1 /* Copyright (c) 1998, 1999, 2000 Thai Open Source Software Center Ltd 2 See the file COPYING for copying permission. 3 */ 4 5 #ifndef XmlParse_INCLUDED 6 #define XmlParse_INCLUDED 1 7 8 #ifdef __VMS 9 /* 0 1 2 3 0 1 2 3 10 1234567890123456789012345678901 1234567890123456789012345678901 */ 11 #define XML_SetProcessingInstructionHandler XML_SetProcessingInstrHandler 12 #define XML_SetUnparsedEntityDeclHandler XML_SetUnparsedEntDeclHandler 13 #define XML_SetStartNamespaceDeclHandler XML_SetStartNamespcDeclHandler 14 #define XML_SetExternalEntityRefHandlerArg XML_SetExternalEntRefHandlerArg 15 #endif 16 17 #include <stdlib.h> 18 #include "expat_external.h" 19 20 struct XML_ParserStruct; 21 typedef struct XML_ParserStruct *XML_Parser; 22 23 /* Should this be defined using stdbool.h when C99 is available? */ 24 typedef unsigned char XML_Bool; 25 #define XML_TRUE ((XML_Bool) 1) 26 #define XML_FALSE ((XML_Bool) 0) 27 28 /* The XML_Status enum gives the possible return values for several 29 API functions. The preprocessor #defines are included so this 30 stanza can be added to code that still needs to support older 31 versions of Expat 1.95.x: 32 33 #ifndef XML_STATUS_OK 34 #define XML_STATUS_OK 1 35 #define XML_STATUS_ERROR 0 36 #endif 37 38 Otherwise, the #define hackery is quite ugly and would have been 39 dropped. 40 */ 41 enum XML_Status { 42 XML_STATUS_ERROR = 0, 43 #define XML_STATUS_ERROR XML_STATUS_ERROR 44 XML_STATUS_OK = 1, 45 #define XML_STATUS_OK XML_STATUS_OK 46 XML_STATUS_SUSPENDED = 2, 47 #define XML_STATUS_SUSPENDED XML_STATUS_SUSPENDED 48 }; 49 50 enum XML_Error { 51 XML_ERROR_NONE, 52 XML_ERROR_NO_MEMORY, 53 XML_ERROR_SYNTAX, 54 XML_ERROR_NO_ELEMENTS, 55 XML_ERROR_INVALID_TOKEN, 56 XML_ERROR_UNCLOSED_TOKEN, 57 XML_ERROR_PARTIAL_CHAR, 58 XML_ERROR_TAG_MISMATCH, 59 XML_ERROR_DUPLICATE_ATTRIBUTE, 60 XML_ERROR_JUNK_AFTER_DOC_ELEMENT, 61 XML_ERROR_PARAM_ENTITY_REF, 62 XML_ERROR_UNDEFINED_ENTITY, 63 XML_ERROR_RECURSIVE_ENTITY_REF, 64 XML_ERROR_ASYNC_ENTITY, 65 XML_ERROR_BAD_CHAR_REF, 66 XML_ERROR_BINARY_ENTITY_REF, 67 XML_ERROR_ATTRIBUTE_EXTERNAL_ENTITY_REF, 68 XML_ERROR_MISPLACED_XML_PI, 69 XML_ERROR_UNKNOWN_ENCODING, 70 XML_ERROR_INCORRECT_ENCODING, 71 XML_ERROR_UNCLOSED_CDATA_SECTION, 72 XML_ERROR_EXTERNAL_ENTITY_HANDLING, 73 XML_ERROR_NOT_STANDALONE, 74 XML_ERROR_UNEXPECTED_STATE, 75 XML_ERROR_ENTITY_DECLARED_IN_PE, 76 XML_ERROR_FEATURE_REQUIRES_XML_DTD, 77 XML_ERROR_CANT_CHANGE_FEATURE_ONCE_PARSING, 78 /* Added in 1.95.7. */ 79 XML_ERROR_UNBOUND_PREFIX, 80 /* Added in 1.95.8. */ 81 XML_ERROR_UNDECLARING_PREFIX, 82 XML_ERROR_INCOMPLETE_PE, 83 XML_ERROR_XML_DECL, 84 XML_ERROR_TEXT_DECL, 85 XML_ERROR_PUBLICID, 86 XML_ERROR_SUSPENDED, 87 XML_ERROR_NOT_SUSPENDED, 88 XML_ERROR_ABORTED, 89 XML_ERROR_FINISHED, 90 XML_ERROR_SUSPEND_PE 91 }; 92 93 enum XML_Content_Type { 94 XML_CTYPE_EMPTY = 1, 95 XML_CTYPE_ANY, 96 XML_CTYPE_MIXED, 97 XML_CTYPE_NAME, 98 XML_CTYPE_CHOICE, 99 XML_CTYPE_SEQ 100 }; 101 102 enum XML_Content_Quant { 103 XML_CQUANT_NONE, 104 XML_CQUANT_OPT, 105 XML_CQUANT_REP, 106 XML_CQUANT_PLUS 107 }; 108 109 /* If type == XML_CTYPE_EMPTY or XML_CTYPE_ANY, then quant will be 110 XML_CQUANT_NONE, and the other fields will be zero or NULL. 111 If type == XML_CTYPE_MIXED, then quant will be NONE or REP and 112 numchildren will contain number of elements that may be mixed in 113 and children point to an array of XML_Content cells that will be 114 all of XML_CTYPE_NAME type with no quantification. 115 116 If type == XML_CTYPE_NAME, then the name points to the name, and 117 the numchildren field will be zero and children will be NULL. The 118 quant fields indicates any quantifiers placed on the name. 119 120 CHOICE and SEQ will have name NULL, the number of children in 121 numchildren and children will point, recursively, to an array 122 of XML_Content cells. 123 124 The EMPTY, ANY, and MIXED types will only occur at top level. 125 */ 126 127 typedef struct XML_cp XML_Content; 128 129 struct XML_cp { 130 enum XML_Content_Type type; 131 enum XML_Content_Quant quant; 132 XML_Char * name; 133 unsigned int numchildren; 134 XML_Content * children; 135 }; 136 137 138 /* This is called for an element declaration. See above for 139 description of the model argument. It's the caller's responsibility 140 to free model when finished with it. 141 */ 142 typedef void (XMLCALL *XML_ElementDeclHandler) (void *userData, 143 const XML_Char *name, 144 XML_Content *model); 145 146 XMLPARSEAPI(void) 147 XML_SetElementDeclHandler(XML_Parser parser, 148 XML_ElementDeclHandler eldecl); 149 150 /* The Attlist declaration handler is called for *each* attribute. So 151 a single Attlist declaration with multiple attributes declared will 152 generate multiple calls to this handler. The "default" parameter 153 may be NULL in the case of the "#IMPLIED" or "#REQUIRED" 154 keyword. The "isrequired" parameter will be true and the default 155 value will be NULL in the case of "#REQUIRED". If "isrequired" is 156 true and default is non-NULL, then this is a "#FIXED" default. 