1 files changed, 127 insertions, 103 deletions

diff --git a/ext/mbstring/php_unicode.c b/ext/mbstring/php_unicode.c
index 8b6a52156a..208c10319e 100644
--- a/ext/mbstring/php_unicode.c
+++ b/ext/mbstring/php_unicode.c
@@ -43,22 +43,10 @@
 #include "mbstring.h"
 #include "php_unicode.h"
 #include "unicode_data.h"
+#include "libmbfl/mbfl/mbfilter_wchar.h"
 
 ZEND_EXTERN_MODULE_GLOBALS(mbstring)
 
-/*
- * A simple array of 32-bit masks for lookup.
- */
-static unsigned long masks32[32] = {
-    0x00000001, 0x00000002, 0x00000004, 0x00000008, 0x00000010, 0x00000020,
-    0x00000040, 0x00000080, 0x00000100, 0x00000200, 0x00000400, 0x00000800,
-    0x00001000, 0x00002000, 0x00004000, 0x00008000, 0x00010000, 0x00020000,
-    0x00040000, 0x00080000, 0x00100000, 0x00200000, 0x00400000, 0x00800000,
-    0x01000000, 0x02000000, 0x04000000, 0x08000000, 0x10000000, 0x20000000,
-    0x40000000, 0x80000000
-};
-
-
 static int prop_lookup(unsigned long code, unsigned long n)
 {
 	long l, r, m;
@@ -98,25 +86,31 @@ static int prop_lookup(unsigned long code, unsigned long n)
 
 }
 
-MBSTRING_API int php_unicode_is_prop(unsigned long code, unsigned long mask1,
-		unsigned long mask2)
+MBSTRING_API int php_unicode_is_prop1(unsigned long code, int prop)
 {
-	unsigned long i;
+	return prop_lookup(code, prop);
+}
 
-	if (mask1 == 0 && mask2 == 0)
-		return 0;
+MBSTRING_API int php_unicode_is_prop(unsigned long code, ...)
+{
+	int result = 0;
+	va_list va;
+	va_start(va, code);
 
-	for (i = 0; mask1 && i < 32; i++) {
-		if ((mask1 & masks32[i]) && prop_lookup(code, i))
-			return 1;
-	}
+	while (1) {
+		int prop = va_arg(va, int);
+		if (prop < 0) {
+			break;
+		}
 
-	for (i = 32; mask2 && i < _ucprop_size; i++) {
-		if ((mask2 & masks32[i & 31]) && prop_lookup(code, i))
-			return 1;
+		if (prop_lookup(code, prop)) {
+			result = 1;
+			break;
+		}
 	}
 
-	return 0;
+	va_end(va);
+	return result;
 }
 
 static unsigned long case_lookup(unsigned long code, long l, long r, int field)
@@ -144,27 +138,22 @@ static unsigned long case_lookup(unsigned long code, long l, long r, int field)
 	return code;
 }
 
-MBSTRING_API unsigned long php_turkish_toupper(unsigned long code, long l, long r, int field)
-{
-	if (code == 0x0069L) {
-		return 0x0130L;
-	}
-	return case_lookup(code, l, r, field);
-}
-
-MBSTRING_API unsigned long php_turkish_tolower(unsigned long code, long l, long r, int field)
-{
-	if (code == 0x0049L) {
-		return 0x0131L;
-	}
-	return case_lookup(code, l, r, field);
-}
-
 MBSTRING_API unsigned long php_unicode_toupper(unsigned long code, enum mbfl_no_encoding enc)
 {
 	int field;
 	long l, r;
 
+	if (code < 0x80) {
+		/* Fast path for ASCII */
+		if (code >= 0x61 && code <= 0x7A) {
+			if (enc == mbfl_no_encoding_8859_9 && code == 0x0069L) {
+				return 0x0130L;
+			}
+			return code - 0x20;
+		}
+		return code;
+	}
+
 	if (php_unicode_is_upper(code))
 		return code;
 
