fbreader/src/formats/rtf/RtfReader.cpp

   1 /*
   2  * Copyright (C) 2004-2008 Geometer Plus <contact@geometerplus.com>
   3  *
   4  * This program is free software; you can redistribute it and/or modify
   5  * it under the terms of the GNU General Public License as published by
   6  * the Free Software Foundation; either version 2 of the License, or
   7  * (at your option) any later version.
   8  *
   9  * This program is distributed in the hope that it will be useful,
  10  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  11  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  12  * GNU General Public License for more details.
  13  *
  14  * You should have received a copy of the GNU General Public License
  15  * along with this program; if not, write to the Free Software
  16  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
  17  * 02110-1301, USA.
  18  */
  19
  20 #include <stdlib.h>
  21
  22 #include <cctype>
  23
  24 #include <ZLFile.h>
  25 #include <ZLInputStream.h>
  26
  27 #include "RtfReader.h"
  28
  29 std::map<std::string, RtfCommand*> RtfReader::ourKeywordMap;
  30
  31 static const int rtfStreamBufferSize = 4096;
  32
  33 RtfReader::RtfReader(const std::string &encoding) : EncodedTextReader(encoding) {
  34 }
  35
  36 RtfReader::~RtfReader() {
  37 }
  38
  39 RtfCommand::~RtfCommand() {
  40 }
  41
  42 void RtfNewParagraphCommand::run(RtfReader &reader, int*) const {
  43         reader.newParagraph();
  44 }
  45
  46 RtfFontPropertyCommand::RtfFontPropertyCommand(RtfReader::FontProperty property) : myProperty(property) {
  47 }
  48
  49 void RtfFontPropertyCommand::run(RtfReader &reader, int *parameter) const {
  50         bool start = (parameter == 0) || (*parameter != 0);
  51         switch (myProperty) {
  52                 case RtfReader::FONT_BOLD:
  53                         if (reader.myState.Bold != start) {
  54                                 reader.myState.Bold = start;
  55                                 reader.setFontProperty(RtfReader::FONT_BOLD);
  56                         }
  57                         break;
  58                 case RtfReader::FONT_ITALIC:
  59                         if (reader.myState.Italic != start) {
  60                                 reader.myState.Italic = start;
  61                                 reader.setFontProperty(RtfReader::FONT_ITALIC);
  62                         }
  63                         break;
  64                 case RtfReader::FONT_UNDERLINED:
  65                         if (reader.myState.Underlined != start) {
  66                                 reader.myState.Underlined = start;
  67                                 reader.setFontProperty(RtfReader::FONT_UNDERLINED);
  68                         }
  69                         break;
  70         }
  71 }
  72
  73 RtfAlignmentCommand::RtfAlignmentCommand(ZLTextAlignmentType alignment) : myAlignment(alignment) {
  74 }
  75
  76 void RtfAlignmentCommand::run(RtfReader &reader, int*) const {
  77         if (reader.myState.Alignment != myAlignment) {
  78                 reader.myState.Alignment = myAlignment;
  79                 reader.setAlignment();
  80         }
  81 }
  82
  83 RtfCharCommand::RtfCharCommand(const std::string &chr) : myChar(chr) {
  84 }
  85
  86 void RtfCharCommand::run(RtfReader &reader, int*) const {
  87         reader.processCharData(myChar.data(), myChar.length(), false);
  88 }
  89
  90 RtfDestinationCommand::RtfDestinationCommand(RtfReader::DestinationType destination) : myDestination(destination) {
  91 }
  92
  93 void RtfDestinationCommand::run(RtfReader &reader, int*) const {
  94         if (reader.myState.Destination == myDestination) {
  95                 return;
  96         }
  97         reader.myState.Destination = myDestination;
  98         if (myDestination == RtfReader::DESTINATION_PICTURE) {
  99                 reader.myState.ReadDataAsHex = true;
 100         }
 101         reader.switchDestination(myDestination, true);
 102 }
 103
 104 void RtfStyleCommand::run(RtfReader &reader, int*) const {
 105         if (reader.myState.Destination == RtfReader::DESTINATION_STYLESHEET) {
 106                 //std::cerr << "Add style index: " << val << "\n";
 107
 108                 //sprintf(style_attributes[0], "%i", val);
 109         } else /*if (myState.Destination == rdsContent)*/ {
 110                 //std::cerr << "Set style index: " << val << "\n";
 111
 112                 //sprintf(style_attributes[0], "%i", val);
 113         }
 114 }
 115
 116 void RtfCodepageCommand::run(RtfReader &reader, int *parameter) const {
 117         if (parameter != 0) {
 118                 reader.setEncoding(*parameter);
 119         }
 120 }
 121
 122 void RtfSpecialCommand::run(RtfReader &reader, int*) const {
 123         reader.mySpecialMode = true;
 124 }
 125
 126 RtfPictureCommand::RtfPictureCommand(const std::string &mimeType) : myMimeType(mimeType) {
 127 }
