xmlfile.c 8.0 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279
  1. /*
  2. __ __ _
  3. ___\ \/ /_ __ __ _| |_
  4. / _ \\ /| '_ \ / _` | __|
  5. | __// \| |_) | (_| | |_
  6. \___/_/\_\ .__/ \__,_|\__|
  7. |_| XML parser
  8. Copyright (c) 1997-2000 Thai Open Source Software Center Ltd
  9. Copyright (c) 2000 Clark Cooper <[email protected]>
  10. Copyright (c) 2002-2003 Fred L. Drake, Jr. <[email protected]>
  11. Copyright (c) 2004-2006 Karl Waclawek <[email protected]>
  12. Copyright (c) 2005-2007 Steven Solie <[email protected]>
  13. Copyright (c) 2016-2021 Sebastian Pipping <[email protected]>
  14. Copyright (c) 2017 Rhodri James <[email protected]>
  15. Copyright (c) 2019 David Loffredo <[email protected]>
  16. Licensed under the MIT license:
  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. The above copyright notice and this permission notice shall be included
  25. in all copies or substantial portions of the Software.
  26. THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
  27. EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
  28. MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN
  29. NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM,
  30. DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
  31. OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
  32. USE OR OTHER DEALINGS IN THE SOFTWARE.
  33. */
  34. #include <stdio.h>
  35. #include <stdlib.h>
  36. #include <stddef.h>
  37. #include <string.h>
  38. #include <fcntl.h>
  39. #ifdef _WIN32
  40. # include "winconfig.h"
  41. #endif
  42. #include <expat_config.h>
  43. #include "expat.h"
  44. #include "internal.h" /* for UNUSED_P only */
  45. #include "xmlfile.h"
  46. #include "xmltchar.h"
  47. #include "filemap.h"
  48. #if defined(_MSC_VER)
  49. # include <io.h>
  50. #endif
  51. #ifdef HAVE_UNISTD_H
  52. # include <unistd.h>
  53. #endif
  54. #ifndef O_BINARY
  55. # ifdef _O_BINARY
  56. # define O_BINARY _O_BINARY
  57. # else
  58. # define O_BINARY 0
  59. # endif
  60. #endif
  61. #ifdef _DEBUG
  62. # define READ_SIZE 16
  63. #else
  64. # define READ_SIZE (1024 * 8)
  65. #endif
  66. typedef struct {
  67. XML_Parser parser;
  68. int *retPtr;
  69. } PROCESS_ARGS;
  70. static int processStream(const XML_Char *filename, XML_Parser parser);
  71. static void
  72. reportError(XML_Parser parser, const XML_Char *filename) {
  73. enum XML_Error code = XML_GetErrorCode(parser);
  74. const XML_Char *message = XML_ErrorString(code);
  75. if (message)
  76. ftprintf(stdout,
  77. T("%s") T(":%") T(XML_FMT_INT_MOD) T("u") T(":%")
  78. T(XML_FMT_INT_MOD) T("u") T(": %s\n"),
  79. filename, XML_GetErrorLineNumber(parser),
  80. XML_GetErrorColumnNumber(parser), message);
  81. else
  82. ftprintf(stderr, T("%s: (unknown message %d)\n"), filename, code);
  83. }
  84. /* This implementation will give problems on files larger than INT_MAX. */
  85. static void
  86. processFile(const void *data, size_t size, const XML_Char *filename,
  87. void *args) {
  88. XML_Parser parser = ((PROCESS_ARGS *)args)->parser;
  89. int *retPtr = ((PROCESS_ARGS *)args)->retPtr;
  90. if (XML_Parse(parser, (const char *)data, (int)size, 1) == XML_STATUS_ERROR) {
  91. reportError(parser, filename);
  92. *retPtr = 0;
  93. } else
  94. *retPtr = 1;
  95. }
  96. #if defined(_WIN32)
  97. static int
  98. isAsciiLetter(XML_Char c) {
  99. return (T('a') <= c && c <= T('z')) || (T('A') <= c && c <= T('Z'));
  100. }
  101. #endif /* _WIN32 */
  102. static const XML_Char *
  103. resolveSystemId(const XML_Char *base, const XML_Char *systemId,
  104. XML_Char **toFree) {
  105. XML_Char *s;
  106. *toFree = 0;
  107. if (! base || *systemId == T('/')
  108. #if defined(_WIN32)
  109. || *systemId == T('\\')
  110. || (isAsciiLetter(systemId[0]) && systemId[1] == T(':'))
  111. #endif
  112. )
  113. return systemId;
  114. *toFree = (XML_Char *)malloc((tcslen(base) + tcslen(systemId) + 2)
  115. * sizeof(XML_Char));
  116. if (! *toFree)
  117. return systemId;
  118. tcscpy(*toFree, base);
  119. s = *toFree;
  120. if (tcsrchr(s, T('/')))
  121. s = tcsrchr(s, T('/')) + 1;
  122. #if defined(_WIN32)
  123. if (tcsrchr(s, T('\\')))
  124. s = tcsrchr(s, T('\\')) + 1;
  125. #endif
  126. tcscpy(s, systemId);
  127. return *toFree;
  128. }
  129. static int
  130. externalEntityRefFilemap(XML_Parser parser, const XML_Char *context,
  131. const XML_Char *base, const XML_Char *systemId,
  132. const XML_Char *publicId) {
  133. int result;
  134. XML_Char *s;
  135. const XML_Char *filename;
  136. XML_Parser entParser = XML_ExternalEntityParserCreate(parser, context, 0);
  137. int filemapRes;
  138. PROCESS_ARGS args;
  139. UNUSED_P(publicId);
  140. args.retPtr = &result;
  141. args.parser = entParser;
  142. filename = resolveSystemId(base, systemId, &s);
  143. XML_SetBase(entParser, filename);
  144. filemapRes = filemap(filename, processFile, &args);
  145. switch (filemapRes) {
  146. case 0:
  147. result = 0;
  148. break;
  149. case 2:
  150. ftprintf(stderr,
  151. T("%s: file too large for memory-mapping")
  152. T(", switching to streaming\n"),
  153. filename);
  154. result = processStream(filename, entParser);
  155. break;
  156. }
  157. free(s);
  158. XML_ParserFree(entParser);
  159. return result;
  160. }
  161. static int
  162. processStream(const XML_Char *filename, XML_Parser parser) {
  163. /* passing NULL for filename means read input from stdin */
  164. int fd = 0; /* 0 is the fileno for stdin */
  165. if (filename != NULL) {
  166. fd = topen(filename, O_BINARY | O_RDONLY);
  167. if (fd < 0) {
  168. tperror(filename);
  169. return 0;
  170. }
  171. }
  172. for (;;) {
  173. int nread;
  174. char *buf = (char *)XML_GetBuffer(parser, READ_SIZE);
  175. if (! buf) {
  176. if (filename != NULL)
  177. close(fd);
  178. ftprintf(stderr, T("%s: out of memory\n"),
  179. filename != NULL ? filename : T("xmlwf"));
  180. return 0;
  181. }
  182. nread = read(fd, buf, READ_SIZE);
  183. if (nread < 0) {
  184. tperror(filename != NULL ? filename : T("STDIN"));
  185. if (filename != NULL)
  186. close(fd);
  187. return 0;
  188. }
  189. if (XML_ParseBuffer(parser, nread, nread == 0) == XML_STATUS_ERROR) {
  190. reportError(parser, filename != NULL ? filename : T("STDIN"));
  191. if (filename != NULL)
  192. close(fd);
  193. return 0;
  194. }
  195. if (nread == 0) {
  196. if (filename != NULL)
  197. close(fd);
  198. break;
  199. ;
  200. }
  201. }
  202. return 1;
  203. }
  204. static int
  205. externalEntityRefStream(XML_Parser parser, const XML_Char *context,
  206. const XML_Char *base, const XML_Char *systemId,
  207. const XML_Char *publicId) {
  208. XML_Char *s;
  209. const XML_Char *filename;
  210. int ret;
  211. XML_Parser entParser = XML_ExternalEntityParserCreate(parser, context, 0);
  212. UNUSED_P(publicId);
  213. filename = resolveSystemId(base, systemId, &s);
  214. XML_SetBase(entParser, filename);
  215. ret = processStream(filename, entParser);
  216. free(s);
  217. XML_ParserFree(entParser);
  218. return ret;
  219. }
  220. int
  221. XML_ProcessFile(XML_Parser parser, const XML_Char *filename, unsigned flags) {
  222. int result;
  223. if (! XML_SetBase(parser, filename)) {
  224. ftprintf(stderr, T("%s: out of memory"), filename);
  225. exit(1);
  226. }
  227. if (flags & XML_EXTERNAL_ENTITIES)
  228. XML_SetExternalEntityRefHandler(parser, (flags & XML_MAP_FILE)
  229. ? externalEntityRefFilemap
  230. : externalEntityRefStream);
  231. if (flags & XML_MAP_FILE) {
  232. int filemapRes;
  233. PROCESS_ARGS args;
  234. args.retPtr = &result;
  235. args.parser = parser;
  236. filemapRes = filemap(filename, processFile, &args);
  237. switch (filemapRes) {
  238. case 0:
  239. result = 0;
  240. break;
  241. case 2:
  242. ftprintf(stderr,
  243. T("%s: file too large for memory-mapping")
  244. T(", switching to streaming\n"),
  245. filename);
  246. result = processStream(filename, parser);
  247. break;
  248. }
  249. } else
  250. result = processStream(filename, parser);
  251. return result;
  252. }