r=yokoyama@netscape.com sr=brendan@mozilla.org

fix bug 7964 : Add ISO-2022-KR decoder (to Unicode) for
Korean emails. Encoder is NOT included since ISO-2022-KR should
not be used for outgoing emails.

intl/uconv/ucvko/nsISO2022KRToUnicode.cpp

@@ -19,7 +19,7 @@
  * Portions created by the Initial Developer are Copyright (C) 1998
  * the Initial Developer. All Rights Reserved.
  *
- * Contributor(s):
+ * Contributor(s): Jungshik Shin <jshin@mailaps.org>
  *
  *
  * Alternatively, the contents of this file may be used under the terms of
@@ -35,3 +35,173 @@
  * the terms of any one of the NPL, the GPL or the LGPL.
  *
  * ***** END LICENSE BLOCK ***** */
+#include "nsISO2022KRToUnicode.h"
+#include "nsUCvKOSupport.h"
+#include "nsICharsetConverterManager.h"
+#include "nsIServiceManager.h"
+
+static NS_DEFINE_CID(kCharsetConverterManagerCID, NS_ICHARSETCONVERTERMANAGER_CID);
+
+NS_IMETHODIMP nsISO2022KRToUnicode::Convert(const char * aSrc, PRInt32 * aSrcLen, PRUnichar * aDest, PRInt32 * aDestLen)
+{
+  const unsigned char* srcEnd = (unsigned char*)aSrc + *aSrcLen;
+  const unsigned char* src =(unsigned char*) aSrc;
+  PRUnichar* destEnd = aDest + *aDestLen;
+  PRUnichar* dest = aDest;
+  while((src < srcEnd))
+  {
+    switch(mState)
+    {
+      case mState_ASCII:
+        if(0x1b == *src) {
+            mLastLegalState = mState;
+            mState = mState_ESC;
+        } 
+        else if(0x0e == *src) { // Shift-Out 
+          mState = mState_KSX1001_1992;
+        } 
+        else if(*src & 0x80) {
+          *dest++ = 0xFFFD;
+          if(dest >= destEnd)
+            goto error1;
+        } 
+        else {
+          *dest++ = (PRUnichar) *src;
+          if(dest >= destEnd)
+            goto error1;
+        }
+        break;
+          
+      case mState_ESC:
+        if('$' == *src) {
+          mState = mState_ESC_24;
+        } 
+        else  {
+          if((dest+2) >= destEnd)
+            goto error1;
+          *dest++ = (PRUnichar) 0x1b;
+          *dest++ = (0x80 & *src) ? 0xFFFD : (PRUnichar) *src;
+          mState =  mLastLegalState;
+        }
+        break;
+
+      case mState_ESC_24: // ESC $
+        if(')' == *src) {
+          mState = mState_ESC_24_29;
+        } 
+        else  {
+          if((dest+3) >= destEnd)
+            goto error1;
+          *dest++ = (PRUnichar) 0x1b;
+          *dest++ = (PRUnichar) '$';
+          *dest++ = (0x80 & *src) ? 0xFFFD : (PRUnichar) *src;
+          mState = mLastLegalState;
+        }
+        break;
+
+      case mState_ESC_24_29: // ESC $ )
+        mState = mLastLegalState;
+        if('C' == *src) {
+          mState = mState_ASCII;
+        } 
+        else  {
+          if((dest+4) >= destEnd)
+            goto error1;
+          *dest++ = (PRUnichar) 0x1b;
+          *dest++ = (PRUnichar) '$';
+          *dest++ = (PRUnichar) ')';
+          *dest++ = (0x80 & *src) ? 0xFFFD : (PRUnichar) *src;
+          mState = mLastLegalState;
