xmlfile.c 8.5 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285
  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) 2016-2025 Sebastian Pipping <[email protected]>
  14. Copyright (c) 2017 Rhodri James <[email protected]>
  15. Copyright (c) 2019 David Loffredo <[email protected]>
  16. Copyright (c) 2021 Donghee Na <[email protected]>
  17. Copyright (c) 2024 Hanno Böck <[email protected]>
  18. Licensed under the MIT license:
  19. Permission is hereby granted, free of charge, to any person obtaining
  20. a copy of this software and associated documentation files (the
  21. "Software"), to deal in the Software without restriction, including
  22. without limitation the rights to use, copy, modify, merge, publish,
  23. distribute, sublicense, and/or sell copies of the Software, and to permit
  24. persons to whom the Software is furnished to do so, subject to the
  25. following conditions:
  26. The above copyright notice and this permission notice shall be included
  27. in all copies or substantial portions of the Software.
  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. #include "expat_config.h"
  37. #include <stdio.h>
  38. #include <stdlib.h>
  39. #include <stddef.h>
  40. #include <string.h>
  41. #include <fcntl.h>
  42. #ifdef _WIN32
  43. # include "winconfig.h"
  44. #endif
  45. #include "expat.h"
  46. #include "internal.h" /* for UNUSED_P only */
  47. #include "xmlfile.h"
  48. #include "xmltchar.h"
  49. #include "filemap.h"
  50. /* Function "read": */
  51. #if defined(_MSC_VER)
  52. # include <io.h>
  53. /* https://msdn.microsoft.com/en-us/library/wyssk1bs(v=vs.100).aspx */
  54. # define EXPAT_read _read
  55. # define EXPAT_read_count_t int
  56. # define EXPAT_read_req_t unsigned int
  57. #else /* POSIX */
  58. # include <unistd.h>
  59. /* https://pubs.opengroup.org/onlinepubs/009695399/functions/read.html */
  60. # define EXPAT_read read
  61. # define EXPAT_read_count_t ssize_t
  62. # define EXPAT_read_req_t size_t
  63. #endif
  64. #ifndef O_BINARY
  65. # ifdef _O_BINARY
  66. # define O_BINARY _O_BINARY
  67. # else
  68. # define O_BINARY 0
  69. # endif
  70. #endif
  71. int g_read_size_bytes = 1024 * 8;
  72. typedef struct {
  73. XML_Parser parser;
  74. int *retPtr;
  75. } PROCESS_ARGS;
  76. static int processStream(const XML_Char *filename, XML_Parser parser);
  77. static void
  78. reportError(XML_Parser parser, const XML_Char *filename) {
  79. enum XML_Error code = XML_GetErrorCode(parser);
  80. const XML_Char *message = XML_ErrorString(code);
  81. if (message)
  82. ftprintf(stdout,
  83. T("%s") T(":%") T(XML_FMT_INT_MOD) T("u") T(":%")
  84. T(XML_FMT_INT_MOD) T("u") T(": %s\n"),
  85. filename, XML_GetCurrentLineNumber(parser),
  86. XML_GetCurrentColumnNumber(parser), message);
  87. else
  88. ftprintf(stderr, T("%s: (unknown message %u)\n"), filename,
  89. (unsigned int)code);
  90. }
  91. /* This implementation will give problems on files larger than INT_MAX. */
  92. static void
  93. processFile(const void *data, size_t size, const XML_Char *filename,
  94. void *args) {
  95. XML_Parser parser = ((PROCESS_ARGS *)args)->parser;
  96. int *retPtr = ((PROCESS_ARGS *)args)->retPtr;
  97. if (XML_Parse(parser, (const char *)data, (int)size, 1) == XML_STATUS_ERROR) {
  98. reportError(parser, filename);
  99. *retPtr = 0;
  100. } else
  101. *retPtr = 1;
  102. }
  103. #if defined(_WIN32)
  104. static int
  105. isAsciiLetter(XML_Char c) {
  106. return (T('a') <= c && c <= T('z')) || (T('A') <= c && c <= T('Z'));
  107. }
  108. #endif /* _WIN32 */
  109. static const XML_Char *
  110. resolveSystemId(const XML_Char *base, const XML_Char *systemId,
  111. XML_Char **toFree) {
  112. XML_Char *s;
  113. *toFree = 0;
  114. if (! base || *systemId == T('/')
  115. #if defined(_WIN32)
  116. || *systemId == T('\\')
  117. || (isAsciiLetter(systemId[0]) && systemId[1] == T(':'))
  118. #endif
  119. )
  120. return systemId;
  121. *toFree = (XML_Char *)malloc((tcslen(base) + tcslen(systemId) + 2)
  122. * sizeof(XML_Char));
  123. if (! *toFree)
  124. return systemId;
  125. tcscpy(*toFree, base);
