2nd arg list */
 		ip = arg->lasti;
diff --git a/awkgram.y b/awkgram.y
index 6b28b520..3ef6382e 100644
--- a/awkgram.y
+++ b/awkgram.y
@@ -1795,6 +1795,7 @@ struct token {
 #	define	RESX		0x0800	/* Bell Labs Research extension */
 #	define	BREAK		0x1000	/* break allowed inside */
 #	define	CONTINUE	0x2000	/* continue allowed inside */
+	
 	NODE *(*ptr)(int);	/* function that implements this keyword */
 };
 
@@ -1852,9 +1853,9 @@ static const struct token tokentab[] = {
 {"for",		Op_K_for,	 LEX_FOR,	BREAK|CONTINUE,	0},
 {"func",	Op_func, LEX_FUNCTION,	NOT_POSIX|NOT_OLD,	0},
 {"function",Op_func, LEX_FUNCTION,	NOT_OLD,	0},
-{"gensub",	Op_builtin,	 LEX_BUILTIN,	GAWKX|A(3)|A(4), do_gensub},
+{"gensub",	Op_sub_builtin,	 LEX_BUILTIN,	GAWKX|A(3)|A(4), 0},
 {"getline",	Op_K_getline_redir,	 LEX_GETLINE,	NOT_OLD,	0},
-{"gsub",	Op_builtin,	 LEX_BUILTIN,	NOT_OLD|A(2)|A(3), do_gsub},
+{"gsub",	Op_sub_builtin,	 LEX_BUILTIN,	NOT_OLD|A(2)|A(3), 0},
 {"if",		Op_K_if,	 LEX_IF,	0,		0},
 {"in",		Op_symbol,	 LEX_IN,	0,		0},
 {"include",  Op_symbol,	 LEX_INCLUDE,	GAWKX,	0},
@@ -1885,7 +1886,7 @@ static const struct token tokentab[] = {
 #endif
 {"strftime",	Op_builtin,	 LEX_BUILTIN,	GAWKX|A(0)|A(1)|A(2)|A(3), do_strftime},
 {"strtonum",	Op_builtin,    LEX_BUILTIN,	GAWKX|A(1),	do_strtonum},
-{"sub",		Op_builtin,	 LEX_BUILTIN,	NOT_OLD|A(2)|A(3), do_sub},
+{"sub",		Op_sub_builtin,	 LEX_BUILTIN,	NOT_OLD|A(2)|A(3), 0},
 {"substr",	Op_builtin,	 LEX_BUILTIN,	A(2)|A(3),	do_substr},
 {"switch",	Op_K_switch,	 LEX_SWITCH,	GAWKX|BREAK,	0},
 {"system",	Op_builtin,	 LEX_BUILTIN,	NOT_OLD|A(1),	do_system},
@@ -3596,8 +3597,6 @@ snode(INSTRUCTION *subn, INSTRUCTION *r)
 		assert(nexp > 0);
 	}		
 
-	r->builtin = tokentab[idx].ptr;
-
 	/* check against how many args. are allowed for this builtin */
 	args_allowed = tokentab[idx].flags & ARGS;
 	if (args_allowed && (args_allowed & A(nexp)) == 0) {
@@ -3606,7 +3605,86 @@ snode(INSTRUCTION *subn, INSTRUCTION *r)
 		return NULL;
 	}
 
