php · realityking · Jun 1, 2013
@@ -69,10 +69,8 @@ PHP_FUNCTION(grapheme_strlen)
 		RETURN_FALSE;
 	}
 
-	ret_len = grapheme_ascii_check(string, string_len);
-
-	if ( ret_len >= 0 )
-		RETURN_LONG(ret_len);
+	if (string_is_ascii(string, string_len))
+		RETURN_LONG(string_len);
 
 	/* convert the string to UTF-16. */
 	status = U_ZERO_ERROR;
@@ -154,7 +152,7 @@ PHP_FUNCTION(grapheme_strpos)
 	}
 
 	/* if it is there, and if the haystack is ascii, we are all done */
-	if ( grapheme_ascii_check(haystack, haystack_len) >= 0 ) {
+	if ( string_is_ascii(haystack, haystack_len) ) {
 
 		RETURN_LONG(found - haystack);
 	}
@@ -211,7 +209,7 @@ PHP_FUNCTION(grapheme_stripos)
 	}
 
 
-	is_ascii = ( grapheme_ascii_check(haystack, haystack_len) >= 0 );
+	is_ascii = ( string_is_ascii(haystack, haystack_len) );
 
 	if ( is_ascii ) {
 		needle_dup = (unsigned char *)estrndup((char *)needle, needle_len);
@@ -229,7 +227,7 @@ PHP_FUNCTION(grapheme_stripos)
 		}
 
 		/* if needle was ascii too, we are all done, otherwise we need to try using Unicode to see what we get */
-		if ( grapheme_ascii_check(needle, needle_len) >= 0 ) {
+		if ( string_is_ascii(needle, needle_len) ) {
 			RETURN_FALSE;
 		}
 	}
@@ -284,7 +282,7 @@ PHP_FUNCTION(grapheme_strrpos)
 		RETURN_FALSE;
 	}
 
-	is_ascii = grapheme_ascii_check(haystack, haystack_len) >= 0;
+	is_ascii = string_is_ascii(haystack, haystack_len);
 
 	if ( is_ascii ) {
 
@@ -297,7 +295,7 @@ PHP_FUNCTION(grapheme_strrpos)
 
 		/* if the needle was ascii too, we are done */
 
-		if (  grapheme_ascii_check(needle, needle_len) >= 0 ) {
+		if (  string_is_ascii(needle, needle_len) ) {
 			RETURN_FALSE;
 		}
 
@@ -354,7 +352,7 @@ PHP_FUNCTION(grapheme_strripos)
 		RETURN_FALSE;
 	}
 
-	is_ascii = grapheme_ascii_check(haystack, haystack_len) >= 0;
+	is_ascii = string_is_ascii(haystack, haystack_len);
 
 	if ( is_ascii ) {
 		unsigned char *needle_dup, *haystack_dup;
@@ -375,7 +373,7 @@ PHP_FUNCTION(grapheme_strripos)
 
 		/* if the needle was ascii too, we are done */
 
-		if (  grapheme_ascii_check(needle, needle_len) >= 0 ) {
+		if (  string_is_ascii(needle, needle_len) ) {
 			RETURN_FALSE;
 		}
 
@@ -430,7 +428,7 @@ PHP_FUNCTION(grapheme_substr)
 
 	/* the offset is 'grapheme count offset' so it still might be invalid - we'll check it later */
 
-	if ( grapheme_ascii_check(str, str_len) >= 0 ) {
+	if ( string_is_ascii(str, str_len) ) {
 		grapheme_substr_ascii((char *)str, str_len, start, length, ZEND_NUM_ARGS(), (char **) &sub_str, &sub_str_len);
 
 		if ( NULL == sub_str ) {
@@ -630,7 +628,7 @@ static void strstr_common_handler(INTERNAL_FUNCTION_PARAMETERS, int f_ignore_cas
 		}
 
 		/* if it is there, and if the haystack is ascii, we are all done */
-		if ( grapheme_ascii_check(haystack, haystack_len) >= 0 ) {
+		if ( string_is_ascii(haystack, haystack_len) ) {
 			size_t found_offset = found - haystack;
 
 			if (part) {
@@ -874,7 +872,7 @@ PHP_FUNCTION(grapheme_extract)
 		(size + 1 because the size-th character might be the beginning of a grapheme cluster)
 	 */
 
-	if ( -1 != grapheme_ascii_check(pstr, size + 1 < str_len ? size + 1 : str_len ) ) {
+	if ( string_is_ascii(pstr, size + 1 < str_len ? size + 1 : str_len ) ) {
         long nsize = ( size < str_len ? size : str_len );
 		if ( NULL != next ) {
 			ZVAL_LONG(next, start+nsize);

@@ -379,20 +379,6 @@ grapheme_strpos_utf16(unsigned char *haystack, int32_t haystack_len, unsigned ch
 
