1 | /* Copyright (c) 1998, 1999 Thai Open Source Software Center Ltd |
2 | See the file COPYING for copying permission. |
3 | */ |
4 | |
5 | #include <stddef.h> |
6 | |
7 | #ifdef COMPILED_FROM_DSP |
8 | #include "winconfig.h" |
9 | #elif defined(MACOS_CLASSIC) |
10 | #include "macconfig.h" |
11 | #elif defined(__amigaos__) |
12 | #include "amigaconfig.h" |
13 | #elif defined(__WATCOMC__) |
14 | #include "watcomconfig.h" |
15 | #else |
16 | #ifdef HAVE_EXPAT_CONFIG_H |
17 | #include <expat_config.h> |
18 | #endif |
19 | #endif /* ndef COMPILED_FROM_DSP */ |
20 | |
21 | #include "expat_external.h" |
22 | #include "internal.h" |
23 | #include "xmltok.h" |
24 | #include "nametab.h" |
25 | |
26 | #ifdef XML_DTD |
27 | #define IGNORE_SECTION_TOK_VTABLE , PREFIX(ignoreSectionTok) |
28 | #else |
29 | #define IGNORE_SECTION_TOK_VTABLE /* as nothing */ |
30 | #endif |
31 | |
32 | #define VTABLE1 \ |
33 | { PREFIX(prologTok), PREFIX(contentTok), \ |
34 | PREFIX(cdataSectionTok) IGNORE_SECTION_TOK_VTABLE }, \ |
35 | { PREFIX(attributeValueTok), PREFIX(entityValueTok) }, \ |
36 | PREFIX(sameName), \ |
37 | PREFIX(nameMatchesAscii), \ |
38 | PREFIX(nameLength), \ |
39 | PREFIX(skipS), \ |
40 | PREFIX(getAtts), \ |
41 | PREFIX(charRefNumber), \ |
42 | PREFIX(predefinedEntityName), \ |
43 | PREFIX(updatePosition), \ |
44 | PREFIX(isPublicId) |
45 | |
46 | #define VTABLE VTABLE1, PREFIX(toUtf8), PREFIX(toUtf16) |
47 | |
48 | #define UCS2_GET_NAMING(pages, hi, lo) \ |
49 | (namingBitmap[(pages[hi] << 3) + ((lo) >> 5)] & (1 << ((lo) & 0x1F))) |
50 | |
51 | /* A 2 byte UTF-8 representation splits the characters 11 bits between |
52 | the bottom 5 and 6 bits of the bytes. We need 8 bits to index into |
53 | pages, 3 bits to add to that index and 5 bits to generate the mask. |
54 | */ |
55 | #define UTF8_GET_NAMING2(pages, byte) \ |
56 | (namingBitmap[((pages)[(((byte)[0]) >> 2) & 7] << 3) \ |
57 | + ((((byte)[0]) & 3) << 1) \ |
58 | + ((((byte)[1]) >> 5) & 1)] \ |
59 | & (1 << (((byte)[1]) & 0x1F))) |
60 | |
61 | /* A 3 byte UTF-8 representation splits the characters 16 bits between |
62 | the bottom 4, 6 and 6 bits of the bytes. We need 8 bits to index |
63 | into pages, 3 bits to add to that index and 5 bits to generate the |
64 | mask. |
65 | */ |
66 | #define UTF8_GET_NAMING3(pages, byte) \ |
67 | (namingBitmap[((pages)[((((byte)[0]) & 0xF) << 4) \ |
68 | + ((((byte)[1]) >> 2) & 0xF)] \ |
69 | << 3) \ |
70 | + ((((byte)[1]) & 3) << 1) \ |
71 | + ((((byte)[2]) >> 5) & 1)] \ |
72 | & (1 << (((byte)[2]) & 0x1F))) |
73 | |
74 | #define UTF8_GET_NAMING(pages, p, n) \ |
75 | ((n) == 2 \ |
76 | ? UTF8_GET_NAMING2(pages, (const unsigned char *)(p)) \ |
77 | : ((n) == 3 \ |
78 | ? UTF8_GET_NAMING3(pages, (const unsigned char *)(p)) \ |
79 | : 0)) |
80 | |
81 | /* Detection of invalid UTF-8 sequences is based on Table 3.1B |
82 | of Unicode 3.2: http://www.unicode.org/unicode/reports/tr28/ |
83 | with the additional restriction of not allowing the Unicode |
84 | code points 0xFFFF and 0xFFFE (sequences EF,BF,BF and EF,BF,BE). |
85 | Implementation details: |
86 | (A & 0x80) == 0 means A < 0x80 |
87 | and |
88 | (A & 0xC0) == 0xC0 means A > 0xBF |
89 | */ |
90 | |
91 | #define UTF8_INVALID2(p) \ |
92 | ((*p) < 0xC2 || ((p)[1] & 0x80) == 0 || ((p)[1] & 0xC0) == 0xC0) |
93 | |
94 | #define UTF8_INVALID3(p) \ |
95 | (((p)[2] & 0x80) == 0 \ |
96 | || \ |
97 | ((*p) == 0xEF && (p)[1] == 0xBF \ |
98 | ? \ |
99 | (p)[2] > 0xBD \ |
100 | : \ |
101 | ((p)[2] & 0xC0) == 0xC0) \ |
102 | || \ |
103 | ((*p) == 0xE0 \ |
104 | ? \ |
105 | (p)[1] < 0xA0 || ((p)[1] & 0xC0) == 0xC0 \ |
106 | : \ |
107 | ((p)[1] & 0x80) == 0 \ |
108 | || \ |
109 | ((*p) == 0xED ? (p)[1] > 0x9F : ((p)[1] & 0xC0) == 0xC0))) |
110 | |
111 | #define UTF8_INVALID4(p) \ |
112 | (((p)[3] & 0x80) == 0 || ((p)[3] & 0xC0) == 0xC0 \ |
113 | || \ |
114 | ((p)[2] & 0x80) == 0 || ((p)[2] & 0xC0) == 0xC0 \ |
115 | || \ |
116 | ((*p) == 0xF0 \ |
117 | ? \ |
118 | (p)[1] < 0x90 || ((p)[1] & 0xC0) == 0xC0 \ |
119 | : \ |
120 | ((p)[1] & 0x80) == 0 \ |
121 | || \ |
122 | ((*p) == 0xF4 ? (p)[1] > 0x8F : ((p)[1] & 0xC0) == 0xC0))) |