+	/* special processing for sub, gsub and gensub */
+
+	if (tokentab[idx].value == Op_sub_builtin) {
+		const char *operator = tokentab[idx].operator;
+
+		r->sub_flags = 0;
+
+		arg = subn->nexti;		/* first arg list */
+		(void) mk_rexp(arg);
+
+		if (strcmp(operator, "gensub") != 0) {
+			/* sub and gsub */
+
+			if (strcmp(operator, "gsub") == 0)
+				r->sub_flags |= GSUB;
+
+			arg = arg->lasti->nexti;	/* 2nd arg list */
+			if (nexp == 2) {
+				INSTRUCTION *expr;
+
+				expr = list_create(instruction(Op_push_i));
+				expr->nexti->memory = mk_number((AWKNUM) 0.0, (PERM|NUMCUR|NUMBER));
+				(void) mk_expression_list(subn,
+						list_append(expr, instruction(Op_field_spec)));
+			}
+
+			arg = arg->lasti->nexti; 	/* third arg list */
+			ip = arg->lasti;
+			if (ip->opcode == Op_push_i) {
+				if (do_lint)
+					lintwarn(_("%s: string literal as last arg of substitute has no effect"),
+						operator);
+				r->sub_flags |=	LITERAL;
+			} else {
+				if (make_assignable(ip) == NULL)
+					yyerror(_("%s third parameter is not a changeable object"),
+						operator);
+				else
+					ip->do_reference = TRUE;
+			}
+
+			r->expr_count = count_expressions(&subn, FALSE);
+			ip = subn->lasti;
+
+			(void) list_append(subn, r);
+
+			/* add after_assign code */
+			if (ip->opcode == Op_push_lhs && ip->memory->type == Node_var && ip->memory->var_assign) {
+				(void) list_append(subn, instruction(Op_var_assign));
+				subn->lasti->memory = ip->memory;
+				subn->lasti->assign_var = ip->memory->var_assign;
+				r->sub_flags |= AFTER_ASSIGN;
+			} else if (ip->opcode == Op_field_spec_lhs) {
+				(void) list_append(subn, instruction(Op_field_assign));
+				subn->lasti->field_assign = (Func_ptr) 0;
+				ip->target_assign = subn->lasti;
+				r->sub_flags |= AFTER_ASSIGN;
+			}
+			return subn;	
+
+		} else {
+			/* gensub */
+
+			r->sub_flags |= GENSUB;
+			if (nexp == 3) {
+				ip = instruction(Op_push_i);
+				ip->memory = mk_number((AWKNUM) 0.0, (PERM|NUMCUR|NUMBER));
+				(void) mk_expression_list(subn,
+						list_append(list_create(ip), instruction(Op_field_spec)));
+			}
+
+			r->expr_count = count_expressions(&subn, FALSE);
+			return list_append(subn, r);
+		}
+	}
+
+	r->builtin = tokentab[idx].ptr;
+
 	/* special case processing for a few builtins */
+
 	if (r->builtin == do_length) {
 		if (nexp == 0) {		
 		    /* no args. Use $0 */
@@ -3648,71 +3726,6 @@ snode(INSTRUCTION *subn, INSTRUCTION *r)
 			if (/*ip == arg->nexti  && */ ip->opcode == Op_push)
 				ip->opcode = Op_push_array;
 		}
-	} else if (r->builtin == do_sub || r->builtin == do_gsub) {
-		int literal = FALSE;
-		
-		arg = subn->nexti;		/* first arg list */
-		(void) mk_rexp(arg);
-
-		arg = arg->lasti->nexti;	/* 2nd arg list */
-		if (nexp == 2) {
-			INSTRUCTION *expr;
-			expr = list_create(instruction(Op_push_i));
-			expr->nexti->memory = mk_number((AWKNUM) 0.0, (PERM|NUMCUR|NUMBER));
-			(void) mk_expression_list(subn,
-					list_append(expr, instruction(Op_field_spec)));
-		}
-
-		arg = arg->lasti->nexti; 	/* third arg list */
-		ip = arg->lasti;
-		if (ip->opcode == Op_push_i) {
-			if (do_lint)
-				lintwarn(_("%s: string literal as last arg of substitute has no effect"),
-					(r->builtin == do_sub) ? "sub" : "gsub");
-			literal = TRUE;
-		} else {
-			if (make_assignable(ip) == NULL)
-				yyerror(_("%s third parameter is not a changeable object"),
-					(r->builtin == do_sub) ? "sub" : "gsub");
-			else
-				ip->do_reference = TRUE;
-		}
-
-		/* kludge: This is one of the few cases
-		 * when we need to know the type of item on stack.
-		 * In case of string literal as the last argument,
-		 * pass 4 as # of args (See sub_common code in builtin.c).
-		 * Other cases like length(array or scalar) seem
-		 * to work out ok.
-		 */
-		
-		r->expr_count = count_expressions(&subn, FALSE) + !!literal;
-		ip = subn->lasti;
-
-		(void) list_append(subn, r);
-
-		/* add after_assign bytecode(s) */
-		if (ip->opcode == Op_push_lhs && ip->memory->type == Node_var && ip->memory->var_assign) {
-			(void) list_append(subn, instruction(Op_var_assign));
-			subn->lasti->memory = ip->memory;
-			subn->lasti->assign_var = ip->memory->var_assign;
-		} else if (ip->opcode == Op_field_spec_lhs) {
-			(void) list_append(subn, instruction(Op_field_assign));
-			subn->lasti->field_assign = (Func_ptr) 0;
-			ip->target_assign = subn->lasti;
-		}
-		return subn;	
-	} else if (r->builtin == do_gensub) {
-		if (nexp == 3) {
-			arg = subn->nexti->lasti->nexti->lasti->nexti;	/* 3rd arg list */
-			ip = instruction(Op_push_i);
-			ip->memory = mk_number((AWKNUM) 0.0, (PERM|NUMCUR|NUMBER));
-			(void) mk_expression_list(subn,
-					list_append(list_create(ip),
-						instruction(Op_field_spec)));
-		}
-		arg = subn->nexti;		/* first arg list */
-		(void) mk_rexp(arg);
 	} else if (r->builtin == do_split) {
 		arg = subn->nexti->lasti->nexti;	/* 2nd arg list */
 		ip = arg->lasti;
diff --git a/builtin.c b/builtin.c
index 1d841dc0..5b75979c 100644
--- a/builtin.c
+++ b/builtin.c
@@ -72,7 +72,6 @@ extern NODE **fields_arr;
 extern int output_is_tty;
 extern FILE *output_fp;
 
