Stop allocating an extra wasted byte at the end of various strings.

15 files changed, 47 insertions, 26 deletions

diff --git a/ChangeLog b/ChangeLog
index a9c7e555..c54a1774 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,16 @@
+2014-12-14         Andrew J. Schorr     <aschorr@telemetry-investments.com>
+
+	* array.c (concat_exp): Do not waste a byte at the end of a string.
+	* awkgram.y (common_exp): Ditto.
+	* builtin.c (do_substr): Ditto.
+	* eval.c (set_OFS): Ditto.
+	* field.c (rebuild_record): Ditto.
+	* gawkapi.h (r_make_string): Ditto.
+	* interpret.h (r_interpret): Ditto for Op_assign_concat.
+	* node.c (r_format_val, r_dupnode, make_str_node, str2wstr, wstr2str):
+	Ditto.
+	* re.c (make_regexp): Ditto.
+
 2014-12-12        Stephen Davies         <sdavies@sdc.com.au>
 
 	Improve comment handling in pretty printing.
diff --git a/array.c b/array.c
index f7993624..59428ba2 100644
--- a/array.c
+++ b/array.c
@@ -414,7 +414,7 @@ concat_exp(int nargs, bool do_subsep)
 	}
 	len += (nargs - 1) * subseplen;
 
-	emalloc(str, char *, len + 2, "concat_exp");
+	emalloc(str, char *, len + 1, "concat_exp");
 
 	r = args_array[nargs];
 	memcpy(str, r->stptr, r->stlen);
diff --git a/awkgram.c b/awkgram.c
index 225cdb4e..4715290d 100644
--- a/awkgram.c
+++ b/awkgram.c
@@ -3398,7 +3398,7 @@ regular_print:
 			n1 = force_string(n1);
 			n2 = force_string(n2);
 			nlen = n1->stlen + n2->stlen;
-			erealloc(n1->stptr, char *, nlen + 2, "constant fold");
+			erealloc(n1->stptr, char *, nlen + 1, "constant fold");
 			memcpy(n1->stptr + n1->stlen, n2->stptr, n2->stlen);
 			n1->stlen = nlen;
 			n1->stptr[nlen] = '\0';
diff --git a/awkgram.y b/awkgram.y
index 6721bcdc..f6a396a5 100644
--- a/awkgram.y
+++ b/awkgram.y
@@ -1442,7 +1442,7 @@ common_exp
 			n1 = force_string(n1);
 			n2 = force_string(n2);
 			nlen = n1->stlen + n2->stlen;
-			erealloc(n1->stptr, char *, nlen + 2, "constant fold");
+			erealloc(n1->stptr, char *, nlen + 1, "constant fold");
 			memcpy(n1->stptr + n1->stlen, n2->stptr, n2->stlen);
 			n1->stlen = nlen;
 			n1->stptr[nlen] = '\0';
diff --git a/builtin.c b/builtin.c
index 21c6ed5c..838ab899 100644
--- a/builtin.c
+++ b/builtin.c
@@ -1833,7 +1833,7 @@ do_substr(int nargs)
 		 * way to do things.
 		 */
 		memset(& mbs, 0, sizeof(mbs));
-		emalloc(substr, char *, (length * gawk_mb_cur_max) + 2, "do_substr");
+		emalloc(substr, char *, (length * gawk_mb_cur_max) + 1, "do_substr");
 		wp = t1->wstptr + indx;
 		for (cp = substr; length > 0; length--) {
 			result = wcrtomb(cp, *wp, & mbs);
diff --git a/eval.c b/eval.c
index 95992704..eca829c1 100644
--- a/eval.c
+++ b/eval.c
@@ -828,9 +828,9 @@ set_OFS()
 	new_ofs_len = OFS_node->var_value->stlen;
 
 	if (OFS == NULL)
-		emalloc(OFS, char *, new_ofs_len + 2, "set_OFS");
+		emalloc(OFS, char *, new_ofs_len + 1, "set_OFS");
 	else if (OFSlen < new_ofs_len)
-		erealloc(OFS, char *, new_ofs_len + 2, "set_OFS");
+		erealloc(OFS, char *, new_ofs_len + 1, "set_OFS");
 
 	memcpy(OFS, OFS_node->var_value->stptr, OFS_node->var_value->stlen);
 	OFSlen = new_ofs_len;
diff --git a/extension/ChangeLog b/extension/ChangeLog
index 41c8a0e4..dca14332 100644
--- a/extension/ChangeLog
+++ b/extension/ChangeLog
@@ -1,3 +1,10 @@
+2014-12-14         Andrew J. Schorr     <aschorr@telemetry-investments.com>
+
+	* readfile.c (read_file_to_buffer): Do not waste a byte at the end of
+	a string.
+	* rwarray.c (read_value): Ditto.
+	* rwarray0.c (read_value): Ditto.
+
 2014-11-23         Arnold D. Robbins     <arnold@skeeve.com>
 