123 | |
124 | static int PTRFASTCALL |
125 | isNever(const ENCODING *enc, const char *p) |
126 | { |
127 | return 0; |
128 | } |
129 | |
130 | static int PTRFASTCALL |
131 | utf8_isName2(const ENCODING *enc, const char *p) |
132 | { |
133 | return UTF8_GET_NAMING2(namePages, (const unsigned char *)p); |
134 | } |
135 | |
136 | static int PTRFASTCALL |
137 | utf8_isName3(const ENCODING *enc, const char *p) |
138 | { |
139 | return UTF8_GET_NAMING3(namePages, (const unsigned char *)p); |
140 | } |
141 | |
142 | #define utf8_isName4 isNever |
143 | |
144 | static int PTRFASTCALL |
145 | utf8_isNmstrt2(const ENCODING *enc, const char *p) |
146 | { |
147 | return UTF8_GET_NAMING2(nmstrtPages, (const unsigned char *)p); |
148 | } |
149 | |
150 | static int PTRFASTCALL |
151 | utf8_isNmstrt3(const ENCODING *enc, const char *p) |
152 | { |
153 | return UTF8_GET_NAMING3(nmstrtPages, (const unsigned char *)p); |
154 | } |
155 | |
156 | #define utf8_isNmstrt4 isNever |
157 | |
158 | static int PTRFASTCALL |
159 | utf8_isInvalid2(const ENCODING *enc, const char *p) |
160 | { |
161 | return UTF8_INVALID2((const unsigned char *)p); |
162 | } |
163 | |
164 | static int PTRFASTCALL |
165 | utf8_isInvalid3(const ENCODING *enc, const char *p) |
166 | { |
167 | return UTF8_INVALID3((const unsigned char *)p); |
168 | } |
169 | |
170 | static int PTRFASTCALL |
171 | utf8_isInvalid4(const ENCODING *enc, const char *p) |
172 | { |
173 | return UTF8_INVALID4((const unsigned char *)p); |
174 | } |
175 | |
176 | struct normal_encoding { |
177 | ENCODING enc; |
178 | unsigned char type[256]; |
179 | #ifdef XML_MIN_SIZE |
180 | int (PTRFASTCALL *byteType)(const ENCODING *, const char *); |
181 | int (PTRFASTCALL *isNameMin)(const ENCODING *, const char *); |
182 | int (PTRFASTCALL *isNmstrtMin)(const ENCODING *, const char *); |
183 | int (PTRFASTCALL *byteToAscii)(const ENCODING *, const char *); |
184 | int (PTRCALL *charMatches)(const ENCODING *, const char *, int); |
185 | #endif /* XML_MIN_SIZE */ |
186 | int (PTRFASTCALL *isName2)(const ENCODING *, const char *); |
187 | int (PTRFASTCALL *isName3)(const ENCODING *, const char *); |
188 | int (PTRFASTCALL *isName4)(const ENCODING *, const char *); |
189 | int (PTRFASTCALL *isNmstrt2)(const ENCODING *, const char *); |
190 | int (PTRFASTCALL *isNmstrt3)(const ENCODING *, const char *); |
191 | int (PTRFASTCALL *isNmstrt4)(const ENCODING *, const char *); |
192 | int (PTRFASTCALL *isInvalid2)(const ENCODING *, const char *); |
193 | int (PTRFASTCALL *isInvalid3)(const ENCODING *, const char *); |
194 | int (PTRFASTCALL *isInvalid4)(const ENCODING *, const char *); |
195 | }; |
196 | |
197 | #define AS_NORMAL_ENCODING(enc) ((const struct normal_encoding *) (enc)) |
198 | |
199 | #ifdef XML_MIN_SIZE |
200 | |
201 | #define STANDARD_VTABLE(E) \ |
202 | E ## byteType, \ |
203 | E ## isNameMin, \ |
204 | E ## isNmstrtMin, \ |
205 | E ## byteToAscii, \ |
206 | E ## charMatches, |
207 | |
208 | #else |
209 | |
210 | #define STANDARD_VTABLE(E) /* as nothing */ |
211 | |
212 | #endif |
213 | |
214 | #define NORMAL_VTABLE(E) \ |
215 | E ## isName2, \ |
216 | E ## isName3, \ |
217 | E ## isName4, \ |
218 | E ## isNmstrt2, \ |
219 | E ## isNmstrt3, \ |
220 | E ## isNmstrt4, \ |
221 | E ## isInvalid2, \ |
222 | E ## isInvalid3, \ |
223 | E ## isInvalid4 |
224 | |
225 | static int FASTCALL checkCharRefNumber(int); |
226 | |
227 | #include "xmltok_impl.h" |
228 | #include "ascii.h" |
229 | |
230 | #ifdef XML_MIN_SIZE |
231 | #define sb_isNameMin isNever |
232 | #define sb_isNmstrtMin isNever |
233 | #endif |
234 | |
235 | #ifdef XML_MIN_SIZE |
236 | #define MINBPC(enc) ((enc)->minBytesPerChar) |
237 | #else |
238 | /* minimum bytes per character */ |
239 | #define MINBPC(enc) 1 |
240 | #endif |
241 | |
242 | #define SB_BYTE_TYPE(enc, p) \ |
243 | (((struct normal_encoding *)(enc))->type[(unsigned char)*(p)]) |
244 | |
245 | #ifdef XML_MIN_SIZE |
246 | static int PTRFASTCALL |
247 | sb_byteType(const ENCODING *enc, const char *p) |
248 | { |
249 | return SB_BYTE_TYPE(enc, p); |
250 | } |
