Author: atsushi
Date: 2006-01-16 04:41:40 -0500 (Mon, 16 Jan 2006)
New Revision: 55626
Modified:
trunk/mcs/class/I18N/CJK/CP51932.cs
trunk/mcs/class/I18N/CJK/ChangeLog
trunk/mcs/class/I18N/CJK/Test/ChangeLog
trunk/mcs/class/I18N/CJK/Test/I18N.CJK.Test.cs
Log:
2006-01-16 Atsushi Enomoto <[EMAIL PROTECTED]>
* CP51932.cs :
In GetChars(), insufficient byte is not preserved.
Under 1.x profile it is just discarded.
Under 2.0 profile it is converted to \u30FB.
Fixed GetCharCount() to match GetChars().
* I18N.CJK.Test.cs : added test for bug #77223, #77224 and #77238.
Modified: trunk/mcs/class/I18N/CJK/CP51932.cs
===================================================================
--- trunk/mcs/class/I18N/CJK/CP51932.cs 2006-01-16 08:52:23 UTC (rev 55625)
+++ trunk/mcs/class/I18N/CJK/CP51932.cs 2006-01-16 09:41:40 UTC (rev 55626)
@@ -66,9 +66,6 @@
// Internal state.
private JISConvert convert;
- // Conversion cache (note that encoding is not thread safe)
- int lastByte;
-
// Constructor.
public CP51932 () : base (EUC_JP_CODE_PAGE)
{
@@ -224,7 +221,6 @@
}
// Get the number of characters needed to decode a byte buffer.
- // TODO: check
public override int GetCharCount (byte [] bytes, int index, int count)
{
// Validate the parameters.
@@ -239,50 +235,96 @@
throw new ArgumentOutOfRangeException
("count", Strings.GetString("ArgRange_Array"));
+
+
// Determine the total length of the converted string.
+ int value = 0;
+ byte[] table0208 = convert.jisx0208ToUnicode;
+ byte[] table0212 = convert.jisx0212ToUnicode;
int length = 0;
- int byteval;
+ int byteval = 0;
int last = 0;
while (count > 0) {
byteval = bytes [index++];
--count;
-
if (last == 0) {
if (byteval == 0x8F) {
if (byteval != 0) {
- // Invalid second byte of a
3-byte character
- // FIXME: What should we do?
+ // Invalid second byte of a
3-byte character.
last = 0;
+ length++;
}
// First byte in a triple-byte sequence
else
last = byteval;
} else if (byteval <= 0x7F) {
// Ordinary ASCII/Latin1/Control
character.
- ++length;
+ length++;
+ } else if (byteval == 0x8E) {
+ // First byte of half-width Katakana
+ last = byteval;
} else if (byteval >= 0xA1 && byteval <= 0xFE) {
// First byte in a double-byte sequence.
last = byteval;
- } else if (byteval == 0x87) {
- // First byte in half-width katakana
} else {
- // Invalid first byte. Let '?'
- ++length;
+ // Invalid first byte.
+ length++;
}
- } else if (last == 0x8F) {
+ }
+ else if (last == 0x8E) {
+ if (byteval >= 0xA1 && byteval <= 0xDF) {
+ value = ((byteval - 0x40) |
+ (last + 0x71) << 8);
+ length++;
+ } else {
+ // Invalid second byte.
+ length++;
+ }
+ last =0;
+ }
+ else if (last == 0x8F) {
// 3-byte character
// FIXME: currently not supported yet
- last = 0;
- ++length;
- } else {
+ last = byteval;
+ }
+ else
+ {
// Second byte in a double-byte sequence.
+ value = (last - 0xA1) * 0x5E;
last = 0;
- ++length;
+ if (byteval >= 0xA1 && byteval <= 0xFE)
+ {
+ value += (byteval - 0xA1);
+ }
+ else
+ {
+ // Invalid second byte.
+ last = 0;
+ length++;
+ continue;
+ }
+
+ value *= 2;
+ value = ((int) (table0208 [value]))
+ | (((int) (table0208 [value + 1])) <<
8);
+ if (value == 0)
+ value = ((int) (table0212 [value]))
+ | (((int) (table0212 [value +
1])) << 8);
+ if (value != 0)
+ length++;
+ else
+ length++;
}
}
+#if NET_2_0
+ // seems like .NET 2.0 adds \u30FB for insufficient
+ // byte seuqence (for Japanese \u30FB makes sense).
+ if (last != 0)
+ length++;
+#endif
- // Return the total length.
+ // Return the final length to the caller.
return length;
}
@@ -319,7 +361,7 @@
int posn = charIndex;
int charLength = chars.Length;
int byteval, value;
- int last = lastByte;
+ int last = 0;
byte[] table0208 = convert.jisx0208ToUnicode;
byte[] table0212 = convert.jisx0212ToUnicode;
@@ -327,22 +369,21 @@
byteval = bytes [byteIndex++];
--byteCount;
if (last == 0) {
- if (posn >= charLength)
- throw new ArgumentException
- (Strings.GetString
-
("Arg_InsufficientSpace"), "chars");
-
if (byteval == 0x8F) {
if (byteval != 0) {
- // Invalid second byte of a
3-byte character
- // FIXME: What should we do?
+ // Invalid second byte of a
3-byte character.
last = 0;
+ if (posn >= charLength)
+ throw Insufficient ();
+ chars [posn++] = '?';
}
// First byte in a triple-byte sequence
else
last = byteval;
} else if (byteval <= 0x7F) {
// Ordinary ASCII/Latin1/Control
character.
