encoding.h 7.9 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227
  1. /*
  2. * Summary: interface for the encoding conversion functions
  3. * Description: interface for the encoding conversion functions needed for
  4. * XML basic encoding and iconv() support.
  5. *
  6. * Related specs are
  7. * rfc2044 (UTF-8 and UTF-16) F. Yergeau Alis Technologies
  8. * [ISO-10646] UTF-8 and UTF-16 in Annexes
  9. * [ISO-8859-1] ISO Latin-1 characters codes.
  10. * [UNICODE] The Unicode Consortium, "The Unicode Standard --
  11. * Worldwide Character Encoding -- Version 1.0", Addison-
  12. * Wesley, Volume 1, 1991, Volume 2, 1992. UTF-8 is
  13. * described in Unicode Technical Report #4.
  14. * [US-ASCII] Coded Character Set--7-bit American Standard Code for
  15. * Information Interchange, ANSI X3.4-1986.
  16. *
  17. * Copy: See Copyright for the status of this software.
  18. *
  19. * Author: Daniel Veillard
  20. */
  21. #ifndef __XML_CHAR_ENCODING_H__
  22. #define __XML_CHAR_ENCODING_H__
  23. #include <libxml/xmlversion.h>
  24. #ifdef LIBXML_ICONV_ENABLED
  25. #include <iconv.h>
  26. #endif
  27. #ifdef __cplusplus
  28. extern "C" {
  29. #endif
  30. /*
  31. * xmlCharEncoding:
  32. *
  33. * Predefined values for some standard encodings.
  34. * Libxml does not do beforehand translation on UTF8 and ISOLatinX.
  35. * It also supports ASCII, ISO-8859-1, and UTF16 (LE and BE) by default.
  36. *
  37. * Anything else would have to be translated to UTF8 before being
  38. * given to the parser itself. The BOM for UTF16 and the encoding
  39. * declaration are looked at and a converter is looked for at that
  40. * point. If not found the parser stops here as asked by the XML REC. A
  41. * converter can be registered by the user using xmlRegisterCharEncodingHandler
  42. * but the current form doesn't allow stateful transcoding (a serious
  43. * problem agreed !). If iconv has been found it will be used
  44. * automatically and allow stateful transcoding, the simplest is then
  45. * to be sure to enable iconv and to provide iconv libs for the encoding
  46. * support needed.
  47. *
  48. * Note that the generic "UTF-16" is not a predefined value. Instead, only
  49. * the specific UTF-16LE and UTF-16BE are present.
  50. */
  51. typedef enum {
  52. XML_CHAR_ENCODING_ERROR= -1, /* No char encoding detected */
  53. XML_CHAR_ENCODING_NONE= 0, /* No char encoding detected */
  54. XML_CHAR_ENCODING_UTF8= 1, /* UTF-8 */
  55. XML_CHAR_ENCODING_UTF16LE= 2, /* UTF-16 little endian */
  56. XML_CHAR_ENCODING_UTF16BE= 3, /* UTF-16 big endian */
  57. XML_CHAR_ENCODING_UCS4LE= 4, /* UCS-4 little endian */
  58. XML_CHAR_ENCODING_UCS4BE= 5, /* UCS-4 big endian */
  59. XML_CHAR_ENCODING_EBCDIC= 6, /* EBCDIC uh! */
  60. XML_CHAR_ENCODING_UCS4_2143=7, /* UCS-4 unusual ordering */
  61. XML_CHAR_ENCODING_UCS4_3412=8, /* UCS-4 unusual ordering */
  62. XML_CHAR_ENCODING_UCS2= 9, /* UCS-2 */
  63. XML_CHAR_ENCODING_8859_1= 10,/* ISO-8859-1 ISO Latin 1 */
  64. XML_CHAR_ENCODING_8859_2= 11,/* ISO-8859-2 ISO Latin 2 */
  65. XML_CHAR_ENCODING_8859_3= 12,/* ISO-8859-3 */
  66. XML_CHAR_ENCODING_8859_4= 13,/* ISO-8859-4 */
  67. XML_CHAR_ENCODING_8859_5= 14,/* ISO-8859-5 */
  68. XML_CHAR_ENCODING_8859_6= 15,/* ISO-8859-6 */
  69. XML_CHAR_ENCODING_8859_7= 16,/* ISO-8859-7 */
  70. XML_CHAR_ENCODING_8859_8= 17,/* ISO-8859-8 */
  71. XML_CHAR_ENCODING_8859_9= 18,/* ISO-8859-9 */
  72. XML_CHAR_ENCODING_2022_JP= 19,/* ISO-2022-JP */
  73. XML_CHAR_ENCODING_SHIFT_JIS=20,/* Shift_JIS */
  74. XML_CHAR_ENCODING_EUC_JP= 21,/* EUC-JP */
  75. XML_CHAR_ENCODING_ASCII= 22 /* pure ASCII */
  76. }
  77. xmlCharEncoding;
  78. /**
  79. * xmlCharEncodingInputFunc:
  80. * @out: a pointer to an array of bytes to store the UTF-8 result
  81. * @outlen: the length of @out
  82. * @in: a pointer to an array of chars in the original encoding
  83. * @inlen: the length of @in
  84. *
  85. * Take a block of chars in the original encoding and try to convert
  86. * it to an UTF-8 block of chars out.
  87. *
  88. * Returns the number of bytes written, -1 if lack of space, or -2
  89. * if the transcoding failed.
  90. * The value of @inlen after return is the number of octets consumed
  91. * if the return value is positive, else unpredictiable.