251 | #define BYTE_TYPE(enc, p) \ |
252 | (AS_NORMAL_ENCODING(enc)->byteType(enc, p)) |
253 | #else |
254 | #define BYTE_TYPE(enc, p) SB_BYTE_TYPE(enc, p) |
255 | #endif |
256 | |
257 | #ifdef XML_MIN_SIZE |
258 | #define BYTE_TO_ASCII(enc, p) \ |
259 | (AS_NORMAL_ENCODING(enc)->byteToAscii(enc, p)) |
260 | static int PTRFASTCALL |
261 | sb_byteToAscii(const ENCODING *enc, const char *p) |
262 | { |
263 | return *p; |
264 | } |
265 | #else |
266 | #define BYTE_TO_ASCII(enc, p) (*(p)) |
267 | #endif |
268 | |
269 | #define IS_NAME_CHAR(enc, p, n) \ |
270 | (AS_NORMAL_ENCODING(enc)->isName ## n(enc, p)) |
271 | #define IS_NMSTRT_CHAR(enc, p, n) \ |
272 | (AS_NORMAL_ENCODING(enc)->isNmstrt ## n(enc, p)) |
273 | #define IS_INVALID_CHAR(enc, p, n) \ |
274 | (AS_NORMAL_ENCODING(enc)->isInvalid ## n(enc, p)) |
275 | |
276 | #ifdef XML_MIN_SIZE |
277 | #define IS_NAME_CHAR_MINBPC(enc, p) \ |
278 | (AS_NORMAL_ENCODING(enc)->isNameMin(enc, p)) |
279 | #define IS_NMSTRT_CHAR_MINBPC(enc, p) \ |
280 | (AS_NORMAL_ENCODING(enc)->isNmstrtMin(enc, p)) |
281 | #else |
282 | #define IS_NAME_CHAR_MINBPC(enc, p) (0) |
283 | #define IS_NMSTRT_CHAR_MINBPC(enc, p) (0) |
284 | #endif |
285 | |
286 | #ifdef XML_MIN_SIZE |
287 | #define CHAR_MATCHES(enc, p, c) \ |
288 | (AS_NORMAL_ENCODING(enc)->charMatches(enc, p, c)) |
289 | static int PTRCALL |
290 | sb_charMatches(const ENCODING *enc, const char *p, int c) |
291 | { |
292 | return *p == c; |
293 | } |
294 | #else |
295 | /* c is an ASCII character */ |
296 | #define CHAR_MATCHES(enc, p, c) (*(p) == c) |
297 | #endif |
298 | |
299 | #define PREFIX(ident) normal_ ## ident |
300 | #define XML_TOK_IMPL_C |
301 | #include "xmltok_impl.c" |
302 | #undef XML_TOK_IMPL_C |
303 | |
304 | #undef MINBPC |
305 | #undef BYTE_TYPE |
306 | #undef BYTE_TO_ASCII |
307 | #undef CHAR_MATCHES |
308 | #undef IS_NAME_CHAR |
309 | #undef IS_NAME_CHAR_MINBPC |
310 | #undef IS_NMSTRT_CHAR |
311 | #undef IS_NMSTRT_CHAR_MINBPC |
312 | #undef IS_INVALID_CHAR |
313 | |
314 | enum { /* UTF8_cvalN is value of masked first byte of N byte sequence */ |
315 | UTF8_cval1 = 0x00, |
316 | UTF8_cval2 = 0xc0, |
317 | UTF8_cval3 = 0xe0, |
318 | UTF8_cval4 = 0xf0 |
319 | }; |
320 | |
321 | static void PTRCALL |
322 | utf8_toUtf8(const ENCODING *enc, |
323 | const char **fromP, const char *fromLim, |
324 | char **toP, const char *toLim) |
325 | { |
326 | char *to; |
327 | const char *from; |
328 | if (fromLim - *fromP > toLim - *toP) { |
329 | /* Avoid copying partial characters. */ |
330 | for (fromLim = *fromP + (toLim - *toP); fromLim > *fromP; fromLim--) |
331 | if (((unsigned char)fromLim[-1] & 0xc0) != 0x80) |
332 | break; |
333 | } |
334 | for (to = *toP, from = *fromP; from != fromLim; from++, to++) |
335 | *to = *from; |
336 | *fromP = from; |
337 | *toP = to; |
338 | } |
339 | |
340 | static void PTRCALL |
341 | utf8_toUtf16(const ENCODING *enc, |
342 | const char **fromP, const char *fromLim, |
343 | unsigned short **toP, const unsigned short *toLim) |
344 | { |
345 | unsigned short *to = *toP; |
346 | const char *from = *fromP; |
347 | while (from != fromLim && to != toLim) { |
348 | switch (((struct normal_encoding *)enc)->type[(unsigned char)*from]) { |
349 | case BT_LEAD2: |
350 | *to++ = (unsigned short)(((from[0] & 0x1f) << 6) | (from[1] & 0x3f)); |
351 | from += 2; |
352 | break; |
353 | case BT_LEAD3: |
354 | *to++ = (unsigned short)(((from[0] & 0xf) << 12) |
355 | | ((from[1] & 0x3f) << 6) | (from[2] & 0x3f)); |
356 | from += 3; |
357 | break; |
358 | case BT_LEAD4: |
359 | { |
360 | unsigned long n; |
361 | if (to + 1 == toLim) |
362 | goto after; |
363 | n = ((from[0] & 0x7) << 18) | ((from[1] & 0x3f) << 12) |
364 | | ((from[2] & 0x3f) << 6) | (from[3] & 0x3f); |
365 | n -= 0x10000; |
366 | to[0] = (unsigned short)((n >> 10) | 0xD800); |
367 | to[1] = (unsigned short)((n & 0x3FF) | 0xDC00); |
368 | to += 2; |
369 | from += 4; |
370 | } |
371 | break; |
372 | default: |
373 | *to++ = *from++; |
374 | break; |
375 | } |
376 | } |
377 | after: |
378 | *fromP = from; |
379 | *toP = to; |
380 | } |
381 | |
382 | #ifdef XML_NS |
