2 * DecodedBitStreamParser.cpp
5 * Created by Christian Brunschen on 20/05/2008.
6 * Copyright 2008 ZXing authors All rights reserved.
8 * Licensed under the Apache License, Version 2.0 (the "License");
9 * you may not use this file except in compliance with the License.
10 * You may obtain a copy of the License at
12 * http://www.apache.org/licenses/LICENSE-2.0
14 * Unless required by applicable law or agreed to in writing, software
15 * distributed under the License is distributed on an "AS IS" BASIS,
16 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
17 * See the License for the specific language governing permissions and
18 * limitations under the License.
21 #include <zxing/qrcode/decoder/DecodedBitStreamParser.h>
27 // Required for compatibility. TODO: test on Symbian
28 #ifdef ZXING_ICONV_CONST
30 #define ICONV_CONST const
34 #define ICONV_CONST /**/
37 using namespace zxing;
44 const char DecodedBitStreamParser::ALPHANUMERIC_CHARS[] = { '0', '1', '2', '3', '4', '5', '6', '7', '8', '9', 'A', 'B',
45 'C', 'D', 'E', 'F', 'G', 'H', 'I', 'J', 'K', 'L', 'M', 'N', 'O', 'P', 'Q', 'R', 'S', 'T', 'U', 'V', 'W', 'X',
46 'Y', 'Z', ' ', '$', '%', '*', '+', '-', '.', '/', ':'
49 const char *DecodedBitStreamParser::ASCII = "ASCII";
50 const char *DecodedBitStreamParser::ISO88591 = "ISO-8859-1";
51 const char *DecodedBitStreamParser::UTF8 = "UTF-8";
52 const char *DecodedBitStreamParser::SHIFT_JIS = "SHIFT_JIS";
53 const char *DecodedBitStreamParser::EUC_JP = "EUC-JP";
55 string DecodedBitStreamParser::convert(const unsigned char *bufIn, size_t nIn, const char *src) {
61 iconv_t cd = iconv_open(UTF8, src);
62 const int maxOut = 4 * nIn + 1;
63 unsigned char* bufOut = new unsigned char[maxOut];
65 ICONV_CONST char *fromPtr = (ICONV_CONST char *)bufIn;
67 char *toPtr = (char *)bufOut;
71 size_t oneway = iconv(cd, &fromPtr, &nFrom, &toPtr, &nTo);
72 if (oneway == (size_t)(-1)) {
75 throw ReaderException("error converting characters");
80 int nResult = maxOut - nTo;
81 bufOut[nResult] = '\0';
82 string result((const char *)bufOut);
86 return string((const char *)bufIn, nIn);
90 string DecodedBitStreamParser::decodeKanjiSegment(Ref<BitSource> bits, int count) {
91 // Each character will require 2 bytes. Read the characters as 2-byte pairs
92 // and decode as Shift_JIS afterwards
93 size_t nBytes = 2 * count;
94 unsigned char* buffer = new unsigned char[nBytes];
97 // Each 13 bits encodes a 2-byte character
99 int twoBytes = bits->readBits(13);
100 int assembledTwoBytes = ((twoBytes / 0x0C0) << 8) | (twoBytes % 0x0C0);
101 if (assembledTwoBytes < 0x01F00) {
102 // In the 0x8140 to 0x9FFC range
103 assembledTwoBytes += 0x08140;
105 // In the 0xE040 to 0xEBBF range
106 assembledTwoBytes += 0x0C140;
108 buffer[offset] = (unsigned char)(assembledTwoBytes >> 8);
109 buffer[offset + 1] = (unsigned char)assembledTwoBytes;
114 string result = convert(buffer, nBytes, SHIFT_JIS);
119 string DecodedBitStreamParser::decodeByteSegment(Ref<BitSource> bits, int count) {
121 unsigned char* readBytes = new unsigned char[nBytes];
122 if (count << 3 > bits->available()) {
124 s << "Count too large: " << count;
126 throw ReaderException(s.str().c_str());
128 for (int i = 0; i < count; i++) {
129 readBytes[i] = (unsigned char)bits->readBits(8);
131 // The spec isn't clear on this mode; see
132 // section 6.4.5: t does not say which encoding to assuming
133 // upon decoding. I have seen ISO-8859-1 used as well as
134 // Shift_JIS -- without anything like an ECI designator to
136 const char *encoding = guessEncoding(readBytes, nBytes);
137 string result = convert(readBytes, nBytes, encoding);
142 string DecodedBitStreamParser::decodeNumericSegment(Ref<BitSource> bits, int count) {
144 unsigned char* bytes = new unsigned char[nBytes];
146 // Read three digits at a time
148 // Each 10 bits encodes three digits
149 int threeDigitsBits = bits->readBits(10);
150 if (threeDigitsBits >= 1000) {
152 s << "Illegal value for 3-digit unit: " << threeDigitsBits;
154 throw ReaderException(s.str().c_str());
156 bytes[i++] = ALPHANUMERIC_CHARS[threeDigitsBits / 100];
157 bytes[i++] = ALPHANUMERIC_CHARS[(threeDigitsBits / 10) % 10];
158 bytes[i++] = ALPHANUMERIC_CHARS[threeDigitsBits % 10];