@@ -175,11 +164,6 @@ MBSTRING_API unsigned long php_unicode_toupper(unsigned long code, enum mbfl_no_
 		field = 2;
 		l = _uccase_len[0];
 		r = (l + _uccase_len[1]) - 3;
-
-		if (enc == mbfl_no_encoding_8859_9) {
-			return php_turkish_toupper(code, l, r, field);
-		}
-
 	} else {
 		/*
 		 * The character is title case.
@@ -196,6 +180,17 @@ MBSTRING_API unsigned long php_unicode_tolower(unsigned long code, enum mbfl_no_
 	int field;
 	long l, r;
 
+	if (code < 0x80) {
+		/* Fast path for ASCII */
+		if (code >= 0x41 && code <= 0x5A) {
+			if (enc == mbfl_no_encoding_8859_9 && code == 0x0049L) {
+				return 0x0131L;
+			}
+			return code + 0x20;
+		}
+		return code;
+	}
+
 	if (php_unicode_is_lower(code))
 		return code;
 
@@ -206,11 +201,6 @@ MBSTRING_API unsigned long php_unicode_tolower(unsigned long code, enum mbfl_no_
 		field = 1;
 		l = 0;
 		r = _uccase_len[0] - 3;
-
-		if (enc == mbfl_no_encoding_8859_9) {
-			return php_turkish_tolower(code, l, r, field);
-		}
-
 	} else {
 		/*
 		 * The character is title case.
@@ -267,71 +257,105 @@ MBSTRING_API unsigned long php_unicode_totitle(unsigned long code, enum mbfl_no_
 	((unsigned char*)(ptr))[3] = (v    ) & 0xff;\
 }
 
-MBSTRING_API char *php_unicode_convert_case(int case_mode, const char *srcstr, size_t srclen, size_t *ret_len,
-		const char *src_encoding)
-{
-	char *unicode, *newstr;
-	size_t unicode_len;
-	unsigned char *unicode_ptr;
-	size_t i;
-	enum mbfl_no_encoding _src_encoding = mbfl_name2no_encoding(src_encoding);
-
-	if (_src_encoding == mbfl_no_encoding_invalid) {
-		php_error_docref(NULL, E_WARNING, "Unknown encoding \"%s\"", src_encoding);
-		return NULL;
-	}
-
-	unicode = php_mb_convert_encoding(srcstr, srclen, "UCS-4BE", src_encoding, &unicode_len);
-	if (unicode == NULL)
-		return NULL;
-
-	unicode_ptr = (unsigned char *)unicode;
+struct convert_case_data {
+	mbfl_convert_filter *next_filter;
+	enum mbfl_no_encoding no_encoding;
+	int case_mode;
+	int title_mode;
+};
 
-	switch(case_mode) {
+static int convert_case_filter(int c, void *void_data)
+{
+	struct convert_case_data *data = (struct convert_case_data *) void_data;
+	switch (data->case_mode) {
 		case PHP_UNICODE_CASE_UPPER:
-			for (i = 0; i < unicode_len; i+=4) {
-				UINT32_TO_BE_ARY(&unicode_ptr[i],
-					php_unicode_toupper(BE_ARY_TO_UINT32(&unicode_ptr[i]), _src_encoding));
-			}
+			c = php_unicode_toupper(c, data->no_encoding);
 			break;
 
 		case PHP_UNICODE_CASE_LOWER:
-			for (i = 0; i < unicode_len; i+=4) {
-				UINT32_TO_BE_ARY(&unicode_ptr[i],
-					php_unicode_tolower(BE_ARY_TO_UINT32(&unicode_ptr[i]), _src_encoding));
-			}
+			c = php_unicode_tolower(c, data->no_encoding);
 			break;
 
-		case PHP_UNICODE_CASE_TITLE: {
-			int mode = 0;
-
-			for (i = 0; i < unicode_len; i+=4) {
-				int res = php_unicode_is_prop(
-					BE_ARY_TO_UINT32(&unicode_ptr[i]),
-					UC_MN|UC_ME|UC_CF|UC_LM|UC_SK|UC_LU|UC_LL|UC_LT|UC_PO|UC_OS, 0);
-				if (mode) {
-					if (res) {
-						UINT32_TO_BE_ARY(&unicode_ptr[i],
-							php_unicode_tolower(BE_ARY_TO_UINT32(&unicode_ptr[i]), _src_encoding));
-					} else {
-						mode = 0;
-					}
+		case PHP_UNICODE_CASE_TITLE:
+		{
+			int res = php_unicode_is_prop(c,
+				UC_MN, UC_ME, UC_CF, UC_LM, UC_SK, UC_LU, UC_LL, UC_LT, UC_PO, UC_OS, -1);
+			if (data->title_mode) {
+				if (res) {
+					c = php_unicode_tolower(c, data->no_encoding);
 				} else {
-					if (res) {
-						mode = 1;
-						UINT32_TO_BE_ARY(&unicode_ptr[i],
-							php_unicode_totitle(BE_ARY_TO_UINT32(&unicode_ptr[i]), _src_encoding));
-					}
+					data->title_mode = 0;
+				}
+			} else {
+				if (res) {
+					data->title_mode = 1;
+					c = php_unicode_totitle(c, data->no_encoding);
 				}
 			}
-		} break;
+			break;
+		}
+	}
+	return (*data->next_filter->filter_function)(c, data->next_filter);
+}
 
+MBSTRING_API char *php_unicode_convert_case(
+		int case_mode, const char *srcstr, size_t srclen, size_t *ret_len,
+		const mbfl_encoding *src_encoding)
+{
+	struct convert_case_data data;
+	mbfl_convert_filter *from_wchar, *to_wchar;
+	mbfl_string result, *result_ptr;
+
+	mbfl_memory_device device;
+	mbfl_memory_device_init(&device, srclen + 1, 0);
+
+	/* encoding -> wchar filter */
+	to_wchar = mbfl_convert_filter_new(src_encoding,
+			&mbfl_encoding_wchar, convert_case_filter, NULL, &data);
+	if (to_wchar == NULL) {
+		mbfl_memory_device_clear(&device);
+		return NULL;
 	}
 
-	newstr = php_mb_convert_encoding(unicode, unicode_len, src_encoding, "UCS-4BE", ret_len);
-	efree(unicode);
+	/* wchar -> encoding filter */
+	from_wchar = mbfl_convert_filter_new(
+			&mbfl_encoding_wchar, src_encoding,
+			mbfl_memory_device_output, NULL, &device);
+	if (from_wchar == NULL) {
+		mbfl_convert_filter_delete(to_wchar);
+		mbfl_memory_device_clear(&device);
+		return NULL;
+	}
+
+	data.next_filter = from_wchar;
+	data.no_encoding = src_encoding->no_encoding;
+	data.case_mode = case_mode;
+	data.title_mode = 0;
+
+	{
+		/* feed data */
+		const unsigned char *p = (const unsigned char *) srcstr;
+		size_t n = srclen;
+		while (n > 0) {
+			if ((*to_wchar->filter_function)(*p++, to_wchar) < 0) {
+				break;
+			}
+			n--;
+		}
+	}
+
+	mbfl_convert_filter_flush(to_wchar);
+	mbfl_convert_filter_flush(from_wchar);
+	result_ptr = mbfl_memory_device_result(&device, &result);
+	mbfl_convert_filter_delete(to_wchar);
+	mbfl_convert_filter_delete(from_wchar);
+
+	if (!result_ptr) {
+		return NULL;
+	}
 
-	return newstr;
+	*ret_len = result.len;
+	return (char *) result.val;
 }