1.1 --- /dev/null Thu Jan 01 00:00:00 1970 +0000
1.2 +++ b/intl/uconv/ucvcn/nsHZToUnicode.cpp Wed Dec 31 06:09:35 2014 +0100
1.3 @@ -0,0 +1,176 @@
1.4 +/* -*- Mode: C; tab-width: 4; indent-tabs-mode: nil; c-basic-offset: 2 -*- */
1.5 +/* This Source Code Form is subject to the terms of the Mozilla Public
1.6 + * License, v. 2.0. If a copy of the MPL was not distributed with this
1.7 + * file, You can obtain one at http://mozilla.org/MPL/2.0/. */
1.8 +/**
1.9 + * A character set converter from HZ to Unicode.
1.10 + *
1.11 + *
1.12 + * @created 08/Sept/1999
1.13 + * @author Yueheng Xu, Yueheng.Xu@intel.com
1.14 + *
1.15 + * Note: in this HZ-GB-2312 converter, we accept a string composed of 7-bit HZ
1.16 + * encoded Chinese chars,as it is defined in RFC1843 available at
1.17 + * http://www.cis.ohio-state.edu/htbin/rfc/rfc1843.html
1.18 + * and RFC1842 available at http://www.cis.ohio-state.edu/htbin/rfc/rfc1842.html.
1.19 + *
1.20 + * Earlier versions of the converter said:
1.21 + * "In an effort to match the similar extended capability of Microsoft
1.22 + * Internet Explorer 5.0. We also accept the 8-bit GB encoded chars
1.23 + * mixed in a HZ string.
1.24 + * But this should not be a recommendedd practice for HTML authors."
1.25 + * However, testing in current versions of IE shows that it only accepts
1.26 + * 8-bit characters when the converter is in GB state, and when in ASCII
1.27 + * state each single 8-bit character is converted to U+FFFD
1.28 + *
1.29 + * The priority of converting are as follows: first convert 8-bit GB code; then,
1.30 + * consume HZ ESC sequences such as '~{', '~}', '~~'; then, depending on the current
1.31 + * state ( default to ASCII state ) of the string, each 7-bit char is converted as an
1.32 + * ASCII, or two 7-bit chars are converted into a Chinese character.
1.33 + */
1.34 +
1.35 +
1.36 +
1.37 +#include "nsHZToUnicode.h"
1.38 +#include "gbku.h"
1.39 +#include "mozilla/Telemetry.h"
1.40 +
1.41 +//----------------------------------------------------------------------
1.42 +// Class nsHZToUnicode [implementation]
1.43 +
1.44 +//----------------------------------------------------------------------
1.45 +// Subclassing of nsTablesDecoderSupport class [implementation]
1.46 +
1.47 +#define HZ_STATE_GB 1
1.48 +#define HZ_STATE_ASCII 2
1.49 +#define HZ_STATE_ODD_BYTE_FLAG 0x80
1.50 +#define HZLEAD1 '~'
1.51 +#define HZLEAD2 '{'
1.52 +#define HZLEAD3 '}'
1.53 +#define HZ_ODD_BYTE_STATE (mHZState & (HZ_STATE_ODD_BYTE_FLAG))
1.54 +#define HZ_ENCODING_STATE (mHZState & ~(HZ_STATE_ODD_BYTE_FLAG))
1.55 +
1.56 +using namespace mozilla;
1.57 +
1.58 +nsHZToUnicode::nsHZToUnicode() : nsBufferDecoderSupport(1)
1.59 +{
1.60 + mHZState = HZ_STATE_ASCII; // per HZ spec, default to ASCII state
1.61 + mRunLength = 0;
1.62 + mOddByte = 0;
1.63 + Telemetry::Accumulate(Telemetry::DECODER_INSTANTIATED_HZ, true);
1.64 +}
1.65 +
1.66 +//Overwriting the ConvertNoBuff() in nsUCvCnSupport.cpp.
