| 1 | /* |
| 2 | Copyright (c) 1998, 1999 Thai Open Source Software Center Ltd |
| 3 | See the file copying.txt for copying permission. |
| 4 | */ |
| 5 | |
| 6 | #include "xmldef.h" |
| 7 | #include "xmltok.h" |
| 8 | #include "nametab.h" |
| 9 | |
| 10 | #ifdef XML_DTD |
| 11 | #define IGNORE_SECTION_TOK_VTABLE , PREFIX(ignoreSectionTok) |
| 12 | #else |
| 13 | #define IGNORE_SECTION_TOK_VTABLE /* as nothing */ |
| 14 | #endif |
| 15 | |
| 16 | #define VTABLE1 \ |
| 17 | { PREFIX(prologTok), PREFIX(contentTok), \ |
| 18 | PREFIX(cdataSectionTok) IGNORE_SECTION_TOK_VTABLE }, \ |
| 19 | { PREFIX(attributeValueTok), PREFIX(entityValueTok) }, \ |
| 20 | PREFIX(sameName), \ |
| 21 | PREFIX(nameMatchesAscii), \ |
| 22 | PREFIX(nameLength), \ |
| 23 | PREFIX(skipS), \ |
| 24 | PREFIX(getAtts), \ |
| 25 | PREFIX(charRefNumber), \ |
| 26 | PREFIX(predefinedEntityName), \ |
| 27 | PREFIX(updatePosition), \ |
| 28 | PREFIX(isPublicId) |
| 29 | |
| 30 | #define VTABLE VTABLE1, PREFIX(toUtf8), PREFIX(toUtf16) |
| 31 | |
| 32 | #define UCS2_GET_NAMING(pages, hi, lo) \ |
| 33 | (namingBitmap[(pages[hi] << 3) + ((lo) >> 5)] & (1 << ((lo) & 0x1F))) |
| 34 | |
| 35 | /* A 2 byte UTF-8 representation splits the characters 11 bits |
| 36 | between the bottom 5 and 6 bits of the bytes. |
| 37 | We need 8 bits to index into pages, 3 bits to add to that index and |
| 38 | 5 bits to generate the mask. */ |
| 39 | #define UTF8_GET_NAMING2(pages, byte) \ |
| 40 | (namingBitmap[((pages)[(((byte)[0]) >> 2) & 7] << 3) \ |
| 41 | + ((((byte)[0]) & 3) << 1) \ |
| 42 | + ((((byte)[1]) >> 5) & 1)] \ |
| 43 | & (1 << (((byte)[1]) & 0x1F))) |
| 44 | |
| 45 | /* A 3 byte UTF-8 representation splits the characters 16 bits |
| 46 | between the bottom 4, 6 and 6 bits of the bytes. |
| 47 | We need 8 bits to index into pages, 3 bits to add to that index and |
| 48 | 5 bits to generate the mask. */ |
| 49 | #define UTF8_GET_NAMING3(pages, byte) \ |
| 50 | (namingBitmap[((pages)[((((byte)[0]) & 0xF) << 4) \ |
| 51 | + ((((byte)[1]) >> 2) & 0xF)] \ |
| 52 | << 3) \ |
| 53 | + ((((byte)[1]) & 3) << 1) \ |
| 54 | + ((((byte)[2]) >> 5) & 1)] \ |
| 55 | & (1 << (((byte)[2]) & 0x1F))) |
| 56 | |
| 57 | #define UTF8_GET_NAMING(pages, p, n) \ |
| 58 | ((n) == 2 \ |
| 59 | ? UTF8_GET_NAMING2(pages, (const unsigned char *)(p)) \ |
| 60 | : ((n) == 3 \ |
| 61 | ? UTF8_GET_NAMING3(pages, (const unsigned char *)(p)) \ |
| 62 | : 0)) |
| 63 | |
| 64 | #define UTF8_INVALID3(p) \ |
| 65 | ((*p) == 0xED \ |
| 66 | ? (((p)[1] & 0x20) != 0) \ |
| 67 | : ((*p) == 0xEF \ |
| 68 | ? ((p)[1] == 0xBF && ((p)[2] == 0xBF || (p)[2] == 0xBE)) \ |
| 69 | : 0)) |
| 70 | |
| 71 | #define UTF8_INVALID4(p) ((*p) == 0xF4 && ((p)[1] & 0x30) != 0) |
| 72 | |
| 73 | static |
| 74 | int isNever(const ENCODING *enc, const char *p) |
| 75 | { |
| 76 | return 0; |
| 77 | } |
| 78 | |
| 79 | static |
| 80 | int utf8_isName2(const ENCODING *enc, const char *p) |
| 81 | { |
| 82 | return UTF8_GET_NAMING2(namePages, (const unsigned char *)p); |
| 83 | } |
| 84 | |
| 85 | static |
| 86 | int utf8_isName3(const ENCODING *enc, const char *p) |
| 87 | { |
| 88 | return UTF8_GET_NAMING3(namePages, (const unsigned char *)p); |
| 89 | } |
| 90 | |
| 91 | #define utf8_isName4 isNever |
| 92 | |
| 93 | static |
| 94 | int utf8_isNmstrt2(const ENCODING *enc, const char *p) |
| 95 | { |
| 96 | return UTF8_GET_NAMING2(nmstrtPages, (const unsigned char *)p); |
| 97 | } |
| 98 | |
| 99 | static |
| 100 | int utf8_isNmstrt3(const ENCODING *enc, const char *p) |
| 101 | { |
| 102 | return UTF8_GET_NAMING3(nmstrtPages, (const unsigned char *)p); |
| 103 | } |
| 104 | |
| 105 | #define utf8_isNmstrt4 isNever |
| 106 | |
| 107 | #define utf8_isInvalid2 isNever |
| 108 | |
| 109 | static |
| 110 | int utf8_isInvalid3(const ENCODING *enc, const char *p) |
| 111 | { |
| 112 | return UTF8_INVALID3((const unsigned char *)p); |
| 113 | } |
| 114 | |
| 115 | static |
| 116 | int utf8_isInvalid4(const ENCODING *enc, const char *p) |
| 117 | { |
| 118 | return UTF8_INVALID4((const unsigned char *)p); |
| 119 | } |
| 120 | |
| 121 | struct normal_encoding { |
| 122 | ENCODING enc; |
| 123 | unsigned char type[256]; |
| 124 | #ifdef XML_MIN_SIZE |
| 125 | int (*byteType)(const ENCODING *, const char *); |
| 126 | int (*isNameMin)(const ENCODING *, const char *); |
| 127 | int (*isNmstrtMin)(const ENCODING *, const char *); |
| 128 | int (*byteToAscii)(const ENCODING *, const char *); |
| 129 | int (*charMatches)(const ENCODING *, const char *, int); |
| 130 | #endif /* XML_MIN_SIZE */ |
| 131 | int (*isName2)(const ENCODING *, const char *); |
| 132 | int (*isName3)(const ENCODING *, const char *); |
| 133 | int (*isName4)(const ENCODING *, const char *); |
| 134 | int (*isNmstrt2)(const ENCODING *, const char *); |
| 135 | int (*isNmstrt3)(const ENCODING *, const char *); |
| 136 | int (*isNmstrt4)(const ENCODING *, const char *); |
| 137 | int (*isInvalid2)(const ENCODING *, const char *); |
| 138 | int (*isInvalid3)(const ENCODING *, const char *); |
| 139 | int (*isInvalid4)(const ENCODING *, const char *); |
| 140 | }; |
| 141 | |
| 142 | #ifdef XML_MIN_SIZE |
| 143 | |
| 144 | #define STANDARD_VTABLE(E) \ |
| 145 | E ## byteType, \ |
| 146 | E ## isNameMin, \ |
| 147 | E ## isNmstrtMin, \ |
| 148 | E ## byteToAscii, \ |
| 149 | E ## charMatches, |
| 150 | |
| 151 | #else |
| 152 | |
| 153 | #define STANDARD_VTABLE(E) /* as nothing */ |
| 154 | |
| 155 | #endif |
| 156 | |
| 157 | #define NORMAL_VTABLE(E) \ |
| 158 | E ## isName2, \ |
| 159 | E ## isName3, \ |
| 160 | E ## isName4, \ |
| 161 | E ## isNmstrt2, \ |
| 162 | E ## isNmstrt3, \ |
| 163 | E ## isNmstrt4, \ |
| 164 | E ## isInvalid2, \ |
| 165 | E ## isInvalid3, \ |
| 166 | E ## isInvalid4 |
| 167 | |
| 168 | static int checkCharRefNumber(int); |
| 169 | |
| 170 | #include "xmltok_impl.h" |
| 171 | #include "ascii.h" |
| 172 | |
| 173 | #ifdef XML_MIN_SIZE |
| 174 | #define sb_isNameMin isNever |
| 175 | #define sb_isNmstrtMin isNever |
| 176 | #endif |
| 177 | |
| 178 | #ifdef XML_MIN_SIZE |
| 179 | #define MINBPC(enc) ((enc)->minBytesPerChar) |
| 180 | #else |
| 181 | /* minimum bytes per character */ |
| 182 | #define MINBPC(enc) 1 |
| 183 | #endif |
| 184 | |
| 185 | #define SB_BYTE_TYPE(enc, p) \ |