157 */ 158 typedef void (XMLCALL *XML_AttlistDeclHandler) ( 159 void *userData, 160 const XML_Char *elname, 161 const XML_Char *attname, 162 const XML_Char *att_type, 163 const XML_Char *dflt, 164 int isrequired); 165 166 XMLPARSEAPI(void) 167 XML_SetAttlistDeclHandler(XML_Parser parser, 168 XML_AttlistDeclHandler attdecl); 169 170 /* The XML declaration handler is called for *both* XML declarations 171 and text declarations. The way to distinguish is that the version 172 parameter will be NULL for text declarations. The encoding 173 parameter may be NULL for XML declarations. The standalone 174 parameter will be -1, 0, or 1 indicating respectively that there 175 was no standalone parameter in the declaration, that it was given 176 as no, or that it was given as yes. 177 */ 178 typedef void (XMLCALL *XML_XmlDeclHandler) (void *userData, 179 const XML_Char *version, 180 const XML_Char *encoding, 181 int standalone); 182 183 XMLPARSEAPI(void) 184 XML_SetXmlDeclHandler(XML_Parser parser, 185 XML_XmlDeclHandler xmldecl); 186 187 188 typedef struct { 189 void *(*malloc_fcn)(size_t size); 190 void *(*realloc_fcn)(void *ptr, size_t size); 191 void (*free_fcn)(void *ptr); 192 } XML_Memory_Handling_Suite; 193 194 /* Constructs a new parser; encoding is the encoding specified by the 195 external protocol or NULL if there is none specified. 196 */ 197 XMLPARSEAPI(XML_Parser) 198 XML_ParserCreate(const XML_Char *encoding); 199 200 /* Constructs a new parser and namespace processor. Element type 201 names and attribute names that belong to a namespace will be 202 expanded; unprefixed attribute names are never expanded; unprefixed 203 element type names are expanded only if there is a default 204 namespace. The expanded name is the concatenation of the namespace 205 URI, the namespace separator character, and the local part of the 206 name. If the namespace separator is '\0' then the namespace URI 207 and the local part will be concatenated without any separator. 208 When a namespace is not declared, the name and prefix will be 209 passed through without expansion. 210 */ 211 XMLPARSEAPI(XML_Parser) 212 XML_ParserCreateNS(const XML_Char *encoding, XML_Char namespaceSeparator); 213 214 215 /* Constructs a new parser using the memory management suite referred to 216 by memsuite. If memsuite is NULL, then use the standard library memory 217 suite. If namespaceSeparator is non-NULL it creates a parser with 218 namespace processing as described above. The character pointed at 219 will serve as the namespace separator. 220 221 All further memory operations used for the created parser will come from 222 the given suite. 223 */ 224 XMLPARSEAPI(XML_Parser) 225 XML_ParserCreate_MM(const XML_Char *encoding, 226 const XML_Memory_Handling_Suite *memsuite, 227 const XML_Char *namespaceSeparator); 228 229 /* Prepare a parser object to be re-used. This is particularly 230 valuable when memory allocation overhead is disproportionatly high, 231 such as when a large number of small documnents need to be parsed. 232 All handlers are cleared from the parser, except for the 233 unknownEncodingHandler. The parser's external state is re-initialized 234 except for the values of ns and ns_triplets. 235 236 Added in Expat 1.95.3. 237 */ 238 XMLPARSEAPI(XML_Bool) 239 XML_ParserReset(XML_Parser parser, const XML_Char *encoding); 240 241 /* atts is array of name/value pairs, terminated by 0; 242 names and values are 0 terminated. 243 */ 244 typedef void (XMLCALL *XML_StartElementHandler) (void *userData, 245 const XML_Char *name, 246 const XML_Char **atts); 247 248 typedef void (XMLCALL *XML_EndElementHandler) (void *userData, 249 const XML_Char *name); 250 251 252 /* s is not 0 terminated. */ 253 typedef void (XMLCALL *XML_CharacterDataHandler) (void *userData, 254 const XML_Char *s, 255 int len); 256 257 /* target and data are 0 terminated */ 258 typedef void (XMLCALL *XML_ProcessingInstructionHandler) ( 259 void *userData, 260 const XML_Char *target, 261 const XML_Char *data); 262 263 /* data is 0 terminated */ 264 typedef void (XMLCALL *XML_CommentHandler) (void *userData, 265 const XML_Char *data); 266 267 typedef void (XMLCALL *XML_StartCdataSectionHandler) (void *userData); 268 typedef void (XMLCALL *XML_EndCdataSectionHandler) (void *userData); 269 270 /* This is called for any characters in the XML document for which 271 there is no applicable handler. This includes both characters that 272 are part of markup which is of a kind that is not reported 273 (comments, markup declarations), or characters that are part of a 274 construct which could be reported but for which no handler has been 275 supplied. The characters are passed exactly as they were in the XML 276 document except that they will be encoded in UTF-8 or UTF-16. 