383 | static const struct normal_encoding utf8_encoding_ns = { |
384 | { VTABLE1, utf8_toUtf8, utf8_toUtf16, 1, 1, 0 }, |
385 | { |
386 | #include "asciitab.h" |
387 | #include "utf8tab.h" |
388 | }, |
389 | STANDARD_VTABLE(sb_) NORMAL_VTABLE(utf8_) |
390 | }; |
391 | #endif |
392 | |
393 | static const struct normal_encoding utf8_encoding = { |
394 | { VTABLE1, utf8_toUtf8, utf8_toUtf16, 1, 1, 0 }, |
395 | { |
396 | #define BT_COLON BT_NMSTRT |
397 | #include "asciitab.h" |
398 | #undef BT_COLON |
399 | #include "utf8tab.h" |
400 | }, |
401 | STANDARD_VTABLE(sb_) NORMAL_VTABLE(utf8_) |
402 | }; |
403 | |
404 | #ifdef XML_NS |
405 | |
406 | static const struct normal_encoding internal_utf8_encoding_ns = { |
407 | { VTABLE1, utf8_toUtf8, utf8_toUtf16, 1, 1, 0 }, |
408 | { |
409 | #include "iasciitab.h" |
410 | #include "utf8tab.h" |
411 | }, |
412 | STANDARD_VTABLE(sb_) NORMAL_VTABLE(utf8_) |
413 | }; |
414 | |
415 | #endif |
416 | |
417 | static const struct normal_encoding internal_utf8_encoding = { |
418 | { VTABLE1, utf8_toUtf8, utf8_toUtf16, 1, 1, 0 }, |
419 | { |
420 | #define BT_COLON BT_NMSTRT |
421 | #include "iasciitab.h" |
422 | #undef BT_COLON |
423 | #include "utf8tab.h" |
424 | }, |
425 | STANDARD_VTABLE(sb_) NORMAL_VTABLE(utf8_) |
426 | }; |
427 | |
428 | static void PTRCALL |
429 | latin1_toUtf8(const ENCODING *enc, |
430 | const char **fromP, const char *fromLim, |
431 | char **toP, const char *toLim) |
432 | { |
433 | for (;;) { |
434 | unsigned char c; |
435 | if (*fromP == fromLim) |
436 | break; |
437 | c = (unsigned char)**fromP; |
438 | if (c & 0x80) { |
439 | if (toLim - *toP < 2) |
440 | break; |
441 | *(*toP)++ = (char)((c >> 6) | UTF8_cval2); |
442 | *(*toP)++ = (char)((c & 0x3f) | 0x80); |
443 | (*fromP)++; |
444 | } |
445 | else { |
446 | if (*toP == toLim) |
447 | break; |
448 | *(*toP)++ = *(*fromP)++; |
449 | } |
450 | } |
451 | } |
452 | |
453 | static void PTRCALL |
454 | latin1_toUtf16(const ENCODING *enc, |
455 | const char **fromP, const char *fromLim, |
456 | unsigned short **toP, const unsigned short *toLim) |
457 | { |
458 | while (*fromP != fromLim && *toP != toLim) |
459 | *(*toP)++ = (unsigned char)*(*fromP)++; |
460 | } |
461 | |
462 | #ifdef XML_NS |
463 | |
464 | static const struct normal_encoding latin1_encoding_ns = { |
465 | { VTABLE1, latin1_toUtf8, latin1_toUtf16, 1, 0, 0 }, |
466 | { |
467 | #include "asciitab.h" |
468 | #include "latin1tab.h" |
469 | }, |
470 | STANDARD_VTABLE(sb_) |
471 | }; |
472 | |
473 | #endif |
474 | |
475 | static const struct normal_encoding latin1_encoding = { |
476 | { VTABLE1, latin1_toUtf8, latin1_toUtf16, 1, 0, 0 }, |
477 | { |
478 | #define BT_COLON BT_NMSTRT |
479 | #include "asciitab.h" |
480 | #undef BT_COLON |
481 | #include "latin1tab.h" |
482 | }, |
483 | STANDARD_VTABLE(sb_) |
484 | }; |
485 | |
486 | static void PTRCALL |
487 | ascii_toUtf8(const ENCODING *enc, |
488 | const char **fromP, const char *fromLim, |
489 | char **toP, const char *toLim) |
490 | { |
491 | while (*fromP != fromLim && *toP != toLim) |
492 | *(*toP)++ = *(*fromP)++; |
493 | } |
494 | |
495 | #ifdef XML_NS |
496 | |
497 | static const struct normal_encoding ascii_encoding_ns = { |
498 | { VTABLE1, ascii_toUtf8, latin1_toUtf16, 1, 1, 0 }, |
499 | { |
500 | #include "asciitab.h" |
501 | /* BT_NONXML == 0 */ |
502 | }, |
503 | STANDARD_VTABLE(sb_) |
504 | }; |
505 | |
506 | #endif |
507 | |
508 | static const struct normal_encoding ascii_encoding = { |
509 | { VTABLE1, ascii_toUtf8, latin1_toUtf16, 1, 1, 0 }, |
510 | { |
511 | #define BT_COLON BT_NMSTRT |
512 | #include "asciitab.h" |
513 | #undef BT_COLON |
514 | /* BT_NONXML == 0 */ |
515 | }, |
516 | STANDARD_VTABLE(sb_) |
517 | }; |
518 | |
519 | static int PTRFASTCALL |
520 | unicode_byte_type(char hi, char lo) |
521 | { |
522 | switch ((unsigned char)hi) { |
523 | case 0xD8: case 0xD9: case 0xDA: case 0xDB: |
524 | return BT_LEAD4; |
525 | case 0xDC: case 0xDD: case 0xDE: case 0xDF: |
526 | return BT_TRAIL; |
527 | case 0xFF: |
528 | switch ((unsigned char)lo) { |
529 | case 0xFF: |
530 | case 0xFE: |