 128
 129 void RtfPictureCommand::run(RtfReader &reader, int*) const {
 130         reader.myNextImageMimeType = myMimeType;
 131 }
 132
 133 void RtfFontResetCommand::run(RtfReader &reader, int*) const {
 134         if (reader.myState.Bold) {
 135                 reader.myState.Bold = false;
 136                 reader.setFontProperty(RtfReader::FONT_BOLD);
 137         }
 138         if (reader.myState.Italic) {
 139                 reader.myState.Italic = false;
 140                 reader.setFontProperty(RtfReader::FONT_ITALIC);
 141         }
 142         if (reader.myState.Underlined) {
 143                 reader.myState.Underlined = false;
 144                 reader.setFontProperty(RtfReader::FONT_UNDERLINED);
 145         }
 146 }
 147
 148 void RtfReader::addAction(const std::string &tag, RtfCommand *command) {
 149         ourKeywordMap.insert(std::pair<std::string,RtfCommand*>(tag, command));
 150 }
 151
 152 void RtfReader::fillKeywordMap() {
 153         if (ourKeywordMap.empty()) {
 154                 addAction("*",  new RtfSpecialCommand());
 155                 addAction("ansicpg",    new RtfCodepageCommand());
 156
 157                 static const char *keywordsToSkip[] = {"buptim", "colortbl", "comment", "creatim", "doccomm", "fonttbl", "footer", "footerf", "footerl", "footerr", "ftncn", "ftnsep", "ftnsepc", "header", "headerf", "headerl", "headerr", "keywords", "operator", "printim", "private1", "revtim", "rxe", "subject", "tc", "txe", "xe", 0};
 158                 RtfCommand *skipCommand = new RtfDestinationCommand(RtfReader::DESTINATION_SKIP);
 159                 for (const char **i = keywordsToSkip; *i != 0; ++i) {
 160                         addAction(*i,   skipCommand);
 161                 }
 162                 addAction("info",       new RtfDestinationCommand(RtfReader::DESTINATION_INFO));
 163                 addAction("title",      new RtfDestinationCommand(RtfReader::DESTINATION_TITLE));
 164                 addAction("author",     new RtfDestinationCommand(RtfReader::DESTINATION_AUTHOR));
 165                 addAction("pict",       new RtfDestinationCommand(RtfReader::DESTINATION_PICTURE));
 166                 addAction("stylesheet", new RtfDestinationCommand(RtfReader::DESTINATION_STYLESHEET));
 167                 addAction("footnote",   new RtfDestinationCommand(RtfReader::DESTINATION_FOOTNOTE));
 168
 169                 RtfCommand *newParagraphCommand = new RtfNewParagraphCommand();
 170                 addAction("\n", newParagraphCommand);
 171                 addAction("\r", newParagraphCommand);
 172                 addAction("par",        newParagraphCommand);
 173
 174                 addAction("\x09",       new RtfCharCommand("\x09"));
 175                 addAction("_",  new RtfCharCommand("-"));
 176                 addAction("\\", new RtfCharCommand("\\"));
 177                 addAction("{",  new RtfCharCommand("{"));
 178                 addAction("}",  new RtfCharCommand("}"));
 179                 addAction("bullet",     new RtfCharCommand("\xE2\x80\xA2"));             // &bullet;
 180                 addAction("endash",     new RtfCharCommand("\xE2\x80\x93"));             // &ndash;
 181                 addAction("emdash",     new RtfCharCommand("\xE2\x80\x94"));             // &mdash;
 182                 addAction("~",  new RtfCharCommand("\xC0\xA0"));                                                        // &nbsp;
 183                 addAction("enspace",    new RtfCharCommand("\xE2\x80\x82"));            // &emsp;
 184                 addAction("emspace",    new RtfCharCommand("\xE2\x80\x83"));            // &ensp;
 185                 addAction("lquote",     new RtfCharCommand("\xE2\x80\x98"));             // &lsquo;
 186                 addAction("rquote",     new RtfCharCommand("\xE2\x80\x99"));             // &rsquo;
 187                 addAction("ldblquote",  new RtfCharCommand("\xE2\x80\x9C"));    // &ldquo;
 188                 addAction("rdblquote",  new RtfCharCommand("\xE2\x80\x9D"));    // &rdquo;
 189
 190                 addAction("jpegblip",   new RtfPictureCommand("image/jpeg"));
 191                 addAction("pngblip",    new RtfPictureCommand("image/png"));
 192
 193                 addAction("s",  new RtfStyleCommand());
 194
 195                 addAction("qc", new RtfAlignmentCommand(ALIGN_CENTER));
 196                 addAction("ql", new RtfAlignmentCommand(ALIGN_LEFT));
 197                 addAction("qr", new RtfAlignmentCommand(ALIGN_RIGHT));
 198                 addAction("qj", new RtfAlignmentCommand(ALIGN_JUSTIFY));
 199                 addAction("pard",       new RtfAlignmentCommand(ALIGN_UNDEFINED));
 200
 201                 addAction("b",  new RtfFontPropertyCommand(RtfReader::FONT_BOLD));
 202                 addAction("i",  new RtfFontPropertyCommand(RtfReader::FONT_ITALIC));
 203                 addAction("u",  new RtfFontPropertyCommand(RtfReader::FONT_UNDERLINED));