277 Line boundaries are not normalized. Note that a byte order mark 278 character is not passed to the default handler. There are no 279 guarantees about how characters are divided between calls to the 280 default handler: for example, a comment might be split between 281 multiple calls. 282 */ 283 typedef void (XMLCALL *XML_DefaultHandler) (void *userData, 284 const XML_Char *s, 285 int len); 286 287 /* This is called for the start of the DOCTYPE declaration, before 288 any DTD or internal subset is parsed. 289 */ 290 typedef void (XMLCALL *XML_StartDoctypeDeclHandler) ( 291 void *userData, 292 const XML_Char *doctypeName, 293 const XML_Char *sysid, 294 const XML_Char *pubid, 295 int has_internal_subset); 296 297 /* This is called for the start of the DOCTYPE declaration when the 298 closing > is encountered, but after processing any external 299 subset. 300 */ 301 typedef void (XMLCALL *XML_EndDoctypeDeclHandler)(void *userData); 302 303 /* This is called for entity declarations. The is_parameter_entity 304 argument will be non-zero if the entity is a parameter entity, zero 305 otherwise. 306 307 For internal entities (<!ENTITY foo "bar">), value will 308 be non-NULL and systemId, publicID, and notationName will be NULL. 309 The value string is NOT nul-terminated; the length is provided in 310 the value_length argument. Since it is legal to have zero-length 311 values, do not use this argument to test for internal entities. 312 313 For external entities, value will be NULL and systemId will be 314 non-NULL. The publicId argument will be NULL unless a public 315 identifier was provided. The notationName argument will have a 316 non-NULL value only for unparsed entity declarations. 317 318 Note that is_parameter_entity can't be changed to XML_Bool, since 319 that would break binary compatibility. 320 */ 321 typedef void (XMLCALL *XML_EntityDeclHandler) ( 322 void *userData, 323 const XML_Char *entityName, 324 int is_parameter_entity, 325 const XML_Char *value, 326 int value_length, 327 const XML_Char *base, 328 const XML_Char *systemId, 329 const XML_Char *publicId, 330 const XML_Char *notationName); 331 332 XMLPARSEAPI(void) 333 XML_SetEntityDeclHandler(XML_Parser parser, 334 XML_EntityDeclHandler handler); 335 336 /* OBSOLETE -- OBSOLETE -- OBSOLETE 337 This handler has been superceded by the EntityDeclHandler above. 338 It is provided here for backward compatibility. 339 340 This is called for a declaration of an unparsed (NDATA) entity. 341 The base argument is whatever was set by XML_SetBase. The 342 entityName, systemId and notationName arguments will never be 343 NULL. The other arguments may be. 344 */ 345 typedef void (XMLCALL *XML_UnparsedEntityDeclHandler) ( 346 void *userData, 347 const XML_Char *entityName, 348 const XML_Char *base, 349 const XML_Char *systemId, 350 const XML_Char *publicId, 351 const XML_Char *notationName); 352 353 /* This is called for a declaration of notation. The base argument is 354 whatever was set by XML_SetBase. The notationName will never be 355 NULL. The other arguments can be. 356 */ 357 typedef void (XMLCALL *XML_NotationDeclHandler) ( 358 void *userData, 359 const XML_Char *notationName, 360 const XML_Char *base, 361 const XML_Char *systemId, 362 const XML_Char *publicId); 363 364 /* When namespace processing is enabled, these are called once for 365 each namespace declaration. The call to the start and end element 366 handlers occur between the calls to the start and end namespace 367 declaration handlers. For an xmlns attribute, prefix will be 368 NULL. For an xmlns="" attribute, uri will be NULL. 369 */ 370 typedef void (XMLCALL *XML_StartNamespaceDeclHandler) ( 371 void *userData, 372 const XML_Char *prefix, 373 const XML_Char *uri); 374 375 typedef void (XMLCALL *XML_EndNamespaceDeclHandler) ( 376 void *userData, 377 const XML_Char *prefix); 378 379 /* This is called if the document is not standalone, that is, it has an 380 external subset or a reference to a parameter entity, but does not 381 have standalone="yes". If this handler returns XML_STATUS_ERROR, 382 then processing will not continue, and the parser will return a 383 XML_ERROR_NOT_STANDALONE error. 384 If parameter entity parsing is enabled, then in addition to the 385 conditions above this handler will only be called if the referenced 386 entity was actually read. 387 */ 388 typedef int (XMLCALL *XML_NotStandaloneHandler) (void *userData); 389 390 /* This is called for a reference to an external parsed general 391 entity. The referenced entity is not automatically parsed. The 392 application can parse it immediately or later using 393 XML_ExternalEntityParserCreate. 