Umasoft
 All Classes Files Functions Variables Typedefs Enumerations Enumerator Groups Pages
xmltok.h
1 /*
2 The contents of this file are subject to the Mozilla Public License
3 Version 1.1 (the "License"); you may not use this file except in
4 compliance with the License. You may obtain a copy of the License at
5 http://www.mozilla.org/MPL/
6 
7 Software distributed under the License is distributed on an "AS IS"
8 basis, WITHOUT WARRANTY OF ANY KIND, either express or implied. See the
9 License for the specific language governing rights and limitations
10 under the License.
11 
12 The Original Code is expat.
13 
14 The Initial Developer of the Original Code is James Clark.
15 Portions created by James Clark are Copyright (C) 1998, 1999
16 James Clark. All Rights Reserved.
17 
18 Contributor(s):
19 
20 Alternatively, the contents of this file may be used under the terms
21 of the GNU General Public License (the "GPL"), in which case the
22 provisions of the GPL are applicable instead of those above. If you
23 wish to allow use of your version of this file only under the terms of
24 the GPL and not to allow others to use your version of this file under
25 the MPL, indicate your decision by deleting the provisions above and
26 replace them with the notice and other provisions required by the
27 GPL. If you do not delete the provisions above, a recipient may use
28 your version of this file under either the MPL or the GPL.
29 */
30 
31 #ifndef XmlTok_INCLUDED
32 #define XmlTok_INCLUDED 1
33 
34 #ifdef __cplusplus
35 extern "C" {
36 #endif
37 
38 #ifndef XMLTOKAPI
39 #define XMLTOKAPI /* as nothing */
40 #endif
41 
42 /* The following token may be returned by XmlContentTok */
43 #define XML_TOK_TRAILING_RSQB -5 /* ] or ]] at the end of the scan; might be start of
44  illegal ]]> sequence */
45 /* The following tokens may be returned by both XmlPrologTok and XmlContentTok */
46 #define XML_TOK_NONE -4 /* The string to be scanned is empty */
47 #define XML_TOK_TRAILING_CR -3 /* A CR at the end of the scan;
48  might be part of CRLF sequence */
49 #define XML_TOK_PARTIAL_CHAR -2 /* only part of a multibyte sequence */
50 #define XML_TOK_PARTIAL -1 /* only part of a token */
51 #define XML_TOK_INVALID 0
52 
53 /* The following tokens are returned by XmlContentTok; some are also
54  returned by XmlAttributeValueTok, XmlEntityTok, XmlCdataSectionTok */
55 
56 #define XML_TOK_START_TAG_WITH_ATTS 1
57 #define XML_TOK_START_TAG_NO_ATTS 2
58 #define XML_TOK_EMPTY_ELEMENT_WITH_ATTS 3 /* empty element tag <e/> */
59 #define XML_TOK_EMPTY_ELEMENT_NO_ATTS 4
60 #define XML_TOK_END_TAG 5
61 #define XML_TOK_DATA_CHARS 6
62 #define XML_TOK_DATA_NEWLINE 7
63 #define XML_TOK_CDATA_SECT_OPEN 8
64 #define XML_TOK_ENTITY_REF 9
65 #define XML_TOK_CHAR_REF 10 /* numeric character reference */
66 
67 /* The following tokens may be returned by both XmlPrologTok and XmlContentTok */
68 #define XML_TOK_PI 11 /* processing instruction */
69 #define XML_TOK_XML_DECL 12 /* XML decl or text decl */
70 #define XML_TOK_COMMENT 13
71 #define XML_TOK_BOM 14 /* Byte order mark */
72 
73 /* The following tokens are returned only by XmlPrologTok */
74 #define XML_TOK_PROLOG_S 15
75 #define XML_TOK_DECL_OPEN 16 /* <!foo */
76 #define XML_TOK_DECL_CLOSE 17 /* > */
77 #define XML_TOK_NAME 18
78 #define XML_TOK_NMTOKEN 19
79 #define XML_TOK_POUND_NAME 20 /* #name */
80 #define XML_TOK_OR 21 /* | */
81 #define XML_TOK_PERCENT 22
82 #define XML_TOK_OPEN_PAREN 23
83 #define XML_TOK_CLOSE_PAREN 24
84 #define XML_TOK_OPEN_BRACKET 25
85 #define XML_TOK_CLOSE_BRACKET 26
86 #define XML_TOK_LITERAL 27
87 #define XML_TOK_PARAM_ENTITY_REF 28
88 #define XML_TOK_INSTANCE_START 29
89 
90 /* The following occur only in element type declarations */
91 #define XML_TOK_NAME_QUESTION 30 /* name? */