-static NODE *sub_common(int nargs, long how_many, int backdigs);
 
 #define POP_TWO_SCALARS(s1, s2) \
 s2 = POP_SCALAR(); \
@@ -2319,7 +2318,7 @@ do_match(int nargs)
 	return make_number((AWKNUM) rstart);
 }
 
-/* sub_common --- the common code (does the work) for sub, gsub, and gensub */
+/* do_sub --- do the work for sub, gsub, and gensub */
 
 /*
  * Gsub can be tricksy; particularly when handling the case of null strings.
@@ -2412,12 +2411,12 @@ do_match(int nargs)
  * NB: `howmany' conflicts with a SunOS 4.x macro in <sys/param.h>.
  */
 
-static NODE *
-sub_common(int nargs, long how_many, int backdigs)
+NODE *
+do_sub(int nargs, unsigned int flags, int *num_matches)
 {
 	char *scan;
 	char *bp, *cp;
-	char *buf;
+	char *buf = NULL;
 	size_t buflen;
 	char *matchend;
 	size_t len;
@@ -2434,38 +2433,77 @@ sub_common(int nargs, long how_many, int backdigs)
 	NODE *s;		/* subst. pattern */
 	NODE *t;		/* string to make sub. in; $0 if none given */
 	NODE *tmp;
-	NODE **lhs;
-	int global = (how_many == -1);
+	NODE **lhs = NULL;
+	long how_many = 1;	/* one substitution for sub, also gensub default */
+	int global;
 	long current;
 	int lastmatchnonzero;
 	char *mb_indices = NULL;
-
-	tmp = PEEK(2);	/* take care of regexp early, in case re_update is fatal */
-	rp = re_update(tmp);
 	
-	/* original string */
-	if (nargs == 4) {	/* kludge: no of items on stack is really 3,
-                         * See snode(..) in awkgram.y
-                         */ 
-		lhs = NULL;
-		t = POP_STRING();
+	if ((flags & GENSUB) != 0) {
+		double d;
+		NODE *t1;
+
+		tmp = PEEK(3);
+		rp = re_update(tmp);
+
+		t = POP_STRING();	/* original string */
+
+		t1 = POP_SCALAR();	/* value of global flag */
+		if ((t1->flags & (STRCUR|STRING)) != 0) {
+			if (t1->stlen > 0 && (t1->stptr[0] == 'g' || t1->stptr[0] == 'G'))
+				how_many = -1;
+			else {
+				d = force_number(t1);
+
+				if ((t1->flags & NUMCUR) != 0)
+					goto set_how_many;
+
+				how_many = 1;
+			}
+		} else {
+			d = force_number(t1);
+set_how_many:
+			if (d < 1)
+				how_many = 1;
+			else if (d < LONG_MAX)
+				how_many = d;
+			else
+				how_many = LONG_MAX;
+			if (d == 0)
+				warning(_("gensub: third argument of 0 treated as 1"));
+		}
+		DEREF(t1);
+
 	} else {
-		lhs = POP_ADDRESS();
-		t = force_string(*lhs);
+
+		/* take care of regexp early, in case re_update is fatal */
+
+		tmp = PEEK(2);
+		rp = re_update(tmp);
+
+		if ((flags & GSUB) != 0)
+			how_many = -1;
+
+		/* original string */
+
+		if ((flags & LITERAL) != 0)
+			t = POP_STRING();
+		else {
+			lhs = POP_ADDRESS();
+			t = force_string(*lhs);
+		}
 	}
 
+	global = (how_many == -1);
 
-	s = POP_STRING();		/* replacement text */
+	s = POP_STRING();	/* replacement text */
 	decr_sp();		/* regexp, already updated above */
 
 	/* do the search early to avoid work on non-match */
 	if (research(rp, t->stptr, 0, t->stlen, RE_NEED_START) == -1 ||
-	    RESTART(rp, t->stptr) > t->stlen) {
-		if (lhs == NULL)
-			DEREF(t);
-		DEREF(s);
-		return make_number((AWKNUM) 0.0);
-	}
+			RESTART(rp, t->stptr) > t->stlen)
+		goto done;
 
 	t->flags |= STRING;
 
@@ -2476,7 +2514,7 @@ sub_common(int nargs, long how_many, int backdigs)
 	repl = s->stptr;
 	replend = repl + s->stlen;
 	repllen = replend - repl;
-	emalloc(buf, char *, buflen + 2, "sub_common");
+	emalloc(buf, char *, buflen + 2, "do_sub");
 	buf[buflen] = '\0';
 	buf[buflen + 1] = '\0';
 	ampersands = 0;
@@ -2490,7 +2528,7 @@ sub_common(int nargs, long how_many, int backdigs)
 	 * for example.
 	 */
 	if (gawk_mb_cur_max > 1 && repllen > 0) {
-		emalloc(mb_indices, char *, repllen * sizeof(char), "sub_common");
+		emalloc(mb_indices, char *, repllen * sizeof(char), "do_sub");
 		index_multibyte_buffer(repl, mb_indices, repllen);
 	}
 
@@ -2500,7 +2538,7 @@ sub_common(int nargs, long how_many, int backdigs)
 			repllen--;
 			ampersands++;
 		} else if (*scan == '\\') {
-			if (backdigs) {	/* gensub, behave sanely */
+			if (flags & GENSUB) {	/* gensub, behave sanely */
 				if (isdigit((unsigned char) scan[1])) {
 					ampersands++;
 					scan++;
@@ -2575,7 +2613,7 @@ sub_common(int nargs, long how_many, int backdigs)
 					&& (gawk_mb_cur_max == 1
 						|| (repllen > 0 && mb_indices[scan - repl] == 1))
 				) {
-					if (backdigs) {	/* gensub, behave sanely */
+					if (flags & GENSUB) {	/* gensub, behave sanely */
 						if (isdigit((unsigned char) scan[1])) {
 							int dig = scan[1] - '0';
 							if (dig < NUMSUBPATS(rp, t->stptr) && SUBPATSTART(rp, tp->stptr, dig) != -1) {
@@ -2619,7 +2657,7 @@ sub_common(int nargs, long how_many, int backdigs)
 		textlen = text + textlen - matchend;
 		text = matchend;
 
-		if ((current >= how_many && !global)
+		if ((current >= how_many && ! global)
 		    || ((long) textlen <= 0 && matchstart == matchend)
 		    || research(rp, t->stptr, text - t->stptr, textlen, RE_NEED_START) == -1)
 			break;
@@ -2628,7 +2666,7 @@ sub_common(int nargs, long how_many, int backdigs)
 	sofar = bp - buf;
 	if (buflen - sofar - textlen - 1) {
 		buflen = sofar + textlen + 2;
-		erealloc(buf, char *, buflen, "sub_common");
+		erealloc(buf, char *, buflen, "do_sub");
 		bp = buf + sofar;
 	}
 	for (scan = matchend; scan < text + textlen; scan++)
@@ -2636,102 +2674,39 @@ sub_common(int nargs, long how_many, int backdigs)
 	*bp = '\0';
 	textlen = bp - buf;
 
