Cleanup (avoid reallocatios and side effects in php_strip_tags)

Zend/zend_operators.c

@@ -2417,6 +2417,34 @@ ZEND_API void ZEND_FASTCALL zend_str_tolower(char *str, size_t length) /* {{{ */
 }
 /* }}} */
 
+ZEND_API char* ZEND_FASTCALL zend_str_tolower_dup_ex(const char *source, size_t length) /* {{{ */
+{
+	register const unsigned char *p = (const unsigned char*)source;
+	register const unsigned char *end = p + length;
+
+	while (p < end) {
+		if (*p != zend_tolower_ascii(*p)) {
+			char *res = (char*)emalloc(length + 1);
+			register unsigned char *r;
+
+			if (p != (const unsigned char*)source) {
+				memcpy(res, source, p - (const unsigned char*)source);
+			}
+			r = (unsigned char*)p + (res - source);
+			while (p < end) {
+				*r = zend_tolower_ascii(*p);
+				p++;
+				r++;
+			}
+			*r = '\0';
+			return res;
+		}
+		p++;
+	}
+	return NULL;
+}
+/* }}} */
+
 ZEND_API zend_string* ZEND_FASTCALL zend_string_tolower(zend_string *str) /* {{{ */
 {
 	register unsigned char *p = (unsigned char*)ZSTR_VAL(str);

Zend/zend_operators.h

@@ -340,6 +340,7 @@ ZEND_API int string_locale_compare_function(zval *result, zval *op1, zval *op2);
 ZEND_API void         ZEND_FASTCALL zend_str_tolower(char *str, size_t length);
 ZEND_API char*        ZEND_FASTCALL zend_str_tolower_copy(char *dest, const char *source, size_t length);
 ZEND_API char*        ZEND_FASTCALL zend_str_tolower_dup(const char *source, size_t length);
+ZEND_API char*        ZEND_FASTCALL zend_str_tolower_dup_ex(const char *source, size_t length);
 ZEND_API zend_string* ZEND_FASTCALL zend_string_tolower(zend_string *str);
 
 ZEND_API int ZEND_FASTCALL zend_binary_zval_strcmp(zval *s1, zval *s2);

ext/standard/file.c

@@ -1084,11 +1084,10 @@ PHPAPI PHP_FUNCTION(fgetss)
 	size_t actual_len, retval_len;
 	char *buf = NULL, *retval;
 	php_stream *stream;
-	zend_string *allowed = NULL;
 	char *allowed_tags=NULL;
 	size_t allowed_tags_len=0;
 
-	if (zend_parse_parameters(ZEND_NUM_ARGS(), "r|lS", &fd, &bytes, &allowed) == FAILURE) {
+	if (zend_parse_parameters(ZEND_NUM_ARGS(), "r|ls", &fd, &bytes, &allowed_tags, &allowed_tags_len) == FAILURE) {
 		RETURN_FALSE;
 	}
 
@@ -1113,24 +1112,8 @@ PHPAPI PHP_FUNCTION(fgetss)
 		RETURN_FALSE;
 	}
 
-	if (allowed != NULL) {
-// TODO: reimplement to avoid reallocation ???
-		if (ZSTR_IS_INTERNED(allowed)) {
-			allowed_tags = estrndup(ZSTR_VAL(allowed), ZSTR_LEN(allowed));
-			allowed_tags_len = ZSTR_LEN(allowed);
-		} else {
-			allowed_tags = ZSTR_VAL(allowed);
-			allowed_tags_len = ZSTR_LEN(allowed);
-		}
-	}
-
 	retval_len = php_strip_tags(retval, actual_len, &stream->fgetss_state, allowed_tags, allowed_tags_len);
 
-// TODO: reimplement to avoid reallocation ???
-	if (allowed && ZSTR_IS_INTERNED(allowed)) {
-		efree(allowed_tags);
-	}
-
 	// TODO: avoid reallocation ???
 	RETVAL_STRINGL(retval, retval_len);
 	efree(retval);

ext/standard/filters.c

@@ -217,7 +217,7 @@ static php_stream_filter_status_t strfilter_strip_tags_filter(
 		bucket = php_stream_bucket_make_writeable(buckets_in->head);
 		consumed = bucket->buflen;
 
-		bucket->buflen = php_strip_tags(bucket->buf, bucket->buflen, &(inst->state), (char *)inst->allowed_tags, inst->allowed_tags_len);
+		bucket->buflen = php_strip_tags(bucket->buf, bucket->buflen, &(inst->state), inst->allowed_tags, inst->allowed_tags_len);
 
 		php_stream_bucket_append(buckets_out, bucket);
 	}

ext/standard/php_string.h

@@ -133,8 +133,8 @@ PHPAPI char *php_stristr(char *s, char *t, size_t s_len, size_t t_len);
 PHPAPI zend_string *php_str_to_str(char *haystack, size_t length, char *needle,
 		size_t needle_len, char *str, size_t str_len);
 PHPAPI zend_string *php_trim(zend_string *str, char *what, size_t what_len, int mode);
-PHPAPI size_t php_strip_tags(char *rbuf, size_t len, int *state, char *allow, size_t allow_len);
+PHPAPI size_t php_strip_tags(char *rbuf, size_t len, int *state, const char *allow, size_t allow_len);
-PHPAPI size_t php_strip_tags_ex(char *rbuf, size_t len, int *stateptr, char *allow, size_t allow_len, zend_bool allow_tag_spaces);
+PHPAPI size_t php_strip_tags_ex(char *rbuf, size_t len, int *stateptr, const char *allow, size_t allow_len, zend_bool allow_tag_spaces);
 PHPAPI void php_implode(const zend_string *delim, zval *arr, zval *return_value);
 PHPAPI void php_explode(const zend_string *delim, zend_string *str, zval *return_value, zend_long limit);
 

ext/standard/string.c

@@ -115,7 +115,7 @@ void register_string_constants(INIT_FUNC_ARGS)
 }
 /* }}} */
 
-int php_tag_find(char *tag, size_t len, char *set);
+int php_tag_find(char *tag, size_t len, const char *set);
 
