Document string termination in header files and remove no-longer-needed string termination logic in various places.

12 files changed, 55 insertions, 58 deletions

diff --git a/ChangeLog b/ChangeLog
index 1eb2fa0a..3284747d 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,5 +1,39 @@
 2016-07-06         Andrew J. Schorr     <aschorr@telemetry-investments.com>
 
+	* awk.h: Modify stptr comment to indicate that all strings are now
+	NUL-terminated.
+	* builtin.c (do_mktime): Remove unnecessary logic to terminate
+	the string with '\0' temporarily.
+	(do_system) Ditto.
+	(nondec2awknum): Add a comment about termination.
+	* eval.c (posix_compare): Remove logic to terminate strings temporarily.
+	(set_ORS): No need to terminate ORS, since the string node is already
+	terminated. What gave us the right to modify that node anyway?
+	(fmt_index): Remove code to terminate string. This seems to have been
+	invalid anyway, since we don't own that memory.
+	(set_TEXTDOMAIN): Do not terminate TEXTDOMAIN string, since the node
+	is already terminated. We didn't have the right to modify that node
+	anyway.
+	* gawkapi.c (node_to_awk_value): Add assert checks to confirm that the
+	string is NUL-terminated.
+	* gawkapi.h: Modify awk_string comment to indicate that strings are
+	always terminated with '\0'.
+	* int_array.c (isinteger): Remove unnecessary logic to terminate string
+	with '\0' temporarily.
+	* interpret.h (Op_push_i): Ditto.
+	* io.c (nextfile): Remove string termination. We didn't own that memory
+	anyway.
+	* mpfr.c (force_mpnum): Remove unnecessary logic to terminate the
+	string with '\0' temporarily.
+	* node.c (r_force_number): Remove NUL termination around strtod call,
+	since we already know that there is either a white space or '\0'
+	character there. Either one will stop strtod.
+	(get_ieee_magic_val): Ditto.
+	* profile.c (pp_number): No need to terminate string returned by
+	r_format_val.
+
+2016-07-06         Andrew J. Schorr     <aschorr@telemetry-investments.com>
+
 	* interpret.h (Op_field_spec): Now that all $n field values are
 	NUL-terminated, there is no reason to call dupnode for $n where n > 0.
 	This saves malloc and copying overhead, thereby more than offsetting the
diff --git a/awk.h b/awk.h
index b210dcdc..09757b9f 100644
--- a/awk.h
+++ b/awk.h
@@ -473,13 +473,7 @@ typedef struct exp_node {
 #define re_cnt	flags
 
 /* Node_val */
-/*
- * Note that the string in stptr may not be NUL-terminated, but it is
- * guaranteed to have at least one extra byte that may be temporarily set
- * to '\0'. This is helpful when calling functions such as strtod that require
- * a NUL-terminated argument. In particular, field values $n for n > 0 and
- * n < NF will not have a NUL terminator, since they point into the $0 buffer.
- */
+/* Note that the string in stptr will always be NUL-terminated. */
 #define stptr	sub.val.sp
 #define stlen	sub.val.slen
 #define valref	sub.val.sref
diff --git a/builtin.c b/builtin.c
index a01018e3..a21df18e 100644
--- a/builtin.c
+++ b/builtin.c
@@ -2035,16 +2035,12 @@ do_mktime(int nargs)
 	int month, day, hour, minute, second, count;
 	int dst = -1; /* default is unknown */
 	time_t then_stamp;
-	char save;
 
 	t1 = POP_SCALAR();
 	if (do_lint && (fixtype(t1)->flags & STRING) == 0)
 		lintwarn(_("mktime: received non-string argument"));
 	t1 = force_string(t1);
 
-	save = t1->stptr[t1->stlen];
-	t1->stptr[t1->stlen] = '\0';
-
 	count = sscanf(t1->stptr, "%ld %d %d %d %d %d %d",
 		        & year, & month, & day,
 			& hour, & minute, & second,
@@ -2058,7 +2054,6 @@ do_mktime(int nargs)
 		|| (month < 1 || month > 12) ))
 			lintwarn(_("mktime: at least one of the values is out of the default range"));
 
-	t1->stptr[t1->stlen] = save;
 	DEREF(t1);
 
 	if (count < 6
@@ -2088,7 +2083,6 @@ do_system(int nargs)
 	NODE *tmp;
 	AWKNUM ret = 0;		/* floating point on purpose, compat Unix awk */
 	char *cmd;
-	char save;
 	int status;
 
 	if (do_sandbox)
@@ -2101,10 +2095,6 @@ do_system(int nargs)
 	cmd = force_string(tmp)->stptr;
 
 	if (cmd && *cmd) {
-		/* insure arg to system is zero-terminated */
-		save = cmd[tmp->stlen];
-		cmd[tmp->stlen] = '\0';
-
 		os_restore_mode(fileno(stdin));
 #ifdef SIGPIPE
 		signal(SIGPIPE, SIG_DFL);
@@ -2148,7 +2138,6 @@ do_system(int nargs)
 		signal(SIGPIPE, SIG_IGN);
 #endif
 
