* DecoderFallback.cs, EncoderFallback.cs, Encoding.cs :
the same "\uFFFD" fix for encoder fallback.
Reduced extra instantiation of those fallbacks.
* EncoderReplacementFallbackTest.cs,
EncoderReplacementFallbackBufferTest.cs : encoder test for
"\uFFFD" replacement.
svn path=/trunk/mcs/; revision=91947
* Encoding.cs : added UTF8UnmarkedUnsafe, for old empty replacement
UTF8.
* BinaryReader.cs, BinaryWriter.cs : use unsafe encoding that has ""
for replacement fallback. Binary serialization regression is fixed.
svn path=/trunk/mcs/; revision=88276
* Encoding.cs : the change broke the build, so reverted part of it.
* EncodingTest.cs :
reverted part of the related patch, thus added NotWorking.
svn path=/trunk/mcs/; revision=72087
* Encoding.cs: In GetEncoding (int) do not allow codepage value below zero
and above 0xffff. Modified NotSupportedException to ArgumentException in
GetEncoding (string).
* corlib_test.dll.sources: Added EncodingTest.cs.
svn path=/trunk/mcs/; revision=72052
* Encoding.cs :
Consider DecoderFallback and EncoderFallback in 2.0 Equals()
and GetHashCode().
* UTF32Encoding.cs, UTF7Encoding.cs :
Fixed GetHashCode() and Equals() as well.
Added several missing overrides in 2.0.
svn path=/trunk/mcs/; revision=64299
Marked 2.0-only members as ComVisible (false).
* Decoder.cs: Marked class ComVisible (true) on 2.0 profile. Marked
2.0-only members as ComVisible (false).
* Encoder.cs: Marked class ComVisible (true) on 2.0 profile. Marked
2.0-only members as ComVisible (false).
* Encoding.cs: Marked class ComVisible (true) on 2.0 profile. Marked
2.0-only members as ComVisible (false).
* NormalizationForm.cs: Marked ComVisible (true).
* StringBuilder.cs: Marked class ComVisible (true) on 2.0 profile.
Marked 2.0-only methods as ComVisible (false).
* UnicodeEncoding.cs: GetString (byte[], int, int) is only available
in 2.0 profile. Marked 2.0-only methods ComVisible (false). Added
missing ctor in 2.0 profile, and marked in MonoTODO.
* UTF7Encoding.cs: Marked class ComVisible (true) on 2.0 profile.
* UTF8Encoding.cs: Marked 2.0-only methods as ComVisible (false).
svn path=/trunk/mcs/; revision=61250
* Encoding.cs : now under 2.0 GetBytes(string,...) dispatches to
byte*-based GetBytes().
* UTF8Encoding.cs : avoid extraneous DecoderFallbackBuffer creation
which came to happen after introducing fallback buffer.
svn path=/trunk/mcs/; revision=54109
* ASCIIEncoding.cs, Latin1Encoding.cs : added EncoderFallback support.
* Encoding.cs : changed default fallback selection. Seems like only
ASCII and GB18030 uses '?' for replacement.
* UTF8Encoding.cs : now that Fallback is read only by default, we
need special setter.
* EncoderReplacementFallbackTest.cs :
Added Latin1 replacement test.
Updated some test results changed from beta2 to RTM.
* DecoderReplacementFallbackTest.cs : the same updates.
svn path=/trunk/mcs/; revision=53350
2005-11-11 Miguel de Icaza <[email protected]>
* Encoding.cs, UnicodeEncoding.cs: A few 2.x methods.
* StringBuilder.cs (Text): Added serialization support in 2.x.
svn path=/trunk/mcs/; revision=52916
Fix line endings of System/Decimal.cs (only actual "change" in text, to force recompile of mscorlib to reflect changes in decimal constant handling in the compiler)
svn path=/trunk/mcs/; revision=44558
* Console.cs: Use new codepage setup.
2004-04-13 Miguel de Icaza <[email protected]>
* Encoding.cs: Use new internal codepage setup.
svn path=/trunk/mcs/; revision=25434
* Encoding.cs: Use an internal variable to track the parameters of
each encoder, since the .NET API does not expose virtual methods
in the child classes, we should not depend on that.
* ASCIIEncoding.cs, UTF7Encoding, UTF8Encoding, UnicodeEncoding:
Initialize the parameters for base class encoding here.
svn path=/trunk/mcs/; revision=19843
* Encoding.cs (UTF8Unmarked): make it also not error out on
invalid input, that is what the Microsoft default is for the
StreamReader and BinaryReader expect (our main consumers).
svn path=/trunk/mcs/; revision=17807
* Encoding.cs (GetEncoding): Add all the aliases documented in the
framework. We were missing a few.
Reorganize the table. Also, convert the input name into the
lower-case - to _ before we compare against our table.
svn path=/trunk/mcs/; revision=14472
* Encoding.cs: g_get_encoding () returns "ASCII". In Default, when the
encoding is not found, default to UTF8Unmarked. Removed comment in
UnixConsoleEncoding, because it's now Default, which gets the
encoding internally using g_get_encoding ().
svn path=/trunk/mcs/; revision=9584
* Encoding.cs (UTF8Unmarked): New static property, used to return
a no-markers UTf8 encoder, used in a few places in the class library.
svn path=/trunk/mcs/; revision=8684
* Console.cs: Specify an encoder, otherwise we will get the UTF8
encoder that by default emits the byte markers.
2002-09-06 Miguel de Icaza <[email protected]>
* StreamReader.cs: Implement detection of byte marks and skipping
of byte marks at the beginning of the stream.
(ReadToEnd): Use buffered read instead of char-by-char
processing.
Correct the default arguments for creating the StreamReader.
2002-09-06 Miguel de Icaza <[email protected]>
* Encoding.cs (UTF8, Unicode): Create with a preamble, that is
what the Microsoft version does.
(UnixConsoleEncoding): New internal method, used to get the
encoding, in the future, this should pull the locale, map to
charset and then code page.
svn path=/trunk/mcs/; revision=7295