394 395 The parser argument is the parser parsing the entity containing the 396 reference; it can be passed as the parser argument to 397 XML_ExternalEntityParserCreate. The systemId argument is the 398 system identifier as specified in the entity declaration; it will 399 not be NULL. 400 401 The base argument is the system identifier that should be used as 402 the base for resolving systemId if systemId was relative; this is 403 set by XML_SetBase; it may be NULL. 404 405 The publicId argument is the public identifier as specified in the 406 entity declaration, or NULL if none was specified; the whitespace 407 in the public identifier will have been normalized as required by 408 the XML spec. 409 410 The context argument specifies the parsing context in the format 411 expected by the context argument to XML_ExternalEntityParserCreate; 412 context is valid only until the handler returns, so if the 413 referenced entity is to be parsed later, it must be copied. 414 context is NULL only when the entity is a parameter entity. 415 416 The handler should return XML_STATUS_ERROR if processing should not 417 continue because of a fatal error in the handling of the external 418 entity. In this case the calling parser will return an 419 XML_ERROR_EXTERNAL_ENTITY_HANDLING error. 420 421 Note that unlike other handlers the first argument is the parser, 422 not userData. 423 */ 424 typedef int (XMLCALL *XML_ExternalEntityRefHandler) ( 425 XML_Parser parser, 426 const XML_Char *context, 427 const XML_Char *base, 428 const XML_Char *systemId, 429 const XML_Char *publicId); 430 431 /* This is called in two situations: 432 1) An entity reference is encountered for which no declaration 433 has been read *and* this is not an error. 434 2) An internal entity reference is read, but not expanded, because 435 XML_SetDefaultHandler has been called. 436 Note: skipped parameter entities in declarations and skipped general 437 entities in attribute values cannot be reported, because 438 the event would be out of sync with the reporting of the 439 declarations or attribute values 440 */ 441 typedef void (XMLCALL *XML_SkippedEntityHandler) ( 442 void *userData, 443 const XML_Char *entityName, 444 int is_parameter_entity); 445 446 /* This structure is filled in by the XML_UnknownEncodingHandler to 447 provide information to the parser about encodings that are unknown 448 to the parser. 449 450 The map[b] member gives information about byte sequences whose 451 first byte is b. 452 453 If map[b] is c where c is >= 0, then b by itself encodes the 454 Unicode scalar value c. 455 456 If map[b] is -1, then the byte sequence is malformed. 457 458 If map[b] is -n, where n >= 2, then b is the first byte of an 459 n-byte sequence that encodes a single Unicode scalar value. 460 461 The data member will be passed as the first argument to the convert 462 function. 463 464 The convert function is used to convert multibyte sequences; s will 465 point to a n-byte sequence where map[(unsigned char)*s] == -n. The 466 convert function must return the Unicode scalar value represented 467 by this byte sequence or -1 if the byte sequence is malformed. 468 469 The convert function may be NULL if the encoding is a single-byte 470 encoding, that is if map[b] >= -1 for all bytes b. 471 472 When the parser is finished with the encoding, then if release is 473 not NULL, it will call release passing it the data member; once 474 release has been called, the convert function will not be called 475 again. 476 477 Expat places certain restrictions on the encodings that are supported 478 using this mechanism. 479 480 1. Every ASCII character that can appear in a well-formed XML document, 481 other than the characters 482 483 $@\^`{}~ 484 485 must be represented by a single byte, and that byte must be the 486 same byte that represents that character in ASCII. 487 488 2. No character may require more than 4 bytes to encode. 489 490 3. All characters encoded must have Unicode scalar values <= 491 0xFFFF, (i.e., characters that would be encoded by surrogates in 492 UTF-16 are not allowed). Note that this restriction doesn't 493 apply to the built-in support for UTF-8 and UTF-16. 494 495 4. No Unicode character may be encoded by more than one distinct 496 sequence of bytes. 497 */ 498 typedef struct { 499 int map[256]; 500 void *data; 501 int (XMLCALL *convert)(void *data, const char *s); 502 void (XMLCALL *release)(void *data); 503 } XML_Encoding; 504 505 /* This is called for an encoding that is unknown to the parser. 506 507 The encodingHandlerData argument is that which was passed as the 508 second argument to XML_SetUnknownEncodingHandler. 509 510 The name argument gives the name of the encoding as specified in 511 the encoding declaration. 512 513 If the callback can provide information about the encoding, it must 514 fill in the XML_Encoding structure, and return XML_STATUS_OK. 515 Otherwise it must return XML_STATUS_ERROR. 