| 186 | (((struct normal_encoding *)(enc))->type[(unsigned char)*(p)]) |
| 187 | |
| 188 | #ifdef XML_MIN_SIZE |
| 189 | static |
| 190 | int sb_byteType(const ENCODING *enc, const char *p) |
| 191 | { |
| 192 | return SB_BYTE_TYPE(enc, p); |
| 193 | } |
| 194 | #define BYTE_TYPE(enc, p) \ |
| 195 | (((const struct normal_encoding *)(enc))->byteType(enc, p)) |
| 196 | #else |
| 197 | #define BYTE_TYPE(enc, p) SB_BYTE_TYPE(enc, p) |
| 198 | #endif |
| 199 | |
| 200 | #ifdef XML_MIN_SIZE |
| 201 | #define BYTE_TO_ASCII(enc, p) \ |
| 202 | (((const struct normal_encoding *)(enc))->byteToAscii(enc, p)) |
| 203 | static |
| 204 | int sb_byteToAscii(const ENCODING *enc, const char *p) |
| 205 | { |
| 206 | return *p; |
| 207 | } |
| 208 | #else |
| 209 | #define BYTE_TO_ASCII(enc, p) (*(p)) |
| 210 | #endif |
| 211 | |
| 212 | #define IS_NAME_CHAR(enc, p, n) \ |
| 213 | (((const struct normal_encoding *)(enc))->isName ## n(enc, p)) |
| 214 | #define IS_NMSTRT_CHAR(enc, p, n) \ |
| 215 | (((const struct normal_encoding *)(enc))->isNmstrt ## n(enc, p)) |
| 216 | #define IS_INVALID_CHAR(enc, p, n) \ |
| 217 | (((const struct normal_encoding *)(enc))->isInvalid ## n(enc, p)) |
| 218 | |
| 219 | #ifdef XML_MIN_SIZE |
| 220 | #define IS_NAME_CHAR_MINBPC(enc, p) \ |
| 221 | (((const struct normal_encoding *)(enc))->isNameMin(enc, p)) |
| 222 | #define IS_NMSTRT_CHAR_MINBPC(enc, p) \ |
| 223 | (((const struct normal_encoding *)(enc))->isNmstrtMin(enc, p)) |
| 224 | #else |
| 225 | #define IS_NAME_CHAR_MINBPC(enc, p) (0) |
| 226 | #define IS_NMSTRT_CHAR_MINBPC(enc, p) (0) |
| 227 | #endif |
| 228 | |
| 229 | #ifdef XML_MIN_SIZE |
| 230 | #define CHAR_MATCHES(enc, p, c) \ |
| 231 | (((const struct normal_encoding *)(enc))->charMatches(enc, p, c)) |
| 232 | static |
| 233 | int sb_charMatches(const ENCODING *enc, const char *p, int c) |
| 234 | { |
| 235 | return *p == c; |
| 236 | } |
| 237 | #else |
| 238 | /* c is an ASCII character */ |
| 239 | #define CHAR_MATCHES(enc, p, c) (*(p) == c) |
| 240 | #endif |
| 241 | |
| 242 | #define PREFIX(ident) normal_ ## ident |
| 243 | #include "xmltok_impl.c" |
| 244 | |
| 245 | #undef MINBPC |
| 246 | #undef BYTE_TYPE |
| 247 | #undef BYTE_TO_ASCII |
| 248 | #undef CHAR_MATCHES |
| 249 | #undef IS_NAME_CHAR |
| 250 | #undef IS_NAME_CHAR_MINBPC |
| 251 | #undef IS_NMSTRT_CHAR |
| 252 | #undef IS_NMSTRT_CHAR_MINBPC |
| 253 | #undef IS_INVALID_CHAR |
| 254 | |
| 255 | enum { /* UTF8_cvalN is value of masked first byte of N byte sequence */ |
| 256 | UTF8_cval1 = 0x00, |
| 257 | UTF8_cval2 = 0xc0, |
| 258 | UTF8_cval3 = 0xe0, |
| 259 | UTF8_cval4 = 0xf0 |
| 260 | }; |
| 261 | |
| 262 | static |
| 263 | void utf8_toUtf8(const ENCODING *enc, |
| 264 | const char **fromP, const char *fromLim, |
| 265 | char **toP, const char *toLim) |
| 266 | { |
| 267 | char *to; |
| 268 | const char *from; |
| 269 | if (fromLim - *fromP > toLim - *toP) { |
| 270 | /* Avoid copying partial characters. */ |
| 271 | for (fromLim = *fromP + (toLim - *toP); fromLim > *fromP; fromLim--) |
| 272 | if (((unsigned char)fromLim[-1] & 0xc0) != 0x80) |
| 273 | break; |
| 274 | } |
| 275 | for (to = *toP, from = *fromP; from != fromLim; from++, to++) |
| 276 | *to = *from; |
| 277 | *fromP = from; |
| 278 | *toP = to; |
| 279 | } |
| 280 | |
| 281 | static |
| 282 | void utf8_toUtf16(const ENCODING *enc, |
| 283 | const char **fromP, const char *fromLim, |
| 284 | unsigned short **toP, const unsigned short *toLim) |
| 285 | { |
| 286 | unsigned short *to = *toP; |
| 287 | const char *from = *fromP; |
| 288 | while (from != fromLim && to != toLim) { |
| 289 | switch (((struct normal_encoding *)enc)->type[(unsigned char)*from]) { |
| 290 | case BT_LEAD2: |
| 291 | *to++ = ((from[0] & 0x1f) << 6) | (from[1] & 0x3f); |
| 292 | from += 2; |
| 293 | break; |
| 294 | case BT_LEAD3: |
| 295 | *to++ = ((from[0] & 0xf) << 12) | ((from[1] & 0x3f) << 6) | (from[2] & 0x3f); |
| 296 | from += 3; |
| 297 | break; |
| 298 | case BT_LEAD4: |
| 299 | { |
| 300 | unsigned long n; |
| 301 | if (to + 1 == toLim) |
| 302 | break; |
| 303 | n = ((from[0] & 0x7) << 18) | ((from[1] & 0x3f) << 12) | ((from[2] & 0x3f) << 6) | (from[3] & 0x3f); |
| 304 | n -= 0x10000; |
| 305 | to[0] = (unsigned short)((n >> 10) | 0xD800); |
| 306 | to[1] = (unsigned short)((n & 0x3FF) | 0xDC00); |
| 307 | to += 2; |
| 308 | from += 4; |
| 309 | } |
| 310 | break; |
| 311 | default: |
| 312 | *to++ = *from++; |
| 313 | break; |
| 314 | } |
| 315 | } |
| 316 | *fromP = from; |
| 317 | *toP = to; |
| 318 | } |
| 319 | |
| 320 | #ifdef XML_NS |
| 321 | static const struct normal_encoding utf8_encoding_ns = { |
| 322 | { VTABLE1, utf8_toUtf8, utf8_toUtf16, 1, 1, 0 }, |
| 323 | { |
| 324 | #include "asciitab.h" |
| 325 | #include "utf8tab.h" |
| 326 | }, |
| 327 | STANDARD_VTABLE(sb_) NORMAL_VTABLE(utf8_) |
| 328 | }; |
| 329 | #endif |
| 330 | |
| 331 | static const struct normal_encoding utf8_encoding = { |
| 332 | { VTABLE1, utf8_toUtf8, utf8_toUtf16, 1, 1, 0 }, |
| 333 | { |
| 334 | #define BT_COLON BT_NMSTRT |
| 335 | #include "asciitab.h" |
| 336 | #undef BT_COLON |
| 337 | #include "utf8tab.h" |
| 338 | }, |
| 339 | STANDARD_VTABLE(sb_) NORMAL_VTABLE(utf8_) |
| 340 | }; |
| 341 | |
| 342 | #ifdef XML_NS |
| 343 | |
| 344 | static const struct normal_encoding internal_utf8_encoding_ns = { |
| 345 | { VTABLE1, utf8_toUtf8, utf8_toUtf16, 1, 1, 0 }, |
| 346 | { |
| 347 | #include "iasciitab.h" |
| 348 | #include "utf8tab.h" |
| 349 | }, |
| 350 | STANDARD_VTABLE(sb_) NORMAL_VTABLE(utf8_) |
| 351 | }; |
| 352 | |
| 353 | #endif |
| 354 | |
| 355 | static const struct normal_encoding internal_utf8_encoding = { |
| 356 | { VTABLE1, utf8_toUtf8, utf8_toUtf16, 1, 1, 0 }, |
| 357 | { |
| 358 | #define BT_COLON BT_NMSTRT |
| 359 | #include "iasciitab.h" |
| 360 | #undef BT_COLON |
| 361 | #include "utf8tab.h" |
| 362 | }, |
| 363 | STANDARD_VTABLE(sb_) NORMAL_VTABLE(utf8_) |
| 364 | }; |
| 365 | |
| 366 | static |
| 367 | void latin1_toUtf8(const ENCODING *enc, |
| 368 | const char **fromP, const char *fromLim, |
| 369 | char **toP, const char *toLim) |
| 370 | { |
| 371 | for (;;) { |
| 372 | unsigned char c; |
| 373 | if (*fromP == fromLim) |
| 374 | break; |
| 375 | c = (unsigned char)**fromP; |
| 376 | if (c & 0x80) { |
| 377 | if (toLim - *toP < 2) |
| 378 | break; |