-	DEREF(s);
-
-	if (lhs != NULL) {
-		if (matches > 0) {
-			unref(*lhs);
-			*lhs = make_str_node(buf, textlen, ALREADY_MALLOCED);	
-		} else
-			efree(buf);
-	} else {
-		efree(buf);
-		DEREF(t);
-	}
-
 	if (mb_indices != NULL)
 		efree(mb_indices);
 
-	return make_number((AWKNUM) matches);
-}
-
-/* do_gsub --- global substitution */
-
-NODE *
-do_gsub(int nargs)
-{
-	return sub_common(nargs, -1, FALSE);
-}
-
-/* do_sub --- single substitution */
-
-NODE *
-do_sub(int nargs)
-{
-	return sub_common(nargs, 1, FALSE);
-}
-
-/* do_gensub --- fix up the tree for sub_common for the gensub function */
-
-NODE *
-do_gensub(int nargs)
-{
-	NODE *t, *tmp, *target, *ret;
-	long how_many = 1;	/* default is one substitution */
-	double d;
-
-	tmp = POP_STRING();	/* target */
-	t = POP_SCALAR();	/* value of global flag */
-
-	/*
-	 * We make copy of the original target string, and pass that
-	 * in to sub_common() as the target to make the substitution in.
-	 * We will then return the result string as the return value of
-	 * this function.
-	 */
-
-	target = make_string(tmp->stptr, tmp->stlen);
-	DEREF(tmp);
-	PUSH_ADDRESS(& target);
-
-	if ((t->flags & (STRCUR|STRING)) != 0) {
-		if (t->stlen > 0 && (t->stptr[0] == 'g' || t->stptr[0] == 'G'))
-			how_many = -1;
-		else {
-			d = force_number(t);
+done:
+	DEREF(s);
 
-			if ((t->flags & NUMCUR) != 0)
-				goto set_how_many;
+	*num_matches = matches;
+	if ((matches == 0 || (flags & LITERAL) != 0) && buf != NULL)
+		efree(buf); 
 
-			how_many = 1;
+	if (flags & GENSUB) {
+		if (matches > 0) {
+			/* return the result string */
+			DEREF(t);
+			return make_str_node(buf, textlen, ALREADY_MALLOCED);	
 		}
-	} else {
-		d = force_number(t);
-set_how_many:
-		if (d < 1)
-			how_many = 1;
-		else if (d < LONG_MAX)
-			how_many = d;
-		else
-			how_many = LONG_MAX;
-		if (d == 0)
-			warning(_("gensub: third argument of 0 treated as 1"));
-	}
-
-	DEREF(t);
 
-	ret = sub_common(3, how_many, TRUE);
-	unref(ret);
+		/* return the original string */
+		return t;
+	}
 
-	/*
-	 * Note that we don't care what sub_common() returns, since the
-	 * easiest thing for the programmer is to return the string, even
-	 * if no substitutions were done.
-	 */
+	/* For a string literal, must not change the original string. */
+	if (flags & LITERAL)
+		DEREF(t);
+	else if (matches > 0) {
+		unref(*lhs);
+		*lhs = make_str_node(buf, textlen, ALREADY_MALLOCED);	
+	}
 
-	return target;
+	return make_number((AWKNUM) matches);
 }
 
+
 /* make_integer - Convert an integer to a number node.  */
 
 static NODE *
diff --git a/debug.c b/debug.c
index 9b9db345..404042c1 100644
--- a/debug.c
+++ b/debug.c
@@ -3740,7 +3740,16 @@ print_instruction(INSTRUCTION *pc, Func_print print_func, FILE *fp, int in_dump)
 		break;
 
 	case Op_var_assign:
-		print_func(fp, "[set_%s]\n", pc->memory->vname);
+		if (pc->assign_var)
+			print_func(fp, "[set_%s()]", pc->memory->vname);
+		print_func(fp, "\n");
+		break;
+
+	case Op_field_assign:
+		if (pc->field_assign)
+			print_func(fp, "[%s]", pc->field_assign == reset_record ?
+					"reset_record()" : "invalidate_field0()");
+		print_func(fp, "\n");
 		break;
 
 	case Op_field_spec_lhs:
@@ -3830,6 +3839,27 @@ print_instruction(INSTRUCTION *pc, Func_print print_func, FILE *fp, int in_dump)
 		                pc->line_range, pc->target_jmp);
 		break;
 
