fixed errors of mbstring in regression tests.

12 files changed, 142 insertions, 84 deletions

diff --git a/ext/mbstring/mbstring.c b/ext/mbstring/mbstring.c
index c1e19b168c..bced5a83d5 100644
--- a/ext/mbstring/mbstring.c
+++ b/ext/mbstring/mbstring.c
@@ -82,8 +82,9 @@ static int php_mbstr_default_identify_list_size = sizeof(php_mbstr_default_ident
 
 static unsigned char third_and_rest_force_ref[] = { 3, BYREF_NONE, BYREF_NONE, BYREF_FORCE_REST };
 static unsigned char second_args_force_ref[] = { 2, BYREF_NONE, BYREF_FORCE };
+#if HAVE_MBREGEX
 static unsigned char third_argument_force_ref[] = { 3, BYREF_NONE, BYREF_NONE, BYREF_FORCE };
-
+#endif
 #if defined(MBSTR_ENC_TRANS)
 SAPI_POST_HANDLER_FUNC(php_mbstr_post_handler);
 
@@ -284,8 +285,13 @@ php_mbstring_parse_encoding_list(const char *value, int value_length, int **retu
 				}
 				p1 = p2 + 1;
 			} while (n < size && p2 != NULL);
-			*return_list = list;
-			*return_size = n;
+			if (n > 0){
+				*return_list = list;
+				*return_size = n;
+			} else {
+				efree(list);
+				*return_list = NULL;
+			}
 		}
 		efree(tmpstr);
 	}
@@ -340,8 +346,13 @@ php_mbstring_parse_encoding_array(zval *array, int **return_list, int *return_si
 				zend_hash_move_forward(target_hash);
 				i--;
 			}
-			*return_list = list;
-			*return_size = n;
+			if (n > 0) {
+				*return_list = list;
+				*return_size = n;
+			} else {
+				efree(list);
+				*return_list = NULL;
+			}
 		}
 	}
 
@@ -913,6 +924,7 @@ PHP_FUNCTION(mb_substitute_character)
 			RETVAL_LONG(MBSTRG(current_filter_illegal_substchar));
 		}
 	} else if (ZEND_NUM_ARGS() == 1 && zend_get_parameters_ex(1, &arg1) != FAILURE) {
+		RETVAL_TRUE;
 		switch (Z_TYPE_PP(arg1)) {
 		case IS_STRING:
 			if (strcasecmp("none", Z_STRVAL_PP(arg1)) == 0) {
@@ -921,17 +933,26 @@ PHP_FUNCTION(mb_substitute_character)
 				MBSTRG(current_filter_illegal_mode) = MBFL_OUTPUTFILTER_ILLEGAL_MODE_LONG;
 			} else {
 				convert_to_long_ex(arg1);
-				MBSTRG(current_filter_illegal_mode) = MBFL_OUTPUTFILTER_ILLEGAL_MODE_CHAR;
-				MBSTRG(current_filter_illegal_substchar) = Z_LVAL_PP(arg1);
+				if (Z_LVAL_PP(arg1)< 0xffff && Z_LVAL_PP(arg1)> 0x0) {
+					MBSTRG(current_filter_illegal_mode) = MBFL_OUTPUTFILTER_ILLEGAL_MODE_CHAR;
+					MBSTRG(current_filter_illegal_substchar) = Z_LVAL_PP(arg1);
+				} else {
+					php_error(E_WARNING, "unknown character.");
+					RETVAL_FALSE;					
+				}
 			}
 			break;
 		default:
 			convert_to_long_ex(arg1);
-			MBSTRG(current_filter_illegal_mode) = MBFL_OUTPUTFILTER_ILLEGAL_MODE_CHAR;
-			MBSTRG(current_filter_illegal_substchar) = Z_LVAL_PP(arg1);
+			if (Z_LVAL_PP(arg1)< 0xffff && Z_LVAL_PP(arg1)> 0x0) {
+				MBSTRG(current_filter_illegal_mode) = MBFL_OUTPUTFILTER_ILLEGAL_MODE_CHAR;
+				MBSTRG(current_filter_illegal_substchar) = Z_LVAL_PP(arg1);
+			} else {
+				php_error(E_WARNING, "unknown character.");
+				RETVAL_FALSE;
+			}
 			break;
 		}
-		RETVAL_TRUE;
 	} else {
 		WRONG_PARAM_COUNT;
 	}
@@ -1521,6 +1542,17 @@ PHP_FUNCTION(mb_strlen)
 	    n < 1 || n > 2) {
 		WRONG_PARAM_COUNT;
 	}
+	if (Z_TYPE_PP(arg1) == IS_ARRAY ||
+		Z_TYPE_PP(arg1) == IS_OBJECT) {
+		php_error(E_NOTICE, "arg1 is invalid.");
+		RETURN_FALSE;
+	}
+	if (( n ==2 && Z_TYPE_PP(arg2) == IS_ARRAY) ||
+		( n ==2 && Z_TYPE_PP(arg2) == IS_OBJECT)) {
+		php_error(E_NOTICE, "arg2 is invalid.");
+		RETURN_FALSE;
+	}
+
 	convert_to_string_ex(arg1);
 	mbfl_string_init(&string);
 	string.no_language = MBSTRG(current_language);
@@ -1553,7 +1585,7 @@ PHP_FUNCTION(mb_strlen)
 PHP_FUNCTION(mb_strpos)
 {
 	pval **arg1, **arg2, **arg3, **arg4;
-	int offset, n;
+	int offset, n, reverse = 0;
 	mbfl_string haystack, needle;
 