531 | return BT_NONXML; |
532 | } |
533 | break; |
534 | } |
535 | return BT_NONASCII; |
536 | } |
537 | |
538 | #define DEFINE_UTF16_TO_UTF8(E) \ |
539 | static void PTRCALL \ |
540 | E ## toUtf8(const ENCODING *enc, \ |
541 | const char **fromP, const char *fromLim, \ |
542 | char **toP, const char *toLim) \ |
543 | { \ |
544 | const char *from; \ |
545 | for (from = *fromP; from != fromLim; from += 2) { \ |
546 | int plane; \ |
547 | unsigned char lo2; \ |
548 | unsigned char lo = GET_LO(from); \ |
549 | unsigned char hi = GET_HI(from); \ |
550 | switch (hi) { \ |
551 | case 0: \ |
552 | if (lo < 0x80) { \ |
553 | if (*toP == toLim) { \ |
554 | *fromP = from; \ |
555 | return; \ |
556 | } \ |
557 | *(*toP)++ = lo; \ |
558 | break; \ |
559 | } \ |
560 | /* fall through */ \ |
561 | case 0x1: case 0x2: case 0x3: \ |
562 | case 0x4: case 0x5: case 0x6: case 0x7: \ |
563 | if (toLim - *toP < 2) { \ |
564 | *fromP = from; \ |
565 | return; \ |
566 | } \ |
567 | *(*toP)++ = ((lo >> 6) | (hi << 2) | UTF8_cval2); \ |
568 | *(*toP)++ = ((lo & 0x3f) | 0x80); \ |
569 | break; \ |
570 | default: \ |
571 | if (toLim - *toP < 3) { \ |
572 | *fromP = from; \ |
573 | return; \ |
574 | } \ |
575 | /* 16 bits divided 4, 6, 6 amongst 3 bytes */ \ |
576 | *(*toP)++ = ((hi >> 4) | UTF8_cval3); \ |
577 | *(*toP)++ = (((hi & 0xf) << 2) | (lo >> 6) | 0x80); \ |
578 | *(*toP)++ = ((lo & 0x3f) | 0x80); \ |
579 | break; \ |
580 | case 0xD8: case 0xD9: case 0xDA: case 0xDB: \ |
581 | if (toLim - *toP < 4) { \ |
582 | *fromP = from; \ |
583 | return; \ |
584 | } \ |
585 | plane = (((hi & 0x3) << 2) | ((lo >> 6) & 0x3)) + 1; \ |
586 | *(*toP)++ = ((plane >> 2) | UTF8_cval4); \ |
587 | *(*toP)++ = (((lo >> 2) & 0xF) | ((plane & 0x3) << 4) | 0x80); \ |
588 | from += 2; \ |
589 | lo2 = GET_LO(from); \ |
590 | *(*toP)++ = (((lo & 0x3) << 4) \ |
591 | | ((GET_HI(from) & 0x3) << 2) \ |
592 | | (lo2 >> 6) \ |
593 | | 0x80); \ |
594 | *(*toP)++ = ((lo2 & 0x3f) | 0x80); \ |
595 | break; \ |
596 | } \ |
597 | } \ |
598 | *fromP = from; \ |
599 | } |
600 | |
601 | #define DEFINE_UTF16_TO_UTF16(E) \ |
602 | static void PTRCALL \ |
603 | E ## toUtf16(const ENCODING *enc, \ |
604 | const char **fromP, const char *fromLim, \ |
605 | unsigned short **toP, const unsigned short *toLim) \ |
606 | { \ |
607 | /* Avoid copying first half only of surrogate */ \ |
608 | if (fromLim - *fromP > ((toLim - *toP) << 1) \ |
609 | && (GET_HI(fromLim - 2) & 0xF8) == 0xD8) \ |
610 | fromLim -= 2; \ |
611 | for (; *fromP != fromLim && *toP != toLim; *fromP += 2) \ |
612 | *(*toP)++ = (GET_HI(*fromP) << 8) | GET_LO(*fromP); \ |
613 | } |
614 | |
615 | #define SET2(ptr, ch) \ |
616 | (((ptr)[0] = ((ch) & 0xff)), ((ptr)[1] = ((ch) >> 8))) |
617 | #define GET_LO(ptr) ((unsigned char)(ptr)[0]) |
618 | #define GET_HI(ptr) ((unsigned char)(ptr)[1]) |
619 | |
620 | DEFINE_UTF16_TO_UTF8(little2_) |
621 | DEFINE_UTF16_TO_UTF16(little2_) |
622 | |
623 | #undef SET2 |
624 | #undef GET_LO |
625 | #undef GET_HI |
626 | |
627 | #define SET2(ptr, ch) \ |
628 | (((ptr)[0] = ((ch) >> 8)), ((ptr)[1] = ((ch) & 0xFF))) |
629 | #define GET_LO(ptr) ((unsigned char)(ptr)[1]) |
630 | #define GET_HI(ptr) ((unsigned char)(ptr)[0]) |
631 | |
632 | DEFINE_UTF16_TO_UTF8(big2_) |
633 | DEFINE_UTF16_TO_UTF16(big2_) |
634 | |
635 | #undef SET2 |
636 | #undef GET_LO |
637 | #undef GET_HI |
638 | |
639 | #define LITTLE2_BYTE_TYPE(enc, p) \ |
640 | ((p)[1] == 0 \ |
641 | ? ((struct normal_encoding *)(enc))->type[(unsigned char)*(p)] \ |
642 | : unicode_byte_type((p)[1], (p)[0])) |
643 | #define LITTLE2_BYTE_TO_ASCII(enc, p) ((p)[1] == 0 ? (p)[0] : -1) |
644 | #define LITTLE2_CHAR_MATCHES(enc, p, c) ((p)[1] == 0 && (p)[0] == c) |
645 | #define LITTLE2_IS_NAME_CHAR_MINBPC(enc, p) \ |
646 | UCS2_GET_NAMING(namePages, (unsigned char)p[1], (unsigned char)p[0]) |
647 | #define LITTLE2_IS_NMSTRT_CHAR_MINBPC(enc, p) \ |
648 | UCS2_GET_NAMING(nmstrtPages, (unsigned char)p[1], (unsigned char)p[0]) |
649 | |
650 | #ifdef XML_MIN_SIZE |
651 | |
652 | static int PTRFASTCALL |
653 | little2_byteType(const ENCODING *enc, const char *p) |
654 | { |