+	case Op_sub_builtin:
+	{
+		const char *fname = "sub";
+		static const struct flagtab values[] = {
+                	{ GSUB, "GSUB" },
+			{ GENSUB, "GENSUB" },
+			{ AFTER_ASSIGN,	"AFTER_ASSIGN" },
+			{ LITERAL, "LITERAL" },
+			{ 0, NULL }
+		};
+
+		if (pc->sub_flags & GSUB)
+			fname = "gsub";
+		else if (pc->sub_flags & GENSUB)
+			fname = "gensub";
+		print_func(fp, "%s [arg_count = %ld] [sub_flags = %s]\n",
+				fname, pc->expr_count,
+				genflags2str(pc->sub_flags, values));
+	}
+		break;
+	
 	case Op_builtin:
 	{
 		const char *fname = getfname(pc->builtin);
diff --git a/eval.c b/eval.c
index 4132474a..bdbd04bd 100644
--- a/eval.c
+++ b/eval.c
@@ -348,6 +348,7 @@ static struct optypetab {
 	{ "Op_K_getline", "getline" },
 	{ "Op_K_nextfile", "nextfile" },
 	{ "Op_builtin", NULL },
+	{ "Op_sub_builtin", NULL },
 	{ "Op_in_array", " in " },
 	{ "Op_func_call", NULL },
 	{ "Op_indirect_func_call", NULL },
@@ -2114,11 +2115,13 @@ post:
 			break;
 
 		case Op_var_assign:
-			pc->assign_var();
+			if (pc->assign_var)
+				pc->assign_var();
 			break;
 
 		case Op_field_assign:
-			pc->field_assign();
+			if (pc->field_assign)
+				pc->field_assign();
 			break;
 
 		case Op_concat:
@@ -2256,7 +2259,34 @@ arrayfor:
 #endif
 				PUSH(r);
 			break;
-			
+
+		case Op_sub_builtin:
+		{
+			/* sub, gsub and gensub */
+ 
+			int matches = 0;
+
+			r = do_sub(pc->expr_count, pc->sub_flags, & matches);
+			PUSH(r);
+
+			if (matches == 0 && (pc->sub_flags & AFTER_ASSIGN) != 0) {
+
+				/* For sub and gsub, must not execute after_assign code;
+				 * If the target is a FIELD, this means no field re-splitting or
+				 * $0 reconstruction. For a special variable as target,
+				 * set_XX routine is not called. 
+				 */
+
+				ni = pc->nexti;
+				assert(ni->opcode == Op_field_assign || ni->opcode == Op_var_assign);
+				if (ni->opcode == Op_field_assign)
+					ni->field_assign = (Func_ptr) 0;
+				else
+					ni->assign_var = (Func_ptr) 0;
+			}
+		}
+			break;
+
 		case Op_K_print:
 			do_print(pc->expr_count, pc->redir_type);
 			break;
diff --git a/profile.c b/profile.c
index cba8be9e..01d1e42f 100644
--- a/profile.c
+++ b/profile.c
@@ -507,6 +507,20 @@ cleanup:
 		case Op_after_endfile:
 			break;
 
+		case Op_sub_builtin:
+		{
+			const char *fname = "sub";
+			if (pc->sub_flags & GSUB)
+				fname = "gsub";
+			else if (pc->sub_flags & GENSUB)
+				fname = "gensub";
+			tmp = pp_list(pc->expr_count, "()", ", ");
+			str = pp_concat(fname, tmp, "");
+			efree(tmp);
+			pp_push(Op_sub_builtin, str, CAN_FREE);
+		}
+			break;
+
 		case Op_builtin:
 		{
 			static char *ext_func = "extension_function()";
diff --git a/test/ChangeLog b/test/ChangeLog
index 2ae225be..60c4f525 100644
--- a/test/ChangeLog
+++ b/test/ChangeLog
@@ -1,3 +1,8 @@
+2011-07-15         Arnold D. Robbins     <arnold@skeeve.com>
+
+	* Makefile.am (gsubtst7): New test.
+	* gsubtst7.awk, gsubtst7.in, gsubtst7.ok: New files.
+
 2011-06-24         Arnold D. Robbins     <arnold@skeeve.com>
 