+        }
+        break;
+
+      case mState_KSX1001_1992:
+        if (0x20 < (PRUint8) *src  && (PRUint8) *src < 0x7f) {
+          mData = (PRUint8) *src;
+          mState = mState_KSX1001_1992_2ndbyte;
+        } 
+        else if (0x0f == *src) { // Shift-In (SI)
+          mState = mState_ASCII;
+        } 
+        else if ((PRUint8) *src == 0x20 || (PRUint8) *src == 0x09) {
+          // Allow space and tab between SO and SI (i.e. in Hangul segment)
+          mState = mState_KSX1001_1992;
+          *dest++ = (PRUnichar) *src;
+          if(dest >= destEnd)
+          goto error1;
+        } 
+        else {         // Everything else is invalid.
+          *dest++ = 0xFFFD;
+          if(dest >= destEnd)
+             goto error1;
+        }
+        break;
+
+      case mState_KSX1001_1992_2ndbyte:
+        if ( 0x20 < (PRUint8) *src && (PRUint8) *src < 0x7f  ) {
+          if (!mEUCKRDecoder) {
+            // creating a delegate converter (EUC-KR)
+            nsresult rv;
+            nsString tmpCharset;
+            nsCOMPtr<nsICharsetConverterManager> ccm = 
+                  do_GetService(kCharsetConverterManagerCID, &rv);
+            if (!NS_FAILED(rv)) {
+              tmpCharset.AssignWithConversion("EUC-KR");
+              rv = ccm->GetUnicodeDecoder(&tmpCharset, &mEUCKRDecoder);
+            }
+          }
+
+          if (!mEUCKRDecoder) {// failed creating a delegate converter
+           *dest++ = 0xFFFD;
+          } 
+          else {              
+            unsigned char ksx[2];
+            PRUnichar uni;
+            PRInt32 ksxLen = 2, uniLen = 1;
+            // mData is the original 1st byte.
+            // *src is the present 2nd byte.
+            // Put 2 bytes (one character) to ksx[] with EUC-KR encoding.
+            ksx[0] = mData | 0x80;
+            ksx[1] = *src | 0x80;
+            // Convert EUC-KR to unicode.
+            mEUCKRDecoder->Convert((const char *)ksx, &ksxLen, &uni, &uniLen);
+            *dest++ = uni;
+          }
+          if(dest >= destEnd)
+            goto error1;
+          mState = mState_KSX1001_1992;
+        } 
+        else {        // Invalid 
+          if ( 0x0f == *src ) {   // Shift-In (SI)
+            mState = mState_ASCII;
+          } 
+          else {
+            mState = mState_KSX1001_1992;
+          }
+          *dest++ = 0xFFFD;
+          if(dest >= destEnd)
+           goto error1;
+        }
+        break;
+
+      case mState_ERROR:
+        mState = mLastLegalState;
+        *dest++ = 0xFFFD;
+        if(dest >= destEnd)
+          goto error1;
+        break;
+
+    } // switch
+    src++;
+    if ( *src == 0x0a || *src == 0x0d )   // if LF/CR, return to US-ASCII unconditionally.
+      mState = mState_ASCII;
+   }
+   *aDestLen = dest - aDest;
+   return NS_OK;
+
+error1:
+   *aDestLen = dest-aDest;
+   *aSrcLen = src-(unsigned char*)aSrc;
+   return NS_OK_UDEC_MOREOUTPUT;
+}
+