  92. * The value of @outlen after return is the number of octets consumed.
  93. */
  94. typedef int (* xmlCharEncodingInputFunc)(unsigned char *out, int *outlen,
  95. const unsigned char *in, int *inlen);
  96. /**
  97. * xmlCharEncodingOutputFunc:
  98. * @out: a pointer to an array of bytes to store the result
  99. * @outlen: the length of @out
  100. * @in: a pointer to an array of UTF-8 chars
  101. * @inlen: the length of @in
  102. *
  103. * Take a block of UTF-8 chars in and try to convert it to another
  104. * encoding.
  105. * Note: a first call designed to produce heading info is called with
  106. * in = NULL. If stateful this should also initialize the encoder state.
  107. *
  108. * Returns the number of bytes written, -1 if lack of space, or -2
  109. * if the transcoding failed.
  110. * The value of @inlen after return is the number of octets consumed
  111. * if the return value is positive, else unpredictiable.
  112. * The value of @outlen after return is the number of octets produced.
  113. */
  114. typedef int (* xmlCharEncodingOutputFunc)(unsigned char *out, int *outlen,
  115. const unsigned char *in, int *inlen);
  116. /*
  117. * Block defining the handlers for non UTF-8 encodings.
  118. * If iconv is supported, there are two extra fields.
  119. */
  120. typedef struct _xmlCharEncodingHandler xmlCharEncodingHandler;
  121. typedef xmlCharEncodingHandler *xmlCharEncodingHandlerPtr;
  122. struct _xmlCharEncodingHandler {
  123. char *name;
  124. xmlCharEncodingInputFunc input;
  125. xmlCharEncodingOutputFunc output;
  126. #ifdef LIBXML_ICONV_ENABLED
  127. iconv_t iconv_in;
  128. iconv_t iconv_out;
  129. #endif /* LIBXML_ICONV_ENABLED */
  130. };
  131. #ifdef __cplusplus
  132. }
  133. #endif
  134. #include <libxml/tree.h>
  135. #ifdef __cplusplus
  136. extern "C" {
  137. #endif
  138. /*
  139. * Interfaces for encoding handlers.
  140. */
  141. XMLPUBFUN void XMLCALL
  142. xmlInitCharEncodingHandlers (void);
  143. XMLPUBFUN void XMLCALL
  144. xmlCleanupCharEncodingHandlers (void);
  145. XMLPUBFUN void XMLCALL
  146. xmlRegisterCharEncodingHandler (xmlCharEncodingHandlerPtr handler);
  147. XMLPUBFUN xmlCharEncodingHandlerPtr XMLCALL
  148. xmlGetCharEncodingHandler (xmlCharEncoding enc);
  149. XMLPUBFUN xmlCharEncodingHandlerPtr XMLCALL
  150. xmlFindCharEncodingHandler (const char *name);
  151. XMLPUBFUN xmlCharEncodingHandlerPtr XMLCALL
  152. xmlNewCharEncodingHandler (const char *name,
  153. xmlCharEncodingInputFunc input,
  154. xmlCharEncodingOutputFunc output);
  155. /*
  156. * Interfaces for encoding names and aliases.
  157. */
  158. XMLPUBFUN int XMLCALL
  159. xmlAddEncodingAlias (const char *name,
  160. const char *alias);
  161. XMLPUBFUN int XMLCALL
  162. xmlDelEncodingAlias (const char *alias);
  163. XMLPUBFUN const char * XMLCALL
  164. xmlGetEncodingAlias (const char *alias);
  165. XMLPUBFUN void XMLCALL
  166. xmlCleanupEncodingAliases (void);
  167. XMLPUBFUN xmlCharEncoding XMLCALL
  168. xmlParseCharEncoding (const char *name);
  169. XMLPUBFUN const char * XMLCALL
  170. xmlGetCharEncodingName (xmlCharEncoding enc);
  171. /*
  172. * Interfaces directly used by the parsers.
  173. */
  174. XMLPUBFUN xmlCharEncoding XMLCALL
  175. xmlDetectCharEncoding (const unsigned char *in,
  176. int len);
  177. XMLPUBFUN int XMLCALL
  178. xmlCharEncOutFunc (xmlCharEncodingHandler *handler,
  179. xmlBufferPtr out,
  180. xmlBufferPtr in);
  181. XMLPUBFUN int XMLCALL
  182. xmlCharEncInFunc (xmlCharEncodingHandler *handler,
  183. xmlBufferPtr out,
  184. xmlBufferPtr in);
  185. XMLPUBFUN int XMLCALL
  186. xmlCharEncFirstLine (xmlCharEncodingHandler *handler,
  187. xmlBufferPtr out,
  188. xmlBufferPtr in);
  189. XMLPUBFUN int XMLCALL
  190. xmlCharEncCloseFunc (xmlCharEncodingHandler *handler);
  191. /*
  192. * Export a few useful functions
  193. */
  194. #ifdef LIBXML_OUTPUT_ENABLED
  195. XMLPUBFUN int XMLCALL
  196. UTF8Toisolat1 (unsigned char *out,
  197. int *outlen,
  198. const unsigned char *in,
  199. int *inlen);
  200. #endif /* LIBXML_OUTPUT_ENABLED */
  201. XMLPUBFUN int XMLCALL
  202. isolat1ToUTF8 (unsigned char *out,
  203. int *outlen,
  204. const unsigned char *in,
  205. int *inlen);
  206. #ifdef __cplusplus
  207. }
  208. #endif
  209. #endif /* __XML_CHAR_ENCODING_H__ */