 /* this is read-only, so it's ok */
 static char hexconvtab[] = "0123456789abcdef";
@@ -4451,26 +4451,15 @@ PHP_FUNCTION(strip_tags)
 	}
 
 	/* To maintain a certain BC, we allow anything for the second parameter and return original string */
-	if (allow != NULL) {
+	if (allow) {
-		convert_to_string_ex(allow);
+		convert_to_string(allow);
-// TODO: reimplement to avoid reallocation ???
+		allowed_tags = Z_STRVAL_P(allow);
-		if (!Z_REFCOUNTED_P(allow)) {
+		allowed_tags_len = Z_STRLEN_P(allow);
-			allowed_tags = estrndup(Z_STRVAL_P(allow), Z_STRLEN_P(allow));
-			allowed_tags_len = Z_STRLEN_P(allow);
-		} else {
-			allowed_tags = Z_STRVAL_P(allow);
-			allowed_tags_len = Z_STRLEN_P(allow);
-		}
 	}
 
 	buf = zend_string_init(ZSTR_VAL(str), ZSTR_LEN(str), 0);
 	ZSTR_LEN(buf) = php_strip_tags_ex(ZSTR_VAL(buf), ZSTR_LEN(str), NULL, allowed_tags, allowed_tags_len, 0);
-
+	RETURN_NEW_STR(buf);
-// TODO: reimplement to avoid reallocation ???
-	if (allow && !Z_REFCOUNTED_P(allow)) {
-		efree(allowed_tags);
-	}
-	RETURN_STR(buf);
 }
 /* }}} */
 
@@ -4608,7 +4597,7 @@ PHP_FUNCTION(parse_str)
  * 0 start tag
  * 1 first non-whitespace char seen
  */
-int php_tag_find(char *tag, size_t len, char *set) {
+int php_tag_find(char *tag, size_t len, const char *set) {
 	char c, *n, *t;
 	int state=0, done=0;
 	char *norm;
@@ -4663,7 +4652,7 @@ int php_tag_find(char *tag, size_t len, char *set) {
 }
 /* }}} */
 
-PHPAPI size_t php_strip_tags(char *rbuf, size_t len, int *stateptr, char *allow, size_t allow_len) /* {{{ */
+PHPAPI size_t php_strip_tags(char *rbuf, size_t len, int *stateptr, const char *allow, size_t allow_len) /* {{{ */
 {
 	return php_strip_tags_ex(rbuf, len, stateptr, allow, allow_len, 0);
 }
@@ -4689,13 +4678,14 @@ PHPAPI size_t php_strip_tags(char *rbuf, size_t len, int *stateptr, char *allow,
 	swm: Added ability to strip <?xml tags without assuming it PHP
 	code.
 */
-PHPAPI size_t php_strip_tags_ex(char *rbuf, size_t len, int *stateptr, char *allow, size_t allow_len, zend_bool allow_tag_spaces)
+PHPAPI size_t php_strip_tags_ex(char *rbuf, size_t len, int *stateptr, const char *allow, size_t allow_len, zend_bool allow_tag_spaces)
 {
 	char *tbuf, *buf, *p, *tp, *rp, c, lc;
 	int br, depth=0, in_q = 0;
 	int state = 0;
 	size_t pos, i = 0;
 	char *allow_free = NULL;
+	const char *allow_actual;
 
 	if (stateptr)
 		state = *stateptr;
@@ -4707,12 +4697,8 @@ PHPAPI size_t php_strip_tags_ex(char *rbuf, size_t len, int *stateptr, char *all
 	rp = rbuf;
 	br = 0;
 	if (allow) {
-//???		if (ZSTR_IS_INTERNED(allow)) {
+		allow_free = zend_str_tolower_dup_ex(allow, allow_len);
-//???			allow_free = allow = zend_str_tolower_dup(allow, allow_len);
+		allow_actual = allow_free ? allow_free : allow;
-//???		} else {
-			allow_free = NULL;
-			php_strtolower(allow, allow_len);
-//???		}
 		tbuf = emalloc(PHP_TAG_BUF_SIZE + 1);
 		tp = tbuf;
 	} else {
@@ -4804,7 +4790,7 @@ PHPAPI size_t php_strip_tags_ex(char *rbuf, size_t len, int *stateptr, char *all
 							}
 							*(tp++) = '>';
 							*tp='\0';
-							if (php_tag_find(tbuf, tp-tbuf, allow)) {
+							if (php_tag_find(tbuf, tp-tbuf, allow_actual)) {
 								memcpy(rp, tbuf, tp-tbuf);
 								rp += tp-tbuf;
 							}