-		cmd[tmp->stlen] = save;
 	}
 	DEREF(tmp);
 	return make_number((AWKNUM) ret);
@@ -3632,6 +3621,11 @@ nondec2awknum(char *str, size_t len, char **endptr)
 			*endptr = str;
 	} else {
 decimal:
+		/*
+		 * Terminating is probably unnecessary, since the caller always
+		 * passes a string ending with '\0' or white space, but it
+		 * seems safest to leave this to avoid future problems.
+		 */
 		save = str[len];
 		str[len] = '\0';
 		retval = strtod(str, endptr);
diff --git a/eval.c b/eval.c
index 2f473569..aaabdb3a 100644
--- a/eval.c
+++ b/eval.c
@@ -493,15 +493,8 @@ static int
 posix_compare(NODE *s1, NODE *s2)
 {
 	int ret = 0;
-	char save1, save2;
 	size_t l = 0;
 
-	save1 = s1->stptr[s1->stlen];
-	s1->stptr[s1->stlen] = '\0';
-
-	save2 = s2->stptr[s2->stlen];
-	s2->stptr[s2->stlen] = '\0';
-
 	if (gawk_mb_cur_max == 1) {
 		if (strlen(s1->stptr) == s1->stlen && strlen(s2->stptr) == s2->stlen)
 			ret = strcoll(s1->stptr, s2->stptr);
@@ -563,8 +556,6 @@ posix_compare(NODE *s1, NODE *s2)
 	}
 #endif
 
-	s1->stptr[s1->stlen] = save1;
-	s2->stptr[s2->stlen] = save2;
 	return ret;
 }
 
@@ -824,7 +815,6 @@ set_ORS()
 	ORS_node->var_value = force_string(ORS_node->var_value);
 	ORS = ORS_node->var_value->stptr;
 	ORSlen = ORS_node->var_value->stlen;
-	ORS[ORSlen] = '\0';
 }
 
 /* fmt_ok --- is the conversion format a valid one? */
@@ -887,7 +877,6 @@ fmt_index(NODE *n)
 		ix++;
 	}
 	/* not found */
-	n->stptr[n->stlen] = '\0';
 	if (do_lint && ! fmt_ok(n))
 		lintwarn(_("bad `%sFMT' specification `%s'"),
 			    n == CONVFMT_node->var_value ? "CONV"
@@ -972,7 +961,6 @@ set_TEXTDOMAIN()
 	tmp = TEXTDOMAIN_node->var_value = force_string(TEXTDOMAIN_node->var_value);
 	TEXTDOMAIN = tmp->stptr;
 	len = tmp->stlen;
-	TEXTDOMAIN[len] = '\0';
 	/*
 	 * Note: don't call textdomain(); this value is for
 	 * the awk program, not for gawk itself.
diff --git a/gawkapi.c b/gawkapi.c
index df69012b..afefa4f6 100644
--- a/gawkapi.c
+++ b/gawkapi.c
@@ -440,6 +440,7 @@ node_to_awk_value(NODE *node, awk_value_t *val, awk_valtype_t wanted)
 			(void) force_string(node);
 			val->str_value.str = node->stptr;
 			val->str_value.len = node->stlen;
+			assert(val->str_value.str[val->str_value.len] == '\0');
 			ret = awk_true;
 			break;
 
@@ -468,6 +469,7 @@ node_to_awk_value(NODE *node, awk_value_t *val, awk_valtype_t wanted)
 				val->val_type = AWK_STRING;
 				val->str_value.str = node->stptr;
 				val->str_value.len = node->stlen;
+				assert(val->str_value.str[val->str_value.len] == '\0');
 				ret = awk_true;
 			} else
 				val->val_type = AWK_UNDEFINED;
diff --git a/gawkapi.h b/gawkapi.h
index 975f82df..10eab1cf 100644
--- a/gawkapi.h
+++ b/gawkapi.h
@@ -279,11 +279,7 @@ enum {
  * be multibyte encoded in the current locale's encoding and character
  * set. Gawk will convert internally to wide characters if necessary.
  *
- * Note that the string may not be terminated with a '\0' character.
- * In particular, this happens for field values $n where n > 0 and n < NF,
- * since the string points directly into the $0 buffer. All other strings,
- * including those created by extensions, should be NUL-terminated. In general
- * though, extension code should not assume that the string is NUL-terminated!
+ * Note that the string will always be terminated with a '\0' character.
  */
 typedef struct awk_string {
 	char *str;	/* data */
diff --git a/int_array.c b/int_array.c
index e7913dea..93e96d1f 100644
--- a/int_array.c
+++ b/int_array.c
@@ -128,7 +128,6 @@ is_integer(NODE *symbol, NODE *subs)
 
