parser/html/nsHtml5Tokenizer.h

   1 /*
   2  * Copyright (c) 2005-2007 Henri Sivonen
   3  * Copyright (c) 2007-2010 Mozilla Foundation
   4  * Portions of comments Copyright 2004-2010 Apple Computer, Inc., Mozilla
   5  * Foundation, and Opera Software ASA.
   6  *
   7  * Permission is hereby granted, free of charge, to any person obtaining a
   8  * copy of this software and associated documentation files (the "Software"),
   9  * to deal in the Software without restriction, including without limitation
  10  * the rights to use, copy, modify, merge, publish, distribute, sublicense,
  11  * and/or sell copies of the Software, and to permit persons to whom the
  12  * Software is furnished to do so, subject to the following conditions:
  13  *
  14  * The above copyright notice and this permission notice shall be included in
  15  * all copies or substantial portions of the Software.
  16  *
  17  * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
  18  * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
  19  * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
  20  * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
  21  * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
  22  * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
  23  * DEALINGS IN THE SOFTWARE.
  24  */
  25
  26 /*
  27  * THIS IS A GENERATED FILE. PLEASE DO NOT EDIT.
  28  * Please edit Tokenizer.java instead and regenerate.
  29  */
  30
  31 #ifndef nsHtml5Tokenizer_h__
  32 #define nsHtml5Tokenizer_h__
  33
  34 #include "prtypes.h"
  35 #include "nsIAtom.h"
  36 #include "nsHtml5AtomTable.h"
  37 #include "nsString.h"
  38 #include "nsINameSpaceManager.h"
  39 #include "nsIContent.h"
  40 #include "nsIDocument.h"
  41 #include "nsTraceRefcnt.h"
  42 #include "jArray.h"
  43 #include "nsHtml5DocumentMode.h"
  44 #include "nsHtml5ArrayCopy.h"
  45 #include "nsHtml5NamedCharacters.h"
  46 #include "nsHtml5NamedCharactersAccel.h"
  47 #include "nsHtml5Atoms.h"
  48 #include "nsHtml5ByteReadable.h"
  49 #include "nsIUnicodeDecoder.h"
  50 #include "nsAHtml5TreeBuilderState.h"
  51 #include "nsHtml5Macros.h"
  52
  53 class nsHtml5StreamParser;
  54
  55 class nsHtml5TreeBuilder;
  56 class nsHtml5MetaScanner;
  57 class nsHtml5AttributeName;
  58 class nsHtml5ElementName;
  59 class nsHtml5HtmlAttributes;
  60 class nsHtml5UTF16Buffer;
  61 class nsHtml5StateSnapshot;
  62 class nsHtml5Portability;
  63
  64
  65 class nsHtml5Tokenizer
  66 {
  67   private:
  68     static PRUnichar LT_GT[];
  69     static PRUnichar LT_SOLIDUS[];
  70     static PRUnichar RSQB_RSQB[];
  71     static PRUnichar REPLACEMENT_CHARACTER[];
  72     static PRUnichar LF[];
  73     static PRUnichar CDATA_LSQB[];
  74     static PRUnichar OCTYPE[];
  75     static PRUnichar UBLIC[];
  76     static PRUnichar YSTEM[];