 	mbfl_string_init(&haystack);
@@ -1595,11 +1627,8 @@ PHP_FUNCTION(mb_strpos)
 
 	convert_to_string_ex(arg1);
 	convert_to_string_ex(arg2);
-	if (offset < 0) {
-		php_error(E_WARNING,"offset is minus value");
-		offset = 0;
-	}
-	if (offset > Z_STRLEN_PP(arg1)) {
+
+	if (offset < 0 || offset > Z_STRLEN_PP(arg1)) {
 		php_error(E_WARNING,"offset not contained in string");
 		RETURN_FALSE;
 	}
@@ -1612,10 +1641,26 @@ PHP_FUNCTION(mb_strpos)
 	needle.val = Z_STRVAL_PP(arg2);
 	needle.len = Z_STRLEN_PP(arg2);
 
-	n = mbfl_strpos(&haystack, &needle, offset, 0);
+	n = mbfl_strpos(&haystack, &needle, offset, reverse);
 	if (n >= 0) {
 		RETVAL_LONG(n);
 	} else {
+		switch (-n) {
+		case 1:
+			break;
+		case 2:
+			php_error(E_WARNING,"needle has not positive length.");
+			break;
+		case 4:
+			php_error(E_WARNING,"unknown encoding or conversion error.");
+			break;
+		case 8:
+			php_error(E_NOTICE,"argument is empty.");
+			break;
+		default:
+			php_error(E_WARNING,"unknown error in mb_strpos.");
+			break;			
+		}
 		RETVAL_FALSE;
 	}
 }
@@ -1945,9 +1990,19 @@ PHP_FUNCTION(mb_strimwidth)
 
 	convert_to_long_ex(arg2);
 	from = Z_LVAL_PP(arg2);
+	if (from < 0 || from > Z_STRLEN_PP(arg1)) {
+		php_error(E_WARNING,"start not contained in string");
+		RETURN_FALSE;
+	}
+
 	convert_to_long_ex(arg3);
 	width = Z_LVAL_PP(arg3);
 