92 #define XML_TOK_NAME_ASTERISK 31 /* name* */
93 #define XML_TOK_NAME_PLUS 32 /* name+ */
94 #define XML_TOK_COND_SECT_OPEN 33 /* <![ */
95 #define XML_TOK_COND_SECT_CLOSE 34 /* ]]> */
96 #define XML_TOK_CLOSE_PAREN_QUESTION 35 /* )? */
97 #define XML_TOK_CLOSE_PAREN_ASTERISK 36 /* )* */
98 #define XML_TOK_CLOSE_PAREN_PLUS 37 /* )+ */
99 #define XML_TOK_COMMA 38
100 
101 /* The following token is returned only by XmlAttributeValueTok */
102 #define XML_TOK_ATTRIBUTE_VALUE_S 39
103 
104 /* The following token is returned only by XmlCdataSectionTok */
105 #define XML_TOK_CDATA_SECT_CLOSE 40
106 
107 /* With namespace processing this is returned by XmlPrologTok
108  for a name with a colon. */
109 #define XML_TOK_PREFIXED_NAME 41
110 
111 #define XML_N_STATES 3
112 #define XML_PROLOG_STATE 0
113 #define XML_CONTENT_STATE 1
114 #define XML_CDATA_SECTION_STATE 2
115 
116 #define XML_N_LITERAL_TYPES 2
117 #define XML_ATTRIBUTE_VALUE_LITERAL 0
118 #define XML_ENTITY_VALUE_LITERAL 1
119 
120 /* The size of the buffer passed to XmlUtf8Encode must be at least this. */
121 #define XML_UTF8_ENCODE_MAX 4
122 /* The size of the buffer passed to XmlUtf16Encode must be at least this. */
123 #define XML_UTF16_ENCODE_MAX 2
124 
125 typedef struct position {
126  /* first line and first column are 0 not 1 */
127  unsigned long lineNumber;
128  unsigned long columnNumber;
129 } POSITION;
130 
131 typedef struct {
132  const char *name;
133  const char *valuePtr;
134  const char *valueEnd;
135  char normalized;
136 } ATTRIBUTE;
137 
138 struct encoding;
139 typedef struct encoding ENCODING;
140 
141 struct encoding {
142  int (*scanners[XML_N_STATES])(const ENCODING *,
143  const char *,
144  const char *,
145  const char **);
146  int (*literalScanners[XML_N_LITERAL_TYPES])(const ENCODING *,
147  const char *,
148  const char *,
149  const char **);
150  int (*sameName)(const ENCODING *,
151  const char *, const char *);
152  int (*nameMatchesAscii)(const ENCODING *,
153  const char *, const char *);
154  int (*nameLength)(const ENCODING *, const char *);
155  const char *(*skipS)(const ENCODING *, const char *);
156  int (*getAtts)(const ENCODING *enc, const char *ptr,
157  int attsMax, ATTRIBUTE *atts);
158  int (*charRefNumber)(const ENCODING *enc, const char *ptr);
159  int (*predefinedEntityName)(const ENCODING *, const char *, const char *);
160  void (*updatePosition)(const ENCODING *,
161  const char *ptr,
162  const char *end,
163  POSITION *);
164  int (*isPublicId)(const ENCODING *enc, const char *ptr, const char *end,
165  const char **badPtr);
166  void (*utf8Convert)(const ENCODING *enc,
167  const char **fromP,
168  const char *fromLim,
169  char **toP,
170  const char *toLim);
171  void (*utf16Convert)(const ENCODING *enc,
172  const char **fromP,
173  const char *fromLim,
174  unsigned short **toP,
175  const unsigned short *toLim);
176  int minBytesPerChar;
177  char isUtf8;
178  char isUtf16;
179 };
180 
181 /*
182 Scan the string starting at ptr until the end of the next complete token,
183 but do not scan past eptr. Return an integer giving the type of token.
184 
185 Return XML_TOK_NONE when ptr == eptr; nextTokPtr will not be set.
186 
187 Return XML_TOK_PARTIAL when the string does not contain a complete token;
188 nextTokPtr will not be set.
189 
190 Return XML_TOK_INVALID when the string does not start a valid token; nextTokPtr
191 will be set to point to the character which made the token invalid.
192 
193 Otherwise the string starts with a valid token; nextTokPtr will be set to point
194 to the character following the end of that token.
195 
196 Each data character counts as a single token, but adjacent data characters
197 may be returned together. Similarly for characters in the prolog outside
198 literals, comments and processing instructions.
