ICU-1864 add BOCU-1

X-SVN-Rev: 8481

icu4c/source/common/Makefile.in

@@ -71,7 +71,7 @@ OBJECTS = putil.o locmap.o mutex.o umutex.o \
 udata.o ucmndata.o udatamem.o umapfile.o filestrm.o \
 uresbund.o uresdata.o resbund.o cwchar.o uloc.o locid.o uhash.o uhash_us.o \
 ucnv.o ucnv_bld.o ucnv_cb.o ucnv_cnv.o ucnv_err.o ucnv_io.o ucnvlat1.o \
-ucnv_utf.o ucnvmbcs.o ucnv2022.o ucnvhz.o ucnv_lmb.o ucnvscsu.o ucnvisci.o \
+ucnv_utf.o ucnvmbcs.o ucnv2022.o ucnvhz.o ucnv_lmb.o ucnvscsu.o ucnvbocu.o ucnvisci.o \
 unistr.o utf_impl.o ustring.o ustrcase.o cstring.o ustrfmt.o ustrtrns.o \
 normlzr.o unorm.o chariter.o schriter.o uchriter.o uiter.o \
 uchar.o uprops.o bidi.o ubidi.o ubidiwrt.o ubidiln.o ushape.o unames.o ucln_cmn.o \

icu4c/source/common/common.dsp

@@ -251,6 +251,17 @@ SOURCE=.\ucnv_utf.c
 # End Source File
 # Begin Source File
 
+SOURCE=.\ucnvbocu.c
+
+!IF  "$(CFG)" == "common - Win32 Release"
+
+!ELSEIF  "$(CFG)" == "common - Win32 Debug"
+
+!ENDIF 
+
+# End Source File
+# Begin Source File
+
 SOURCE=.\ucnvhz.c
 # End Source File
 # Begin Source File

icu4c/source/common/ucnv.c

@@ -1276,7 +1276,7 @@ ucnv_detectUnicodeSignature( const char* source,
         *signatureLength=2;
         return  "UTF-16BE";
     } else if(start[0] == '\xFF' && start[1] == '\xFE') {
-        if(start[2] == '\x00' && start[3] =='\x00'){
+        if(start[2] == '\x00' && start[3] =='\x00') {
             *signatureLength=4;
             return "UTF-32LE";
         } else {
@@ -1286,14 +1286,17 @@ ucnv_detectUnicodeSignature( const char* source,
     } else if(start[0] == '\xEF' && start[1] == '\xBB' && start[2] == '\xBF') {
         *signatureLength=3;
         return  "UTF-8";
-    }else if(start[0] == '\x0E' && start[1] == '\xFE' && start[2] == '\xFF'){
-        *signatureLength=3;
-        return "SCSU";
-    }else if(start[0] == '\x00' && start[1] == '\x00' && 
-            start[2] == '\xFE' && start[3]=='\xFF'){
+    } else if(start[0] == '\x00' && start[1] == '\x00' && 
+              start[2] == '\xFE' && start[3]=='\xFF') {
         *signatureLength=4;
         return  "UTF-32BE";
-    }else{
+    } else if(start[0] == '\x0E' && start[1] == '\xFE' && start[2] == '\xFF') {
+        *signatureLength=3;
+        return "SCSU";
+    } else if(start[0] == '\xFC' && start[1] == '\xEE' && start[2] == '\x27') {
+        *signatureLength=3;
+        return "BOCU-1";
+    } else {
         *signatureLength=0;
         return NULL;
     }

icu4c/source/common/ucnv_bld.c

@@ -49,7 +49,7 @@ converterData[UCNV_NUMBER_OF_SUPPORTED_CONVERTER_TYPES]={
     NULL, &_ISO2022Data, 
     &_LMBCSData1,&_LMBCSData2, &_LMBCSData3, &_LMBCSData4, &_LMBCSData5, &_LMBCSData6,
     &_LMBCSData8,&_LMBCSData11,&_LMBCSData16,&_LMBCSData17,&_LMBCSData18,&_LMBCSData19,
-    &_HZData, &_SCSUData, &_ISCIIData, &_ASCIIData, &_UTF7Data
+    &_HZData, &_SCSUData, &_ISCIIData, &_ASCIIData, &_UTF7Data, &_Bocu1Data
 };
 
 static struct {
@@ -93,7 +93,8 @@ static struct {
   { "SCSU", UCNV_SCSU },
   { "ISCII", UCNV_ISCII },
   { "US-ASCII", UCNV_US_ASCII },
-  { "UTF-7", UCNV_UTF7 }
+  { "UTF-7", UCNV_UTF7 },
+  { "BOCU-1", UCNV_BOCU1 }
 };
 
 

icu4c/source/common/ucnv_cnv.h

@@ -153,7 +153,7 @@ extern const UConverterSharedData
     _ISO2022Data, 
     _LMBCSData1,_LMBCSData2, _LMBCSData3, _LMBCSData4, _LMBCSData5, _LMBCSData6,
     _LMBCSData8,_LMBCSData11,_LMBCSData16,_LMBCSData17,_LMBCSData18,_LMBCSData19,
-    _HZData,_ISCIIData, _SCSUData, _ASCIIData, _UTF7Data;
+    _HZData,_ISCIIData, _SCSUData, _ASCIIData, _UTF7Data, _Bocu1Data;
 
 U_CDECL_END
 

icu4c/source/common/unicode/ucnv.h

@@ -82,6 +82,7 @@ typedef enum {
     UCNV_ISCII,
     UCNV_US_ASCII,
     UCNV_UTF7,
+    UCNV_BOCU1,
 
     /* Number of converter types for which we have conversion routines. */
     UCNV_NUMBER_OF_SUPPORTED_CONVERTER_TYPES