162 // Two digits left over to read, encoded in 7 bits
163 int twoDigitsBits = bits->readBits(7);
164 if (twoDigitsBits >= 100) {
166 s << "Illegal value for 2-digit unit: " << twoDigitsBits;
168 throw ReaderException(s.str().c_str());
170 bytes[i++] = ALPHANUMERIC_CHARS[twoDigitsBits / 10];
171 bytes[i++] = ALPHANUMERIC_CHARS[twoDigitsBits % 10];
172 } else if (count == 1) {
173 // One digit left over to read
174 int digitBits = bits->readBits(4);
175 if (digitBits >= 10) {
177 s << "Illegal value for digit unit: " << digitBits;
179 throw ReaderException(s.str().c_str());
181 bytes[i++] = ALPHANUMERIC_CHARS[digitBits];
183 string result = convert(bytes, nBytes, ASCII);
188 string DecodedBitStreamParser::decodeAlphanumericSegment(Ref<BitSource> bits, int count) {
190 unsigned char* bytes = new unsigned char[nBytes];
192 // Read two characters at a time
194 int nextTwoCharsBits = bits->readBits(11);
195 bytes[i++] = ALPHANUMERIC_CHARS[nextTwoCharsBits / 45];
196 bytes[i++] = ALPHANUMERIC_CHARS[nextTwoCharsBits % 45];
200 bytes[i++] = ALPHANUMERIC_CHARS[bits->readBits(6)];
202 string result = convert(bytes, nBytes, ASCII);
208 DecodedBitStreamParser::guessEncoding(unsigned char *bytes, int length) {
209 // Does it start with the UTF-8 byte order mark? then guess it's UTF-8
210 if (length > 3 && bytes[0] == (unsigned char)0xEF && bytes[1] == (unsigned char)0xBB && bytes[2]
211 == (unsigned char)0xBF) {
214 // For now, merely tries to distinguish ISO-8859-1, UTF-8 and Shift_JIS,
215 // which should be by far the most common encodings. ISO-8859-1
216 // should not have bytes in the 0x80 - 0x9F range, while Shift_JIS
217 // uses this as a first byte of a two-byte character. If we see this
218 // followed by a valid second byte in Shift_JIS, assume it is Shift_JIS.
219 // If we see something else in that second byte, we'll make the risky guess
221 bool canBeISO88591 = true;
222 bool lastWasPossibleDoubleByteStart = false;
223 for (int i = 0; i < length; i++) {
224 int value = bytes[i] & 0xFF;
225 if (value >= 0x80 && value <= 0x9F && i < length - 1) {
226 canBeISO88591 = false;
227 // ISO-8859-1 shouldn't use this, but before we decide it is Shift_JIS,
228 // just double check that it is followed by a byte that's valid in
229 // the Shift_JIS encoding
230 if (lastWasPossibleDoubleByteStart) {
231 // If we just checked this and the last byte for being a valid double-byte
232 // char, don't check starting on this byte. If the this and the last byte
233 // formed a valid pair, then this shouldn't be checked to see if it starts
234 // a double byte pair of course.
235 lastWasPossibleDoubleByteStart = false;
237 // ... otherwise do check to see if this plus the next byte form a valid
238 // double byte pair encoding a character.
239 lastWasPossibleDoubleByteStart = true;
240 int nextValue = bytes[i + 1] & 0xFF;
241 if ((value & 0x1) == 0) {
242 // if even, next value should be in [0x9F,0xFC]
243 // if not, we'll guess UTF-8
244 if (nextValue < 0x9F || nextValue > 0xFC) {
248 // if odd, next value should be in [0x40,0x9E]
249 // if not, we'll guess UTF-8
250 if (nextValue < 0x40 || nextValue > 0x9E) {
257 return canBeISO88591 ? ISO88591 : SHIFT_JIS;
260 string DecodedBitStreamParser::decode(ArrayRef<unsigned char> bytes, Version *version) {
262 Ref<BitSource> bits(new BitSource(bytes));
263 Mode *mode = &Mode::TERMINATOR;
265 // While still another segment to read...
266 if (bits->available() < 4) {
267 // OK, assume we're done. Really, a TERMINATOR mode should have been recorded here
268 mode = &Mode::TERMINATOR;
270 mode = &Mode::forBits(bits->readBits(4)); // mode is encoded by 4 bits
272 if (mode != &Mode::TERMINATOR) {
273 // How many characters will follow, encoded in this mode?
274 int count = bits->readBits(mode->getCharacterCountBits(version));
275 if (mode == &Mode::NUMERIC) {
276 result = decodeNumericSegment(bits, count);
277 } else if (mode == &Mode::ALPHANUMERIC) {
278 result = decodeAlphanumericSegment(bits, count);
279 } else if (mode == &Mode::BYTE) {
280 result = decodeByteSegment(bits, count);
281 } else if (mode == &Mode::KANJI) {
282 result = decodeKanjiSegment(bits, count);
284 throw ReaderException("Unsupported mode indicator");
287 } while (mode != &Mode::TERMINATOR);