/src/ftk_xml_parser.c
C | 563 lines | 462 code | 71 blank | 30 comment | 108 complexity | 1d8d14ec6199398a417758016a78ff18 MD5 | raw file
1/* 2 * File: ftk_xml_parser.c 3 * Author: Li XianJing <xianjimli@hotmail.com> 4 * Brief: xml parser 5 * 6 * Copyright (c) Li XianJing 7 * 8 * Licensed under the Academic Free License version 2.1 9 * 10 * This program is free software; you can redistribute it and/or modify 11 * it under the terms of the GNU General Public License as published by 12 * the Free Software Foundation; either version 2 of the License, or 13 * (at your option) any later version. 14 * 15 * This program is distributed in the hope that it will be useful, 16 * but WITHOUT ANY WARRANTY; without even the implied warranty of 17 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the 18 * GNU General Public License for more details. 19 * 20 * You should have received a copy of the GNU General Public License 21 * along with this program; if not, write to the Free Software 22 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA 23 */ 24 25/* 26 * History: 27 * ================================================================ 28 * 2009-05-30 Li XianJing <xianjimli@hotmail.com> created. 29 * 30 */ 31 32#include "ftk_allocator.h" 33#include "ftk_xml_parser.h" 34 35struct _FtkXmlParser 36{ 37 const char* read_ptr; 38 39 int attrs_nr; 40 char* attrs[2*MAX_ATTR_NR+1]; 41 42 char* buffer; 43 int buffer_used; 44 int buffer_total; 45 46 FtkXmlBuilder* builder; 47}; 48 49static const char* strtrim(char* str); 50static void ftk_xml_parser_parse_entity(FtkXmlParser* thiz); 51static void ftk_xml_parser_parse_start_tag(FtkXmlParser* thiz); 52static void ftk_xml_parser_parse_end_tag(FtkXmlParser* thiz); 53static void ftk_xml_parser_parse_comment(FtkXmlParser* thiz); 54static void ftk_xml_parser_parse_pi(FtkXmlParser* thiz); 55static void ftk_xml_parser_parse_text(FtkXmlParser* thiz); 56static void ftk_xml_parser_reset_buffer(FtkXmlParser* thiz); 57 58FtkXmlParser* ftk_xml_parser_create(void) 59{ 60 return (FtkXmlParser*)FTK_ZALLOC(sizeof(FtkXmlParser)); 61} 62 63void ftk_xml_parser_set_builder(FtkXmlParser* thiz, FtkXmlBuilder* builder) 64{ 65 thiz->builder = builder; 66 67 return; 68} 69 70void ftk_xml_parser_parse(FtkXmlParser* thiz, const char* xml, int length) 71{ 72 int i = 0; 73 enum _State 74 { 75 STAT_NONE, 76 STAT_AFTER_LT, 77 STAT_START_TAG, 78 STAT_END_TAG, 79 STAT_TEXT, 80 STAT_PRE_COMMENT1, 81 STAT_PRE_COMMENT2, 82 STAT_COMMENT, 83 STAT_PROCESS_INSTRUCTION, 84 }state = STAT_NONE; 85 86 thiz->read_ptr = xml; 87 88 for(; *thiz->read_ptr != '\0' && (thiz->read_ptr - xml) < length; thiz->read_ptr++, i++) 89 { 90 char c = thiz->read_ptr[0]; 91 92 switch(state) 93 { 94 case STAT_NONE: 95 { 96 97 if(c == '<') 98 { 99 ftk_xml_parser_reset_buffer(thiz); 100 state = STAT_AFTER_LT; 101 } 102 else if(!isspace(c)) 103 { 104 state = STAT_TEXT; 105 } 106 break; 107 } 108 case STAT_AFTER_LT: 109 { 110 if(c == '?') 