| 379 | *(*toP)++ = ((c >> 6) | UTF8_cval2); |
| 380 | *(*toP)++ = ((c & 0x3f) | 0x80); |
| 381 | (*fromP)++; |
| 382 | } |
| 383 | else { |
| 384 | if (*toP == toLim) |
| 385 | break; |
| 386 | *(*toP)++ = *(*fromP)++; |
| 387 | } |
| 388 | } |
| 389 | } |
| 390 | |
| 391 | static |
| 392 | void latin1_toUtf16(const ENCODING *enc, |
| 393 | const char **fromP, const char *fromLim, |
| 394 | unsigned short **toP, const unsigned short *toLim) |
| 395 | { |
| 396 | while (*fromP != fromLim && *toP != toLim) |
| 397 | *(*toP)++ = (unsigned char)*(*fromP)++; |
| 398 | } |
| 399 | |
| 400 | #ifdef XML_NS |
| 401 | |
| 402 | static const struct normal_encoding latin1_encoding_ns = { |
| 403 | { VTABLE1, latin1_toUtf8, latin1_toUtf16, 1, 0, 0 }, |
| 404 | { |
| 405 | #include "asciitab.h" |
| 406 | #include "latin1tab.h" |
| 407 | }, |
| 408 | STANDARD_VTABLE(sb_) |
| 409 | }; |
| 410 | |
| 411 | #endif |
| 412 | |
| 413 | static const struct normal_encoding latin1_encoding = { |
| 414 | { VTABLE1, latin1_toUtf8, latin1_toUtf16, 1, 0, 0 }, |
| 415 | { |
| 416 | #define BT_COLON BT_NMSTRT |
| 417 | #include "asciitab.h" |
| 418 | #undef BT_COLON |
| 419 | #include "latin1tab.h" |
| 420 | }, |
| 421 | STANDARD_VTABLE(sb_) |
| 422 | }; |
| 423 | |
| 424 | static |
| 425 | void ascii_toUtf8(const ENCODING *enc, |
| 426 | const char **fromP, const char *fromLim, |
| 427 | char **toP, const char *toLim) |
| 428 | { |
| 429 | while (*fromP != fromLim && *toP != toLim) |
| 430 | *(*toP)++ = *(*fromP)++; |
| 431 | } |
| 432 | |
| 433 | #ifdef XML_NS |
| 434 | |
| 435 | static const struct normal_encoding ascii_encoding_ns = { |
| 436 | { VTABLE1, ascii_toUtf8, latin1_toUtf16, 1, 1, 0 }, |
| 437 | { |
| 438 | #include "asciitab.h" |
| 439 | /* BT_NONXML == 0 */ |
| 440 | }, |
| 441 | STANDARD_VTABLE(sb_) |
| 442 | }; |
| 443 | |
| 444 | #endif |
| 445 | |
| 446 | static const struct normal_encoding ascii_encoding = { |
| 447 | { VTABLE1, ascii_toUtf8, latin1_toUtf16, 1, 1, 0 }, |
| 448 | { |
| 449 | #define BT_COLON BT_NMSTRT |
| 450 | #include "asciitab.h" |
| 451 | #undef BT_COLON |
| 452 | /* BT_NONXML == 0 */ |
| 453 | }, |
| 454 | STANDARD_VTABLE(sb_) |
| 455 | }; |
| 456 | |
| 457 | static int unicode_byte_type(char hi, char lo) |
| 458 | { |
| 459 | switch ((unsigned char)hi) { |
| 460 | case 0xD8: case 0xD9: case 0xDA: case 0xDB: |
| 461 | return BT_LEAD4; |
| 462 | case 0xDC: case 0xDD: case 0xDE: case 0xDF: |
| 463 | return BT_TRAIL; |
| 464 | case 0xFF: |
| 465 | switch ((unsigned char)lo) { |
| 466 | case 0xFF: |
| 467 | case 0xFE: |
| 468 | return BT_NONXML; |
| 469 | } |
| 470 | break; |
| 471 | } |
| 472 | return BT_NONASCII; |
| 473 | } |
| 474 | |
| 475 | #define DEFINE_UTF16_TO_UTF8(E) \ |
| 476 | static \ |
| 477 | void E ## toUtf8(const ENCODING *enc, \ |
| 478 | const char **fromP, const char *fromLim, \ |
| 479 | char **toP, const char *toLim) \ |
| 480 | { \ |
| 481 | const char *from; \ |
| 482 | for (from = *fromP; from != fromLim; from += 2) { \ |
| 483 | int plane; \ |
| 484 | unsigned char lo2; \ |
| 485 | unsigned char lo = GET_LO(from); \ |
| 486 | unsigned char hi = GET_HI(from); \ |
| 487 | switch (hi) { \ |
| 488 | case 0: \ |
| 489 | if (lo < 0x80) { \ |
| 490 | if (*toP == toLim) { \ |
| 491 | *fromP = from; \ |
| 492 | return; \ |
| 493 | } \ |
| 494 | *(*toP)++ = lo; \ |
| 495 | break; \ |
| 496 | } \ |
| 497 | /* fall through */ \ |
| 498 | case 0x1: case 0x2: case 0x3: \ |
| 499 | case 0x4: case 0x5: case 0x6: case 0x7: \ |
| 500 | if (toLim - *toP < 2) { \ |
| 501 | *fromP = from; \ |
| 502 | return; \ |
| 503 | } \ |
| 504 | *(*toP)++ = ((lo >> 6) | (hi << 2) | UTF8_cval2); \ |
| 505 | *(*toP)++ = ((lo & 0x3f) | 0x80); \ |
| 506 | break; \ |
| 507 | default: \ |
| 508 | if (toLim - *toP < 3) { \ |
| 509 | *fromP = from; \ |
| 510 | return; \ |
| 511 | } \ |
| 512 | /* 16 bits divided 4, 6, 6 amongst 3 bytes */ \ |
| 513 | *(*toP)++ = ((hi >> 4) | UTF8_cval3); \ |
| 514 | *(*toP)++ = (((hi & 0xf) << 2) | (lo >> 6) | 0x80); \ |
| 515 | *(*toP)++ = ((lo & 0x3f) | 0x80); \ |
| 516 | break; \ |
| 517 | case 0xD8: case 0xD9: case 0xDA: case 0xDB: \ |
| 518 | if (toLim - *toP < 4) { \ |
| 519 | *fromP = from; \ |
| 520 | return; \ |
| 521 | } \ |
| 522 | plane = (((hi & 0x3) << 2) | ((lo >> 6) & 0x3)) + 1; \ |
| 523 | *(*toP)++ = ((plane >> 2) | UTF8_cval4); \ |
| 524 | *(*toP)++ = (((lo >> 2) & 0xF) | ((plane & 0x3) << 4) | 0x80); \ |
| 525 | from += 2; \ |
| 526 | lo2 = GET_LO(from); \ |
| 527 | *(*toP)++ = (((lo & 0x3) << 4) \ |
| 528 | | ((GET_HI(from) & 0x3) << 2) \ |
| 529 | | (lo2 >> 6) \ |
| 530 | | 0x80); \ |
| 531 | *(*toP)++ = ((lo2 & 0x3f) | 0x80); \ |
| 532 | break; \ |
| 533 | } \ |
| 534 | } \ |
| 535 | *fromP = from; \ |
| 536 | } |
| 537 | |
| 538 | #define DEFINE_UTF16_TO_UTF16(E) \ |
| 539 | static \ |
| 540 | void E ## toUtf16(const ENCODING *enc, \ |
| 541 | const char **fromP, const char *fromLim, \ |
| 542 | unsigned short **toP, const unsigned short *toLim) \ |
| 543 | { \ |
| 544 | /* Avoid copying first half only of surrogate */ \ |
| 545 | if (fromLim - *fromP > ((toLim - *toP) << 1) \ |
| 546 | && (GET_HI(fromLim - 2) & 0xF8) == 0xD8) \ |
| 547 | fromLim -= 2; \ |
| 548 | for (; *fromP != fromLim && *toP != toLim; *fromP += 2) \ |
| 549 | *(*toP)++ = (GET_HI(*fromP) << 8) | GET_LO(*fromP); \ |
| 550 | } |
| 551 | |
| 552 | #define SET2(ptr, ch) \ |
| 553 | (((ptr)[0] = ((ch) & 0xff)), ((ptr)[1] = ((ch) >> 8))) |
| 554 | #define GET_LO(ptr) ((unsigned char)(ptr)[0]) |
| 555 | #define GET_HI(ptr) ((unsigned char)(ptr)[1]) |
| 556 | |
| 557 | DEFINE_UTF16_TO_UTF8(little2_) |
| 558 | DEFINE_UTF16_TO_UTF16(little2_) |
| 559 | |
| 560 | #undef SET2 |
| 561 | #undef GET_LO |
| 562 | #undef GET_HI |
| 563 | |
| 564 | #define SET2(ptr, ch) \ |
| 565 | (((ptr)[0] = ((ch) >> 8)), ((ptr)[1] = ((ch) & 0xFF))) |
| 566 | #define GET_LO(ptr) ((unsigned char)(ptr)[1]) |
| 567 | #define GET_HI(ptr) ((unsigned char)(ptr)[0]) |
| 568 | |
| 569 | DEFINE_UTF16_TO_UTF8(big2_) |
| 570 | DEFINE_UTF16_TO_UTF16(big2_) |
| 571 | |
| 572 | #undef SET2 |
| 573 | #undef GET_LO |
| 574 | #undef GET_HI |
| 575 | |
| 576 | #define LITTLE2_BYTE_TYPE(enc, p) \ |
| 577 | ((p)[1] == 0 \ |
| 578 | ? ((struct normal_encoding *)(enc))->type[(unsigned char)*(p)] \ |
