michael@0: /* -*- Mode: C; tab-width: 4; indent-tabs-mode: nil; c-basic-offset: 2 -*- */
michael@0: /* This Source Code Form is subject to the terms of the Mozilla Public
michael@0:  * License, v. 2.0. If a copy of the MPL was not distributed with this
michael@0:  * file, You can obtain one at http://mozilla.org/MPL/2.0/. */
michael@0: #include "nsISO2022KRToUnicode.h"
michael@0: #include "nsUCSupport.h"
michael@0: #include "nsICharsetConverterManager.h"
michael@0: #include "nsServiceManagerUtils.h"
michael@0: 
michael@0: NS_IMETHODIMP nsISO2022KRToUnicode::Convert(const char * aSrc, int32_t * aSrcLen, char16_t * aDest, int32_t * aDestLen)
michael@0: {
michael@0:   static NS_DEFINE_CID(kCharsetConverterManagerCID, NS_ICHARSETCONVERTERMANAGER_CID);
michael@0: 
michael@0:   const unsigned char* srcEnd = (unsigned char*)aSrc + *aSrcLen;
michael@0:   const unsigned char* src =(unsigned char*) aSrc;
michael@0:   char16_t* destEnd = aDest + *aDestLen;
michael@0:   char16_t* dest = aDest;
michael@0:   while((src < srcEnd))
michael@0:   {
michael@0:     // if LF/CR, return to US-ASCII unconditionally.
michael@0:     if ( *src == 0x0a || *src == 0x0d )
michael@0:       mState = mState_Init;
michael@0: 
michael@0:     switch(mState)
michael@0:     {
michael@0:       case mState_Init:
michael@0:         if(0x1b == *src) {
michael@0:           mLastLegalState = mState_ASCII;
michael@0:           mState = mState_ESC;
michael@0:           break;
michael@0:         }
michael@0:         mState = mState_ASCII;
michael@0:         // fall through
michael@0: 
michael@0:       case mState_ASCII:
michael@0:         if(0x0e == *src) { // Shift-Out 
michael@0:           mState = mState_KSX1001_1992;
michael@0:           mRunLength = 0;
michael@0:         } 
michael@0:         else if(*src & 0x80) {
michael@0:           if (CHECK_OVERRUN(dest, destEnd, 1))
michael@0:             goto error1;
michael@0:           *dest++ = 0xFFFD;
michael@0:         } 
michael@0:         else {
michael@0:           if (CHECK_OVERRUN(dest, destEnd, 1))
michael@0:             goto error1;
michael@0:           *dest++ = (char16_t) *src;
michael@0:         }
michael@0:         break;
michael@0:           
michael@0:       case mState_ESC:
michael@0:         if('$' == *src) {
michael@0:           mState = mState_ESC_24;
michael@0:         } 
michael@0:         else  {
michael@0:           if (CHECK_OVERRUN(dest, destEnd, 2))
michael@0:             goto error1;
michael@0:           *dest++ = (char16_t) 0x1b;
michael@0:           *dest++ = (0x80 & *src) ? 0xFFFD : (char16_t) *src;
michael@0:           mState =  mLastLegalState;
michael@0:         }
michael@0:         break;
michael@0: 
michael@0:       case mState_ESC_24: // ESC $
michael@0:         if(')' == *src) {
michael@0:           mState = mState_ESC_24_29;
michael@0:         } 
michael@0:         else  {
michael@0:           if (CHECK_OVERRUN(dest, destEnd, 3))
michael@0:             goto error1;
michael@0:           *dest++ = (char16_t) 0x1b;
michael@0:           *dest++ = (char16_t) '$';
michael@0:           *dest++ = (0x80 & *src) ? 0xFFFD : (char16_t) *src;
michael@0:           mState = mLastLegalState;
michael@0:         }
michael@0:         break;
michael@0: 
michael@0:       case mState_ESC_24_29: // ESC $ )
michael@0:         mState = mLastLegalState;
michael@0:         if('C' == *src) {
michael@0:           mState = mState_ASCII;
michael@0:           mRunLength = 0;
michael@0:         } 
michael@0:         else  {
michael@0:           if (CHECK_OVERRUN(dest, destEnd, 4))
michael@0:             goto error1;
michael@0:           *dest++ = (char16_t) 0x1b;
michael@0:           *dest++ = (char16_t) '$';
michael@0:           *dest++ = (char16_t) ')';
michael@0:           *dest++ = (0x80 & *src) ? 0xFFFD : (char16_t) *src;
michael@0:           mState = mLastLegalState;
michael@0:         }
michael@0:         break;
michael@0: 
michael@0:       case mState_KSX1001_1992:
michael@0:         if (0x20 < (uint8_t) *src  && (uint8_t) *src < 0x7f) {
michael@0:           mData = (uint8_t) *src;
michael@0:           mState = mState_KSX1001_1992_2ndbyte;
michael@0:         } 
michael@0:         else if (0x0f == *src) { // Shift-In (SI)
michael@0:           mState = mState_ASCII;
michael@0:           if (mRunLength == 0) {
michael@0:             if (CHECK_OVERRUN(dest, destEnd, 1))
michael@0:               goto error1;
michael@0:             *dest++ = 0xFFFD;
michael@0:           }
michael@0:           mRunLength = 0;
michael@0:         } 
michael@0:         else if ((uint8_t) *src == 0x20 || (uint8_t) *src == 0x09) {
michael@0:           // Allow space and tab between SO and SI (i.e. in Hangul segment)
michael@0:           if (CHECK_OVERRUN(dest, destEnd, 1))
michael@0:             goto error1;
michael@0:           mState = mState_KSX1001_1992;
michael@0:           *dest++ = (char16_t) *src;
michael@0:           ++mRunLength;
michael@0:         } 
michael@0:         else {         // Everything else is invalid.