199 */
200 
201 
202 #define XmlTok(enc, state, ptr, end, nextTokPtr) \
203  (((enc)->scanners[state])(enc, ptr, end, nextTokPtr))
204 
205 #define XmlPrologTok(enc, ptr, end, nextTokPtr) \
206  XmlTok(enc, XML_PROLOG_STATE, ptr, end, nextTokPtr)
207 
208 #define XmlContentTok(enc, ptr, end, nextTokPtr) \
209  XmlTok(enc, XML_CONTENT_STATE, ptr, end, nextTokPtr)
210 
211 #define XmlCdataSectionTok(enc, ptr, end, nextTokPtr) \
212  XmlTok(enc, XML_CDATA_SECTION_STATE, ptr, end, nextTokPtr)
213 
214 /* This is used for performing a 2nd-level tokenization on
215 the content of a literal that has already been returned by XmlTok. */
216 
217 #define XmlLiteralTok(enc, literalType, ptr, end, nextTokPtr) \
218  (((enc)->literalScanners[literalType])(enc, ptr, end, nextTokPtr))
219 
220 #define XmlAttributeValueTok(enc, ptr, end, nextTokPtr) \
221  XmlLiteralTok(enc, XML_ATTRIBUTE_VALUE_LITERAL, ptr, end, nextTokPtr)
222 
223 #define XmlEntityValueTok(enc, ptr, end, nextTokPtr) \
224  XmlLiteralTok(enc, XML_ENTITY_VALUE_LITERAL, ptr, end, nextTokPtr)
225 
226 #define XmlSameName(enc, ptr1, ptr2) (((enc)->sameName)(enc, ptr1, ptr2))
227 
228 #define XmlNameMatchesAscii(enc, ptr1, ptr2) \
229  (((enc)->nameMatchesAscii)(enc, ptr1, ptr2))
230 
231 #define XmlNameLength(enc, ptr) \
232  (((enc)->nameLength)(enc, ptr))
233 
234 #define XmlSkipS(enc, ptr) \
235  (((enc)->skipS)(enc, ptr))
236 
237 #define XmlGetAttributes(enc, ptr, attsMax, atts) \
238  (((enc)->getAtts)(enc, ptr, attsMax, atts))
239 
240 #define XmlCharRefNumber(enc, ptr) \
241  (((enc)->charRefNumber)(enc, ptr))
242 
243 #define XmlPredefinedEntityName(enc, ptr, end) \
244  (((enc)->predefinedEntityName)(enc, ptr, end))
245 
246 #define XmlUpdatePosition(enc, ptr, end, pos) \
247  (((enc)->updatePosition)(enc, ptr, end, pos))
248 
249 #define XmlIsPublicId(enc, ptr, end, badPtr) \
250  (((enc)->isPublicId)(enc, ptr, end, badPtr))
251 
252 #define XmlUtf8Convert(enc, fromP, fromLim, toP, toLim) \
253  (((enc)->utf8Convert)(enc, fromP, fromLim, toP, toLim))
254 
255 #define XmlUtf16Convert(enc, fromP, fromLim, toP, toLim) \
256  (((enc)->utf16Convert)(enc, fromP, fromLim, toP, toLim))
257 
258 typedef struct {
259  ENCODING initEnc;
260  const ENCODING **encPtr;
261 } INIT_ENCODING;
262 
263 int XMLTOKAPI XmlParseXmlDecl(int isGeneralTextEntity,
264  const ENCODING *enc,
265  const char *ptr,
266  const char *end,
267  const char **badPtr,
268  const char **versionPtr,
269  const char **encodingNamePtr,
270  const ENCODING **namedEncodingPtr,
271  int *standalonePtr);
272 
273 int XMLTOKAPI XmlInitEncoding(INIT_ENCODING *, const ENCODING **, const char *name);
274 const ENCODING XMLTOKAPI *XmlGetUtf8InternalEncoding();
275 const ENCODING XMLTOKAPI *XmlGetUtf16InternalEncoding();
276 int XMLTOKAPI XmlUtf8Encode(int charNumber, char *buf);
277 int XMLTOKAPI XmlUtf16Encode(int charNumber, unsigned short *buf);
278 
279 int XMLTOKAPI XmlSizeOfUnknownEncoding();
280 ENCODING XMLTOKAPI *
281 XmlInitUnknownEncoding(void *mem,
282  int *table,
283  int (*conv)(void *userData, const char *p),
284  void *userData);
285 
286 int XMLTOKAPI XmlParseXmlDeclNS(int isGeneralTextEntity,
287  const ENCODING *enc,
288  const char *ptr,
289  const char *end,
290  const char **badPtr,
291  const char **versionPtr,
292  const char **encodingNamePtr,
293  const ENCODING **namedEncodingPtr,
294  int *standalonePtr);
295 int XMLTOKAPI XmlInitEncodingNS(INIT_ENCODING *, const ENCODING **, const char *name);
296 const ENCODING XMLTOKAPI *XmlGetUtf8InternalEncodingNS();
297 const ENCODING XMLTOKAPI *XmlGetUtf16InternalEncodingNS();
298 ENCODING XMLTOKAPI *
299 XmlInitUnknownEncodingNS(void *mem,
300  int *table,
301  int (*conv)(void *userData, const char *p),
302  void *userData);
303 #ifdef __cplusplus
304 }
305 #endif
306 
307 #endif /* not XmlTok_INCLUDED */