| 579 | : unicode_byte_type((p)[1], (p)[0])) |
| 580 | #define LITTLE2_BYTE_TO_ASCII(enc, p) ((p)[1] == 0 ? (p)[0] : -1) |
| 581 | #define LITTLE2_CHAR_MATCHES(enc, p, c) ((p)[1] == 0 && (p)[0] == c) |
| 582 | #define LITTLE2_IS_NAME_CHAR_MINBPC(enc, p) \ |
| 583 | UCS2_GET_NAMING(namePages, (unsigned char)p[1], (unsigned char)p[0]) |
| 584 | #define LITTLE2_IS_NMSTRT_CHAR_MINBPC(enc, p) \ |
| 585 | UCS2_GET_NAMING(nmstrtPages, (unsigned char)p[1], (unsigned char)p[0]) |
| 586 | |
| 587 | #ifdef XML_MIN_SIZE |
| 588 | |
| 589 | static |
| 590 | int little2_byteType(const ENCODING *enc, const char *p) |
| 591 | { |
| 592 | return LITTLE2_BYTE_TYPE(enc, p); |
| 593 | } |
| 594 | |
| 595 | static |
| 596 | int little2_byteToAscii(const ENCODING *enc, const char *p) |
| 597 | { |
| 598 | return LITTLE2_BYTE_TO_ASCII(enc, p); |
| 599 | } |
| 600 | |
| 601 | static |
| 602 | int little2_charMatches(const ENCODING *enc, const char *p, int c) |
| 603 | { |
| 604 | return LITTLE2_CHAR_MATCHES(enc, p, c); |
| 605 | } |
| 606 | |
| 607 | static |
| 608 | int little2_isNameMin(const ENCODING *enc, const char *p) |
| 609 | { |
| 610 | return LITTLE2_IS_NAME_CHAR_MINBPC(enc, p); |
| 611 | } |
| 612 | |
| 613 | static |
| 614 | int little2_isNmstrtMin(const ENCODING *enc, const char *p) |
| 615 | { |
| 616 | return LITTLE2_IS_NMSTRT_CHAR_MINBPC(enc, p); |
| 617 | } |
| 618 | |
| 619 | #undef VTABLE |
| 620 | #define VTABLE VTABLE1, little2_toUtf8, little2_toUtf16 |
| 621 | |
| 622 | #else /* not XML_MIN_SIZE */ |
| 623 | |
| 624 | #undef PREFIX |
| 625 | #define PREFIX(ident) little2_ ## ident |
| 626 | #define MINBPC(enc) 2 |
| 627 | /* CHAR_MATCHES is guaranteed to have MINBPC bytes available. */ |
| 628 | #define BYTE_TYPE(enc, p) LITTLE2_BYTE_TYPE(enc, p) |
| 629 | #define BYTE_TO_ASCII(enc, p) LITTLE2_BYTE_TO_ASCII(enc, p) |
| 630 | #define CHAR_MATCHES(enc, p, c) LITTLE2_CHAR_MATCHES(enc, p, c) |
| 631 | #define IS_NAME_CHAR(enc, p, n) 0 |
| 632 | #define IS_NAME_CHAR_MINBPC(enc, p) LITTLE2_IS_NAME_CHAR_MINBPC(enc, p) |
| 633 | #define IS_NMSTRT_CHAR(enc, p, n) (0) |
| 634 | #define IS_NMSTRT_CHAR_MINBPC(enc, p) LITTLE2_IS_NMSTRT_CHAR_MINBPC(enc, p) |
| 635 | |
| 636 | #include "xmltok_impl.c" |
| 637 | |
| 638 | #undef MINBPC |
| 639 | #undef BYTE_TYPE |
| 640 | #undef BYTE_TO_ASCII |
| 641 | #undef CHAR_MATCHES |
| 642 | #undef IS_NAME_CHAR |
| 643 | #undef IS_NAME_CHAR_MINBPC |
| 644 | #undef IS_NMSTRT_CHAR |
| 645 | #undef IS_NMSTRT_CHAR_MINBPC |
| 646 | #undef IS_INVALID_CHAR |
| 647 | |
| 648 | #endif /* not XML_MIN_SIZE */ |
| 649 | |
| 650 | #ifdef XML_NS |
| 651 | |
| 652 | static const struct normal_encoding little2_encoding_ns = { |
| 653 | { VTABLE, 2, 0, |
| 654 | #if XML_BYTE_ORDER == 12 |
| 655 | 1 |
| 656 | #else |
| 657 | 0 |
| 658 | #endif |
| 659 | }, |
| 660 | { |
| 661 | #include "asciitab.h" |
| 662 | #include "latin1tab.h" |
| 663 | }, |
| 664 | STANDARD_VTABLE(little2_) |
| 665 | }; |
| 666 | |
| 667 | #endif |
| 668 | |
| 669 | static const struct normal_encoding little2_encoding = { |
| 670 | { VTABLE, 2, 0, |
| 671 | #if XML_BYTE_ORDER == 12 |
| 672 | 1 |
| 673 | #else |
| 674 | 0 |
| 675 | #endif |
| 676 | }, |
| 677 | { |
| 678 | #define BT_COLON BT_NMSTRT |
| 679 | #include "asciitab.h" |
| 680 | #undef BT_COLON |
| 681 | #include "latin1tab.h" |
| 682 | }, |
| 683 | STANDARD_VTABLE(little2_) |
| 684 | }; |
| 685 | |
| 686 | #if XML_BYTE_ORDER != 21 |
| 687 | |
| 688 | #ifdef XML_NS |
| 689 | |
| 690 | static const struct normal_encoding internal_little2_encoding_ns = { |
| 691 | { VTABLE, 2, 0, 1 }, |
| 692 | { |
| 693 | #include "iasciitab.h" |
| 694 | #include "latin1tab.h" |
| 695 | }, |
| 696 | STANDARD_VTABLE(little2_) |
| 697 | }; |
| 698 | |
| 699 | #endif |
| 700 | |
| 701 | static const struct normal_encoding internal_little2_encoding = { |
| 702 | { VTABLE, 2, 0, 1 }, |
| 703 | { |
| 704 | #define BT_COLON BT_NMSTRT |
| 705 | #include "iasciitab.h" |
| 706 | #undef BT_COLON |
| 707 | #include "latin1tab.h" |
| 708 | }, |
| 709 | STANDARD_VTABLE(little2_) |
| 710 | }; |
| 711 | |
| 712 | #endif |
| 713 | |
| 714 | |
| 715 | #define BIG2_BYTE_TYPE(enc, p) \ |
| 716 | ((p)[0] == 0 \ |
| 717 | ? ((struct normal_encoding *)(enc))->type[(unsigned char)(p)[1]] \ |
| 718 | : unicode_byte_type((p)[0], (p)[1])) |
| 719 | #define BIG2_BYTE_TO_ASCII(enc, p) ((p)[0] == 0 ? (p)[1] : -1) |
| 720 | #define BIG2_CHAR_MATCHES(enc, p, c) ((p)[0] == 0 && (p)[1] == c) |
| 721 | #define BIG2_IS_NAME_CHAR_MINBPC(enc, p) \ |
| 722 | UCS2_GET_NAMING(namePages, (unsigned char)p[0], (unsigned char)p[1]) |
| 723 | #define BIG2_IS_NMSTRT_CHAR_MINBPC(enc, p) \ |
| 724 | UCS2_GET_NAMING(nmstrtPages, (unsigned char)p[0], (unsigned char)p[1]) |
| 725 | |
| 726 | #ifdef XML_MIN_SIZE |
| 727 | |
| 728 | static |
| 729 | int big2_byteType(const ENCODING *enc, const char *p) |
| 730 | { |
| 731 | return BIG2_BYTE_TYPE(enc, p); |
| 732 | } |
| 733 | |
| 734 | static |
| 735 | int big2_byteToAscii(const ENCODING *enc, const char *p) |
| 736 | { |
| 737 | return BIG2_BYTE_TO_ASCII(enc, p); |
| 738 | } |
| 739 | |
| 740 | static |
| 741 | int big2_charMatches(const ENCODING *enc, const char *p, int c) |
| 742 | { |
| 743 | return BIG2_CHAR_MATCHES(enc, p, c); |
| 744 | } |
| 745 | |
| 746 | static |
| 747 | int big2_isNameMin(const ENCODING *enc, const char *p) |
| 748 | { |
| 749 | return BIG2_IS_NAME_CHAR_MINBPC(enc, p); |
| 750 | } |
| 751 | |
| 752 | static |
| 753 | int big2_isNmstrtMin(const ENCODING *enc, const char *p) |
| 754 | { |
| 755 | return BIG2_IS_NMSTRT_CHAR_MINBPC(enc, p); |
| 756 | } |
| 757 | |
| 758 | #undef VTABLE |
| 759 | #define VTABLE VTABLE1, big2_toUtf8, big2_toUtf16 |
| 760 | |
| 761 | #else /* not XML_MIN_SIZE */ |
| 762 | |
| 763 | #undef PREFIX |
| 764 | #define PREFIX(ident) big2_ ## ident |
| 765 | #define MINBPC(enc) 2 |
| 766 | /* CHAR_MATCHES is guaranteed to have MINBPC bytes available. */ |
| 767 | #define BYTE_TYPE(enc, p) BIG2_BYTE_TYPE(enc, p) |
| 768 | #define BYTE_TO_ASCII(enc, p) BIG2_BYTE_TO_ASCII(enc, p) |
| 769 | #define CHAR_MATCHES(enc, p, c) BIG2_CHAR_MATCHES(enc, p, c) |
| 770 | #define IS_NAME_CHAR(enc, p, n) 0 |
| 771 | #define IS_NAME_CHAR_MINBPC(enc, p) BIG2_IS_NAME_CHAR_MINBPC(enc, p) |
| 772 | #define IS_NMSTRT_CHAR(enc, p, n) (0) |
| 773 | #define IS_NMSTRT_CHAR_MINBPC(enc, p) BIG2_IS_NMSTRT_CHAR_MINBPC(enc, p) |
| 774 | |
| 775 | #include "xmltok_impl.c" |
| 776 | |
| 777 | #undef MINBPC |
| 778 | #undef BYTE_TYPE |
| 779 | #undef BYTE_TO_ASCII |
| 780 | #undef CHAR_MATCHES |
| 781 | #undef IS_NAME_CHAR |
| 782 | #undef IS_NAME_CHAR_MINBPC |
| 783 | #undef IS_NMSTRT_CHAR |
| 784 | #undef IS_NMSTRT_CHAR_MINBPC |
| 785 | #undef IS_INVALID_CHAR |
| 786 | |
| 787 | #endif /* not XML_MIN_SIZE */ |
| 788 | |
| 789 | #ifdef XML_NS |
| 790 | |
| 791 | static const struct normal_encoding big2_encoding_ns = { |
| 792 | { VTABLE, 2, 0, |
| 793 | #if XML_BYTE_ORDER == 21 |
| 794 | 1 |
| 795 | #else |
| 796 | 0 |
| 797 | #endif |
| 798 | }, |
| 799 | { |
| 800 | #include "asciitab.h" |
| 801 | #include "latin1tab.h" |
| 802 | }, |
| 803 | STANDARD_VTABLE(big2_) |
| 804 | }; |
| 805 | |
| 806 | #endif |
| 807 | |
| 808 | static const struct normal_encoding big2_encoding = { |
| 809 | { VTABLE, 2, 0, |
| 810 | #if XML_BYTE_ORDER == 21 |
| 811 | 1 |
| 812 | #else |
| 813 | 0 |
| 814 | #endif |
| 815 | }, |
| 816 | { |
| 817 | #define BT_COLON BT_NMSTRT |
| 818 | #include "asciitab.h" |
| 819 | #undef BT_COLON |
| 820 | #include "latin1tab.h" |
| 821 | }, |
| 822 | STANDARD_VTABLE(big2_) |
| 823 | }; |
| 824 | |
| 825 | #if XML_BYTE_ORDER != 12 |
| 826 | |
| 827 | #ifdef XML_NS |
| 828 | |
| 829 | static const struct normal_encoding internal_big2_encoding_ns = { |
| 830 | { VTABLE, 2, 0, 1 }, |
| 831 | { |
| 832 | #include "iasciitab.h" |
| 833 | #include "latin1tab.h" |
| 834 | }, |
| 835 | STANDARD_VTABLE(big2_) |
| 836 | }; |
| 837 | |
| 838 | #endif |
| 839 | |
| 840 | static const struct normal_encoding internal_big2_encoding = { |
| 841 | { VTABLE, 2, 0, 1 }, |
| 842 | { |
| 843 | #define BT_COLON BT_NMSTRT |
| 844 | #include "iasciitab.h" |
| 845 | #undef BT_COLON |
| 846 | #include "latin1tab.h" |
| 847 | }, |
| 848 | STANDARD_VTABLE(big2_) |
| 849 | }; |
| 850 | |
| 851 | #endif |
| 852 | |
| 853 | #undef PREFIX |
| 854 | |
| 855 | static |
| 856 | int streqci(const char *s1, const char *s2) |
| 857 | { |
| 858 | for (;;) { |
| 859 | char c1 = *s1++; |
| 860 | char c2 = *s2++; |
| 861 | if (ASCII_a <= c1 && c1 <= ASCII_z) |
| 862 | c1 += ASCII_A - ASCII_a; |
| 863 | if (ASCII_a <= c2 && c2 <= ASCII_z) |
| 864 | c2 += ASCII_A - ASCII_a; |
| 865 | if (c1 != c2) |
| 866 | return 0; |
| 867 | if (!c1) |
| 868 | break; |
| 869 | } |
| 870 | return 1; |
| 871 | } |
| 872 | |
| 873 | static |
| 874 | void initUpdatePosition(const ENCODING *enc, const char *ptr, |
| 875 | const char *end, POSITION *pos) |
| 876 | { |
| 877 | normal_updatePosition(&utf8_encoding.enc, ptr, end, pos); |
| 878 | } |
| 879 | |
| 880 | static |
| 881 | int toAscii(const ENCODING *enc, const char *ptr, const char *end) |
| 882 | { |
| 883 | char buf[1]; |
| 884 | char *p = buf; |
| 885 | XmlUtf8Convert(enc, &ptr, end, &p, p + 1); |
| 886 | if (p == buf) |
| 887 | return -1; |
| 888 | else |
| 889 | return buf[0]; |
| 890 | } |
| 891 | |
| 892 | static |
| 893 | int isSpace(int c) |
| 894 | { |
| 895 | switch (c) { |
| 896 | case 0x20: |
| 897 | case 0xD: |
| 898 | case 0xA: |
| 899 | case 0x9: |
| 900 | return 1; |
| 901 | } |
| 902 | return 0; |
| 903 | } |
| 904 | |
| 905 | /* Return 1 if there's just optional white space |
| 906 | or there's an S followed by name=val. */ |
| 907 | static |
| 908 | int parsePseudoAttribute(const ENCODING *enc, |
| 909 | const char *ptr, |
| 910 | const char *end, |
| 911 | const char **namePtr, |
| 912 | const char **nameEndPtr, |
| 913 | const char **valPtr, |
| 914 | const char **nextTokPtr) |
| 915 | { |
| 916 | int c; |
| 917 | char open; |
| 918 | if (ptr == end) { |
| 919 | *namePtr = 0; |
| 920 | return 1; |
| 921 | } |
| 922 | if (!isSpace(toAscii(enc, ptr, end))) { |
| 923 | *nextTokPtr = ptr; |
| 924 | return 0; |
| 925 | } |
| 926 | do { |
| 927 | ptr += enc->minBytesPerChar; |
| 928 | } while (isSpace(toAscii(enc, ptr, end))); |
| 929 | if (ptr == end) { |
| 930 | *namePtr = 0; |
| 931 | return 1; |
| 932 | } |
| 933 | *namePtr = ptr; |
| 934 | for (;;) { |
| 935 | c = toAscii(enc, ptr, end); |
| 936 | if (c == -1) { |
| 937 | *nextTokPtr = ptr; |
| 938 | return 0; |
| 939 | } |
| 940 | if (c == ASCII_EQUALS) { |
| 941 | *nameEndPtr = ptr; |
| 942 | break; |
| 943 | } |
| 944 | if (isSpace(c)) { |
| 945 | *nameEndPtr = ptr; |
| 946 | do { |
| 947 | ptr += enc->minBytesPerChar; |
| 948 | } while (isSpace(c = toAscii(enc, ptr, end))); |
| 949 | if (c != ASCII_EQUALS) { |
| 950 | *nextTokPtr = ptr; |
| 951 | return 0; |
| 952 | } |
| 953 | break; |
| 954 | } |
| 955 | ptr += enc->minBytesPerChar; |
| 956 | } |
| 957 | if (ptr == *namePtr) { |
| 958 | *nextTokPtr = ptr; |
| 959 | return 0; |
| 960 | } |
| 961 | ptr += enc->minBytesPerChar; |
| 962 | c = toAscii(enc, ptr, end); |
| 963 | while (isSpace(c)) { |
| 964 | ptr += enc->minBytesPerChar; |
| 965 | c = toAscii(enc, ptr, end); |
| 966 | } |
| 967 | if (c != ASCII_QUOT && c != ASCII_APOS) { |
| 968 | *nextTokPtr = ptr; |
| 969 | return 0; |
| 970 | } |
| 971 | open = c; |
| 972 | ptr += enc->minBytesPerChar; |
| 973 | *valPtr = ptr; |
| 974 | for (;; ptr += enc->minBytesPerChar) { |
| 975 | c = toAscii(enc, ptr, end); |
| 976 | if (c == open) |
| 977 | break; |
| 978 | if (!(ASCII_a <= c && c <= ASCII_z) |
| 979 | && !(ASCII_A <= c && c <= ASCII_Z) |
| 980 | && !(ASCII_0 <= c && c <= ASCII_9) |
| 981 | && c != ASCII_PERIOD |
| 982 | && c != ASCII_MINUS |
| 983 | && c != ASCII_UNDERSCORE) { |
| 984 | *nextTokPtr = ptr; |
| 985 | return 0; |
| 986 | } |
| 987 | } |
| 988 | *nextTokPtr = ptr + enc->minBytesPerChar; |
| 989 | return 1; |
| 990 | } |
| 991 | |
| 992 | static const char KW_version[] = { |
| 993 | ASCII_v, ASCII_e, ASCII_r, ASCII_s, ASCII_i, ASCII_o, ASCII_n, '\0' |
| 994 | }; |
| 995 | |
| 996 | static const char KW_encoding[] = { |
| 997 | ASCII_e, ASCII_n, ASCII_c, ASCII_o, ASCII_d, ASCII_i, ASCII_n, ASCII_g, '\0' |
| 998 | }; |
| 999 | |
| 1000 | static const char KW_standalone[] = { |
| 1001 | ASCII_s, ASCII_t, ASCII_a, ASCII_n, ASCII_d, ASCII_a, ASCII_l, ASCII_o, ASCII_n, ASCII_e, '\0' |
| 1002 | }; |
| 1003 | |
| 1004 | static const char KW_yes[] = { |
| 1005 | ASCII_y, ASCII_e, ASCII_s, '\0' |
| 1006 | }; |
| 1007 | |
| 1008 | static const char KW_no[] = { |
