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 Clark Cooper <coopercc@users.sourceforge.net> 11 Copyright (c) 2002 Fred L. Drake, Jr. <fdrake@users.sourceforge.net> 12 Copyright (c) 2002-2005 Karl Waclawek <karl@waclawek.net> 13 Copyright (c) 2016-2024 Sebastian Pipping <sebastian@pipping.org> 14 Copyright (c) 2017 Rhodri James <rhodri@wildebeest.org.uk> 15 Licensed under the MIT license: 16 17 Permission is hereby granted, free of charge, to any person obtaining 18 a copy of this software and associated documentation files (the 19 "Software"), to deal in the Software without restriction, including 20 without limitation the rights to use, copy, modify, merge, publish, 21 distribute, sublicense, and/or sell copies of the Software, and to permit 22 persons to whom the Software is furnished to do so, subject to the 23 following conditions: 24 25 The above copyright notice and this permission notice shall be included 26 in all copies or substantial portions of the Software. 27 28 THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, 29 EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF 30 MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN 31 NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, 32 DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR 33 OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE 34 USE OR OTHER DEALINGS IN THE SOFTWARE. 35 */ 36 37 #include "../../lv_conf_internal.h" 38 #if LV_USE_XML 39 40 #ifndef XmlTok_INCLUDED 41 #define XmlTok_INCLUDED 1 42 43 #ifdef __cplusplus 44 extern "C" { 45 #endif 46 47 /* The following token may be returned by XmlContentTok */ 48 #define XML_TOK_TRAILING_RSQB \ 49 -5 /* ] or ]] at the end of the scan; might be \ 50 start of illegal ]]> sequence */ 51 /* The following tokens may be returned by both XmlPrologTok and 52 XmlContentTok. 53 */ 54 #define XML_TOK_NONE -4 /* The string to be scanned is empty */ 55 #define XML_TOK_TRAILING_CR \ 56 -3 /* A CR at the end of the scan; \ 57 might be part of CRLF sequence */ 58 #define XML_TOK_PARTIAL_CHAR -2 /* only part of a multibyte sequence */ 59 #define XML_TOK_PARTIAL -1 /* only part of a token */ 60 #define XML_TOK_INVALID 0 61 62 /* The following tokens are returned by XmlContentTok; some are also 63 returned by XmlAttributeValueTok, XmlEntityTok, XmlCdataSectionTok. 64 */ 65 #define XML_TOK_START_TAG_WITH_ATTS 1 66 #define XML_TOK_START_TAG_NO_ATTS 2 67 #define XML_TOK_EMPTY_ELEMENT_WITH_ATTS 3 /* empty element tag <e/> */ 68 #define XML_TOK_EMPTY_ELEMENT_NO_ATTS 4 69 #define XML_TOK_END_TAG 5 70 #define XML_TOK_DATA_CHARS 6 71 #define XML_TOK_DATA_NEWLINE 7 72 #define XML_TOK_CDATA_SECT_OPEN 8 73 #define XML_TOK_ENTITY_REF 9 74 #define XML_TOK_CHAR_REF 10 /* numeric character reference */ 75 76 /* The following tokens may be returned by both XmlPrologTok and 77 XmlContentTok. 78 */ 79 #define XML_TOK_PI 11 /* processing instruction */ 80 #define XML_TOK_XML_DECL 12 /* XML decl or text decl */ 81 #define XML_TOK_COMMENT 13 82 #define XML_TOK_BOM 14 /* Byte order mark */ 83 84 /* The following tokens are returned only by XmlPrologTok */ 85 #define XML_TOK_PROLOG_S 15 86 #define XML_TOK_DECL_OPEN 16 /* <!foo */ 87 #define XML_TOK_DECL_CLOSE 17 /* > */ 88 #define XML_TOK_NAME 18 89 #define XML_TOK_NMTOKEN 19 90 #define XML_TOK_POUND_NAME 20 /* #name */ 91 #define XML_TOK_OR 21 /* | */ 92 #define XML_TOK_PERCENT 22 93 #define XML_TOK_OPEN_PAREN 23 94 #define XML_TOK_CLOSE_PAREN 24 95 #define XML_TOK_OPEN_BRACKET 25 96 #define XML_TOK_CLOSE_BRACKET 26 97 #define XML_TOK_LITERAL 27 98 #define XML_TOK_PARAM_ENTITY_REF 28 99 #define XML_TOK_INSTANCE_START 29 100 101 /* The following occur only in element type declarations */ 102 #define XML_TOK_NAME_QUESTION 30 /* name? */ 103 #define XML_TOK_NAME_ASTERISK 31 /* name* */ 104 #define XML_TOK_NAME_PLUS 32 /* name+ */ 105 #define XML_TOK_COND_SECT_OPEN 33 /* <![ */ 106 #define XML_TOK_COND_SECT_CLOSE 34 /* ]]> */ 107 #define XML_TOK_CLOSE_PAREN_QUESTION 35 /* )? */ 108 #define XML_TOK_CLOSE_PAREN_ASTERISK 36 /* )* */ 109 #define XML_TOK_CLOSE_PAREN_PLUS 37 /* )+ */ 110 #define XML_TOK_COMMA 38 111 112 /* The following token is returned only by XmlAttributeValueTok */ 113 #define XML_TOK_ATTRIBUTE_VALUE_S 39 114 115 /* The following token is returned only by XmlCdataSectionTok */ 116 #define XML_TOK_CDATA_SECT_CLOSE 40 117 118 /* With namespace processing this is returned by XmlPrologTok for a 119 name with a colon. 