516 517 If info does not describe a suitable encoding, then the parser will 518 return an XML_UNKNOWN_ENCODING error. 519 */ 520 typedef int (XMLCALL *XML_UnknownEncodingHandler) ( 521 void *encodingHandlerData, 522 const XML_Char *name, 523 XML_Encoding *info); 524 525 XMLPARSEAPI(void) 526 XML_SetElementHandler(XML_Parser parser, 527 XML_StartElementHandler start, 528 XML_EndElementHandler end); 529 530 XMLPARSEAPI(void) 531 XML_SetStartElementHandler(XML_Parser parser, 532 XML_StartElementHandler handler); 533 534 XMLPARSEAPI(void) 535 XML_SetEndElementHandler(XML_Parser parser, 536 XML_EndElementHandler handler); 537 538 XMLPARSEAPI(void) 539 XML_SetCharacterDataHandler(XML_Parser parser, 540 XML_CharacterDataHandler handler); 541 542 XMLPARSEAPI(void) 543 XML_SetProcessingInstructionHandler(XML_Parser parser, 544 XML_ProcessingInstructionHandler handler); 545 XMLPARSEAPI(void) 546 XML_SetCommentHandler(XML_Parser parser, 547 XML_CommentHandler handler); 548 549 XMLPARSEAPI(void) 550 XML_SetCdataSectionHandler(XML_Parser parser, 551 XML_StartCdataSectionHandler start, 552 XML_EndCdataSectionHandler end); 553 554 XMLPARSEAPI(void) 555 XML_SetStartCdataSectionHandler(XML_Parser parser, 556 XML_StartCdataSectionHandler start); 557 558 XMLPARSEAPI(void) 559 XML_SetEndCdataSectionHandler(XML_Parser parser, 560 XML_EndCdataSectionHandler end); 561 562 /* This sets the default handler and also inhibits expansion of 563 internal entities. These entity references will be passed to the 564 default handler, or to the skipped entity handler, if one is set. 565 */ 566 XMLPARSEAPI(void) 567 XML_SetDefaultHandler(XML_Parser parser, 568 XML_DefaultHandler handler); 569 570 /* This sets the default handler but does not inhibit expansion of 571 internal entities. The entity reference will not be passed to the 572 default handler. 573 */ 574 XMLPARSEAPI(void) 575 XML_SetDefaultHandlerExpand(XML_Parser parser, 576 XML_DefaultHandler handler); 577 578 XMLPARSEAPI(void) 579 XML_SetDoctypeDeclHandler(XML_Parser parser, 580 XML_StartDoctypeDeclHandler start, 581 XML_EndDoctypeDeclHandler end); 582 583 XMLPARSEAPI(void) 584 XML_SetStartDoctypeDeclHandler(XML_Parser parser, 585 XML_StartDoctypeDeclHandler start); 586 587 XMLPARSEAPI(void) 588 XML_SetEndDoctypeDeclHandler(XML_Parser parser, 589 XML_EndDoctypeDeclHandler end); 590 591 XMLPARSEAPI(void) 592 XML_SetUnparsedEntityDeclHandler(XML_Parser parser, 593 XML_UnparsedEntityDeclHandler handler); 594 595 XMLPARSEAPI(void) 596 XML_SetNotationDeclHandler(XML_Parser parser, 597 XML_NotationDeclHandler handler); 598 599 XMLPARSEAPI(void) 600 XML_SetNamespaceDeclHandler(XML_Parser parser, 601 XML_StartNamespaceDeclHandler start, 602 XML_EndNamespaceDeclHandler end); 603 604 XMLPARSEAPI(void) 605 XML_SetStartNamespaceDeclHandler(XML_Parser parser, 606 XML_StartNamespaceDeclHandler start); 607 608 XMLPARSEAPI(void) 609 XML_SetEndNamespaceDeclHandler(XML_Parser parser, 610 XML_EndNamespaceDeclHandler end); 611 612 XMLPARSEAPI(void) 613 XML_SetNotStandaloneHandler(XML_Parser parser, 614 XML_NotStandaloneHandler handler); 615 616 XMLPARSEAPI(void) 617 XML_SetExternalEntityRefHandler(XML_Parser parser, 618 XML_ExternalEntityRefHandler handler); 619 620 /* If a non-NULL value for arg is specified here, then it will be 621 passed as the first argument to the external entity ref handler 622 instead of the parser object. 623 */ 624 XMLPARSEAPI(void) 625 XML_SetExternalEntityRefHandlerArg(XML_Parser parser, 626 void *arg); 627 628 XMLPARSEAPI(void) 629 XML_SetSkippedEntityHandler(XML_Parser parser, 630 XML_SkippedEntityHandler handler); 631 632 XMLPARSEAPI(void) 633 XML_SetUnknownEncodingHandler(XML_Parser parser, 634 XML_UnknownEncodingHandler handler, 635 void *encodingHandlerData); 636 637 /* This can be called within a handler for a start element, end 638 element, processing instruction or character data. It causes the 639 corresponding markup to be passed to the default handler. 640 */ 641 XMLPARSEAPI(void) 642 XML_DefaultCurrent(XML_Parser parser); 643 644 /* If do_nst is non-zero, and namespace processing is in effect, and 645 a name has a prefix (i.e. an explicit namespace qualifier) then 646 that name is returned as a triplet in a single string separated by 647 the separator character specified when the parser was created: URI 648 + sep + local_name + sep + prefix. 649 650 If do_nst is zero, then namespace information is returned in the 651 default manner (URI + sep + local_name) whether or not the name 652 has a prefix. 653 654 Note: Calling XML_SetReturnNSTriplet after XML_Parse or 655 XML_ParseBuffer has no effect. 