  77     static staticJArray<PRUnichar,PRInt32> TITLE_ARR;
  78     static staticJArray<PRUnichar,PRInt32> SCRIPT_ARR;
  79     static staticJArray<PRUnichar,PRInt32> STYLE_ARR;
  80     static staticJArray<PRUnichar,PRInt32> PLAINTEXT_ARR;
  81     static staticJArray<PRUnichar,PRInt32> XMP_ARR;
  82     static staticJArray<PRUnichar,PRInt32> TEXTAREA_ARR;
  83     static staticJArray<PRUnichar,PRInt32> IFRAME_ARR;
  84     static staticJArray<PRUnichar,PRInt32> NOEMBED_ARR;
  85     static staticJArray<PRUnichar,PRInt32> NOSCRIPT_ARR;
  86     static staticJArray<PRUnichar,PRInt32> NOFRAMES_ARR;
  87   protected:
  88     nsHtml5TreeBuilder* tokenHandler;
  89     nsHtml5StreamParser* encodingDeclarationHandler;
  90     PRBool lastCR;
  91     PRInt32 stateSave;
  92   private:
  93     PRInt32 returnStateSave;
  94   protected:
  95     PRInt32 index;
  96   private:
  97     PRBool forceQuirks;
  98     PRUnichar additional;
  99     PRInt32 entCol;
 100     PRInt32 firstCharKey;
 101     PRInt32 lo;
 102     PRInt32 hi;
 103     PRInt32 candidate;
 104     PRInt32 strBufMark;
 105     PRInt32 prevValue;
 106   protected:
 107     PRInt32 value;
 108   private:
 109     PRBool seenDigits;
 110   protected:
 111     PRInt32 cstart;
 112   private:
 113     nsString* publicId;
 114     nsString* systemId;
 115     autoJArray<PRUnichar,PRInt32> strBuf;
 116     PRInt32 strBufLen;
 117     autoJArray<PRUnichar,PRInt32> longStrBuf;
 118     PRInt32 longStrBufLen;
 119     autoJArray<PRUnichar,PRInt32> bmpChar;
 120     autoJArray<PRUnichar,PRInt32> astralChar;
 121   protected:
 122     nsHtml5ElementName* endTagExpectation;
 123   private:
 124     jArray<PRUnichar,PRInt32> endTagExpectationAsArray;
 125   protected:
 126     PRBool endTag;
 127   private:
 128     nsHtml5ElementName* tagName;
 129   protected:
 130     nsHtml5AttributeName* attributeName;
 131   private:
 132     nsIAtom* doctypeName;
 133     nsString* publicIdentifier;
 134     nsString* systemIdentifier;
 135     nsHtml5HtmlAttributes* attributes;
 136     PRInt32 mappingLangToXmlLang;
 137     PRBool shouldSuspend;
 138   protected:
 139     PRBool confident;
 140   private:
 141     PRInt32 line;
 142     nsHtml5AtomTable* interner;
 143   public:
 144     nsHtml5Tokenizer(nsHtml5TreeBuilder* tokenHandler);
 145     void setInterner(nsHtml5AtomTable* interner);
 146     void initLocation(nsString* newPublicId, nsString* newSystemId);
 147     void setStateAndEndTagExpectation(PRInt32 specialTokenizerState, nsIAtom* endTagExpectation);
 148     void setStateAndEndTagExpectation(PRInt32 specialTokenizerState, nsHtml5ElementName* endTagExpectation);
 149   private:
 150     void endTagExpectationToArray();
 151   public:
 152     void setLineNumber(PRInt32 line);
 153     inline PRInt32 getLineNumber()
 154     {
 155       return line;
 156     }
 157
 158     nsHtml5HtmlAttributes* emptyAttributes();
 159   private:
 160     inline void clearStrBufAndAppend(PRUnichar c)
 161     {
 162       strBuf[0] = c;
 163       strBufLen = 1;
 164     }
 165
 166     inline void clearStrBuf()
 167     {
 168       strBufLen = 0;
 169     }
 170
 171     void appendStrBuf(PRUnichar c);
 172   protected:
 173     nsString* strBufToString();
 174   private:
 175     void strBufToDoctypeName();
 176     void emitStrBuf();
 177     inline void clearLongStrBuf()
 178     {
 179       longStrBufLen = 0;
 180     }
 181
 182     inline void clearLongStrBufAndAppend(PRUnichar c)
 183     {
 184       longStrBuf[0] = c;
 185       longStrBufLen = 1;
 186     }
 187
 188     void appendLongStrBuf(PRUnichar c);
 189     inline void appendSecondHyphenToBogusComment()
 190     {
 191       appendLongStrBuf('-');
 192     }
 193
 194     inline void adjustDoubleHyphenAndAppendToLongStrBufAndErr(PRUnichar c)
 195     {
 196
 197       appendLongStrBuf(c);
 198     }
 199
 200     void appendLongStrBuf(PRUnichar* buffer, PRInt32 offset, PRInt32 length);
 201     inline void appendStrBufToLongStrBuf()
 202     {
 203       appendLongStrBuf(strBuf, 0, strBufLen);
 204     }
 205
 206     nsString* longStrBufToString();
 207     void emitComment(PRInt32 provisionalHyphens, PRInt32 pos);
 208   protected:
 209     void flushChars(PRUnichar* buf, PRInt32 pos);
 210   private:
 211     void resetAttributes();
 212     void strBufToElementNameString();
 213     PRInt32 emitCurrentTagToken(PRBool selfClosing, PRInt32 pos);
 214     void attributeNameComplete();
 215     void addAttributeWithoutValue();
 216     void addAttributeWithValue();
 217   public:
 218     void start();
 219     PRBool tokenizeBuffer(nsHtml5UTF16Buffer* buffer);
 220   private:
 221     PRInt32 stateLoop(PRInt32 state, PRUnichar c, PRInt32 pos, PRUnichar* buf, PRBool reconsume, PRInt32 returnState, PRInt32 endPos);
 222     void initDoctypeFields();
 223     inline void adjustDoubleHyphenAndAppendToLongStrBufCarriageReturn()
 224     {
 225       silentCarriageReturn();
 226       adjustDoubleHyphenAndAppendToLongStrBufAndErr('\n');
 227     }
 228
 229     inline void adjustDoubleHyphenAndAppendToLongStrBufLineFeed()
 230     {
 231       silentLineFeed();
 232       adjustDoubleHyphenAndAppendToLongStrBufAndErr('\n');
 233     }
 234
 235     inline void appendLongStrBufLineFeed()
 236     {
 237       silentLineFeed();
 238       appendLongStrBuf('\n');
 239     }
 240
 241     inline void appendLongStrBufCarriageReturn()
 242     {
 243       silentCarriageReturn();
 244       appendLongStrBuf('\n');
 245     }
 246
 247   protected:
 248     inline void silentCarriageReturn()
 249     {
 250       ++line;
 251       lastCR = PR_TRUE;
 252     }
 253
 254     inline void silentLineFeed()
 255     {
 256       ++line;
 257     }
 258
 259   private:
 260     void emitCarriageReturn(PRUnichar* buf, PRInt32 pos);
 261     void emitReplacementCharacter(PRUnichar* buf, PRInt32 pos);
 262     void setAdditionalAndRememberAmpersandLocation(PRUnichar add);
 263     void bogusDoctype();
 264     void bogusDoctypeWithoutQuirks();
 265     void emitOrAppendStrBuf(PRInt32 returnState);
 266     void handleNcrValue(PRInt32 returnState);
 267   public:
 268     void eof();
 269   private:
 270     void emitDoctypeToken(PRInt32 pos);
 271   protected:
 272     inline PRUnichar checkChar(PRUnichar* buf, PRInt32 pos)
 273     {
 274       return buf[pos];
 275     }
 276
 277   public:
 278     void internalEncodingDeclaration(nsString* internalCharset);
 279   private:
 280     void emitOrAppendTwo(const PRUnichar* val, PRInt32 returnState);
 281     void emitOrAppendOne(const PRUnichar* val, PRInt32 returnState);
 282   public:
 283     void end();
 284     void requestSuspension();
 285     PRBool isInDataState();
 286     void resetToDataState();
 287     void loadState(nsHtml5Tokenizer* other);
 288     void initializeWithoutStarting();
 289     void setEncodingDeclarationHandler(nsHtml5StreamParser* encodingDeclarationHandler);
 290     ~nsHtml5Tokenizer();
 291     static void initializeStatics();
 292     static void releaseStatics();
 293 };
 294
 295 #define NS_HTML5TOKENIZER_DATA_AND_RCDATA_MASK ~1
 296 #define NS_HTML5TOKENIZER_DATA 0
 297 #define NS_HTML5TOKENIZER_RCDATA 1
 298 #define NS_HTML5TOKENIZER_SCRIPT_DATA 2
 299 #define NS_HTML5TOKENIZER_RAWTEXT 3
 300 #define NS_HTML5TOKENIZER_SCRIPT_DATA_ESCAPED 4
 301 #define NS_HTML5TOKENIZER_ATTRIBUTE_VALUE_DOUBLE_QUOTED 5
 302 #define NS_HTML5TOKENIZER_ATTRIBUTE_VALUE_SINGLE_QUOTED 6
 303 #define NS_HTML5TOKENIZER_ATTRIBUTE_VALUE_UNQUOTED 7
 304 #define NS_HTML5TOKENIZER_PLAINTEXT 8
 305 #define NS_HTML5TOKENIZER_TAG_OPEN 9
 306 #define NS_HTML5TOKENIZER_CLOSE_TAG_OPEN 10
 307 #define NS_HTML5TOKENIZER_TAG_NAME 11
 308 #define NS_HTML5TOKENIZER_BEFORE_ATTRIBUTE_NAME 12
 309 #define NS_HTML5TOKENIZER_ATTRIBUTE_NAME 13
 310 #define NS_HTML5TOKENIZER_AFTER_ATTRIBUTE_NAME 14
 311 #define NS_HTML5TOKENIZER_BEFORE_ATTRIBUTE_VALUE 15
 312 #define NS_HTML5TOKENIZER_AFTER_ATTRIBUTE_VALUE_QUOTED 16
 313 #define NS_HTML5TOKENIZER_BOGUS_COMMENT 17
 314 #define NS_HTML5TOKENIZER_MARKUP_DECLARATION_OPEN 18
 315 #define NS_HTML5TOKENIZER_DOCTYPE 19
 316 #define NS_HTML5TOKENIZER_BEFORE_DOCTYPE_NAME 20
 317 #define NS_HTML5TOKENIZER_DOCTYPE_NAME 21
 318 #define NS_HTML5TOKENIZER_AFTER_DOCTYPE_NAME 22
 319 #define NS_HTML5TOKENIZER_BEFORE_DOCTYPE_PUBLIC_IDENTIFIER 23
 320 #define NS_HTML5TOKENIZER_DOCTYPE_PUBLIC_IDENTIFIER_DOUBLE_QUOTED 24
 321 #define NS_HTML5TOKENIZER_DOCTYPE_PUBLIC_IDENTIFIER_SINGLE_QUOTED 25
 322 #define NS_HTML5TOKENIZER_AFTER_DOCTYPE_PUBLIC_IDENTIFIER 26