+	if (width < 0) {
+		php_error(E_WARNING,"width has negative value");
+		RETURN_FALSE;
+	}
+
 	if (ZEND_NUM_ARGS() >= 4) {
 		convert_to_string_ex(arg4);
 		marker.val = Z_STRVAL_PP(arg4);
diff --git a/ext/mbstring/php_mbregex.c b/ext/mbstring/php_mbregex.c
index b71b9e4882..ecf57fa0b1 100644
--- a/ext/mbstring/php_mbregex.c
+++ b/ext/mbstring/php_mbregex.c
@@ -1,33 +1,19 @@
-/* 
- * PHP3 Internationalization support program.
- *
- * Copyright (c) 1999,2000 by the PHP3 internationalization team.
- * All rights reserved.
- *
- * This program is free software. You can use, redistribute and/or modify
- * without fee under the terms of the GNU General Public License version 2
- * as published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it will be useful, but
- * WITHOUT ANY WARRANTY including implied or express warranty of
- * marchantability or fitness for a particular purpose.
- *
- * Currently, the "PHP3 internationalization team" has no relationship with
- * the "PHP Development Team". But we hope these code will be integrated
- * into the PHP3, and it will be distributed as a part of PHP3.
- *
- * See README_i18n for more detail.
- *
- * Authors:
- *    Hironori Sato <satoh@jpnnet.com>
- *    Shigeru Kanemoto <sgk@happysize.co.jp>
- *    Tsukada Takuya <tsukada@fminn.nagano.nagano.jp>
- */
-
-/* 
- * PHP4 multibyte regular expression module
- * Authors:
- *    Tsukada Takuya <tsukada@fminn.nagano.nagano.jp>
+/*
+   +----------------------------------------------------------------------+
+   | PHP Version 4                                                        |
+   +----------------------------------------------------------------------+
+   | Copyright (c) 2001 The PHP Group                                     |
+   +----------------------------------------------------------------------+
+   | This source file is subject to version 2.02 of the PHP license,      |
+   | that is bundled with this package in the file LICENSE, and is        |
+   | available at through the world-wide-web at                           |
+   | http://www.php.net/license/2_02.txt.                                 |
+   | If you did not receive a copy of the PHP license and are unable to   |
+   | obtain it through the world-wide-web, please send a note to          |
+   | license@php.net so we can mail you a copy immediately.               |
+   +----------------------------------------------------------------------+
+   | Author: Tsukada Takuya <tsukada@fminn.nagano.nagano.jp>              |
+   +----------------------------------------------------------------------+
  */
 
 /* $Id$ */
diff --git a/ext/mbstring/tests/005.inc b/ext/mbstring/tests/005.inc
index 2aaaf8866d..5d486eb2a7 100644
--- a/ext/mbstring/tests/005.inc
+++ b/ext/mbstring/tests/005.inc
@@ -32,7 +32,7 @@ $r = mb_detect_order('BAD_NAME');
 ($r === FALSE) ? print "OK_BAD_STR\n" : print "NG_BAD_STR\n";
 print implode(', ', mb_detect_order()) . "\n";
 
-$a[] = 'BAD_NAME';
+$a = array('BAD_NAME');
 $r = mb_detect_order($a);
 ($r ===	FALSE) ? print "OK_BAD_ARRAY\n" : print "NG_BAD_ARRAY\n";
 print implode(', ', mb_detect_order()) . "\n";
diff --git a/ext/mbstring/tests/006.inc b/ext/mbstring/tests/006.inc
index 7da2af0708..749b019990 100644
--- a/ext/mbstring/tests/006.inc
+++ b/ext/mbstring/tests/006.inc
@@ -28,6 +28,5 @@ print mb_substitute_character() . "\n";
 print "== INVALID PARAMETER ==\n";
 $r = mb_substitute_character('BAD_NAME');
 ($r === FALSE) ? print "OK_BAD_NAME\n" : print("NG_BAD_NAME: ".gettype($r)." $r\n");
-print mb_substitute_character() . "\n";
 