| 1009 | ASCII_n, ASCII_o, '\0' |
| 1010 | }; |
| 1011 | |
| 1012 | static |
| 1013 | int doParseXmlDecl(const ENCODING *(*encodingFinder)(const ENCODING *, |
| 1014 | const char *, |
| 1015 | const char *), |
| 1016 | int isGeneralTextEntity, |
| 1017 | const ENCODING *enc, |
| 1018 | const char *ptr, |
| 1019 | const char *end, |
| 1020 | const char **badPtr, |
| 1021 | const char **versionPtr, |
| 1022 | const char **encodingName, |
| 1023 | const ENCODING **encoding, |
| 1024 | int *standalone) |
| 1025 | { |
| 1026 | const char *val = 0; |
| 1027 | const char *name = 0; |
| 1028 | const char *nameEnd = 0; |
| 1029 | ptr += 5 * enc->minBytesPerChar; |
| 1030 | end -= 2 * enc->minBytesPerChar; |
| 1031 | if (!parsePseudoAttribute(enc, ptr, end, &name, &nameEnd, &val, &ptr) || !name) { |
| 1032 | *badPtr = ptr; |
| 1033 | return 0; |
| 1034 | } |
| 1035 | if (!XmlNameMatchesAscii(enc, name, nameEnd, KW_version)) { |
| 1036 | if (!isGeneralTextEntity) { |
| 1037 | *badPtr = name; |
| 1038 | return 0; |
| 1039 | } |
| 1040 | } |
| 1041 | else { |
| 1042 | if (versionPtr) |
| 1043 | *versionPtr = val; |
| 1044 | if (!parsePseudoAttribute(enc, ptr, end, &name, &nameEnd, &val, &ptr)) { |
| 1045 | *badPtr = ptr; |
| 1046 | return 0; |
| 1047 | } |
| 1048 | if (!name) { |
| 1049 | if (isGeneralTextEntity) { |
| 1050 | /* a TextDecl must have an EncodingDecl */ |
| 1051 | *badPtr = ptr; |
| 1052 | return 0; |
| 1053 | } |
| 1054 | return 1; |
| 1055 | } |
| 1056 | } |
| 1057 | if (XmlNameMatchesAscii(enc, name, nameEnd, KW_encoding)) { |
| 1058 | int c = toAscii(enc, val, end); |
| 1059 | if (!(ASCII_a <= c && c <= ASCII_z) && !(ASCII_A <= c && c <= ASCII_Z)) { |
| 1060 | *badPtr = val; |
| 1061 | return 0; |
| 1062 | } |
| 1063 | if (encodingName) |
| 1064 | *encodingName = val; |
| 1065 | if (encoding) |
| 1066 | *encoding = encodingFinder(enc, val, ptr - enc->minBytesPerChar); |
| 1067 | if (!parsePseudoAttribute(enc, ptr, end, &name, &nameEnd, &val, &ptr)) { |
| 1068 | *badPtr = ptr; |
| 1069 | return 0; |
| 1070 | } |
| 1071 | if (!name) |
| 1072 | return 1; |
| 1073 | } |
| 1074 | if (!XmlNameMatchesAscii(enc, name, nameEnd, KW_standalone) || isGeneralTextEntity) { |
| 1075 | *badPtr = name; |
| 1076 | return 0; |
| 1077 | } |
| 1078 | if (XmlNameMatchesAscii(enc, val, ptr - enc->minBytesPerChar, KW_yes)) { |
| 1079 | if (standalone) |
| 1080 | *standalone = 1; |
| 1081 | } |
| 1082 | else if (XmlNameMatchesAscii(enc, val, ptr - enc->minBytesPerChar, KW_no)) { |
| 1083 | if (standalone) |
| 1084 | *standalone = 0; |
| 1085 | } |
| 1086 | else { |
| 1087 | *badPtr = val; |
| 1088 | return 0; |
| 1089 | } |
| 1090 | while (isSpace(toAscii(enc, ptr, end))) |
| 1091 | ptr += enc->minBytesPerChar; |
| 1092 | if (ptr != end) { |
| 1093 | *badPtr = ptr; |
| 1094 | return 0; |
| 1095 | } |
| 1096 | return 1; |
| 1097 | } |
| 1098 | |
| 1099 | static |
| 1100 | int checkCharRefNumber(int result) |
| 1101 | { |
| 1102 | switch (result >> 8) { |
| 1103 | case 0xD8: case 0xD9: case 0xDA: case 0xDB: |
| 1104 | case 0xDC: case 0xDD: case 0xDE: case 0xDF: |
| 1105 | return -1; |
| 1106 | case 0: |
| 1107 | if (latin1_encoding.type[result] == BT_NONXML) |
| 1108 | return -1; |
| 1109 | break; |
| 1110 | case 0xFF: |
| 1111 | if (result == 0xFFFE || result == 0xFFFF) |
| 1112 | return -1; |
| 1113 | break; |
| 1114 | } |
| 1115 | return result; |
| 1116 | } |
| 1117 | |
| 1118 | int XmlUtf8Encode(int c, char *buf) |
| 1119 | { |
| 1120 | enum { |
| 1121 | /* minN is minimum legal resulting value for N byte sequence */ |
| 1122 | min2 = 0x80, |
| 1123 | min3 = 0x800, |
| 1124 | min4 = 0x10000 |
| 1125 | }; |
| 1126 | |
| 1127 | if (c < 0) |
| 1128 | return 0; |
| 1129 | if (c < min2) { |
| 1130 | buf[0] = (c | UTF8_cval1); |
| 1131 | return 1; |
| 1132 | } |
| 1133 | if (c < min3) { |
| 1134 | buf[0] = ((c >> 6) | UTF8_cval2); |
| 1135 | buf[1] = ((c & 0x3f) | 0x80); |
| 1136 | return 2; |
| 1137 | } |
| 1138 | if (c < min4) { |
| 1139 | buf[0] = ((c >> 12) | UTF8_cval3); |
| 1140 | buf[1] = (((c >> 6) & 0x3f) | 0x80); |
| 1141 | buf[2] = ((c & 0x3f) | 0x80); |
| 1142 | return 3; |
| 1143 | } |
| 1144 | if (c < 0x110000) { |
| 1145 | buf[0] = ((c >> 18) | UTF8_cval4); |
| 1146 | buf[1] = (((c >> 12) & 0x3f) | 0x80); |
| 1147 | buf[2] = (((c >> 6) & 0x3f) | 0x80); |
| 1148 | buf[3] = ((c & 0x3f) | 0x80); |
| 1149 | return 4; |
| 1150 | } |
| 1151 | return 0; |
| 1152 | } |
| 1153 | |
| 1154 | int XmlUtf16Encode(int charNum, unsigned short *buf) |
| 1155 | { |
| 1156 | if (charNum < 0) |
| 1157 | return 0; |
| 1158 | if (charNum < 0x10000) { |
| 1159 | buf[0] = charNum; |
| 1160 | return 1; |
| 1161 | } |
| 1162 | if (charNum < 0x110000) { |
| 1163 | charNum -= 0x10000; |
| 1164 | buf[0] = (charNum >> 10) + 0xD800; |
| 1165 | buf[1] = (charNum & 0x3FF) + 0xDC00; |
| 1166 | return 2; |
| 1167 | } |
| 1168 | return 0; |
| 1169 | } |
| 1170 | |
| 1171 | struct unknown_encoding { |
| 1172 | struct normal_encoding normal; |
| 1173 | int (*convert)(void *userData, const char *p); |
| 1174 | void *userData; |
| 1175 | unsigned short utf16[256]; |
| 1176 | char utf8[256][4]; |
| 1177 | }; |
| 1178 | |
| 1179 | int XmlSizeOfUnknownEncoding(void) |
| 1180 | { |
| 1181 | return sizeof(struct unknown_encoding); |
| 1182 | } |
| 1183 | |
| 1184 | static |
| 1185 | int unknown_isName(const ENCODING *enc, const char *p) |
| 1186 | { |
| 1187 | int c = ((const struct unknown_encoding *)enc) |
| 1188 | ->convert(((const struct unknown_encoding *)enc)->userData, p); |
| 1189 | if (c & ~0xFFFF) |
| 1190 | return 0; |
| 1191 | return UCS2_GET_NAMING(namePages, c >> 8, c & 0xFF); |
| 1192 | } |
| 1193 | |
| 1194 | static |
| 1195 | int unknown_isNmstrt(const ENCODING *enc, const char *p) |
| 1196 | { |
| 1197 | int c = ((const struct unknown_encoding *)enc) |
| 1198 | ->convert(((const struct unknown_encoding *)enc)->userData, p); |
| 1199 | if (c & ~0xFFFF) |
| 1200 | return 0; |
| 1201 | return UCS2_GET_NAMING(nmstrtPages, c >> 8, c & 0xFF); |
| 1202 | } |
| 1203 | |
| 1204 | static |
| 1205 | int unknown_isInvalid(const ENCODING *enc, const char *p) |
| 1206 | { |
| 1207 | int c = ((const struct unknown_encoding *)enc) |
| 1208 | ->convert(((const struct unknown_encoding *)enc)->userData, p); |