1.67 +NS_IMETHODIMP nsHZToUnicode::ConvertNoBuff(
1.68 + const char* aSrc,
1.69 + int32_t * aSrcLength,
1.70 + char16_t *aDest,
1.71 + int32_t * aDestLength)
1.72 +{
1.73 + int32_t i=0;
1.74 + int32_t iSrcLength = *aSrcLength;
1.75 + int32_t iDestlen = 0;
1.76 + *aSrcLength=0;
1.77 + nsresult res = NS_OK;
1.78 + char oddByte = mOddByte;
1.79 +
1.80 + for (i=0; i<iSrcLength; i++) {
1.81 + if (iDestlen >= (*aDestLength)) {
1.82 + res = NS_OK_UDEC_MOREOUTPUT;
1.83 + break;
1.84 + }
1.85 +
1.86 + char srcByte = *aSrc++;
1.87 + (*aSrcLength)++;
1.88 +
1.89 + if (!HZ_ODD_BYTE_STATE) {
1.90 + if (srcByte == HZLEAD1 ||
1.91 + (HZ_ENCODING_STATE == HZ_STATE_GB &&
1.92 + (UINT8_IN_RANGE(0x21, srcByte, 0x7E) ||
1.93 + UINT8_IN_RANGE(0x81, srcByte, 0xFE)))) {
1.94 + oddByte = srcByte;
1.95 + mHZState |= HZ_STATE_ODD_BYTE_FLAG;
1.96 + } else {
1.97 + *aDest++ = (srcByte & 0x80) ? UCS2_NO_MAPPING :
1.98 + CAST_CHAR_TO_UNICHAR(srcByte);
1.99 + iDestlen++;
1.100 + }
1.101 + } else {
1.102 + if (oddByte & 0x80) {
1.103 + // Accept legal 8-bit GB 2312-80 sequences in GB mode only
1.104 + NS_ASSERTION(HZ_ENCODING_STATE == HZ_STATE_GB,
1.105 + "Invalid lead byte in ASCII mode");
1.106 + *aDest++ = (UINT8_IN_RANGE(0x81, oddByte, 0xFE) &&
1.107 + UINT8_IN_RANGE(0x40, srcByte, 0xFE)) ?
1.108 + mUtil.GBKCharToUnicode(oddByte, srcByte) : UCS2_NO_MAPPING;
1.109 + mRunLength++;
1.110 + iDestlen++;
1.111 + // otherwise, it is a 7-bit byte
1.112 + // The source will be an ASCII or a 7-bit HZ code depending on oddByte
1.113 + } else if (oddByte == HZLEAD1) { // if it is lead by '~'
1.114 + switch (srcByte) {
1.115 + case HZLEAD2:
1.116 + // we got a '~{'
1.117 + // we are switching to HZ state
1.118 + mHZState = HZ_STATE_GB;
1.119 + mRunLength = 0;
1.120 + break;
1.121 +
1.122 + case HZLEAD3:
1.123 + // we got a '~}'
1.124 + // we are switching to ASCII state
1.125 + mHZState = HZ_STATE_ASCII;
1.126 + if (mRunLength == 0) {
1.127 + *aDest++ = UCS2_NO_MAPPING;
1.128 + iDestlen++;
1.129 + }
1.130 + mRunLength = 0;
1.131 + break;
1.132 +
1.133 + case HZLEAD1:
1.134 + // we got a '~~', process like an ASCII, but no state change
1.135 + *aDest++ = CAST_CHAR_TO_UNICHAR(srcByte);
1.136 + iDestlen++;
1.137 + mRunLength++;
1.138 + break;
1.139 +
1.140 + default:
1.141 + // Undefined ESC sequence '~X': treat as an error if X is a
1.142 + // printable character or we are in ASCII mode, and resynchronize
1.143 + // on the second character.
1.144 + //
1.145 + // N.B. For compatibility with other implementations, we treat '~\n'
1.146 + // as an illegal sequence even though RFC1843 permits it, and for
1.147 + // the same reason we pass through control characters including '\n'
1.148 + // and ' ' even in GB mode.
1.149 + if (srcByte > 0x20 || HZ_ENCODING_STATE == HZ_STATE_ASCII) {
1.150 + *aDest++ = UCS2_NO_MAPPING;
1.151 + iDestlen++;
1.152 + }
1.153 + aSrc--;
1.154 + (*aSrcLength)--;
1.155 + i--;
1.156 + break;
1.157 + }
1.158 + } else if (HZ_ENCODING_STATE == HZ_STATE_GB) {
1.159 + *aDest++ = (UINT8_IN_RANGE(0x21, oddByte, 0x7E) &&
1.160 + UINT8_IN_RANGE(0x21, srcByte, 0x7E)) ?
1.161 + mUtil.GBKCharToUnicode(oddByte|0x80, srcByte|0x80) :
1.162 + UCS2_NO_MAPPING;
1.163 + mRunLength++;
1.164 + iDestlen++;
1.165 + } else {
1.166 + NS_NOTREACHED("2-byte sequence that we don't know how to handle");
1.167 + *aDest++ = UCS2_NO_MAPPING;
1.168 + iDestlen++;
1.169 + }
1.170 + oddByte = 0;
1.171 + mHZState &= ~HZ_STATE_ODD_BYTE_FLAG;
1.172 + }
1.173 + } // for loop
1.174 + mOddByte = HZ_ODD_BYTE_STATE ? oddByte : 0;
1.175 + *aDestLength = iDestlen;
1.176 + return res;
1.177 +}
1.178 +
1.179 +