 	/* must be a STRING */
 	char *cp = subs->stptr, *cpend, *ptr;
-	char save;
 	size_t len = subs->stlen;
 
 	if (len == 0 || (! isdigit((unsigned char) *cp) && *cp != '-'))
@@ -151,12 +150,9 @@ is_integer(NODE *symbol, NODE *subs)
 	}
 
 	cpend = cp + len;
-	save = *cpend;
-	*cpend = '\0';
 
 	errno = 0;
 	l = strtol(cp, & ptr, 10);
-	*cpend = save;
 	if (errno != 0 || ptr != cpend)
 		return NULL;
 
diff --git a/interpret.h b/interpret.h
index 106367f7..9d7b423e 100644
--- a/interpret.h
+++ b/interpret.h
@@ -135,13 +135,10 @@ top:
 		case Op_push_i:
 			m = pc->memory;
 			if (! do_traditional && (m->flags & INTLSTR) != 0) {
-				char *orig, *trans, save;
+				char *orig, *trans;
 
-				save = m->stptr[m->stlen];
-				m->stptr[m->stlen] = '\0';
 				orig = m->stptr;
 				trans = dgettext(TEXTDOMAIN, orig);
-				m->stptr[m->stlen] = save;
 				m = make_string(trans, strlen(trans));
 			} else
 				UPREF(m);
diff --git a/io.c b/io.c
index b9bce694..5ad7913a 100644
--- a/io.c
+++ b/io.c
@@ -480,7 +480,6 @@ nextfile(IOBUF **curfile, bool skipping)
 		if (arg == NULL || arg->stlen == 0)
 			continue;
 		arg = force_string(arg);
-		arg->stptr[arg->stlen] = '\0';
 		if (! do_traditional) {
 			unref(ARGIND_node->var_value);
 			ARGIND_node->var_value = make_number((AWKNUM) i);
diff --git a/mpfr.c b/mpfr.c
index 0bb5b435..cf7a7c21 100644
--- a/mpfr.c
+++ b/mpfr.c
@@ -275,7 +275,6 @@ static int
 force_mpnum(NODE *n, int do_nondec, int use_locale)
 {
 	char *cp, *cpend, *ptr, *cp1;
-	char save;
 	int tval, base = 10;
 
 	if (n->stlen == 0) {
@@ -292,9 +291,6 @@ force_mpnum(NODE *n, int do_nondec, int use_locale)
 		return false;
 	}
 
-	save = *cpend;
-	*cpend = '\0';
-
 	if (*cp == '+' || *cp == '-')
 		cp1 = cp + 1;
 	else
@@ -329,7 +325,6 @@ done:
 	/* trailing space is OK for NUMBER */
 	while (ptr < cpend && isspace((unsigned char) *ptr))
 		ptr++;
-	*cpend = save;
 	if (errno == 0 && ptr == cpend)
 		return true;
 	errno = 0;
diff --git a/node.c b/node.c
index a435e01d..2cd88331 100644
--- a/node.c
+++ b/node.c
@@ -59,7 +59,6 @@ r_force_number(NODE *n)
 {
 	char *cp;
 	char *cpend;
-	char save;
 	char *ptr;
 	extern double strtod();
 
@@ -133,10 +132,13 @@ r_force_number(NODE *n)
 		/* nondec2awknum() saves and restores the byte after the string itself */
 		n->numbr = nondec2awknum(cp, cpend - cp, &ptr);
 	} else {
-		save = *cpend;
-		*cpend = '\0';
+		/*
+		 * There is no need to set *cpend to '\0' because it is either
+		 * pointing to white space or the '\0' at the end of the string.
+		 * In either case, strtod should terminate on that character
+		 * or earlier due to non-numeric characters.
+		 */
 		n->numbr = (AWKNUM) strtod((const char *) cp, &ptr);
-		*cpend = save;
 	}
 
 	if (errno == 0) {
@@ -941,13 +943,14 @@ get_ieee_magic_val(char *val)
 	static bool first = true;
 	static AWKNUM inf;
 	static AWKNUM nan;
-	char save;
 
 	char *ptr;
-	save = val[4];
-	val[4] = '\0';
+	/*
+	 * There is no need to set val[4] to '\0' because it is either white
+	 * space or the NUL character at the end of the string. Either way,
+	 * strtod should terminate on that character.
+	 */
 	AWKNUM v = strtod(val, &ptr);
-	val[4] = save;
 
 	if (val == ptr) { /* Older strtod implementations don't support inf or nan. */
 		if (first) {
diff --git a/profile.c b/profile.c
index be8977e8..aa8a152b 100644
--- a/profile.c
+++ b/profile.c
@@ -1541,7 +1541,6 @@ pp_number(NODE *n)
 
 		s = r_format_val("%.6g", 0, s);
 
-		s->stptr[s->stlen] = '\0';
 		str = s->stptr;
 
 		freenode(s);