 204                 addAction("plain",      new RtfFontResetCommand());
 205         }
 206 }
 207
 208 bool RtfReader::parseDocument() {
 209         enum {
 210                 READ_NORMAL_DATA,
 211                 READ_BINARY_DATA,
 212                 READ_HEX_SYMBOL,
 213                 READ_KEYWORD,
 214                 READ_KEYWORD_PARAMETER
 215         } parserState = READ_NORMAL_DATA;
 216
 217         std::string keyword;
 218         std::string parameterString;
 219         std::string hexString;
 220         int imageStartOffset = -1;
 221
 222         while (!myIsInterrupted) {
 223                 const char *ptr = myStreamBuffer;
 224                 const char *end = myStreamBuffer + myStream->read(myStreamBuffer, rtfStreamBufferSize);
 225                 if (ptr == end) {
 226                         break;
 227                 }
 228                 const char *dataStart = ptr;
 229                 bool readNextChar = true;
 230                 while (ptr != end) {
 231                         switch (parserState) {
 232                                 case READ_BINARY_DATA:
 233                                         // TODO: optimize
 234                                         processCharData(ptr, 1);
 235                                         --myBinaryDataSize;
 236                                         if (myBinaryDataSize == 0) {
 237                                                 parserState = READ_NORMAL_DATA;
 238                                         }
 239                                         break;
 240                                 case READ_NORMAL_DATA:
 241                                         switch (*ptr) {
 242                                                 case '{':
 243                                                         if (ptr > dataStart) {
 244                                                                 processCharData(dataStart, ptr - dataStart);
 245                                                         }
 246                                                         dataStart = ptr + 1;
 247                                                         myStateStack.push(myState);
 248                                                         myState.ReadDataAsHex = false;
 249                                                         break;
 250                                                 case '}':
 251                                                 {
 252                                                         if (ptr > dataStart) {
 253                                                                 processCharData(dataStart, ptr - dataStart);
 254                                                         }
 255                                                         dataStart = ptr + 1;
 256
 257                                                         if (imageStartOffset >= 0) {
 258                                                                 int imageSize = myStream->offset() + (ptr - end) - imageStartOffset;
 259                                                                 insertImage(myNextImageMimeType, myFileName, imageStartOffset, imageSize);
 260                                                                 imageStartOffset = -1;
 261                                                         }
 262
 263                                                         if (myStateStack.empty()) {
 264                                                                 return false;
 265                                                         }
 266
 267                                                         if (myState.Destination != myStateStack.top().Destination) {
 268                                                                 switchDestination(myState.Destination, false);
 269                                                                 switchDestination(myStateStack.top().Destination, true);
 270                                                         }
 271
 272                                                         bool oldItalic = myState.Italic;
 273                                                         bool oldBold = myState.Bold;
 274                                                         bool oldUnderlined = myState.Underlined;
 275                                                         ZLTextAlignmentType oldAlignment = myState.Alignment;
 276                                                         myState = myStateStack.top();
 277                                                         myStateStack.pop();
 278
 279                                                         if (myState.Italic != oldItalic) {
 280                                                                 setFontProperty(RtfReader::FONT_ITALIC);
 281                                                         }
 282                                                         if (myState.Bold != oldBold) {