 ?>
diff --git a/ext/mbstring/tests/006.phpt b/ext/mbstring/tests/006.phpt
index da4f543f21..22b3e4a4fd 100644
--- a/ext/mbstring/tests/006.phpt
+++ b/ext/mbstring/tests/006.phpt
@@ -14,6 +14,6 @@ long
 OK_NONE
 none
 == INVALID PARAMETER ==
-ERR:
+ERR: Warning
 OK_BAD_NAME
 
diff --git a/ext/mbstring/tests/007.phpt b/ext/mbstring/tests/007.phpt
index 447243ab15..08c8b8409a 100644
--- a/ext/mbstring/tests/007.phpt
+++ b/ext/mbstring/tests/007.phpt
@@ -5,7 +5,7 @@ mb_output_handler() (EUC-JP)
 --POST--
 --GET--
 --FILE--
-<?php include('skipif.inc'); ?>
+<?php include('007.inc'); ?>
 --EXPECT--
 テスト用日本語文字列。このモジュールはPHPにマルチバイト関数を提供します。
 
diff --git a/ext/mbstring/tests/009.inc b/ext/mbstring/tests/009.inc
index aabc8daed0..efdb9f8503 100644
--- a/ext/mbstring/tests/009.inc
+++ b/ext/mbstring/tests/009.inc
@@ -33,6 +33,7 @@ print  mb_strlen($sjis) . "\n";
 // JIS
 // Note: either convert_encoding or strlen has problem
 echo "== JIS ==\n";
+mb_internal_encoding('EUC-JP') or print("mb_internal_encoding() failed\n");
 $jis = mb_convert_encoding($euc_jp, 'JIS');
 print  mb_strlen($jis,'JIS') . "\n";
 mb_internal_encoding('JIS')  or print("mb_internal_encoding() failed\n");
@@ -41,6 +42,7 @@ print  mb_strlen($jis) . "\n";
 // UTF-8
 // Note: either convert_encoding or strlen has problem
 echo "== UTF-8 ==\n";
+mb_internal_encoding('EUC-JP') or print("mb_internal_encoding() failed\n");
 $utf8 = mb_convert_encoding($euc_jp, 'UTF-8');
 print  mb_strlen($utf8,'UTF-8') . "\n";
 mb_internal_encoding('UTF-8')  or print("mb_internal_encoding() failed\n");
diff --git a/ext/mbstring/tests/010.inc b/ext/mbstring/tests/010.inc
index f4493a5482..2f6ccc406f 100644
--- a/ext/mbstring/tests/010.inc
+++ b/ext/mbstring/tests/010.inc
@@ -26,24 +26,32 @@ print  mb_strpos($euc_jp, 0, 15,       'EUC-JP') . "\n";
 // Note: PHP Warning - offset is negative.
 // Note: For offset(-15). It does not return position of latter string. (ie the same result as -50)
 echo "== NEGATIVE OFFSET ==\n";
-print  mb_strpos($euc_jp,'日本語', -15, 'EUC-JP') . "\n";
-print  mb_strpos($euc_jp, '0', -15,     'EUC-JP') . "\n";
-print  mb_strpos($euc_jp, 3, -15,       'EUC-JP') . "\n";
-print  mb_strpos($euc_jp, 0, -15,       'EUC-JP') . "\n";
-print  mb_strpos($euc_jp,'日本語', -50, 'EUC-JP') . "\n";
-print  mb_strpos($euc_jp, '0', -50,     'EUC-JP') . "\n";
-print  mb_strpos($euc_jp, 3, -50,       'EUC-JP') . "\n";
-print  mb_strpos($euc_jp, 0, -50,       'EUC-JP') . "\n";
+$r = mb_strpos($euc_jp,'日本語', -15, 'EUC-JP');
+($r === FALSE) ? print "OK_NEGATIVE_OFFSET\n" : print "NG_NEGATIVE_OFFSET\n";
+$r = mb_strpos($euc_jp, '0', -15,     'EUC-JP');
+($r === FALSE) ? print "OK_NEGATIVE_OFFSET\n" : print "NG_NEGATIVE_OFFSET\n";
+$r = mb_strpos($euc_jp, 3, -15,       'EUC-JP');
+($r === FALSE) ? print "OK_NEGATIVE_OFFSET\n" : print "NG_NEGATIVE_OFFSET\n";
+$r = mb_strpos($euc_jp, 0, -15,       'EUC-JP');
+($r === FALSE) ? print "OK_NEGATIVE_OFFSET\n" : print "NG_NEGATIVE_OFFSET\n";
+$r = mb_strpos($euc_jp,'日本語', -50, 'EUC-JP');
+($r === FALSE) ? print "OK_NEGATIVE_OFFSET\n" : print "NG_NEGATIVE_OFFSET\n";
+$r = mb_strpos($euc_jp, '0', -50,     'EUC-JP');
+($r === FALSE) ? print "OK_NEGATIVE_OFFSET\n" : print "NG_NEGATIVE_OFFSET\n";
+$r = mb_strpos($euc_jp, 3, -50,       'EUC-JP');
+($r === FALSE) ? print "OK_NEGATIVE_OFFSET\n" : print "NG_NEGATIVE_OFFSET\n";
+$r = mb_strpos($euc_jp, 0, -50,       'EUC-JP');
+($r === FALSE) ? print "OK_NEGATIVE_OFFSET\n" : print "NG_NEGATIVE_OFFSET\n";
 