120 */ 121 #define XML_TOK_PREFIXED_NAME 41 122 123 #ifdef XML_DTD 124 # define XML_TOK_IGNORE_SECT 42 125 #endif /* XML_DTD */ 126 127 #ifdef XML_DTD 128 # define XML_N_STATES 4 129 #else /* not XML_DTD */ 130 # define XML_N_STATES 3 131 #endif /* not XML_DTD */ 132 133 #define XML_PROLOG_STATE 0 134 #define XML_CONTENT_STATE 1 135 #define XML_CDATA_SECTION_STATE 2 136 #ifdef XML_DTD 137 # define XML_IGNORE_SECTION_STATE 3 138 #endif /* XML_DTD */ 139 140 #define XML_N_LITERAL_TYPES 2 141 #define XML_ATTRIBUTE_VALUE_LITERAL 0 142 #define XML_ENTITY_VALUE_LITERAL 1 143 144 /* The size of the buffer passed to XmlUtf8Encode must be at least this. */ 145 #define XML_UTF8_ENCODE_MAX 4 146 /* The size of the buffer passed to XmlUtf16Encode must be at least this. */ 147 #define XML_UTF16_ENCODE_MAX 2 148 149 typedef struct position { 150 /* first line and first column are 0 not 1 */ 151 XML_Size lineNumber; 152 XML_Size columnNumber; 153 } POSITION; 154 155 typedef struct { 156 const char *name; 157 const char *valuePtr; 158 const char *valueEnd; 159 char normalized; 160 } ATTRIBUTE; 161 162 struct encoding; 163 typedef struct encoding ENCODING; 164 165 typedef int(PTRCALL *SCANNER)(const ENCODING *, const char *, const char *, 166 const char **); 167 168 enum XML_Convert_Result { 169 XML_CONVERT_COMPLETED = 0, 170 XML_CONVERT_INPUT_INCOMPLETE = 1, 171 XML_CONVERT_OUTPUT_EXHAUSTED 172 = 2 /* and therefore potentially input remaining as well */ 173 }; 174 175 struct encoding { 176 SCANNER scanners[XML_N_STATES]; 177 SCANNER literalScanners[XML_N_LITERAL_TYPES]; 178 int(PTRCALL *nameMatchesAscii)(const ENCODING *, const char *, const char *, 179 const char *); 180 int(PTRFASTCALL *nameLength)(const ENCODING *, const char *); 181 const char *(PTRFASTCALL *skipS)(const ENCODING *, const char *); 182 int(PTRCALL *getAtts)(const ENCODING *enc, const char *ptr, int attsMax, 183 ATTRIBUTE *atts); 184 int(PTRFASTCALL *charRefNumber)(const ENCODING *enc, const char *ptr); 185 int(PTRCALL *predefinedEntityName)(const ENCODING *, const char *, 186 const char *); 187 void(PTRCALL *updatePosition)(const ENCODING *, const char *ptr, 188 const char *end, POSITION *); 189 int(PTRCALL *isPublicId)(const ENCODING *enc, const char *ptr, 190 const char *end, const char **badPtr); 191 enum XML_Convert_Result(PTRCALL *utf8Convert)(const ENCODING *enc, 192 const char **fromP, 193 const char *fromLim, char **toP, 194 const char *toLim); 195 enum XML_Convert_Result(PTRCALL *utf16Convert)(const ENCODING *enc, 196 const char **fromP, 197 const char *fromLim, 198 unsigned short **toP, 199 const unsigned short *toLim); 200 int minBytesPerChar; 201 char isUtf8; 202 char isUtf16; 203 }; 204 205 /* Scan the string starting at ptr until the end of the next complete 206 token, but do not scan past eptr. Return an integer giving the 207 type of token. 208 209 Return XML_TOK_NONE when ptr == eptr; nextTokPtr will not be set. 210 211 Return XML_TOK_PARTIAL when the string does not contain a complete 212 token; nextTokPtr will not be set. 213 214 Return XML_TOK_INVALID when the string does not start a valid 215 token; nextTokPtr will be set to point to the character which made 216 the token invalid. 217 218 Otherwise the string starts with a valid token; nextTokPtr will be 219 set to point to the character following the end of that token. 220 221 Each data character counts as a single token, but adjacent data 222 characters may be returned together. Similarly for characters in 223 the prolog outside literals, comments and processing instructions. 