 283                                                                 setFontProperty(RtfReader::FONT_BOLD);
 284                                                         }
 285                                                         if (myState.Underlined != oldUnderlined) {
 286                                                                 setFontProperty(RtfReader::FONT_UNDERLINED);
 287                                                         }
 288                                                         if (myState.Alignment != oldAlignment) {
 289                                                                 setAlignment();
 290                                                         }
 291
 292                                                         break;
 293                                                 }
 294                                                 case '\\':
 295                                                         if (ptr > dataStart) {
 296                                                                 processCharData(dataStart, ptr - dataStart);
 297                                                         }
 298                                                         dataStart = ptr + 1;
 299                                                         keyword.erase();
 300                                                         parserState = READ_KEYWORD;
 301                                                         break;
 302                                                 case 0x0d:
 303                                                 case 0x0a:                      // cr and lf are noise characters...
 304                                                         if (ptr > dataStart) {
 305                                                                 processCharData(dataStart, ptr - dataStart);
 306                                                         }
 307                                                         dataStart = ptr + 1;
 308                                                         break;
 309                                                 default:
 310                                                         if (myState.ReadDataAsHex) {
 311                                                                 if (imageStartOffset == -1) {
 312                                                                         imageStartOffset = myStream->offset() + (ptr - end);
 313                                                                 }
 314                                                         }
 315                                                         break;
 316                                         }
 317                                         break;
 318                                 case READ_HEX_SYMBOL:
 319                                         hexString += *ptr;
 320                                         if (hexString.size() == 2) {
 321                                                 char ch = strtol(hexString.c_str(), 0, 16);
 322                                                 hexString.erase();
 323                                                 processCharData(&ch, 1);
 324                                                 parserState = READ_NORMAL_DATA;
 325                                                 dataStart = ptr + 1;
 326                                         }
 327                                         break;
 328                                 case READ_KEYWORD:
 329                                         if (!isalpha(*ptr)) {
 330                                                 if ((ptr == dataStart) && (keyword.empty())) {
 331                                                         if (*ptr == '\'') {
 332                                                                 parserState = READ_HEX_SYMBOL;
 333                                                         } else {
 334                                                                 keyword = *ptr;
 335                                                                 processKeyword(keyword);
 336                                                                 parserState = READ_NORMAL_DATA;
 337                                                         }
 338                                                         dataStart = ptr + 1;
 339                                                 } else {
 340                                                         keyword.append(dataStart, ptr - dataStart);
 341                                                         if ((*ptr == '-') || isdigit(*ptr)) {
 342                                                                 dataStart = ptr;
 343                                                                 parserState = READ_KEYWORD_PARAMETER;
 344                                                         } else {
 345                                                                 readNextChar = *ptr == ' ';