 // Out of range - should return false
 print ("== OUT OF RANGE ==\n");
-$r =  mb_strpos($euc_jp,'日本語', 40, 'EUC-JP') . "\n";
+$r =  mb_strpos($euc_jp,'日本語', 40, 'EUC-JP');
 ($r === FALSE) ? print "OK_OUT_RANGE\n"     : print "NG_OUT_RANGE\n";
-$r =  mb_strpos($euc_jp, '0', 40,     'EUC-JP') . "\n";
+$r =  mb_strpos($euc_jp, '0', 40,     'EUC-JP');
 ($r === FALSE) ? print "OK_OUT_RANGE\n"     : print "NG_OUT_RANGE\n";
-$r =  mb_strpos($euc_jp, 3, 40,       'EUC-JP') . "\n";
+$r =  mb_strpos($euc_jp, 3, 40,       'EUC-JP');
 ($r === FALSE) ? print "OK_OUT_RANGE\n"     : print "NG_OUT_RANGE\n";
-$r =   mb_strpos($euc_jp, 0, 40,       'EUC-JP') . "\n";
+$r =   mb_strpos($euc_jp, 0, 40,       'EUC-JP');
 ($r === FALSE) ? print "OK_OUT_RANGE\n"     : print "NG_OUT_RANGE\n";
 // Note: Returned NULL string
 // echo gettype($r). ' val '. $r ."\n"; 
diff --git a/ext/mbstring/tests/010.phpt b/ext/mbstring/tests/010.phpt
index 880176dfa3..8b3a81f8ce 100644
--- a/ext/mbstring/tests/010.phpt
+++ b/ext/mbstring/tests/010.phpt
@@ -17,14 +17,22 @@ mb_strpos()
 33
 30
 == NEGATIVE OFFSET ==
-34
-30
-33
-30
-10
-0
-3
-0
+ERR: Warning
+OK_NEGATIVE_OFFSET
+ERR: Warning
+OK_NEGATIVE_OFFSET
+ERR: Warning
+OK_NEGATIVE_OFFSET
+ERR: Warning
+OK_NEGATIVE_OFFSET
+ERR: Warning
+OK_NEGATIVE_OFFSET
+ERR: Warning
+OK_NEGATIVE_OFFSET
+ERR: Warning
+OK_NEGATIVE_OFFSET
+ERR: Warning
+OK_NEGATIVE_OFFSET
 == OUT OF RANGE ==
 OK_OUT_RANGE
 OK_OUT_RANGE
diff --git a/ext/mbstring/tests/014.inc b/ext/mbstring/tests/014.inc
index 14e61633a3..afcbae28b8 100644
--- a/ext/mbstring/tests/014.inc
+++ b/ext/mbstring/tests/014.inc
@@ -11,12 +11,13 @@ print  "1: ". mb_strimwidth($euc_jp,  0, 15,'...','EUC-JP') . "\n";
 print  "2: ". mb_strimwidth($euc_jp,  0, 100,'...','EUC-JP') . "\n";
 print  "3: ". mb_strimwidth($euc_jp, 15, 100,'...','EUC-JP') . "\n";
 // Note: Did not start form -22 offset. Staring from 0.
-print  "4: ". mb_strimwidth($euc_jp,-22, 100,'...','EUC-JP') . "\n";
+$str = mb_strimwidth($euc_jp,-22, 100,'...','EUC-JP');
+($str === FALSE) ? print "4 OK\n" : print "NG: $str\n";
 