655 | return LITTLE2_BYTE_TYPE(enc, p); |
656 | } |
657 | |
658 | static int PTRFASTCALL |
659 | little2_byteToAscii(const ENCODING *enc, const char *p) |
660 | { |
661 | return LITTLE2_BYTE_TO_ASCII(enc, p); |
662 | } |
663 | |
664 | static int PTRCALL |
665 | little2_charMatches(const ENCODING *enc, const char *p, int c) |
666 | { |
667 | return LITTLE2_CHAR_MATCHES(enc, p, c); |
668 | } |
669 | |
670 | static int PTRFASTCALL |
671 | little2_isNameMin(const ENCODING *enc, const char *p) |
672 | { |
673 | return LITTLE2_IS_NAME_CHAR_MINBPC(enc, p); |
674 | } |
675 | |
676 | static int PTRFASTCALL |
677 | little2_isNmstrtMin(const ENCODING *enc, const char *p) |
678 | { |
679 | return LITTLE2_IS_NMSTRT_CHAR_MINBPC(enc, p); |
680 | } |
681 | |
682 | #undef VTABLE |
683 | #define VTABLE VTABLE1, little2_toUtf8, little2_toUtf16 |
684 | |
685 | #else /* not XML_MIN_SIZE */ |
686 | |
687 | #undef PREFIX |
688 | #define PREFIX(ident) little2_ ## ident |
689 | #define MINBPC(enc) 2 |
690 | /* CHAR_MATCHES is guaranteed to have MINBPC bytes available. */ |
691 | #define BYTE_TYPE(enc, p) LITTLE2_BYTE_TYPE(enc, p) |
692 | #define BYTE_TO_ASCII(enc, p) LITTLE2_BYTE_TO_ASCII(enc, p) |
693 | #define CHAR_MATCHES(enc, p, c) LITTLE2_CHAR_MATCHES(enc, p, c) |
694 | #define IS_NAME_CHAR(enc, p, n) 0 |
695 | #define IS_NAME_CHAR_MINBPC(enc, p) LITTLE2_IS_NAME_CHAR_MINBPC(enc, p) |
696 | #define IS_NMSTRT_CHAR(enc, p, n) (0) |
697 | #define IS_NMSTRT_CHAR_MINBPC(enc, p) LITTLE2_IS_NMSTRT_CHAR_MINBPC(enc, p) |
698 | |
699 | #define XML_TOK_IMPL_C |
700 | #include "xmltok_impl.c" |
701 | #undef XML_TOK_IMPL_C |
702 | |
703 | #undef MINBPC |
704 | #undef BYTE_TYPE |
705 | #undef BYTE_TO_ASCII |
706 | #undef CHAR_MATCHES |
707 | #undef IS_NAME_CHAR |
708 | #undef IS_NAME_CHAR_MINBPC |
709 | #undef IS_NMSTRT_CHAR |
710 | #undef IS_NMSTRT_CHAR_MINBPC |
711 | #undef IS_INVALID_CHAR |
712 | |
713 | #endif /* not XML_MIN_SIZE */ |
714 | |
715 | #ifdef XML_NS |
716 | |
717 | static const struct normal_encoding little2_encoding_ns = { |
718 | { VTABLE, 2, 0, |
719 | #if BYTEORDER == 1234 |
720 | 1 |
721 | #else |
722 | 0 |
723 | #endif |
724 | }, |
725 | { |
726 | #include "asciitab.h" |
727 | #include "latin1tab.h" |
728 | }, |
729 | STANDARD_VTABLE(little2_) |
730 | }; |
731 | |
732 | #endif |
733 | |
734 | static const struct normal_encoding little2_encoding = { |
735 | { VTABLE, 2, 0, |
736 | #if BYTEORDER == 1234 |
737 | 1 |
738 | #else |
739 | 0 |
740 | #endif |
741 | }, |
742 | { |
743 | #define BT_COLON BT_NMSTRT |
744 | #include "asciitab.h" |
745 | #undef BT_COLON |
746 | #include "latin1tab.h" |
747 | }, |
748 | STANDARD_VTABLE(little2_) |
749 | }; |
750 | |
751 | #if BYTEORDER != 4321 |
752 | |
753 | #ifdef XML_NS |
754 | |
755 | static const struct normal_encoding internal_little2_encoding_ns = { |
756 | { VTABLE, 2, 0, 1 }, |
757 | { |
758 | #include "iasciitab.h" |
759 | #include "latin1tab.h" |
760 | }, |
761 | STANDARD_VTABLE(little2_) |
762 | }; |
763 | |
764 | #endif |
765 | |
766 | static const struct normal_encoding internal_little2_encoding = { |
767 | { VTABLE, 2, 0, 1 }, |
768 | { |
769 | #define BT_COLON BT_NMSTRT |
770 | #include "iasciitab.h" |
771 | #undef BT_COLON |
772 | #include "latin1tab.h" |
773 | }, |
774 | STANDARD_VTABLE(little2_) |
775 | }; |
776 | |
777 | #endif |
778 | |
779 | |
780 | #define BIG2_BYTE_TYPE(enc, p) \ |
781 | ((p)[0] == 0 \ |
782 | ? ((struct normal_encoding *)(enc))->type[(unsigned char)(p)[1]] \ |
783 | : unicode_byte_type((p)[0], (p)[1])) |
784 | #define BIG2_BYTE_TO_ASCII(enc, p) ((p)[0] == 0 ? (p)[1] : -1) |
785 | #define BIG2_CHAR_MATCHES(enc, p, c) ((p)[0] == 0 && (p)[1] == c) |
786 | #define BIG2_IS_NAME_CHAR_MINBPC(enc, p) \ |
787 | UCS2_GET_NAMING(namePages, (unsigned char)p[0], (unsigned char)p[1]) |
788 | #define BIG2_IS_NMSTRT_CHAR_MINBPC(enc, p) \ |
789 | UCS2_GET_NAMING(nmstrtPages, (unsigned char)p[0], (unsigned char)p[1]) |
790 | |
791 | #ifdef XML_MIN_SIZE |
792 | |
793 | static int PTRFASTCALL |
794 | big2_byteType(const ENCODING *enc, const char *p) |
795 | { |
796 | return BIG2_BYTE_TYPE(enc, p); |