 	* inplace.c (do_inplace_begin): Jump through hoops to silence
diff --git a/extension/readfile.c b/extension/readfile.c
index d4b4aef9..7673589f 100644
--- a/extension/readfile.c
+++ b/extension/readfile.c
@@ -82,8 +82,8 @@ read_file_to_buffer(int fd, const struct stat *sbuf)
 		goto done;
 	}
 
-	emalloc(text, char *, sbuf->st_size + 2, "do_readfile");
-	memset(text, '\0', sbuf->st_size + 2);
+	emalloc(text, char *, sbuf->st_size + 1, "do_readfile");
+	memset(text, '\0', sbuf->st_size + 1);
 
 	if ((ret = read(fd, text, sbuf->st_size)) != sbuf->st_size) {
 		update_ERRNO_int(errno);
diff --git a/extension/rwarray.c b/extension/rwarray.c
index aa05a0d5..5bead973 100644
--- a/extension/rwarray.c
+++ b/extension/rwarray.c
@@ -468,8 +468,8 @@ read_value(FILE *fp, awk_value_t *value)
 		len = ntohl(len);
 		value->val_type = AWK_STRING;
 		value->str_value.len = len;
-		value->str_value.str = gawk_malloc(len + 2);
-		memset(value->str_value.str, '\0', len + 2);
+		value->str_value.str = gawk_malloc(len + 1);
+		memset(value->str_value.str, '\0', len + 1);
 
 		if (fread(value->str_value.str, 1, len, fp) != (ssize_t) len) {
 			gawk_free(value->str_value.str);
diff --git a/extension/rwarray0.c b/extension/rwarray0.c
index e2de3cf5..ec3663cc 100644
--- a/extension/rwarray0.c
+++ b/extension/rwarray0.c
@@ -452,8 +452,8 @@ read_value(int fd, awk_value_t *value)
 		len = ntohl(len);
 		value->val_type = AWK_STRING;
 		value->str_value.len = len;
-		value->str_value.str = malloc(len + 2);
-		memset(value->str_value.str, '\0', len + 2);
+		value->str_value.str = malloc(len + 1);
+		memset(value->str_value.str, '\0', len + 1);
 
 		if (read(fd, value->str_value.str, len) != (ssize_t) len) {
 			free(value->str_value.str);
diff --git a/field.c b/field.c
index 6a7c6b1d..c23f0b10 100644
--- a/field.c
+++ b/field.c
@@ -163,7 +163,7 @@ rebuild_record()
 	tlen += (NF - 1) * OFSlen;
 	if ((long) tlen < 0)
 		tlen = 0;
-	emalloc(ops, char *, tlen + 2, "rebuild_record");
+	emalloc(ops, char *, tlen + 1, "rebuild_record");
 	cops = ops;
 	ops[0] = '\0';
 	for (i = 1;  i <= NF; i++) {
diff --git a/gawkapi.h b/gawkapi.h
index d8215450..b9dc2c0b 100644
--- a/gawkapi.h
+++ b/gawkapi.h
@@ -791,7 +791,7 @@ r_make_string(const gawk_api_t *api,	/* needed for emalloc */
 	result->str_value.len = length;
 
 	if (duplicate) {
-		emalloc(cp, char *, length + 2, "r_make_string");
+		emalloc(cp, char *, length + 1, "r_make_string");
 		memcpy(cp, string, length);
 		cp[length] = '\0';
 		result->str_value.str = cp;
diff --git a/interpret.h b/interpret.h
index 3a9cab37..665f6faa 100644
--- a/interpret.h
+++ b/interpret.h
@@ -709,7 +709,7 @@ mod:
 			if (t1 != t2 && t1->valref == 1 && (t1->flags & MPFN) == 0) {
 				size_t nlen = t1->stlen + t2->stlen;
 
-				erealloc(t1->stptr, char *, nlen + 2, "r_interpret");
+				erealloc(t1->stptr, char *, nlen + 1, "r_interpret");
 				memcpy(t1->stptr + t1->stlen, t2->stptr, t2->stlen);
 				t1->stlen = nlen;
 				t1->stptr[nlen] = '\0';
@@ -719,7 +719,7 @@ mod:
 					size_t wlen = t1->wstlen + t2->wstlen;
 
 					erealloc(t1->wstptr, wchar_t *,
-							sizeof(wchar_t) * (wlen + 2), "r_interpret");
+							sizeof(wchar_t) * (wlen + 1), "r_interpret");
 					memcpy(t1->wstptr + t1->wstlen, t2->wstptr, t2->wstlen);
 					t1->wstlen = wlen;
 					t1->wstptr[wlen] = L'\0';
@@ -730,9 +730,10 @@ mod:
 				size_t nlen = t1->stlen + t2->stlen;  
 				char *p;
 