| 1209 | return (c & ~0xFFFF) || checkCharRefNumber(c) < 0; |
| 1210 | } |
| 1211 | |
| 1212 | static |
| 1213 | void unknown_toUtf8(const ENCODING *enc, |
| 1214 | const char **fromP, const char *fromLim, |
| 1215 | char **toP, const char *toLim) |
| 1216 | { |
| 1217 | char buf[XML_UTF8_ENCODE_MAX]; |
| 1218 | for (;;) { |
| 1219 | const char *utf8; |
| 1220 | int n; |
| 1221 | if (*fromP == fromLim) |
| 1222 | break; |
| 1223 | utf8 = ((const struct unknown_encoding *)enc)->utf8[(unsigned char)**fromP]; |
| 1224 | n = *utf8++; |
| 1225 | if (n == 0) { |
| 1226 | int c = ((const struct unknown_encoding *)enc) |
| 1227 | ->convert(((const struct unknown_encoding *)enc)->userData, *fromP); |
| 1228 | n = XmlUtf8Encode(c, buf); |
| 1229 | if (n > toLim - *toP) |
| 1230 | break; |
| 1231 | utf8 = buf; |
| 1232 | *fromP += ((const struct normal_encoding *)enc)->type[(unsigned char)**fromP] |
| 1233 | - (BT_LEAD2 - 2); |
| 1234 | } |
| 1235 | else { |
| 1236 | if (n > toLim - *toP) |
| 1237 | break; |
| 1238 | (*fromP)++; |
| 1239 | } |
| 1240 | do { |
| 1241 | *(*toP)++ = *utf8++; |
| 1242 | } while (--n != 0); |
| 1243 | } |
| 1244 | } |
| 1245 | |
| 1246 | static |
| 1247 | void unknown_toUtf16(const ENCODING *enc, |
| 1248 | const char **fromP, const char *fromLim, |
| 1249 | unsigned short **toP, const unsigned short *toLim) |
| 1250 | { |
| 1251 | while (*fromP != fromLim && *toP != toLim) { |
| 1252 | unsigned short c |
| 1253 | = ((const struct unknown_encoding *)enc)->utf16[(unsigned char)**fromP]; |
| 1254 | if (c == 0) { |
| 1255 | c = (unsigned short)((const struct unknown_encoding *)enc) |
| 1256 | ->convert(((const struct unknown_encoding *)enc)->userData, *fromP); |
| 1257 | *fromP += ((const struct normal_encoding *)enc)->type[(unsigned char)**fromP] |
| 1258 | - (BT_LEAD2 - 2); |
| 1259 | } |
| 1260 | else |
| 1261 | (*fromP)++; |
| 1262 | *(*toP)++ = c; |
| 1263 | } |
| 1264 | } |
| 1265 | |
| 1266 | ENCODING * |
| 1267 | XmlInitUnknownEncoding(void *mem, |
| 1268 | int *table, |
| 1269 | int (*convert)(void *userData, const char *p), |
| 1270 | void *userData) |
| 1271 | { |
| 1272 | int i; |
| 1273 | struct unknown_encoding *e = mem; |
| 1274 | for (i = 0; i < (int)sizeof(struct normal_encoding); i++) |
| 1275 | ((char *)mem)[i] = ((char *)&latin1_encoding)[i]; |
| 1276 | for (i = 0; i < 128; i++) |
| 1277 | if (latin1_encoding.type[i] != BT_OTHER |
| 1278 | && latin1_encoding.type[i] != BT_NONXML |
| 1279 | && table[i] != i) |
| 1280 | return 0; |
| 1281 | for (i = 0; i < 256; i++) { |
| 1282 | int c = table[i]; |
| 1283 | if (c == -1) { |
| 1284 | e->normal.type[i] = BT_MALFORM; |
| 1285 | /* This shouldn't really get used. */ |
| 1286 | e->utf16[i] = 0xFFFF; |
| 1287 | e->utf8[i][0] = 1; |
| 1288 | e->utf8[i][1] = 0; |
| 1289 | } |
| 1290 | else if (c < 0) { |
| 1291 | if (c < -4) |
| 1292 | return 0; |
| 1293 | e->normal.type[i] = BT_LEAD2 - (c + 2); |
| 1294 | e->utf8[i][0] = 0; |
| 1295 | e->utf16[i] = 0; |
| 1296 | } |
| 1297 | else if (c < 0x80) { |
| 1298 | if (latin1_encoding.type[c] != BT_OTHER |
| 1299 | && latin1_encoding.type[c] != BT_NONXML |
| 1300 | && c != i) |
| 1301 | return 0; |
| 1302 | e->normal.type[i] = latin1_encoding.type[c]; |
| 1303 | e->utf8[i][0] = 1; |
| 1304 | e->utf8[i][1] = (char)c; |
| 1305 | e->utf16[i] = c == 0 ? 0xFFFF : c; |
| 1306 | } |
| 1307 | else if (checkCharRefNumber(c) < 0) { |
| 1308 | e->normal.type[i] = BT_NONXML; |
| 1309 | /* This shouldn't really get used. */ |
| 1310 | e->utf16[i] = 0xFFFF; |
| 1311 | e->utf8[i][0] = 1; |
| 1312 | e->utf8[i][1] = 0; |
| 1313 | } |
| 1314 | else { |
| 1315 | if (c > 0xFFFF) |
| 1316 | return 0; |
| 1317 | if (UCS2_GET_NAMING(nmstrtPages, c >> 8, c & 0xff)) |
| 1318 | e->normal.type[i] = BT_NMSTRT; |
| 1319 | else if (UCS2_GET_NAMING(namePages, c >> 8, c & 0xff)) |
| 1320 | e->normal.type[i] = BT_NAME; |
| 1321 | else |
| 1322 | e->normal.type[i] = BT_OTHER; |
| 1323 | e->utf8[i][0] = (char)XmlUtf8Encode(c, e->utf8[i] + 1); |
| 1324 | e->utf16[i] = c; |
| 1325 | } |
| 1326 | } |
| 1327 | e->userData = userData; |
| 1328 | e->convert = convert; |
| 1329 | if (convert) { |
| 1330 | e->normal.isName2 = unknown_isName; |
| 1331 | e->normal.isName3 = unknown_isName; |
| 1332 | e->normal.isName4 = unknown_isName; |
| 1333 | e->normal.isNmstrt2 = unknown_isNmstrt; |
| 1334 | e->normal.isNmstrt3 = unknown_isNmstrt; |
| 1335 | e->normal.isNmstrt4 = unknown_isNmstrt; |
| 1336 | e->normal.isInvalid2 = unknown_isInvalid; |
| 1337 | e->normal.isInvalid3 = unknown_isInvalid; |
| 1338 | e->normal.isInvalid4 = unknown_isInvalid; |
| 1339 | } |
| 1340 | e->normal.enc.utf8Convert = unknown_toUtf8; |
| 1341 | e->normal.enc.utf16Convert = unknown_toUtf16; |
| 1342 | return &(e->normal.enc); |
| 1343 | } |
| 1344 | |
| 1345 | /* If this enumeration is changed, getEncodingIndex and encodings |
| 1346 | must also be changed. */ |
| 1347 | enum { |
| 1348 | UNKNOWN_ENC = -1, |
| 1349 | ISO_8859_1_ENC = 0, |
| 1350 | US_ASCII_ENC, |
| 1351 | UTF_8_ENC, |
| 1352 | UTF_16_ENC, |
| 1353 | UTF_16BE_ENC, |
| 1354 | UTF_16LE_ENC, |
| 1355 | /* must match encodingNames up to here */ |
| 1356 | NO_ENC |
| 1357 | }; |
| 1358 | |
| 1359 | static const char KW_ISO_8859_1[] = { |
| 1360 | ASCII_I, ASCII_S, ASCII_O, ASCII_MINUS, ASCII_8, ASCII_8, ASCII_5, ASCII_9, ASCII_MINUS, ASCII_1, '\0' |
| 1361 | }; |
| 1362 | static const char KW_US_ASCII[] = { |
| 1363 | ASCII_U, ASCII_S, ASCII_MINUS, ASCII_A, ASCII_S, ASCII_C, ASCII_I, ASCII_I, '\0' |
| 1364 | }; |
| 1365 | static const char KW_UTF_8[] = { |
| 1366 | ASCII_U, ASCII_T, ASCII_F, ASCII_MINUS, ASCII_8, '\0' |
| 1367 | }; |
| 1368 | static const char KW_UTF_16[] = { |
| 1369 | ASCII_U, ASCII_T, ASCII_F, ASCII_MINUS, ASCII_1, ASCII_6, '\0' |
| 1370 | }; |
| 1371 | static const char KW_UTF_16BE[] = { |
| 1372 | ASCII_U, ASCII_T, ASCII_F, ASCII_MINUS, ASCII_1, ASCII_6, ASCII_B, ASCII_E, '\0' |
| 1373 | }; |
| 1374 | static const char KW_UTF_16LE[] = { |
| 1375 | ASCII_U, ASCII_T, ASCII_F, ASCII_MINUS, ASCII_1, ASCII_6, ASCII_L, ASCII_E, '\0' |
| 1376 | }; |
| 1377 | |
| 1378 | static |
| 1379 | int getEncodingIndex(const char *name) |
| 1380 | { |
| 1381 | static const char *encodingNames[] = { |
| 1382 | KW_ISO_8859_1, |
| 1383 | KW_US_ASCII, |
| 1384 | KW_UTF_8, |
| 1385 | KW_UTF_16, |