 323 #define NS_HTML5TOKENIZER_BEFORE_DOCTYPE_SYSTEM_IDENTIFIER 27
 324 #define NS_HTML5TOKENIZER_DOCTYPE_SYSTEM_IDENTIFIER_DOUBLE_QUOTED 28
 325 #define NS_HTML5TOKENIZER_DOCTYPE_SYSTEM_IDENTIFIER_SINGLE_QUOTED 29
 326 #define NS_HTML5TOKENIZER_AFTER_DOCTYPE_SYSTEM_IDENTIFIER 30
 327 #define NS_HTML5TOKENIZER_BOGUS_DOCTYPE 31
 328 #define NS_HTML5TOKENIZER_COMMENT_START 32
 329 #define NS_HTML5TOKENIZER_COMMENT_START_DASH 33
 330 #define NS_HTML5TOKENIZER_COMMENT 34
 331 #define NS_HTML5TOKENIZER_COMMENT_END_DASH 35
 332 #define NS_HTML5TOKENIZER_COMMENT_END 36
 333 #define NS_HTML5TOKENIZER_COMMENT_END_BANG 37
 334 #define NS_HTML5TOKENIZER_NON_DATA_END_TAG_NAME 38
 335 #define NS_HTML5TOKENIZER_MARKUP_DECLARATION_HYPHEN 39
 336 #define NS_HTML5TOKENIZER_MARKUP_DECLARATION_OCTYPE 40
 337 #define NS_HTML5TOKENIZER_DOCTYPE_UBLIC 41
 338 #define NS_HTML5TOKENIZER_DOCTYPE_YSTEM 42
 339 #define NS_HTML5TOKENIZER_AFTER_DOCTYPE_PUBLIC_KEYWORD 43
 340 #define NS_HTML5TOKENIZER_BETWEEN_DOCTYPE_PUBLIC_AND_SYSTEM_IDENTIFIERS 44
 341 #define NS_HTML5TOKENIZER_AFTER_DOCTYPE_SYSTEM_KEYWORD 45
 342 #define NS_HTML5TOKENIZER_CONSUME_CHARACTER_REFERENCE 46
 343 #define NS_HTML5TOKENIZER_CONSUME_NCR 47
 344 #define NS_HTML5TOKENIZER_CHARACTER_REFERENCE_TAIL 48
 345 #define NS_HTML5TOKENIZER_HEX_NCR_LOOP 49
 346 #define NS_HTML5TOKENIZER_DECIMAL_NRC_LOOP 50
 347 #define NS_HTML5TOKENIZER_HANDLE_NCR_VALUE 51
 348 #define NS_HTML5TOKENIZER_HANDLE_NCR_VALUE_RECONSUME 52
 349 #define NS_HTML5TOKENIZER_CHARACTER_REFERENCE_HILO_LOOKUP 53
 350 #define NS_HTML5TOKENIZER_SELF_CLOSING_START_TAG 54
 351 #define NS_HTML5TOKENIZER_CDATA_START 55
 352 #define NS_HTML5TOKENIZER_CDATA_SECTION 56
 353 #define NS_HTML5TOKENIZER_CDATA_RSQB 57
 354 #define NS_HTML5TOKENIZER_CDATA_RSQB_RSQB 58
 355 #define NS_HTML5TOKENIZER_SCRIPT_DATA_LESS_THAN_SIGN 59
 356 #define NS_HTML5TOKENIZER_SCRIPT_DATA_ESCAPE_START 60
 357 #define NS_HTML5TOKENIZER_SCRIPT_DATA_ESCAPE_START_DASH 61
 358 #define NS_HTML5TOKENIZER_SCRIPT_DATA_ESCAPED_DASH 62
 359 #define NS_HTML5TOKENIZER_SCRIPT_DATA_ESCAPED_DASH_DASH 63
 360 #define NS_HTML5TOKENIZER_BOGUS_COMMENT_HYPHEN 64
 361 #define NS_HTML5TOKENIZER_RAWTEXT_RCDATA_LESS_THAN_SIGN 65
 362 #define NS_HTML5TOKENIZER_SCRIPT_DATA_ESCAPED_LESS_THAN_SIGN 66
 363 #define NS_HTML5TOKENIZER_SCRIPT_DATA_DOUBLE_ESCAPE_START 67
 364 #define NS_HTML5TOKENIZER_SCRIPT_DATA_DOUBLE_ESCAPED 68
 365 #define NS_HTML5TOKENIZER_SCRIPT_DATA_DOUBLE_ESCAPED_LESS_THAN_SIGN 69
 366 #define NS_HTML5TOKENIZER_SCRIPT_DATA_DOUBLE_ESCAPED_DASH 70
 367 #define NS_HTML5TOKENIZER_SCRIPT_DATA_DOUBLE_ESCAPED_DASH_DASH 71
 368 #define NS_HTML5TOKENIZER_SCRIPT_DATA_DOUBLE_ESCAPE_END 72
 369 #define NS_HTML5TOKENIZER_LEAD_OFFSET (0xD800 - (0x10000 >> 10))
 370 #define NS_HTML5TOKENIZER_BUFFER_GROW_BY 1024
 371
 372
 373 #endif
 374