 /* }}} */
 
-/* {{{ grapheme_ascii_check: ASCII check */
-int grapheme_ascii_check(const unsigned char *day, int32_t len)
-{
-	int ret_len = len;
-	while ( len-- ) {
-	if ( *day++ > 0x7f )
-		return -1;
-	}
-
-	return ret_len;
-}
-
-/* }}} */
-
 /* {{{ grapheme_split_string: find and optionally return grapheme boundaries */
 int grapheme_split_string(const UChar *text, int32_t text_length, int boundary_array[], int boundary_array_len TSRMLS_DC )
 {

@@ -32,8 +32,6 @@ grapheme_strrpos_utf16(unsigned char *haystack, int32_t haystack_len, unsigned c
 int
 grapheme_strpos_utf16(unsigned char *haystack, int32_t haystack_len, unsigned char*needle, int32_t needle_len, int32_t offset, int *puchar_pos, int f_ignore_case TSRMLS_DC);
 
-int grapheme_ascii_check(const unsigned char *day, int32_t len);
-
 int grapheme_split_string(const UChar *text, int32_t text_length, int boundary_array[], int boundary_array_len TSRMLS_DC );
 
 int32_t

@@ -2494,6 +2494,10 @@ ZEND_BEGIN_ARG_INFO_EX(arginfo_substr_compare, 0, 0, 3)
 	ZEND_ARG_INFO(0, length)
 	ZEND_ARG_INFO(0, case_sensitivity)
 ZEND_END_ARG_INFO()
+
+ZEND_BEGIN_ARG_INFO_EX(arginfo_str_is_ascii, ZEND_SEND_BY_VAL, ZEND_RETURN_VALUE, 1)
+	ZEND_ARG_INFO(0, str)
+ZEND_END_ARG_INFO()
 /* }}} */
 /* {{{ syslog.c */
 #ifdef HAVE_SYSLOG_H
@@ -2771,6 +2775,7 @@ const zend_function_entry basic_functions[] = { /* {{{ */
 	PHP_FE(str_split,														arginfo_str_split)
 	PHP_FE(strpbrk,															arginfo_strpbrk)
 	PHP_FE(substr_compare,													arginfo_substr_compare)
+	PHP_FE(str_is_ascii,													arginfo_str_is_ascii)
 
 #ifdef HAVE_STRCOLL
 	PHP_FE(strcoll,															arginfo_strcoll)

@@ -93,6 +93,7 @@ PHP_FUNCTION(str_word_count);
 PHP_FUNCTION(str_split);
 PHP_FUNCTION(strpbrk);
 PHP_FUNCTION(substr_compare);
+PHP_FUNCTION(str_is_ascii);
 #ifdef HAVE_STRCOLL
 PHP_FUNCTION(strcoll);
 #endif
@@ -147,6 +148,8 @@ PHPAPI int string_natural_compare_function_ex(zval *result, zval *op1, zval *op2
 PHPAPI int string_natural_compare_function(zval *result, zval *op1, zval *op2 TSRMLS_DC);
 PHPAPI int string_natural_case_compare_function(zval *result, zval *op1, zval *op2 TSRMLS_DC);
 
+PHPAPI zend_bool string_is_ascii(const unsigned char *str, size_t len);
+
 #ifndef HAVE_STRERROR
 PHPAPI char *php_strerror(int errnum);
 #define strerror php_strerror

@@ -5607,6 +5607,38 @@ PHP_FUNCTION(substr_compare)
 }
 /* }}} */
 
+/* {{{ string_is_ascii
+*/
+PHPAPI zend_bool string_is_ascii(const unsigned char *str, size_t str_len)
+{
+	while (str_len--) {
+		if (*str++ > 0x7f) {
+			return 0;
+		}
+	}
+
+	return 1;
+}
+/* }}} */
+
+/* {{{ proto bool str_is_ascii(string str)
+   Checks whether the given string contains only ASCII characters */
+PHP_FUNCTION(str_is_ascii)
+{
+	unsigned char *str = NULL;
+	int str_len = 0;
+	zend_bool ret = 0;
+
+	if (zend_parse_parameters(ZEND_NUM_ARGS() TSRMLS_CC, "s", &str, &str_len) == FAILURE) {
+		return;
+	}
+
+	ret = string_is_ascii(str, str_len);
+
+	RETURN_BOOL(ret);
+}
+/* }}} str_is_ascii */
+
 /*
  * Local variables:
  * tab-width: 4

@@ -0,0 +1,12 @@
+--TEST--
+str_is_ascii() function basic functionality
+--FILE--
+<?php
+var_dump(str_is_ascii(''));
+var_dump(str_is_ascii('ABC 123'));
+var_dump(str_is_ascii('Iñtërnâtiônàlizætiøn'));
+?>
+--EXPECT--
+bool(true)
+bool(true)
+bool(false)
-Original file line number
+Diff line change
@@ Expand Up @@
     int
     grapheme_strpos_utf16(unsigned char *haystack, int32_t haystack_len, unsigned char*needle, int32_t needle_len, int32_t offset, int *puchar_pos, int f_ignore_case TSRMLS_DC);
-    int grapheme_ascii_check(const unsigned char *day, int32_t len);
     int grapheme_split_string(const UChar *text, int32_t text_length, int boundary_array[], int boundary_array_len TSRMLS_DC );
     int32_t
@@ Expand Down @@