 $str = mb_strimwidth($euc_jp, 100, -10,'...','EUC-JP');
-($str === "") ? print "5 OK\n" : print "NG: $str\n";
+($str === FALSE) ? print "5 OK\n" : print "NG: $str\n";
 
 $str = mb_strimwidth($euc_jp, -100, 10,'...','EUC-JP');
-($str !== "") ?	print "6 OK: $str\n" : print "NG: $str\n";
+($str === FALSE) ? print "6 OK\n" : print "NG: $str\n";
 
 ?>
diff --git a/ext/mbstring/tests/014.phpt b/ext/mbstring/tests/014.phpt
index bc27326b32..db37e9f8bc 100644
--- a/ext/mbstring/tests/014.phpt
+++ b/ext/mbstring/tests/014.phpt
@@ -10,8 +10,11 @@ mb_strimwidth()
 1: 0123この文字...
 2: 0123この文字列は日本語です。EUC-JPを使っています。日本語は面倒臭い。
 3: 。EUC-JPを使っています。日本語は面倒臭い。
-4: EUC-JPを使っています。日本語は面倒臭い。
+ERR: Warning
+4 OK
+ERR: Warning
 5 OK
-6 OK: 0123こ...
+ERR: Warning
+6 OK
 
 
diff --git a/ext/mbstring/tests/016.inc b/ext/mbstring/tests/016.inc
index 321bbeba83..bea679e502 100644
--- a/ext/mbstring/tests/016.inc
+++ b/ext/mbstring/tests/016.inc
@@ -35,8 +35,7 @@ print("JIS: ".base64_encode($s)."\n"); // JIS
 // Using Encoding List Array 
 echo "== ARRAY ENCODING LIST ==\n";
 
-$a = array(0=>'UTF-8',1=>'EUC-JP', 2=>'SJIS', 3=>'JIS');
-
+$a = array(0=>'JIS', 1=>'UTF-8', 2=>'EUC-JP', 3=>'SJIS');
 $s = $jis;
 $s = mb_convert_encoding($s, 'EUC-JP', $a);
 print("EUC-JP: $s\n"); // EUC-JP
@@ -53,19 +52,16 @@ print("JIS: ".base64_encode($s)."\n"); // JIS
 // Using Detect Order 
 echo "== DETECT ORDER ==\n";
 
-mb_detect_order('auto');
-
-
 $s = $jis;
-$s = mb_convert_encoding($s, 'EUC-JP');
+$s = mb_convert_encoding($s, 'EUC-JP', 'auto');
 print("EUC-JP: $s\n"); // EUC-JP
 
 $s = $euc_jp;
-$s = mb_convert_encoding($s, 'SJIS');
+$s = mb_convert_encoding($s, 'SJIS', 'auto');
 print("SJIS: ".base64_encode($s)."\n");  // SJIS
 
 $s = $euc_jp;
-$s = mb_convert_encoding($s, 'JIS');
+$s = mb_convert_encoding($s, 'JIS', 'auto');
 print("JIS: ".base64_encode($s)."\n"); // JIS