intl/uconv/ucvko/nsISO2022KRToUnicode.h

@@ -35,3 +35,59 @@
  * the terms of any one of the NPL, the GPL or the LGPL.
  *
  * ***** END LICENSE BLOCK ***** */
+#ifndef nsISO2022KRToUnicode_h__
+#define nsISO2022KRToUnicode_h__
+#include "nsISupports.h"
+#include "nsUCvKOSupport.h"
+
+
+ 
+class nsISO2022KRToUnicode : public nsBasicDecoderSupport
+{
+public:
+  nsISO2022KRToUnicode()
+  { 
+    mState = mState_ASCII;
+    mLastLegalState = mState_ASCII;
+    mData = 0;
+    mEUCKRDecoder = nsnull;
+  };
+
+  virtual ~nsISO2022KRToUnicode()
+  {
+    NS_IF_RELEASE(mEUCKRDecoder);
+  };
+
+  NS_IMETHOD Convert(const char * aSrc, PRInt32 * aSrcLength,
+     PRUnichar * aDest, PRInt32 * aDestLength) ;
+  
+  NS_IMETHOD GetMaxLength(const char * aSrc, PRInt32 aSrcLength,
+     PRInt32 * aDestLength) 
+  {
+    *aDestLength = aSrcLength;
+    return NS_OK;
+  };
+
+  NS_IMETHOD Reset()
+  {
+    mState = mState_ASCII;
+    mLastLegalState = mState_ASCII;
+    return NS_OK;
+  };
+
+private:
+  enum {
+    mState_ASCII,
+    mState_ESC,
+    mState_ESC_24,
+    mState_ESC_24_29,
+    mState_KSX1001_1992,
+    mState_KSX1001_1992_2ndbyte,
+    mState_ERROR
+  } mState, mLastLegalState;
+
+  PRUint8 mData;
+
+  nsIUnicodeDecoder *mEUCKRDecoder;
+};
+#endif // nsISO2022KRToUnicode_h__

intl/uconv/ucvko/nsUCvKoModule.cpp

@@ -62,6 +62,7 @@
 #include "nsUnicodeToJohabNoAscii.h"
 #include "nsCP949ToUnicode.h"
 #include "nsUnicodeToCP949.h"
+#include "nsISO2022KRToUnicode.h"
 
 //----------------------------------------------------------------------------
 // Global functions and data [declaration]
@@ -97,6 +98,7 @@ NS_UCONV_REG_UNREG(nsUnicodeToJohab, "Unicode", "x-johab",  NS_UNICODETOJOHAB_CI
 NS_UCONV_REG_UNREG(nsUnicodeToJohabNoAscii, "Unicode", "x-johab-noascii",  NS_UNICODETOJOHABNOASCII_CID);
 NS_UCONV_REG_UNREG(nsCP949ToUnicode, "x-windows-949", "Unicode" , NS_CP949TOUNICODE_CID);
 NS_UCONV_REG_UNREG(nsUnicodeToCP949, "Unicode", "x-windows-949",  NS_UNICODETOCP949_CID);
+NS_UCONV_REG_UNREG(nsISO2022KRToUnicode, "ISO-2022-KR", "Unicode" , NS_ISO2022KRTOUNICODE_CID);
 
 NS_GENERIC_FACTORY_CONSTRUCTOR(nsEUCKRToUnicode);
 NS_GENERIC_FACTORY_CONSTRUCTOR(nsUnicodeToEUCKR);
@@ -107,6 +109,7 @@ NS_GENERIC_FACTORY_CONSTRUCTOR(nsUnicodeToJohab);
 NS_GENERIC_FACTORY_CONSTRUCTOR(nsUnicodeToJohabNoAscii);
 NS_GENERIC_FACTORY_CONSTRUCTOR(nsCP949ToUnicode);
 NS_GENERIC_FACTORY_CONSTRUCTOR(nsUnicodeToCP949);
+NS_GENERIC_FACTORY_CONSTRUCTOR(nsISO2022KRToUnicode);
 
 static nsModuleComponentInfo components[] = 
 {
@@ -163,6 +166,12 @@ static nsModuleComponentInfo components[] =
     NS_UNICODEENCODER_CONTRACTID_BASE "x-windows-949",
     nsUnicodeToCP949Constructor,
     nsUnicodeToCP949RegSelf, nsUnicodeToCP949UnRegSelf
+  },
+  { 
+    DECODER_NAME_BASE "ISO-2022-KR" , NS_ISO2022KRTOUNICODE_CID, 
+    NS_UNICODEDECODER_CONTRACTID_BASE "ISO-2022-KR",
+    nsISO2022KRToUnicodeConstructor ,
+    nsISO2022KRToUnicodeRegSelf , nsISO2022KRToUnicodeUnRegSelf 
   }
 };