797 | } |
798 | |
799 | static int PTRFASTCALL |
800 | big2_byteToAscii(const ENCODING *enc, const char *p) |
801 | { |
802 | return BIG2_BYTE_TO_ASCII(enc, p); |
803 | } |
804 | |
805 | static int PTRCALL |
806 | big2_charMatches(const ENCODING *enc, const char *p, int c) |
807 | { |
808 | return BIG2_CHAR_MATCHES(enc, p, c); |
809 | } |
810 | |
811 | static int PTRFASTCALL |
812 | big2_isNameMin(const ENCODING *enc, const char *p) |
813 | { |
814 | return BIG2_IS_NAME_CHAR_MINBPC(enc, p); |
815 | } |
816 | |
817 | static int PTRFASTCALL |
818 | big2_isNmstrtMin(const ENCODING *enc, const char *p) |
819 | { |
820 | return BIG2_IS_NMSTRT_CHAR_MINBPC(enc, p); |
821 | } |
822 | |
823 | #undef VTABLE |
824 | #define VTABLE VTABLE1, big2_toUtf8, big2_toUtf16 |
825 | |
826 | #else /* not XML_MIN_SIZE */ |
827 | |
828 | #undef PREFIX |
829 | #define PREFIX(ident) big2_ ## ident |
830 | #define MINBPC(enc) 2 |
831 | /* CHAR_MATCHES is guaranteed to have MINBPC bytes available. */ |
832 | #define BYTE_TYPE(enc, p) BIG2_BYTE_TYPE(enc, p) |
833 | #define BYTE_TO_ASCII(enc, p) BIG2_BYTE_TO_ASCII(enc, p) |
834 | #define CHAR_MATCHES(enc, p, c) BIG2_CHAR_MATCHES(enc, p, c) |
835 | #define IS_NAME_CHAR(enc, p, n) 0 |
836 | #define IS_NAME_CHAR_MINBPC(enc, p) BIG2_IS_NAME_CHAR_MINBPC(enc, p) |
837 | #define IS_NMSTRT_CHAR(enc, p, n) (0) |
838 | #define IS_NMSTRT_CHAR_MINBPC(enc, p) BIG2_IS_NMSTRT_CHAR_MINBPC(enc, p) |
839 | |
840 | #define XML_TOK_IMPL_C |
841 | #include "xmltok_impl.c" |
842 | #undef XML_TOK_IMPL_C |
843 | |
844 | #undef MINBPC |
845 | #undef BYTE_TYPE |
846 | #undef BYTE_TO_ASCII |
847 | #undef CHAR_MATCHES |
848 | #undef IS_NAME_CHAR |
849 | #undef IS_NAME_CHAR_MINBPC |
850 | #undef IS_NMSTRT_CHAR |
851 | #undef IS_NMSTRT_CHAR_MINBPC |
852 | #undef IS_INVALID_CHAR |
853 | |
854 | #endif /* not XML_MIN_SIZE */ |
855 | |
856 | #ifdef XML_NS |
857 | |
858 | static const struct normal_encoding big2_encoding_ns = { |
859 | { VTABLE, 2, 0, |
860 | #if BYTEORDER == 4321 |
861 | 1 |
862 | #else |
863 | 0 |
864 | #endif |
865 | }, |
866 | { |
867 | #include "asciitab.h" |
868 | #include "latin1tab.h" |
869 | }, |
870 | STANDARD_VTABLE(big2_) |
871 | }; |
872 | |
873 | #endif |
874 | |
875 | static const struct normal_encoding big2_encoding = { |
876 | { VTABLE, 2, 0, |
877 | #if BYTEORDER == 4321 |
878 | 1 |
879 | #else |
880 | 0 |
881 | #endif |
882 | }, |
883 | { |
884 | #define BT_COLON BT_NMSTRT |
885 | #include "asciitab.h" |
886 | #undef BT_COLON |
887 | #include "latin1tab.h" |
888 | }, |
889 | STANDARD_VTABLE(big2_) |
890 | }; |
891 | |
892 | #if BYTEORDER != 1234 |
893 | |
894 | #ifdef XML_NS |
895 | |
896 | static const struct normal_encoding internal_big2_encoding_ns = { |
897 | { VTABLE, 2, 0, 1 }, |
898 | { |
899 | #include "iasciitab.h" |
900 | #include "latin1tab.h" |
901 | }, |
902 | STANDARD_VTABLE(big2_) |
903 | }; |
904 | |
905 | #endif |
906 | |
907 | static const struct normal_encoding internal_big2_encoding = { |
908 | { VTABLE, 2, 0, 1 }, |
909 | { |
910 | #define BT_COLON BT_NMSTRT |
911 | #include "iasciitab.h" |
912 | #undef BT_COLON |
913 | #include "latin1tab.h" |
914 | }, |
915 | STANDARD_VTABLE(big2_) |
916 | }; |
917 | |
918 | #endif |
919 | |
920 | #undef PREFIX |
921 | |
922 | static int FASTCALL |
923 | streqci(const char *s1, const char *s2) |
924 | { |
925 | for (;;) { |
926 | char c1 = *s1++; |
927 | char c2 = *s2++; |
928 | if (ASCII_a <= c1 && c1 <= ASCII_z) |
929 | c1 += ASCII_A - ASCII_a; |
930 | if (ASCII_a <= c2 && c2 <= ASCII_z) |
931 | c2 += ASCII_A - ASCII_a; |
932 | if (c1 != c2) |
933 | return 0; |
934 | if (!c1) |
935 | break; |
936 | } |
937 | return 1; |
938 | } |
939 | |
940 | static void PTRCALL |
941 | initUpdatePosition(const ENCODING *enc, const char *ptr, |
942 | const char *end, POSITION *pos) |
943 | { |
944 | normal_updatePosition(&utf8_encoding.enc, ptr, end, pos); |
945 | } |
946 | |
947 | static int |
948 | toAscii(const ENCODING *enc, const char *ptr, const char *end) |
949 | { |
950 | char buf[1]; |
951 | char *p = buf; |
952 | XmlUtf8Convert(enc, &ptr, end, &p, p + 1); |