| 1386 | KW_UTF_16BE, |
| 1387 | KW_UTF_16LE, |
| 1388 | }; |
| 1389 | int i; |
| 1390 | if (name == 0) |
| 1391 | return NO_ENC; |
| 1392 | for (i = 0; i < (int)(sizeof(encodingNames)/sizeof(encodingNames[0])); i++) |
| 1393 | if (streqci(name, encodingNames[i])) |
| 1394 | return i; |
| 1395 | return UNKNOWN_ENC; |
| 1396 | } |
| 1397 | |
| 1398 | /* For binary compatibility, we store the index of the encoding specified |
| 1399 | at initialization in the isUtf16 member. */ |
| 1400 | |
| 1401 | #define INIT_ENC_INDEX(enc) ((int)(enc)->initEnc.isUtf16) |
| 1402 | #define SET_INIT_ENC_INDEX(enc, i) ((enc)->initEnc.isUtf16 = (char)i) |
| 1403 | |
| 1404 | /* This is what detects the encoding. |
| 1405 | encodingTable maps from encoding indices to encodings; |
| 1406 | INIT_ENC_INDEX(enc) is the index of the external (protocol) specified encoding; |
| 1407 | state is XML_CONTENT_STATE if we're parsing an external text entity, |
| 1408 | and XML_PROLOG_STATE otherwise. |
| 1409 | */ |
| 1410 | |
| 1411 | |
| 1412 | static |
| 1413 | int initScan(const ENCODING **encodingTable, |
| 1414 | const INIT_ENCODING *enc, |
| 1415 | int state, |
| 1416 | const char *ptr, |
| 1417 | const char *end, |
| 1418 | const char **nextTokPtr) |
| 1419 | { |
| 1420 | const ENCODING **encPtr; |
| 1421 | |
| 1422 | if (ptr == end) |
| 1423 | return XML_TOK_NONE; |
| 1424 | encPtr = enc->encPtr; |
| 1425 | if (ptr + 1 == end) { |
| 1426 | /* only a single byte available for auto-detection */ |
| 1427 | #ifndef XML_DTD /* FIXME */ |
| 1428 | /* a well-formed document entity must have more than one byte */ |
| 1429 | if (state != XML_CONTENT_STATE) |
| 1430 | return XML_TOK_PARTIAL; |
| 1431 | #endif |
| 1432 | /* so we're parsing an external text entity... */ |
| 1433 | /* if UTF-16 was externally specified, then we need at least 2 bytes */ |
| 1434 | switch (INIT_ENC_INDEX(enc)) { |
| 1435 | case UTF_16_ENC: |
| 1436 | case UTF_16LE_ENC: |
| 1437 | case UTF_16BE_ENC: |
| 1438 | return XML_TOK_PARTIAL; |
| 1439 | } |
| 1440 | switch ((unsigned char)*ptr) { |
| 1441 | case 0xFE: |
| 1442 | case 0xFF: |
| 1443 | case 0xEF: /* possibly first byte of UTF-8 BOM */ |
| 1444 | if (INIT_ENC_INDEX(enc) == ISO_8859_1_ENC |
| 1445 | && state == XML_CONTENT_STATE) |
| 1446 | break; |
| 1447 | /* fall through */ |
| 1448 | case 0x00: |
| 1449 | case 0x3C: |
| 1450 | return XML_TOK_PARTIAL; |
| 1451 | } |
| 1452 | } |
| 1453 | else { |
| 1454 | switch (((unsigned char)ptr[0] << 8) | (unsigned char)ptr[1]) { |
| 1455 | case 0xFEFF: |
| 1456 | if (INIT_ENC_INDEX(enc) == ISO_8859_1_ENC |
| 1457 | && state == XML_CONTENT_STATE) |
| 1458 | break; |
| 1459 | *nextTokPtr = ptr + 2; |
| 1460 | *encPtr = encodingTable[UTF_16BE_ENC]; |
| 1461 | return XML_TOK_BOM; |
| 1462 | /* 00 3C is handled in the default case */ |
| 1463 | case 0x3C00: |
| 1464 | if ((INIT_ENC_INDEX(enc) == UTF_16BE_ENC |
| 1465 | || INIT_ENC_INDEX(enc) == UTF_16_ENC) |
| 1466 | && state == XML_CONTENT_STATE) |
| 1467 | break; |
| 1468 | *encPtr = encodingTable[UTF_16LE_ENC]; |
| 1469 | return XmlTok(*encPtr, state, ptr, end, nextTokPtr); |
| 1470 | case 0xFFFE: |
| 1471 | if (INIT_ENC_INDEX(enc) == ISO_8859_1_ENC |
| 1472 | && state == XML_CONTENT_STATE) |
| 1473 | break; |
| 1474 | *nextTokPtr = ptr + 2; |
| 1475 | *encPtr = encodingTable[UTF_16LE_ENC]; |
| 1476 | return XML_TOK_BOM; |
| 1477 | case 0xEFBB: |
| 1478 | /* Maybe a UTF-8 BOM (EF BB BF) */ |
| 1479 | /* If there's an explicitly specified (external) encoding |
| 1480 | of ISO-8859-1 or some flavour of UTF-16 |
| 1481 | and this is an external text entity, |
| 1482 | don't look for the BOM, |
| 1483 | because it might be a legal data. */ |
| 1484 | if (state == XML_CONTENT_STATE) { |
| 1485 | int e = INIT_ENC_INDEX(enc); |
| 1486 | if (e == ISO_8859_1_ENC || e == UTF_16BE_ENC || e == UTF_16LE_ENC || e == UTF_16_ENC) |
| 1487 | break; |
| 1488 | } |
| 1489 | if (ptr + 2 == end) |
| 1490 | return XML_TOK_PARTIAL; |
| 1491 | if ((unsigned char)ptr[2] == 0xBF) { |
| 1492 | *encPtr = encodingTable[UTF_8_ENC]; |
| 1493 | return XML_TOK_BOM; |
| 1494 | } |
| 1495 | break; |
| 1496 | default: |
| 1497 | if (ptr[0] == '\0') { |
| 1498 | /* 0 isn't a legal data character. Furthermore a document entity can only |
| 1499 | start with ASCII characters. So the only way this can fail to be big-endian |
| 1500 | UTF-16 if it it's an external parsed general entity that's labelled as |
| 1501 | UTF-16LE. */ |
| 1502 | if (state == XML_CONTENT_STATE && INIT_ENC_INDEX(enc) == UTF_16LE_ENC) |
| 1503 | break; |
| 1504 | *encPtr = encodingTable[UTF_16BE_ENC]; |
| 1505 | return XmlTok(*encPtr, state, ptr, end, nextTokPtr); |
| 1506 | } |
| 1507 | else if (ptr[1] == '\0') { |
| 1508 | /* We could recover here in the case: |
| 1509 | - parsing an external entity |
| 1510 | - second byte is 0 |
| 1511 | - no externally specified encoding |
| 1512 | - no encoding declaration |
| 1513 | by assuming UTF-16LE. But we don't, because this would mean when |
| 1514 | presented just with a single byte, we couldn't reliably determine |
| 1515 | whether we needed further bytes. */ |
| 1516 | if (state == XML_CONTENT_STATE) |
| 1517 | break; |
| 1518 | *encPtr = encodingTable[UTF_16LE_ENC]; |
| 1519 | return XmlTok(*encPtr, state, ptr, end, nextTokPtr); |
| 1520 | } |
| 1521 | break; |
| 1522 | } |
| 1523 | } |
| 1524 | *encPtr = encodingTable[INIT_ENC_INDEX(enc)]; |
| 1525 | return XmlTok(*encPtr, state, ptr, end, nextTokPtr); |
| 1526 | } |
| 1527 | |
| 1528 | |
| 1529 | #define NS(x) x |
| 1530 | #define ns(x) x |
| 1531 | #include "xmltok_ns.c" |
| 1532 | #undef NS |
| 1533 | #undef ns |
| 1534 | |
| 1535 | #ifdef XML_NS |
| 1536 | |
| 1537 | #define NS(x) x ## NS |
| 1538 | #define ns(x) x ## _ns |
| 1539 | |
| 1540 | #include "xmltok_ns.c" |
| 1541 | |
| 1542 | #undef NS |
| 1543 | #undef ns |
| 1544 | |
| 1545 | ENCODING * |
| 1546 | XmlInitUnknownEncodingNS(void *mem, |
| 1547 | int *table, |
| 1548 | int (*convert)(void *userData, const char *p), |
| 1549 | void *userData) |
| 1550 | { |
| 1551 | ENCODING *enc = XmlInitUnknownEncoding(mem, table, convert, userData); |
| 1552 | if (enc) |
| 1553 | ((struct normal_encoding *)enc)->type[ASCII_COLON] = BT_COLON; |
| 1554 | return enc; |
| 1555 | } |
| 1556 | |
| 1557 | #endif /* XML_NS */ |