656 */ 657 658 XMLPARSEAPI(void) 659 XML_SetReturnNSTriplet(XML_Parser parser, int do_nst); 660 661 /* This value is passed as the userData argument to callbacks. */ 662 XMLPARSEAPI(void) 663 XML_SetUserData(XML_Parser parser, void *userData); 664 665 /* Returns the last value set by XML_SetUserData or NULL. */ 666 #define XML_GetUserData(parser) (*(void **)(parser)) 667 668 /* This is equivalent to supplying an encoding argument to 669 XML_ParserCreate. On success XML_SetEncoding returns non-zero, 670 zero otherwise. 671 Note: Calling XML_SetEncoding after XML_Parse or XML_ParseBuffer 672 has no effect and returns XML_STATUS_ERROR. 673 */ 674 XMLPARSEAPI(enum XML_Status) 675 XML_SetEncoding(XML_Parser parser, const XML_Char *encoding); 676 677 /* If this function is called, then the parser will be passed as the 678 first argument to callbacks instead of userData. The userData will 679 still be accessible using XML_GetUserData. 680 */ 681 XMLPARSEAPI(void) 682 XML_UseParserAsHandlerArg(XML_Parser parser); 683 684 /* If useDTD == XML_TRUE is passed to this function, then the parser 685 will assume that there is an external subset, even if none is 686 specified in the document. In such a case the parser will call the 687 externalEntityRefHandler with a value of NULL for the systemId 688 argument (the publicId and context arguments will be NULL as well). 689 Note: For the purpose of checking WFC: Entity Declared, passing 690 useDTD == XML_TRUE will make the parser behave as if the document 691 had a DTD with an external subset. 692 Note: If this function is called, then this must be done before 693 the first call to XML_Parse or XML_ParseBuffer, since it will 694 have no effect after that. Returns 695 XML_ERROR_CANT_CHANGE_FEATURE_ONCE_PARSING. 696 Note: If the document does not have a DOCTYPE declaration at all, 697 then startDoctypeDeclHandler and endDoctypeDeclHandler will not 698 be called, despite an external subset being parsed. 699 Note: If XML_DTD is not defined when Expat is compiled, returns 700 XML_ERROR_FEATURE_REQUIRES_XML_DTD. 701 */ 702 XMLPARSEAPI(enum XML_Error) 703 XML_UseForeignDTD(XML_Parser parser, XML_Bool useDTD); 704 705 706 /* Sets the base to be used for resolving relative URIs in system 707 identifiers in declarations. Resolving relative identifiers is 708 left to the application: this value will be passed through as the 709 base argument to the XML_ExternalEntityRefHandler, 710 XML_NotationDeclHandler and XML_UnparsedEntityDeclHandler. The base 711 argument will be copied. Returns XML_STATUS_ERROR if out of memory, 712 XML_STATUS_OK otherwise. 713 */ 714 XMLPARSEAPI(enum XML_Status) 715 XML_SetBase(XML_Parser parser, const XML_Char *base); 716 717 XMLPARSEAPI(const XML_Char *) 718 XML_GetBase(XML_Parser parser); 719 720 /* Returns the number of the attribute/value pairs passed in last call 721 to the XML_StartElementHandler that were specified in the start-tag 722 rather than defaulted. Each attribute/value pair counts as 2; thus 723 this correspondds to an index into the atts array passed to the 724 XML_StartElementHandler. 725 */ 726 XMLPARSEAPI(int) 727 XML_GetSpecifiedAttributeCount(XML_Parser parser); 728 729 /* Returns the index of the ID attribute passed in the last call to 730 XML_StartElementHandler, or -1 if there is no ID attribute. Each 731 attribute/value pair counts as 2; thus this correspondds to an 732 index into the atts array passed to the XML_StartElementHandler. 733 */ 734 XMLPARSEAPI(int) 735 XML_GetIdAttributeIndex(XML_Parser parser); 736 737 /* Parses some input. Returns XML_STATUS_ERROR if a fatal error is 738 detected. The last call to XML_Parse must have isFinal true; len 739 may be zero for this call (or any other). 740 741 Though the return values for these functions has always been 742 described as a Boolean value, the implementation, at least for the 743 1.95.x series, has always returned exactly one of the XML_Status 744 values. 745 */ 746 XMLPARSEAPI(enum XML_Status) 747 XML_Parse(XML_Parser parser, const char *s, int len, int isFinal); 748 749 XMLPARSEAPI(void *) 750 XML_GetBuffer(XML_Parser parser, int len); 751 752 XMLPARSEAPI(enum XML_Status) 753 XML_ParseBuffer(XML_Parser parser, int len, int isFinal); 754 755 /* Stops parsing, causing XML_Parse() or XML_ParseBuffer() to return. 756 Must be called from within a call-back handler, except when aborting 757 (resumable = 0) an already suspended parser. Some call-backs may 758 still follow because they would otherwise get lost. Examples: 759 - endElementHandler() for empty elements when stopped in 760 startElementHandler(), 761 - endNameSpaceDeclHandler() when stopped in endElementHandler(), 762 and possibly others. 763 764 Can be called from most handlers, including DTD related call-backs, 765 except when parsing an external parameter entity and resumable != 0. 766 Returns XML_STATUS_OK when successful, XML_STATUS_ERROR otherwise. 767 Possible error codes: 768 - XML_ERROR_SUSPENDED: when suspending an already suspended parser. 769 - XML_ERROR_FINISHED: when the parser has already finished. 770 - XML_ERROR_SUSPEND_PE: when suspending while parsing an external PE. 771 772 When resumable != 0 (true) then parsing is suspended, that is, 773 XML_Parse() and XML_ParseBuffer() return XML_STATUS_SUSPENDED. 774 Otherwise, parsing is aborted, that is, XML_Parse() and XML_ParseBuffer() 775 return XML_STATUS_ERROR with error code XML_ERROR_ABORTED. 776 777 *Note*: 778 This will be applied to the current parser instance only, that is, if 779 there is a parent parser then it will continue parsing when the 780 externalEntityRefHandler() returns. It is up to the implementation of 781 the externalEntityRefHandler() to call XML_StopParser() on the parent 782 parser (recursively), if one wants to stop parsing altogether. 783 784 When suspended, parsing can be resumed by calling XML_ResumeParser(). 785 */ 786 XMLPARSEAPI(enum XML_Status) 787 XML_StopParser(XML_Parser parser, XML_Bool resumable); 788 789 /* Resumes parsing after it has been suspended with XML_StopParser(). 790 Must not be called from within a handler call-back. Returns same 791 status codes as XML_Parse() or XML_ParseBuffer(). 792 Additional error code XML_ERROR_NOT_SUSPENDED possible. 793 794 *Note*: 795 This must be called on the most deeply nested child parser instance 796 first, and on its parent parser only after the child parser has finished, 797 to be applied recursively until the document entity's parser is restarted. 798 That is, the parent parser will not resume by itself and it is up to the 799 application to call XML_ResumeParser() on it at the appropriate moment. 800 */ 801 XMLPARSEAPI(enum XML_Status) 802 XML_ResumeParser(XML_Parser parser); 803 804 enum XML_Parsing { 805 XML_INITIALIZED, 806 XML_PARSING, 807 XML_FINISHED, 808 XML_SUSPENDED 809 }; 810 811 typedef struct { 812 enum XML_Parsing parsing; 813 XML_Bool finalBuffer; 814 } XML_ParsingStatus; 815 816 /* Returns status of parser with respect to being initialized, parsing, 817 finished, or suspended and processing the final buffer. 818 XXX XML_Parse() and XML_ParseBuffer() should return XML_ParsingStatus, 819 XXX with XML_FINISHED_OK or XML_FINISHED_ERROR replacing XML_FINISHED 820 */ 821 XMLPARSEAPI(void) 822 XML_GetParsingStatus(XML_Parser parser, XML_ParsingStatus *status); 823 824 /* Creates an XML_Parser object that can parse an external general 825 entity; context is a '\0'-terminated string specifying the parse 826 context; encoding is a '\0'-terminated string giving the name of 827 the externally specified encoding, or NULL if there is no 828 externally specified encoding. The context string consists of a 829 sequence of tokens separated by formfeeds (\f); a token consisting 830 of a name specifies that the general entity of the name is open; a 831 token of the form prefix=uri specifies the namespace for a 832 particular prefix; a token of the form =uri specifies the default 833 namespace. This can be called at any point after the first call to 834 an ExternalEntityRefHandler so longer as the parser has not yet 835 been freed. The new parser is completely independent and may 836 safely be used in a separate thread. The handlers and userData are 837 initialized from the parser argument. Returns NULL if out of memory. 838 Otherwise returns a new XML_Parser object. 839 */ 840 XMLPARSEAPI(XML_Parser) 841 XML_ExternalEntityParserCreate(XML_Parser parser, 842 const XML_Char *context, 843 const XML_Char *encoding); 844 845 enum XML_ParamEntityParsing { 846 XML_PARAM_ENTITY_PARSING_NEVER, 847 XML_PARAM_ENTITY_PARSING_UNLESS_STANDALONE, 848 XML_PARAM_ENTITY_PARSING_ALWAYS 849 }; 850 851 /* Controls parsing of parameter entities (including the external DTD 852 subset). If parsing of parameter entities is enabled, then 853 references to external parameter entities (including the external 854 DTD subset) will be passed to the handler set with 855 XML_SetExternalEntityRefHandler. The context passed will be 0. 856 857 Unlike external general entities, external parameter entities can 858 only be parsed synchronously. If the external parameter entity is 859 to be parsed, it must be parsed during the call to the external 860 entity ref handler: the complete sequence of 861 XML_ExternalEntityParserCreate, XML_Parse/XML_ParseBuffer and 862 XML_ParserFree calls must be made during this call. After 863 XML_ExternalEntityParserCreate has been called to create the parser 864 for the external parameter entity (context must be 0 for this 865 call), it is illegal to make any calls on the old parser until 866 XML_ParserFree has been called on the newly created parser. 867 If the library has been compiled without support for parameter 868 entity parsing (ie without XML_DTD being defined), then 869 XML_SetParamEntityParsing will return 0 if parsing of parameter 870 entities is requested; otherwise it will return non-zero. 871 Note: If XML_SetParamEntityParsing is called after XML_Parse or 872 XML_ParseBuffer, then it has no effect and will always return 0. 873 */ 874 XMLPARSEAPI(int) 875 XML_SetParamEntityParsing(XML_Parser parser, 876 enum XML_ParamEntityParsing parsing); 877 878 /* If XML_Parse or XML_ParseBuffer have returned XML_STATUS_ERROR, then 879 XML_GetErrorCode returns information about the error. 880 */ 881 XMLPARSEAPI(enum XML_Error) 882 XML_GetErrorCode(XML_Parser parser); 883 884 /* These functions return information about the current parse 885 location. They may be called from any callback called to report 886 some parse event; in this case the location is the location of the 887 first of the sequence of characters that generated the event. When 888 called from callbacks generated by declarations in the document 889 prologue, the location identified isn't as neatly defined, but will 890 be within the relevant markup. When called outside of the callback 891 functions, the position indicated will be just past the last parse 892 event (regardless of whether there was an associated callback). 893 894 They may also be called after returning from a call to XML_Parse 895 or XML_ParseBuffer. If the return value is XML_STATUS_ERROR then 896 the location is the location of the character at which the error 897 was detected; otherwise the location is the location of the last 898 parse event, as described above. 899 */ 900 XMLPARSEAPI(int) XML_GetCurrentLineNumber(XML_Parser parser); 901 XMLPARSEAPI(int) XML_GetCurrentColumnNumber(XML_Parser parser); 902 XMLPARSEAPI(long) XML_GetCurrentByteIndex(XML_Parser parser); 903 904 /* Return the number of bytes in the current event. 905 Returns 0 if the event is in an internal entity. 906 */ 907 XMLPARSEAPI(int) 908 XML_GetCurrentByteCount(XML_Parser parser); 909 910 /* If XML_CONTEXT_BYTES is defined, returns the input buffer, sets 911 the integer pointed to by offset to the offset within this buffer 912 of the current parse position, and sets the integer pointed to by size 913 to the size of this buffer (the number of input bytes). Otherwise 914 returns a NULL pointer. Also returns a NULL pointer if a parse isn't 915 active. 916 917 NOTE: The character pointer returned should not be used outside 918 the handler that makes the call. 919 */ 920 XMLPARSEAPI(const char *) 921 XML_GetInputContext(XML_Parser parser, 922 int *offset, 923 int *size); 924 925 /* For backwards compatibility with previous versions. */ 926 #define XML_GetErrorLineNumber XML_GetCurrentLineNumber 927 #define XML_GetErrorColumnNumber XML_GetCurrentColumnNumber 928 #define XML_GetErrorByteIndex XML_GetCurrentByteIndex 929 930 /* Frees the content model passed to the element declaration handler */ 931 XMLPARSEAPI(void) 932 XML_FreeContentModel(XML_Parser parser, XML_Content *model); 933 934 /* Exposing the memory handling functions used in Expat */ 935 XMLPARSEAPI(void *) 936 XML_MemMalloc(XML_Parser parser, size_t size); 937 938 XMLPARSEAPI(void *) 939 XML_MemRealloc(XML_Parser parser, void *ptr, size_t size); 940 941 XMLPARSEAPI(void) 942 XML_MemFree(XML_Parser parser, void *ptr); 943 944 /* Frees memory used by the parser. */ 945 XMLPARSEAPI(void) 946 XML_ParserFree(XML_Parser parser); 947 948 /* Returns a string describing the error. */ 949 XMLPARSEAPI(const XML_LChar *) 950 XML_ErrorString(enum XML_Error code); 951 952 /* Return a string containing the version number of this expat */ 953 XMLPARSEAPI(const XML_LChar *) 954 XML_ExpatVersion(void); 955 956 typedef struct { 957 int major; 958 int minor; 959 int micro; 960 } XML_Expat_Version; 961 962 /* Return an XML_Expat_Version structure containing numeric version 963 number information for this version of expat. 964 */ 965 XMLPARSEAPI(XML_Expat_Version) 966 XML_ExpatVersionInfo(void); 967 968 /* Added in Expat 1.95.5. */ 969 enum XML_FeatureEnum { 970 XML_FEATURE_END = 0, 971 XML_FEATURE_UNICODE, 972 XML_FEATURE_UNICODE_WCHAR_T, 973 XML_FEATURE_DTD, 974 XML_FEATURE_CONTEXT_BYTES, 975 XML_FEATURE_MIN_SIZE, 976 XML_FEATURE_SIZEOF_XML_CHAR, 977 XML_FEATURE_SIZEOF_XML_LCHAR 978 /* Additional features must be added to the end of this enum. */ 979 }; 980 981 typedef struct { 982 enum XML_FeatureEnum feature; 983 const XML_LChar *name; 984 long int value; 985 } XML_Feature; 986 987 XMLPARSEAPI(const XML_Feature *) 988 XML_GetFeatureList(void); 989 990 991 /* Expat follows the GNU/Linux convention of odd number minor version for 992 beta/development releases and even number minor version for stable 993 releases. Micro is bumped with each release, and set to 0 with each 994 change to major or minor version. 995 */ 996 #define XML_MAJOR_VERSION 1 997 #define XML_MINOR_VERSION 95 998 #define XML_MICRO_VERSION 8 999 1000 #ifdef __cplusplus 1001 } 1002 #endif 1003 1004 #endif /* not XmlParse_INCLUDED */ 1005