 	* Makefile.am (EXTRA_DIST): Add ChangeLog.0.
diff --git a/test/Makefile.am b/test/Makefile.am
index 2d7bf34f..12f64a62 100644
--- a/test/Makefile.am
+++ b/test/Makefile.am
@@ -318,6 +318,9 @@ EXTRA_DIST = \
 	gsubtst5.ok \
 	gsubtst6.awk \
 	gsubtst6.ok \
+	gsubtst7.awk \
+	gsubtst7.in \
+	gsubtst7.ok \
 	gtlnbufv.awk \
 	hex.awk \
 	hex.ok \
@@ -768,6 +771,7 @@ BASIC_TESTS = \
 	fordel forref forsimp fsbs fsrs fsspcoln fstabplus funsemnl funsmnam \
 	funstack getline getline2 getline3 getlnbuf getnr2tb getnr2tm \
 	gsubasgn gsubtest gsubtst2 gsubtst3 gsubtst4 gsubtst5 gsubtst6 \
+	gsubtst7 \
 	hex hsprint inputred intest intprec iobug1 leaddig leadnl litoct \
 	longsub longwrds manglprm math membug1 messages minusstr mmap8k \
 	mtchi18n nasty nasty2 negexp negrange nested nfldstr nfneg \
diff --git a/test/Makefile.in b/test/Makefile.in
index 04ea041c..eee32eba 100644
--- a/test/Makefile.in
+++ b/test/Makefile.in
@@ -503,6 +503,9 @@ EXTRA_DIST = \
 	gsubtst5.ok \
 	gsubtst6.awk \
 	gsubtst6.ok \
+	gsubtst7.awk \
+	gsubtst7.in \
+	gsubtst7.ok \
 	gtlnbufv.awk \
 	hex.awk \
 	hex.ok \
@@ -953,6 +956,7 @@ BASIC_TESTS = \
 	fordel forref forsimp fsbs fsrs fsspcoln fstabplus funsemnl funsmnam \
 	funstack getline getline2 getline3 getlnbuf getnr2tb getnr2tm \
 	gsubasgn gsubtest gsubtst2 gsubtst3 gsubtst4 gsubtst5 gsubtst6 \
+	gsubtst7 \
 	hex hsprint inputred intest intprec iobug1 leaddig leadnl litoct \
 	longsub longwrds manglprm math membug1 messages minusstr mmap8k \
 	mtchi18n nasty nasty2 negexp negrange nested nfldstr nfneg \
@@ -2077,6 +2081,11 @@ gsubtst5:
 	@AWKPATH=$(srcdir) $(AWK) -f $@.awk  < $(srcdir)/$@.in >_$@ 2>&1 || echo EXIT CODE: $$? >>_$@
 	@-$(CMP) $(srcdir)/$@.ok _$@ && rm -f _$@
 
+gsubtst7:
+	@echo gsubtst7
+	@AWKPATH=$(srcdir) $(AWK) -f $@.awk  < $(srcdir)/$@.in >_$@ 2>&1 || echo EXIT CODE: $$? >>_$@
+	@-$(CMP) $(srcdir)/$@.ok _$@ && rm -f _$@
+
 hex:
 	@echo hex
 	@AWKPATH=$(srcdir) $(AWK) -f $@.awk  >_$@ 2>&1 || echo EXIT CODE: $$? >>_$@
diff --git a/test/Maketests b/test/Maketests
index 9a16eb7c..9f364038 100644
--- a/test/Maketests
+++ b/test/Maketests
@@ -350,6 +350,11 @@ gsubtst5:
 	@AWKPATH=$(srcdir) $(AWK) -f $@.awk  < $(srcdir)/$@.in >_$@ 2>&1 || echo EXIT CODE: $$? >>_$@
 	@-$(CMP) $(srcdir)/$@.ok _$@ && rm -f _$@
 
+gsubtst7:
+	@echo gsubtst7
+	@AWKPATH=$(srcdir) $(AWK) -f $@.awk  < $(srcdir)/$@.in >_$@ 2>&1 || echo EXIT CODE: $$? >>_$@
+	@-$(CMP) $(srcdir)/$@.ok _$@ && rm -f _$@
+
 hex:
 	@echo hex
 	@AWKPATH=$(srcdir) $(AWK) -f $@.awk  >_$@ 2>&1 || echo EXIT CODE: $$? >>_$@