224 */ 225 226 #define XmlTok(enc, state, ptr, end, nextTokPtr) \ 227 (((enc)->scanners[state])(enc, ptr, end, nextTokPtr)) 228 229 #define XmlPrologTok(enc, ptr, end, nextTokPtr) \ 230 XmlTok(enc, XML_PROLOG_STATE, ptr, end, nextTokPtr) 231 232 #define XmlContentTok(enc, ptr, end, nextTokPtr) \ 233 XmlTok(enc, XML_CONTENT_STATE, ptr, end, nextTokPtr) 234 235 #define XmlCdataSectionTok(enc, ptr, end, nextTokPtr) \ 236 XmlTok(enc, XML_CDATA_SECTION_STATE, ptr, end, nextTokPtr) 237 238 #ifdef XML_DTD 239 240 # define XmlIgnoreSectionTok(enc, ptr, end, nextTokPtr) \ 241 XmlTok(enc, XML_IGNORE_SECTION_STATE, ptr, end, nextTokPtr) 242 243 #endif /* XML_DTD */ 244 245 /* This is used for performing a 2nd-level tokenization on the content 246 of a literal that has already been returned by XmlTok. 247 */ 248 #define XmlLiteralTok(enc, literalType, ptr, end, nextTokPtr) \ 249 (((enc)->literalScanners[literalType])(enc, ptr, end, nextTokPtr)) 250 251 #define XmlAttributeValueTok(enc, ptr, end, nextTokPtr) \ 252 XmlLiteralTok(enc, XML_ATTRIBUTE_VALUE_LITERAL, ptr, end, nextTokPtr) 253 254 #define XmlEntityValueTok(enc, ptr, end, nextTokPtr) \ 255 XmlLiteralTok(enc, XML_ENTITY_VALUE_LITERAL, ptr, end, nextTokPtr) 256 257 #define XmlNameMatchesAscii(enc, ptr1, end1, ptr2) \ 258 (((enc)->nameMatchesAscii)(enc, ptr1, end1, ptr2)) 259 260 #define XmlNameLength(enc, ptr) (((enc)->nameLength)(enc, ptr)) 261 262 #define XmlSkipS(enc, ptr) (((enc)->skipS)(enc, ptr)) 263 264 #define XmlGetAttributes(enc, ptr, attsMax, atts) \ 265 (((enc)->getAtts)(enc, ptr, attsMax, atts)) 266 267 #define XmlCharRefNumber(enc, ptr) (((enc)->charRefNumber)(enc, ptr)) 268 269 #define XmlPredefinedEntityName(enc, ptr, end) \ 270 (((enc)->predefinedEntityName)(enc, ptr, end)) 271 272 #define XmlUpdatePosition(enc, ptr, end, pos) \ 273 (((enc)->updatePosition)(enc, ptr, end, pos)) 274 275 #define XmlIsPublicId(enc, ptr, end, badPtr) \ 276 (((enc)->isPublicId)(enc, ptr, end, badPtr)) 277 278 #define XmlUtf8Convert(enc, fromP, fromLim, toP, toLim) \ 279 (((enc)->utf8Convert)(enc, fromP, fromLim, toP, toLim)) 280 281 #define XmlUtf16Convert(enc, fromP, fromLim, toP, toLim) \ 282 (((enc)->utf16Convert)(enc, fromP, fromLim, toP, toLim)) 283 284 typedef struct { 285 ENCODING initEnc; 286 const ENCODING **encPtr; 287 } INIT_ENCODING; 288 289 int XmlParseXmlDecl(int isGeneralTextEntity, const ENCODING *enc, 290 const char *ptr, const char *end, const char **badPtr, 291 const char **versionPtr, const char **versionEndPtr, 292 const char **encodingNamePtr, 293 const ENCODING **namedEncodingPtr, int *standalonePtr); 294 295 int XmlInitEncoding(INIT_ENCODING *p, const ENCODING **encPtr, 296 const char *name); 297 const ENCODING *XmlGetUtf8InternalEncoding(void); 298 const ENCODING *XmlGetUtf16InternalEncoding(void); 299 int FASTCALL XmlUtf8Encode(int charNumber, char *buf); 300 int FASTCALL XmlUtf16Encode(int charNumber, unsigned short *buf); 301 int XmlSizeOfUnknownEncoding(void); 302 303 typedef int(XMLCALL *CONVERTER)(void *userData, const char *p); 304 305 ENCODING *XmlInitUnknownEncoding(void *mem, int *table, CONVERTER convert, 306 void *userData); 307 308 int XmlParseXmlDeclNS(int isGeneralTextEntity, const ENCODING *enc, 309 const char *ptr, const char *end, const char **badPtr, 310 const char **versionPtr, const char **versionEndPtr, 311 const char **encodingNamePtr, 312 const ENCODING **namedEncodingPtr, int *standalonePtr); 313 314 int XmlInitEncodingNS(INIT_ENCODING *p, const ENCODING **encPtr, 315 const char *name); 316 const ENCODING *XmlGetUtf8InternalEncodingNS(void); 317 const ENCODING *XmlGetUtf16InternalEncodingNS(void); 318 ENCODING *XmlInitUnknownEncodingNS(void *mem, int *table, CONVERTER convert, 319 void *userData); 320 #ifdef __cplusplus 321 } 322 #endif 323 324 #endif /* not XmlTok_INCLUDED */ 325 326 #endif /* LV_USE_XML */ 327 328