  126. s = *toFree;
  127. if (tcsrchr(s, T('/')))
  128. s = tcsrchr(s, T('/')) + 1;
  129. #if defined(_WIN32)
  130. if (tcsrchr(s, T('\\')))
  131. s = tcsrchr(s, T('\\')) + 1;
  132. #endif
  133. tcscpy(s, systemId);
  134. return *toFree;
  135. }
  136. static int
  137. externalEntityRefFilemap(XML_Parser parser, const XML_Char *context,
  138. const XML_Char *base, const XML_Char *systemId,
  139. const XML_Char *publicId) {
  140. int result;
  141. XML_Char *s;
  142. const XML_Char *filename;
  143. XML_Parser entParser = XML_ExternalEntityParserCreate(parser, context, 0);
  144. int filemapRes;
  145. PROCESS_ARGS args;
  146. UNUSED_P(publicId);
  147. args.retPtr = &result;
  148. args.parser = entParser;
  149. filename = resolveSystemId(base, systemId, &s);
  150. XML_SetBase(entParser, filename);
  151. filemapRes = filemap(filename, processFile, &args);
  152. switch (filemapRes) {
  153. case 0:
  154. result = 0;
  155. break;
  156. case 2:
  157. ftprintf(stderr,
  158. T("%s: file too large for memory-mapping")
  159. T(", switching to streaming\n"),
  160. filename);
  161. result = processStream(filename, entParser);
  162. break;
  163. }
  164. free(s);
  165. XML_ParserFree(entParser);
  166. return result;
  167. }
  168. static int
  169. processStream(const XML_Char *filename, XML_Parser parser) {
  170. /* passing NULL for filename means read input from stdin */
  171. int fd = 0; /* 0 is the fileno for stdin */
  172. if (filename != NULL) {
  173. fd = topen(filename, O_BINARY | O_RDONLY);
  174. if (fd < 0) {
  175. tperror(filename);
  176. return 0;
  177. }
  178. }
  179. for (;;) {
  180. EXPAT_read_count_t nread;
  181. char *buf = (char *)XML_GetBuffer(parser, g_read_size_bytes);
  182. if (! buf) {
  183. if (filename != NULL)
  184. close(fd);
  185. ftprintf(stderr, T("%s: out of memory\n"),
  186. filename != NULL ? filename : T("xmlwf"));
  187. return 0;
  188. }
  189. nread = EXPAT_read(fd, buf, (EXPAT_read_req_t)g_read_size_bytes);
  190. if (nread < 0) {
  191. tperror(filename != NULL ? filename : T("STDIN"));
  192. if (filename != NULL)
  193. close(fd);
  194. return 0;
  195. }
  196. if (XML_ParseBuffer(parser, (int)nread, nread == 0) == XML_STATUS_ERROR) {
  197. reportError(parser, filename != NULL ? filename : T("STDIN"));
  198. if (filename != NULL)
  199. close(fd);
  200. return 0;
  201. }
  202. if (nread == 0) {
  203. if (filename != NULL)
  204. close(fd);
  205. break;
  206. ;
  207. }
  208. }
  209. return 1;
  210. }
  211. static int
  212. externalEntityRefStream(XML_Parser parser, const XML_Char *context,
  213. const XML_Char *base, const XML_Char *systemId,
  214. const XML_Char *publicId) {
  215. XML_Char *s;
  216. const XML_Char *filename;
  217. int ret;
  218. XML_Parser entParser = XML_ExternalEntityParserCreate(parser, context, 0);
  219. UNUSED_P(publicId);
  220. filename = resolveSystemId(base, systemId, &s);
  221. XML_SetBase(entParser, filename);
  222. ret = processStream(filename, entParser);
  223. free(s);
  224. XML_ParserFree(entParser);
  225. return ret;
  226. }
  227. int
  228. XML_ProcessFile(XML_Parser parser, const XML_Char *filename, unsigned flags) {
  229. int result;
  230. if (! XML_SetBase(parser, filename)) {
  231. ftprintf(stderr, T("%s: out of memory"), filename);
  232. exit(1);
  233. }
  234. if (flags & XML_EXTERNAL_ENTITIES)
  235. XML_SetExternalEntityRefHandler(parser, (flags & XML_MAP_FILE)
  236. ? externalEntityRefFilemap
  237. : externalEntityRefStream);
  238. if (flags & XML_MAP_FILE) {
  239. int filemapRes;
  240. PROCESS_ARGS args;
  241. args.retPtr = &result;
  242. args.parser = parser;
  243. filemapRes = filemap(filename, processFile, &args);
  244. switch (filemapRes) {
  245. case 0:
  246. result = 0;
  247. break;
  248. case 2:
  249. ftprintf(stderr,
  250. T("%s: file too large for memory-mapping")
  251. T(", switching to streaming\n"),
  252. filename);
  253. result = processStream(filename, parser);
  254. break;
  255. }
  256. } else
  257. result = processStream(filename, parser);
  258. return result;
  259. }