michael@0:           if (CHECK_OVERRUN(dest, destEnd, 1))
michael@0:             goto error1;
michael@0:           *dest++ = 0xFFFD;
michael@0:         }
michael@0:         break;
michael@0: 
michael@0:       case mState_KSX1001_1992_2ndbyte:
michael@0:         if ( 0x20 < (uint8_t) *src && (uint8_t) *src < 0x7f  ) {
michael@0:           if (!mEUCKRDecoder) {
michael@0:             // creating a delegate converter (EUC-KR)
michael@0:             nsresult rv;
michael@0:             nsCOMPtr<nsICharsetConverterManager> ccm = 
michael@0:                   do_GetService(kCharsetConverterManagerCID, &rv);
michael@0:             if (NS_SUCCEEDED(rv)) {
michael@0:               rv = ccm->GetUnicodeDecoderRaw("EUC-KR", &mEUCKRDecoder);
michael@0:             }
michael@0:           }
michael@0: 
michael@0:           if (!mEUCKRDecoder) {// failed creating a delegate converter
michael@0:            *dest++ = 0xFFFD;
michael@0:           } 
michael@0:           else {              
michael@0:             if (CHECK_OVERRUN(dest, destEnd, 1))
michael@0:               goto error1;
michael@0:             unsigned char ksx[2];
michael@0:             char16_t uni;
michael@0:             int32_t ksxLen = 2, uniLen = 1;
michael@0:             // mData is the original 1st byte.
michael@0:             // *src is the present 2nd byte.
michael@0:             // Put 2 bytes (one character) to ksx[] with EUC-KR encoding.
michael@0:             ksx[0] = mData | 0x80;
michael@0:             ksx[1] = *src | 0x80;
michael@0:             // Convert EUC-KR to unicode.
michael@0:             mEUCKRDecoder->Convert((const char *)ksx, &ksxLen, &uni, &uniLen);
michael@0:             *dest++ = uni;
michael@0:             ++mRunLength;
michael@0:           }
michael@0:           mState = mState_KSX1001_1992;
michael@0:         } 
michael@0:         else {        // Invalid 
michael@0:           if ( 0x0f == *src ) {   // Shift-In (SI)
michael@0:             mState = mState_ASCII;
michael@0:           } 
michael@0:           else {
michael@0:             mState = mState_KSX1001_1992;
michael@0:           }
michael@0:           if (CHECK_OVERRUN(dest, destEnd, 1))
michael@0:             goto error1;
michael@0:           *dest++ = 0xFFFD;
michael@0:         }
michael@0:         break;
michael@0: 
michael@0:       case mState_ERROR:
michael@0:         mState = mLastLegalState;
michael@0:         if (CHECK_OVERRUN(dest, destEnd, 1))
michael@0:           goto error1;
michael@0:         *dest++ = 0xFFFD;
michael@0:         break;
michael@0: 
michael@0:     } // switch
michael@0:     src++;
michael@0:   }
michael@0:   *aDestLen = dest - aDest;
michael@0:   return NS_OK;
michael@0: 
michael@0: error1:
michael@0:   *aDestLen = dest-aDest;
michael@0:   *aSrcLen = src-(unsigned char*)aSrc;
michael@0:   return NS_OK_UDEC_MOREOUTPUT;
michael@0: }
michael@0: