1 /* 2 __ __ _ 3 ___\ \/ /_ __ __ _| |_ 4 / _ \\ /| '_ \ / _` | __| 5 | __// \| |_) | (_| | |_ 6 \___/_/\_\ .__/ \__,_|\__| 7 |_| XML parser 8 9 Copyright (c) 1997-2000 Thai Open Source Software Center Ltd 10 Copyright (c) 2000-2017 Expat development team 11 Licensed under the MIT license: 12 13 Permission is hereby granted, free of charge, to any person obtaining 14 a copy of this software and associated documentation files (the 15 "Software"), to deal in the Software without restriction, including 16 without limitation the rights to use, copy, modify, merge, publish, 17 distribute, sublicense, and/or sell copies of the Software, and to permit 18 persons to whom the Software is furnished to do so, subject to the 19 following conditions: 20 21 The above copyright notice and this permission notice shall be included 22 in all copies or substantial portions of the Software. 23 24 THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, 25 EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF 26 MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN 27 NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, 28 DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR 29 OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE 30 USE OR OTHER DEALINGS IN THE SOFTWARE. 31 */ 32 33 #ifndef XmlTok_INCLUDED 34 #define XmlTok_INCLUDED 1 35 36 #ifdef __cplusplus 37 extern "C" { 38 #endif 39 40 /* The following token may be returned by XmlContentTok */ 41 #define XML_TOK_TRAILING_RSQB \ 42 -5 /* ] or ]] at the end of the scan; might be \ 43 start of illegal ]]> sequence */ 44 /* The following tokens may be returned by both XmlPrologTok and 45 XmlContentTok. 46 */ 47 #define XML_TOK_NONE -4 /* The string to be scanned is empty */ 48 #define XML_TOK_TRAILING_CR \ 49 -3 /* A CR at the end of the scan; \ 50 might be part of CRLF sequence */ 51 #define XML_TOK_PARTIAL_CHAR -2 /* only part of a multibyte sequence */ 52 #define XML_TOK_PARTIAL -1 /* only part of a token */ 53 #define XML_TOK_INVALID 0 54 55 /* The following tokens are returned by XmlContentTok; some are also 56 returned by XmlAttributeValueTok, XmlEntityTok, XmlCdataSectionTok. 57 */ 58 #define XML_TOK_START_TAG_WITH_ATTS 1 59 #define XML_TOK_START_TAG_NO_ATTS 2 60 #define XML_TOK_EMPTY_ELEMENT_WITH_ATTS 3 /* empty element tag <e/> */ 61 #define XML_TOK_EMPTY_ELEMENT_NO_ATTS 4 62 #define XML_TOK_END_TAG 5 63 #define XML_TOK_DATA_CHARS 6 64 #define XML_TOK_DATA_NEWLINE 7 65 #define XML_TOK_CDATA_SECT_OPEN 8 66 #define XML_TOK_ENTITY_REF 9 67 #define XML_TOK_CHAR_REF 10 /* numeric character reference */ 68 69 /* The following tokens may be returned by both XmlPrologTok and 70 XmlContentTok. 71 */ 72 #define XML_TOK_PI 11 /* processing instruction */ 73 #define XML_TOK_XML_DECL 12 /* XML decl or text decl */ 74 #define XML_TOK_COMMENT 13 75 #define XML_TOK_BOM 14 /* Byte order mark */ 76 77 /* The following tokens are returned only by XmlPrologTok */ 78 #define XML_TOK_PROLOG_S 15 79 #define XML_TOK_DECL_OPEN 16 /* <!foo */ 80 #define XML_TOK_DECL_CLOSE 17 /* > */ 81 #define XML_TOK_NAME 18 82 #define XML_TOK_NMTOKEN 19 83 #define XML_TOK_POUND_NAME 20 /* #name */ 84 #define XML_TOK_OR 21 /* | */ 85 #define XML_TOK_PERCENT 22 86 #define XML_TOK_OPEN_PAREN 23 87 #define XML_TOK_CLOSE_PAREN 24 88 #define XML_TOK_OPEN_BRACKET 25 89 #define XML_TOK_CLOSE_BRACKET 26 90 #define XML_TOK_LITERAL 27 91 #define XML_TOK_PARAM_ENTITY_REF 28 92 #define XML_TOK_INSTANCE_START 29 93 94 /* The following occur only in element type declarations */ 95 #define XML_TOK_NAME_QUESTION 30 /* name? */ 96 #define XML_TOK_NAME_ASTERISK 31 /* name* */ 97 #define XML_TOK_NAME_PLUS 32 /* name+ */ 98 #define XML_TOK_COND_SECT_OPEN 33 /* <![ */ 99 #define XML_TOK_COND_SECT_CLOSE 34 /* ]]> */ 100 #define XML_TOK_CLOSE_PAREN_QUESTION 35 /* )? */ 101 #define XML_TOK_CLOSE_PAREN_ASTERISK 36 /* )* */ 102 #define XML_TOK_CLOSE_PAREN_PLUS 37 /* )+ */ 103 #define XML_TOK_COMMA 38 104 105 /* The following token is returned only by XmlAttributeValueTok */ 106 #define XML_TOK_ATTRIBUTE_VALUE_S 39 107 108 /* The following token is returned only by XmlCdataSectionTok */ 109 #define XML_TOK_CDATA_SECT_CLOSE 40 110 111 /* With namespace processing this is returned by XmlPrologTok for a 112 name with a colon. 113 */ 114 #define XML_TOK_PREFIXED_NAME 41 115 116 #ifdef XML_DTD 117 # define XML_TOK_IGNORE_SECT 42 118 #endif /* XML_DTD */ 119 120 #ifdef XML_DTD 121 # define XML_N_STATES 4 122 #else /* not XML_DTD */ 123 # define XML_N_STATES 3 124 #endif /* not XML_DTD */ 125 126 #define XML_PROLOG_STATE 0 127 #define XML_CONTENT_STATE 1 128 #define XML_CDATA_SECTION_STATE 2 129 #ifdef XML_DTD 130 # define XML_IGNORE_SECTION_STATE 3 131 #endif /* XML_DTD */ 132 133 #define XML_N_LITERAL_TYPES 2 134 #define XML_ATTRIBUTE_VALUE_LITERAL 0 135 #define XML_ENTITY_VALUE_LITERAL 1 136 137 /* The size of the buffer passed to XmlUtf8Encode must be at least this. */ 138 #define XML_UTF8_ENCODE_MAX 4 139 /* The size of the buffer passed to XmlUtf16Encode must be at least this. */ 140 #define XML_UTF16_ENCODE_MAX 2 141 142 typedef struct position { 143 /* first line and first column are 0 not 1 */ 144 XML_Size lineNumber; 145 XML_Size columnNumber; 146 } POSITION; 147 148 typedef struct { 149 const char *name; 150 const char *valuePtr; 151 const char *valueEnd; 152 char normalized; 153 } ATTRIBUTE; 154 155 struct encoding; 156 typedef struct encoding ENCODING; 157 158 typedef int(PTRCALL *SCANNER)(const ENCODING *, const char *, const char *, 159 const char **); 160 161 enum XML_Convert_Result { 162 XML_CONVERT_COMPLETED = 0, 163 XML_CONVERT_INPUT_INCOMPLETE = 1, 164 XML_CONVERT_OUTPUT_EXHAUSTED 165 = 2 /* and therefore potentially input remaining as well */ 166 }; 167 168 struct encoding { 169 SCANNER scanners[XML_N_STATES]; 170 SCANNER literalScanners[XML_N_LITERAL_TYPES]; 171 int(PTRCALL *nameMatchesAscii)(const ENCODING *, const char *, const char *, 172 const char *); 173 int(PTRFASTCALL *nameLength)(const ENCODING *, const char *); 174 const char *(PTRFASTCALL *skipS)(const ENCODING *, const char *); 175 int(PTRCALL *getAtts)(const ENCODING *enc, const char *ptr, int attsMax, 176 ATTRIBUTE *atts); 177 int(PTRFASTCALL *charRefNumber)(const ENCODING *enc, const char *ptr); 178 int(PTRCALL *predefinedEntityName)(const ENCODING *, const char *, 179 const char *); 180 void(PTRCALL *updatePosition)(const ENCODING *, const char *ptr, 181 const char *end, POSITION *); 182 int(PTRCALL *isPublicId)(const ENCODING *enc, const char *ptr, 183 const char *end, const char **badPtr); 184 enum XML_Convert_Result(PTRCALL *utf8Convert)(const ENCODING *enc, 185 const char **fromP, 186 const char *fromLim, char **toP, 187 const char *toLim); 188 enum XML_Convert_Result(PTRCALL *utf16Convert)(const ENCODING *enc, 189 const char **fromP, 190 const char *fromLim, 191 unsigned short **toP, 192 const unsigned short *toLim); 193 int minBytesPerChar; 194 char isUtf8; 195 char isUtf16; 196 }; 197 198 /* Scan the string starting at ptr until the end of the next complete 199 token, but do not scan past eptr. Return an integer giving the 200 type of token. 201 202 Return XML_TOK_NONE when ptr == eptr; nextTokPtr will not be set. 203 204 Return XML_TOK_PARTIAL when the string does not contain a complete 205 token; nextTokPtr will not be set. 206 207 Return XML_TOK_INVALID when the string does not start a valid 208 token; nextTokPtr will be set to point to the character which made 209 the token invalid. 210 211 Otherwise the string starts with a valid token; nextTokPtr will be 212 set to point to the character following the end of that token. 213 214 Each data character counts as a single token, but adjacent data 215 characters may be returned together. Similarly for characters in 216 the prolog outside literals, comments and processing instructions. 217 */ 218 219 #define XmlTok(enc, state, ptr, end, nextTokPtr) \ 220 (((enc)->scanners[state])(enc, ptr, end, nextTokPtr)) 221 222 #define XmlPrologTok(enc, ptr, end, nextTokPtr) \ 223 XmlTok(enc, XML_PROLOG_STATE, ptr, end, nextTokPtr) 224 225 #define XmlContentTok(enc, ptr, end, nextTokPtr) \ 226 XmlTok(enc, XML_CONTENT_STATE, ptr, end, nextTokPtr) 227 228 #define XmlCdataSectionTok(enc, ptr, end, nextTokPtr) \ 229 XmlTok(enc, XML_CDATA_SECTION_STATE, ptr, end, nextTokPtr) 230 231 #ifdef XML_DTD 232 233 # define XmlIgnoreSectionTok(enc, ptr, end, nextTokPtr) \ 234 XmlTok(enc, XML_IGNORE_SECTION_STATE, ptr, end, nextTokPtr) 235 236 #endif /* XML_DTD */ 237 238 /* This is used for performing a 2nd-level tokenization on the content 239 of a literal that has already been returned by XmlTok. 240 */ 241 #define XmlLiteralTok(enc, literalType, ptr, end, nextTokPtr) \ 242 (((enc)->literalScanners[literalType])(enc, ptr, end, nextTokPtr)) 243 244 #define XmlAttributeValueTok(enc, ptr, end, nextTokPtr) \ 245 XmlLiteralTok(enc, XML_ATTRIBUTE_VALUE_LITERAL, ptr, end, nextTokPtr) 246 247 #define XmlEntityValueTok(enc, ptr, end, nextTokPtr) \ 248 XmlLiteralTok(enc, XML_ENTITY_VALUE_LITERAL, ptr, end, nextTokPtr) 249 250 #define XmlNameMatchesAscii(enc, ptr1, end1, ptr2) \ 251 (((enc)->nameMatchesAscii)(enc, ptr1, end1, ptr2)) 252 253 #define XmlNameLength(enc, ptr) (((enc)->nameLength)(enc, ptr)) 254 255 #define XmlSkipS(enc, ptr) (((enc)->skipS)(enc, ptr)) 256 257 #define XmlGetAttributes(enc, ptr, attsMax, atts) \ 258 (((enc)->getAtts)(enc, ptr, attsMax, atts)) 259 260 #define XmlCharRefNumber(enc, ptr) (((enc)->charRefNumber)(enc, ptr)) 261 262 #define XmlPredefinedEntityName(enc, ptr, end) \ 263 (((enc)->predefinedEntityName)(enc, ptr, end)) 264 265 #define XmlUpdatePosition(enc, ptr, end, pos) \ 266 (((enc)->updatePosition)(enc, ptr, end, pos)) 267 268 #define XmlIsPublicId(enc, ptr, end, badPtr) \ 269 (((enc)->isPublicId)(enc, ptr, end, badPtr)) 270 271 #define XmlUtf8Convert(enc, fromP, fromLim, toP, toLim) \ 272 (((enc)->utf8Convert)(enc, fromP, fromLim, toP, toLim)) 273 274 #define XmlUtf16Convert(enc, fromP, fromLim, toP, toLim) \ 275 (((enc)->utf16Convert)(enc, fromP, fromLim, toP, toLim)) 276 277 typedef struct { 278 ENCODING initEnc; 279 const ENCODING **encPtr; 280 } INIT_ENCODING; 281 282 int XmlParseXmlDecl(int isGeneralTextEntity, const ENCODING *enc, 283 const char *ptr, const char *end, const char **badPtr, 284 const char **versionPtr, const char **versionEndPtr, 285 const char **encodingNamePtr, 286 const ENCODING **namedEncodingPtr, int *standalonePtr); 287 288 int XmlInitEncoding(INIT_ENCODING *, const ENCODING **, const char *name); 289 const ENCODING *XmlGetUtf8InternalEncoding(void); 290 const ENCODING *XmlGetUtf16InternalEncoding(void); 291 int FASTCALL XmlUtf8Encode(int charNumber, char *buf); 292 int FASTCALL XmlUtf16Encode(int charNumber, unsigned short *buf); 293 int XmlSizeOfUnknownEncoding(void); 294 295 typedef int(XMLCALL *CONVERTER)(void *userData, const char *p); 296 297 ENCODING *XmlInitUnknownEncoding(void *mem, int *table, CONVERTER convert, 298 void *userData); 299 300 int XmlParseXmlDeclNS(int isGeneralTextEntity, const ENCODING *enc, 301 const char *ptr, const char *end, const char **badPtr, 302 const char **versionPtr, const char **versionEndPtr, 303 const char **encodingNamePtr, 304 const ENCODING **namedEncodingPtr, int *standalonePtr); 305 306 int XmlInitEncodingNS(INIT_ENCODING *, const ENCODING **, const char *name); 307 const ENCODING *XmlGetUtf8InternalEncodingNS(void); 308 const ENCODING *XmlGetUtf16InternalEncodingNS(void); 309 ENCODING *XmlInitUnknownEncodingNS(void *mem, int *table, CONVERTER convert, 310 void *userData); 311 #ifdef __cplusplus 312 } 313 #endif 314 315 #endif /* not XmlTok_INCLUDED */ 316