111 { 112 state = STAT_PROCESS_INSTRUCTION; 113 } 114 else if(c == '/') 115 { 116 state = STAT_END_TAG; 117 } 118 else if(c == '!') 119 { 120 state = STAT_PRE_COMMENT1; 121 } 122 else if(isalpha(c) || c == '_') 123 { 124 state = STAT_START_TAG; 125 } 126 else 127 { 128 ftk_xml_builder_on_error(thiz->builder, 0, 0, "unexpected char"); 129 } 130 break; 131 } 132 case STAT_START_TAG: 133 { 134 ftk_xml_parser_parse_start_tag(thiz); 135 state = STAT_NONE; 136 break; 137 } 138 case STAT_END_TAG: 139 { 140 ftk_xml_parser_parse_end_tag(thiz); 141 state = STAT_NONE; 142 break; 143 } 144 case STAT_PROCESS_INSTRUCTION: 145 { 146 ftk_xml_parser_parse_pi(thiz); 147 state = STAT_NONE; 148 break; 149 } 150 case STAT_TEXT: 151 { 152 ftk_xml_parser_parse_text(thiz); 153 state = STAT_NONE; 154 break; 155 } 156 case STAT_PRE_COMMENT1: 157 { 158 if(c == '-') 159 { 160 state = STAT_PRE_COMMENT2; 161 } 162 else 163 { 164 ftk_xml_builder_on_error(thiz->builder, 0, 0, "expected \'-\'"); 165 } 166 break; 167 } 168 case STAT_PRE_COMMENT2: 169 { 170 if(c == '-') 171 { 172 state = STAT_COMMENT; 173 } 174 else 175 { 176 ftk_xml_builder_on_error(thiz->builder, 0, 0, "expected \'-\'"); 177 } 178 } 179 case STAT_COMMENT: 180 { 181 ftk_xml_parser_parse_comment(thiz); 182 state = STAT_NONE; 183 break; 184 } 185 default:break; 186 } 187 188 if(*thiz->read_ptr == '\0') 189 { 190 break; 191 } 192 } 193 194 return; 195} 196 197static void ftk_xml_parser_reset_buffer(FtkXmlParser* thiz) 198{ 199 thiz->buffer_used = 0; 200 thiz->attrs_nr = 0; 201 thiz->attrs[0] = NULL; 202 203 return; 204} 205 206static int ftk_xml_parser_strdup(FtkXmlParser* thiz, const char* start, int length) 207{ 208 int offset = -1; 209 210 if((thiz->buffer_used + length) >= thiz->buffer_total) 211 { 212 int length = thiz->buffer_total+(thiz->buffer_total>>1) + 128; 213 char* buffer = (char*)FTK_REALLOC(thiz->buffer, length); 214 if(buffer != NULL) 215 { 216 thiz->buffer = buffer; 217 thiz->buffer_total = length; 218 } 219 } 220 221 if((thiz->buffer_used + length) >= thiz->buffer_total) 222 { 223 return offset; 224 } 225 226 offset = thiz->buffer_used; 227 ftk_strncpy(thiz->buffer + offset, start, length); 228 thiz->buffer[offset + length] = '\0'; 229 strtrim(thiz->buffer+offset); 230 thiz->buffer_used += length + 1; 231 232 return offset; 233} 234 235static void ftk_xml_parser_parse_attrs(FtkXmlParser* thiz, char end_char) 236{ 237 int i = 0; 238 enum _State 239 { 240 STAT_PRE_KEY, 241 STAT_KEY, 242 STAT_PRE_VALUE, 243 STAT_VALUE, 244 STAT_END, 245 }state = STAT_PRE_KEY; 246 247 char value_end = '\"'; 248 const char* start = thiz->read_ptr; 249 250 thiz->attrs_nr = 0; 251 for(; *thiz->read_ptr != '\0' && thiz->attrs_nr < MAX_ATTR_NR; thiz->read_ptr++) 252 { 253 char c = *thiz->read_ptr; 254 255 switch(state) 256 { 257 case STAT_PRE_KEY: 258 { 259 if(c == end_char || c == '>') 260 { 261 state = STAT_END; 262 } 263 else if(!isspace(c)) 264 { 265 state = STAT_KEY; 266 start = thiz->read_ptr; 267 } 268 } 269 case STAT_KEY: 270 { 271 if(c == '=') 272 { 273 thiz->attrs[thiz->attrs_nr++] = (char*)ftk_xml_parser_strdup(thiz, start, thiz->read_ptr - start); 274 state = STAT_PRE_VALUE; 275 } 276 277 break; 278 } 279 case STAT_PRE_VALUE: 280 { 281 if(c == '\"' || c == '\'') 282 { 283 state = STAT_VALUE; 284 value_end = c; 285 start = thiz->read_ptr + 1; 286 } 287 break; 288 } 289 case STAT_VALUE: 290 { 291 if(c == value_end) 292 { 293 thiz->attrs[thiz->attrs_nr++] = (char*)ftk_xml_parser_strdup(thiz, start, thiz->read_ptr - start); 294 state = STAT_PRE_KEY; 295 } 296 } 297 default:break; 298 } 299 300 if(state == STAT_END) 301 { 302 break; 303 } 304 } 305 306 for(i = 0; i < thiz->attrs_nr; i++) 307 { 308 thiz->attrs[i] = thiz->buffer + (int)(thiz->attrs[i]); 309 } 310 thiz->attrs[thiz->attrs_nr] = NULL; 311 312 return; 313} 314 315static void ftk_xml_parser_parse_start_tag(FtkXmlParser* thiz) 316{ 317 enum _State 318 { 319 STAT_NAME, 320 STAT_ATTR, 321 STAT_END, 322 }state = STAT_NAME; 323 324 char* tag_name = NULL; 325 const char* start = thiz->read_ptr - 1; 326 327 for(; *thiz->read_ptr != '\0'; thiz->read_ptr++) 328 { 329 char c = *thiz->read_ptr; 330 331 switch(state) 332 { 333 case STAT_NAME: 334 { 335 if(isspace(c) || c == '>' || c == '/') 336 { 337 tag_name = (char*)ftk_xml_parser_strdup(thiz, start, thiz->read_ptr - start); 338 state = (c != '>' && c != '/') ? STAT_ATTR : STAT_END; 339 } 340 break; 341 } 342 case STAT_ATTR: 343 { 344 ftk_xml_parser_parse_attrs(thiz, '/'); 345 state = STAT_END; 346 347 break; 348 } 349 default:break; 350 } 351 352 if(state == STAT_END) 353 { 354 break; 355 } 356 } 357 358 tag_name = thiz->buffer + (int)tag_name; 359 ftk_xml_builder_on_start_element(thiz->builder, tag_name, (const char**)thiz->attrs); 360 361 if(thiz->read_ptr[0] == '/') 362 { 363 ftk_xml_builder_on_end_element(thiz->builder, tag_name); 364 } 365 366 for(; *thiz->read_ptr != '>' && *thiz->read_ptr != '\0'; thiz->read_ptr++); 367 368 return; 369} 370 371static void ftk_xml_parser_parse_end_tag(FtkXmlParser* thiz) 372{ 373 char* tag_name = NULL; 374 const char* start = thiz->read_ptr; 375 for(; *thiz->read_ptr != '\0'; thiz->read_ptr++) 376 { 377 if(*thiz->read_ptr == '>') 378 { 379 tag_name = thiz->buffer + ftk_xml_parser_strdup(thiz, start, thiz->read_ptr-start); 380 ftk_xml_builder_on_end_element(thiz->builder, tag_name); 381 382 break; 383 } 384 } 385 386 return; 387} 388 389static void ftk_xml_parser_parse_comment(FtkXmlParser* thiz) 390{ 391 enum _State 392 { 393 STAT_COMMENT, 394 STAT_MINUS1, 395 STAT_MINUS2, 396 }state = STAT_COMMENT; 397 398 const char* start = ++thiz->read_ptr; 399 for(; *thiz->read_ptr != '\0'; thiz->read_ptr++) 400 { 401 char c = *thiz->read_ptr; 402 403 switch(state) 404 { 405 case STAT_COMMENT: 406 { 407 if(c == '-') 408 { 409 state = STAT_MINUS1; 410 } 411 break; 412 } 413 case STAT_MINUS1: 414 { 415 if(c == '-') 416 { 417 state = STAT_MINUS2; 418 } 419 else 420 { 421 state = STAT_COMMENT; 422 } 423 break; 424 } 425 case STAT_MINUS2: 426 { 427 if(c == '>') 428 { 429 ftk_xml_builder_on_comment(thiz->builder, start, thiz->read_ptr-start-2); 430 return; 431 } 432 } 433 default:break; 434 } 435 } 436 437 return; 438} 439 440static void ftk_xml_parser_parse_pi(FtkXmlParser* thiz) 441{ 442 enum _State 443 { 444 STAT_NAME, 445 STAT_ATTR, 446 STAT_END 447 }state = STAT_NAME; 448 449 char* tag_name = NULL; 450 const char* start = thiz->read_ptr; 451 452 for(; *thiz->read_ptr != '\0'; thiz->read_ptr++) 453 { 454 char c = *thiz->read_ptr; 455 456 switch(state) 457 { 458 case STAT_NAME: 459 { 460 if(isspace(c) || c == '>') 461 { 462 tag_name = (char*)ftk_xml_parser_strdup(thiz, start, thiz->read_ptr - start); 463 state = c != '>' ? STAT_ATTR : STAT_END; 464 } 465 466 break; 467 } 468 case STAT_ATTR: 469 { 470 ftk_xml_parser_parse_attrs(thiz, '?'); 471 state = STAT_END; 472 break; 473 } 474 default:break; 475 } 476 477 if(state == STAT_END) 478 { 479 break; 480 } 481 } 482 483 tag_name = thiz->buffer + (int)tag_name; 484 ftk_xml_builder_on_pi_element(thiz->builder, tag_name, (const char**)thiz->attrs); 485 486 for(; *thiz->read_ptr != '>' && *thiz->read_ptr != '\0'; thiz->read_ptr++); 487 488 return; 489} 490 491static void ftk_xml_parser_parse_text(FtkXmlParser* thiz) 492{ 493 const char* start = thiz->read_ptr - 1; 494 for(; *thiz->read_ptr != '\0'; thiz->read_ptr++) 495 { 496 char c = *thiz->read_ptr; 497 498 if(c == '<') 499 { 500 if(thiz->read_ptr > start) 501 { 502 ftk_xml_builder_on_text(thiz->builder, start, thiz->read_ptr-start); 503 } 504 thiz->read_ptr--; 505 return; 506 } 507 else if(c == '&') 508 { 509 ftk_xml_parser_parse_entity(thiz); 510 } 511 } 512 513 return; 514} 515 516static void ftk_xml_parser_parse_entity(FtkXmlParser* thiz) 517{ 518 /*TODO*/ 519 520 return; 521} 522 523void ftk_xml_parser_destroy(FtkXmlParser* thiz) 524{ 525 if(thiz != NULL) 526 { 527 FTK_FREE(thiz->buffer); 528 FTK_FREE(thiz); 529 } 530 531 return; 532} 533 534static const char* strtrim(char* str) 535{ 536 char* p = NULL; 537 538 p = str + strlen(str) - 1; 539 540 while(p != str && isspace(*p)) 541 { 542 *p = '\0'; 543 p--; 544 } 545 546 p = str; 547 while(*p != '\0' && isspace(*p)) p++; 548 549 if(p != str) 550 { 551 char* s = p; 552 char* d = str; 553 while(*s != '\0') 554 { 555 *d = *s; 556 d++; 557 s++; 558 } 559 *d = '\0'; 560 } 561 562 return str; 563}