-				emalloc(p, char *, nlen + 2, "r_interpret");
+				emalloc(p, char *, nlen + 1, "r_interpret");
 				memcpy(p, t1->stptr, t1->stlen);
 				memcpy(p + t1->stlen, t2->stptr, t2->stlen);
+				/* N.B. No NUL-termination required, since make_str_node will do it. */
 				unref(*lhs);
 				t1 = *lhs = make_str_node(p, nlen, ALREADY_MALLOCED); 
 			}
diff --git a/node.c b/node.c
index 507d0650..869e4cbc 100644
--- a/node.c
+++ b/node.c
@@ -252,7 +252,7 @@ r_format_val(const char *format, int index, NODE *s)
 	}
 	if (s->stptr != NULL)
 		efree(s->stptr);
-	emalloc(s->stptr, char *, s->stlen + 2, "format_val");
+	emalloc(s->stptr, char *, s->stlen + 1, "format_val");
 	memcpy(s->stptr, sp, s->stlen + 1);
 no_malloc:
 	s->flags |= STRCUR;
@@ -290,12 +290,12 @@ r_dupnode(NODE *n)
 	r->wstlen = 0;
 
 	if ((n->flags & STRCUR) != 0) {
-		emalloc(r->stptr, char *, n->stlen + 2, "r_dupnode");
+		emalloc(r->stptr, char *, n->stlen + 1, "r_dupnode");
 		memcpy(r->stptr, n->stptr, n->stlen);
 		r->stptr[n->stlen] = '\0';
 		if ((n->flags & WSTRCUR) != 0) {
 			r->wstlen = n->wstlen;
-			emalloc(r->wstptr, wchar_t *, sizeof(wchar_t) * (n->wstlen + 2), "r_dupnode");
+			emalloc(r->wstptr, wchar_t *, sizeof(wchar_t) * (n->wstlen + 1), "r_dupnode");
 			memcpy(r->wstptr, n->wstptr, n->wstlen * sizeof(wchar_t));
 			r->wstptr[n->wstlen] = L'\0';
 			r->flags |= WSTRCUR;
@@ -368,7 +368,7 @@ make_str_node(const char *s, size_t len, int flags)
 	if ((flags & ALREADY_MALLOCED) != 0)
 		r->stptr = (char *) s;
 	else {
-		emalloc(r->stptr, char *, len + 2, "make_str_node");
+		emalloc(r->stptr, char *, len + 1, "make_str_node");
 		memcpy(r->stptr, s, len);
 	}
 	r->stptr[len] = '\0';
@@ -672,7 +672,7 @@ str2wstr(NODE *n, size_t **ptr)
 	 * realloc the wide string down in size.
 	 */
 
-	emalloc(n->wstptr, wchar_t *, sizeof(wchar_t) * (n->stlen + 2), "str2wstr");
+	emalloc(n->wstptr, wchar_t *, sizeof(wchar_t) * (n->stlen + 1), "str2wstr");
 	wsp = n->wstptr;
 
 	/*
@@ -746,7 +746,7 @@ str2wstr(NODE *n, size_t **ptr)
 	n->flags |= WSTRCUR;
 #define ARBITRARY_AMOUNT_TO_GIVE_BACK 100
 	if (n->stlen - n->wstlen > ARBITRARY_AMOUNT_TO_GIVE_BACK)
-		erealloc(n->wstptr, wchar_t *, sizeof(wchar_t) * (n->wstlen + 2), "str2wstr");
+		erealloc(n->wstptr, wchar_t *, sizeof(wchar_t) * (n->wstlen + 1), "str2wstr");
 
 	return n;
 }
@@ -773,7 +773,7 @@ wstr2str(NODE *n)
 	memset(& mbs, 0, sizeof(mbs));
 
 	length = n->wstlen;
-	emalloc(newval, char *, (length * gawk_mb_cur_max) + 2, "wstr2str");
+	emalloc(newval, char *, (length * gawk_mb_cur_max) + 1, "wstr2str");
 
 	wp = n->wstptr;
 	for (cp = newval; length > 0; length--) {
diff --git a/re.c b/re.c
index edb5bc48..98baf5ca 100644
--- a/re.c
+++ b/re.c
@@ -75,10 +75,10 @@ make_regexp(const char *s, size_t len, bool ignorecase, bool dfa, bool canfatal)
 	 * from that. 
 	 */
 	if (buf == NULL) {
-		emalloc(buf, char *, len + 2, "make_regexp");
+		emalloc(buf, char *, len + 1, "make_regexp");
 		buflen = len;
 	} else if (len > buflen) {
-		erealloc(buf, char *, len + 2, "make_regexp");
+		erealloc(buf, char *, len + 1, "make_regexp");
 		buflen = len;
 	}
 	dest = buf;