 346                                                                 processKeyword(keyword);
 347                                                                 parserState = READ_NORMAL_DATA;
 348                                                                 dataStart = readNextChar ? ptr + 1 : ptr;
 349                                                         }
 350                                                 }
 351                                         }
 352                                         break;
 353                                 case READ_KEYWORD_PARAMETER:
 354                                         if (!isdigit(*ptr)) {
 355                                                 parameterString.append(dataStart, ptr - dataStart);
 356                                                 int parameter = atoi(parameterString.c_str());
 357                                                 parameterString.erase();
 358                                                 readNextChar = *ptr == ' ';
 359                                                 if ((keyword == "bin") && (parameter > 0)) {
 360                                                         myBinaryDataSize = parameter;
 361                                                         parserState = READ_BINARY_DATA;
 362                                                 } else {
 363                                                         processKeyword(keyword, &parameter);
 364                                                         parserState = READ_NORMAL_DATA;
 365                                                 }
 366                                                 dataStart = readNextChar ? ptr + 1 : ptr;
 367                                         }
 368                                         break;
 369                         }
 370                         if (readNextChar) {
 371                                 ++ptr;
 372                         } else {
 373                                 readNextChar = true;
 374                         }
 375                 }
 376                 if (dataStart < end) {
 377                         switch (parserState) {
 378                                 case READ_NORMAL_DATA:
 379                                         processCharData(dataStart, end - dataStart);
 380                                 case READ_KEYWORD:
 381                                         keyword.append(dataStart, end - dataStart);
 382                                         break;
 383                                 case READ_KEYWORD_PARAMETER:
 384                                         parameterString.append(dataStart, end - dataStart);
 385                                         break;
 386                                 default:
 387                                         break;
 388                         }
 389                 }
 390         }
 391
 392         return myIsInterrupted || myStateStack.empty();
 393 }
 394
 395 void RtfReader::processKeyword(const std::string &keyword, int *parameter) {
 396         bool wasSpecialMode = mySpecialMode;
 397         mySpecialMode = false;
 398         if (myState.Destination == RtfReader::DESTINATION_SKIP) {
 399                 return;
 400         }
 401
 402         std::map<std::string, RtfCommand*>::const_iterator it = ourKeywordMap.find(keyword);
 403
 404         if (it == ourKeywordMap.end()) {
 405                 if (wasSpecialMode)
 406                         myState.Destination = RtfReader::DESTINATION_SKIP;
 407                 return;
 408         }
 409
 410         it->second->run(*this, parameter);
 411 }
 412
 413
 414 void RtfReader::processCharData(const char *data, size_t len, bool convert) {
 415         if (myState.Destination != RtfReader::DESTINATION_SKIP) {
 416                 addCharData(data, len, convert);
 417         }
 418 }
 419
 420 void RtfReader::interrupt() {
 421         myIsInterrupted = true;
 422 }
 423
 424 bool RtfReader::readDocument(const std::string &fileName) {
 425         myFileName = fileName;
 426         myStream = ZLFile(fileName).inputStream();
 427         if (myStream.isNull() || !myStream->open()) {
 428                         return false;
 429         }
 430
 431         fillKeywordMap();
 432
 433         myStreamBuffer = new char[rtfStreamBufferSize];
 434
 435         myIsInterrupted = false;
 436
 437         mySpecialMode = false;
 438
 439         myState.Alignment = ALIGN_UNDEFINED;
 440         myState.Italic = false;
 441         myState.Bold = false;
 442         myState.Underlined = false;
 443         myState.Destination = RtfReader::DESTINATION_NONE;
 444         myState.ReadDataAsHex = false;
 445
 446         bool code = parseDocument();
 447
 448         while (!myStateStack.empty()) {
 449                 myStateStack.pop();
 450         }
 451
 452         delete[] myStreamBuffer;
 453         myStream->close();
 454
 455         return code;
 456 }