+ if (posn >= charLength)
+ throw Insufficient ();
chars [posn++] = (char) byteval;
} else if (byteval == 0x8E) {
// First byte of half-width Katakana
@@ -352,6 +393,8 @@
last = byteval;
} else {
// Invalid first byte.
+ if (posn >= charLength)
+ throw Insufficient ();
chars [posn++] = '?';
}
}
@@ -359,9 +402,13 @@
if (byteval >= 0xA1 && byteval <= 0xDF) {
value = ((byteval - 0x40) |
(last + 0x71) << 8);
+ if (posn >= charLength)
+ throw Insufficient ();
chars [posn++] = (char) value;
} else {
// Invalid second byte.
+ if (posn >= charLength)
+ throw Insufficient ();
chars [posn++] = '?';
}
last =0;
@@ -383,7 +430,9 @@
else
{
// Invalid second byte.
- lastByte = 0;
+ last = 0;
+ if (posn >= charLength)
+ throw Insufficient ();
chars [posn++] = '?';
continue;
}
@@ -394,18 +443,35 @@
if (value == 0)
value = ((int) (table0212 [value]))
| (((int) (table0212 [value +
1])) << 8);
+ if (posn >= charLength)
+ throw Insufficient ();
if (value != 0)
chars [posn++] = (char)value;
else
chars [posn++] = '?';
}
}
- lastByte = last;
+#if NET_2_0
+ if (last != 0) {
+ // seems like .NET 2.0 adds \u30FB for insufficient
+ // byte seuqence (for Japanese \u30FB makes sense).
+ if (posn >= charLength)
+ throw Insufficient ();
+ chars [posn++] = '\u30FB';
+ }
+#endif
// Return the final length to the caller.
return posn - charIndex;
}
+ Exception Insufficient ()
+ {
+ throw new ArgumentException
+ (Strings.GetString
+ ("Arg_InsufficientSpace"), "chars");
+ }
+
// Get the maximum number of bytes needed to encode a
// specified number of characters.
public override int GetMaxByteCount(int charCount)
Modified: trunk/mcs/class/I18N/CJK/ChangeLog
===================================================================
--- trunk/mcs/class/I18N/CJK/ChangeLog 2006-01-16 08:52:23 UTC (rev 55625)
+++ trunk/mcs/class/I18N/CJK/ChangeLog 2006-01-16 09:41:40 UTC (rev 55626)
@@ -1,3 +1,11 @@
+2006-01-16 Atsushi Enomoto <[EMAIL PROTECTED]>
+
+ * CP51932.cs :
+ In GetChars(), insufficient byte is not preserved.
+ Under 1.x profile it is just discarded.
+ Under 2.0 profile it is converted to \u30FB.
+ Fixed GetCharCount() to match GetChars().
+
2006-01-13 Atsushi Enomoto <[EMAIL PROTECTED]>
* CP932.cs : GetByteCount() returned insufficient count at PrivateUse
Modified: trunk/mcs/class/I18N/CJK/Test/ChangeLog
===================================================================
--- trunk/mcs/class/I18N/CJK/Test/ChangeLog 2006-01-16 08:52:23 UTC (rev
55625)
+++ trunk/mcs/class/I18N/CJK/Test/ChangeLog 2006-01-16 09:41:40 UTC (rev
55626)
@@ -1,3 +1,7 @@
+2006-01-16 Atsushi Enomoto <[EMAIL PROTECTED]>
+
+ * I18N.CJK.Test.cs : added test for bug #77223, #77224 and #77238.
+
2005-12-06 Atsushi Enomoto <[EMAIL PROTECTED]>
* I18N.CJK.Test.cs : w00t, Now all test are working, including
Modified: trunk/mcs/class/I18N/CJK/Test/I18N.CJK.Test.cs
===================================================================
--- trunk/mcs/class/I18N/CJK/Test/I18N.CJK.Test.cs 2006-01-16 08:52:23 UTC
(rev 55625)
+++ trunk/mcs/class/I18N/CJK/Test/I18N.CJK.Test.cs 2006-01-16 09:41:40 UTC
(rev 55626)
@@ -148,7 +148,7 @@
AssertDecode ("Test/texts/japanese-utf8.txt",
"Test/texts/japanese-51932.txt", 51932);
}
- // ISI-2022-JP
+ // ISO-2022-JP
[Test]
public void CP50220_Encode ()
@@ -192,6 +192,36 @@
AssertDecode ("Test/texts/japanese-utf8.txt",
"Test/texts/japanese-50222.txt", 50222);
}
+ [Test]
+ public void Bug77723 ()
+ {
+ Encoding e = Encoding.GetEncoding (51932);
+ for (int i = 0; i < 0x10000; i++)
+ e.GetBytes (new char [] { (char)i });
+ }
+
+ [Test]
+ public void Bug77224 ()
+ {
+ Encoding e = Encoding.GetEncoding (932);
+ for (int i = 0; i < 0x10000; i++)
+ e.GetBytes (new char [] {(char) i});
+ }
+
+ [Test]
+ public void Bug77238 ()
+ {
+ Encoding e = Encoding.GetEncoding (51932);
+ byte [] bytes = new byte [2];
+ for (int i0 = 0; i0 < 0x100; i0++) {
+ bytes [0] = (byte) i0;
+ for (int i1 = 0; i1 < 0x100; i1++) {
+ bytes [1] = (byte) i1;
+ e.GetChars (bytes);
+ }
+ }
+ }
+
#endregion
#region Korean
_______________________________________________
Mono-patches maillist - [email protected]
http://lists.ximian.com/mailman/listinfo/mono-patches