953 | if (p == buf) |
954 | return -1; |
955 | else |
956 | return buf[0]; |
957 | } |
958 | |
959 | static int FASTCALL |
960 | isSpace(int c) |
961 | { |
962 | switch (c) { |
963 | case 0x20: |
964 | case 0xD: |
965 | case 0xA: |
966 | case 0x9: |
967 | return 1; |
968 | } |
969 | return 0; |
970 | } |
971 | |
972 | /* Return 1 if there's just optional white space or there's an S |
973 | followed by name=val. |
974 | */ |
975 | static int |
976 | parsePseudoAttribute(const ENCODING *enc, |
977 | const char *ptr, |
978 | const char *end, |
979 | const char **namePtr, |
980 | const char **nameEndPtr, |
981 | const char **valPtr, |
982 | const char **nextTokPtr) |
983 | { |
984 | int c; |
985 | char open; |
986 | if (ptr == end) { |
987 | *namePtr = NULL; |
988 | return 1; |
989 | } |
990 | if (!isSpace(toAscii(enc, ptr, end))) { |
991 | *nextTokPtr = ptr; |
992 | return 0; |
993 | } |
994 | do { |
995 | ptr += enc->minBytesPerChar; |
996 | } while (isSpace(toAscii(enc, ptr, end))); |
997 | if (ptr == end) { |
998 | *namePtr = NULL; |
999 | return 1; |
1000 | } |
1001 | *namePtr = ptr; |
1002 | for (;;) { |
1003 | c = toAscii(enc, ptr, end); |
1004 | if (c == -1) { |
1005 | *nextTokPtr = ptr; |
1006 | return 0; |
1007 | } |
1008 | if (c == ASCII_EQUALS) { |
1009 | *nameEndPtr = ptr; |
1010 | break; |
1011 | } |
1012 | if (isSpace(c)) { |
1013 | *nameEndPtr = ptr; |
1014 | do { |
1015 | ptr += enc->minBytesPerChar; |
1016 | } while (isSpace(c = toAscii(enc, ptr, end))); |
1017 | if (c != ASCII_EQUALS) { |
1018 | *nextTokPtr = ptr; |
1019 | return 0; |
1020 | } |
1021 | break; |
1022 | } |
1023 | ptr += enc->minBytesPerChar; |
1024 | } |
1025 | if (ptr == *namePtr) { |
1026 | *nextTokPtr = ptr; |
1027 | return 0; |
1028 | } |
1029 | ptr += enc->minBytesPerChar; |
1030 | c = toAscii(enc, ptr, end); |
1031 | while (isSpace(c)) { |
1032 | ptr += enc->minBytesPerChar; |
1033 | c = toAscii(enc, ptr, end); |
1034 | } |
1035 | if (c != ASCII_QUOT && c != ASCII_APOS) { |
1036 | *nextTokPtr = ptr; |
1037 | return 0; |
1038 | } |
1039 | open = (char)c; |
1040 | ptr += enc->minBytesPerChar; |
1041 | *valPtr = ptr; |
1042 | for (;; ptr += enc->minBytesPerChar) { |
1043 | c = toAscii(enc, ptr, end); |
1044 | if (c == open) |
1045 | break; |
1046 | if (!(ASCII_a <= c && c <= ASCII_z) |
1047 | && !(ASCII_A <= c && c <= ASCII_Z) |
1048 | && !(ASCII_0 <= c && c <= ASCII_9) |
1049 | && c != ASCII_PERIOD |
1050 | && c != ASCII_MINUS |
1051 | && c != ASCII_UNDERSCORE) { |
1052 | *nextTokPtr = ptr; |
1053 | return 0; |
1054 | } |
1055 | } |
1056 | *nextTokPtr = ptr + enc->minBytesPerChar; |
1057 | return 1; |
1058 | } |
1059 | |
1060 | static const char KW_version[] = { |
1061 | ASCII_v, ASCII_e, ASCII_r, ASCII_s, ASCII_i, ASCII_o, ASCII_n, '\0' |
1062 | }; |
1063 | |
1064 | static const char KW_encoding[] = { |
1065 | ASCII_e, ASCII_n, ASCII_c, ASCII_o, ASCII_d, ASCII_i, ASCII_n, ASCII_g, '\0' |
1066 | }; |
1067 | |
1068 | static const char KW_standalone[] = { |
1069 | ASCII_s, ASCII_t, ASCII_a, ASCII_n, ASCII_d, ASCII_a, ASCII_l, ASCII_o, |
1070 | ASCII_n, ASCII_e, '\0' |
1071 | }; |
1072 | |
1073 | static const char KW_yes[] = { |
1074 | ASCII_y, ASCII_e, ASCII_s, '\0' |
1075 | }; |
1076 | |
1077 | static const char KW_no[] = { |
1078 | ASCII_n, ASCII_o, '\0' |
1079 | }; |
1080 | |
1081 | static int |
1082 | doParseXmlDecl(const ENCODING *(*encodingFinder)(const ENCODING *, |
1083 | const char *, |
1084 | const char *), |
1085 | int isGeneralTextEntity, |
1086 | const ENCODING *enc, |
1087 | const char *ptr, |
1088 | const char *end, |
1089 | const char **badPtr, |
1090 | const char **versionPtr, |
1091 | const char **versionEndPtr, |
1092 | const char **encodingName, |
1093 | const ENCODING **encoding, |
1094 | int *standalone) |
1095 | { |
1096 | const char *val = NULL; |
1097 | const char *name = NULL; |
1098 | const char *nameEnd = NULL; |
1099 | ptr += 5 * enc->minBytesPerChar; |
1100 | end -= 2 * enc->minBytesPerChar; |
1101 | if (!parsePseudoAttribute(enc, ptr, end, &name, &nameEnd, &val, &ptr) |
1102 | || !name) { |
1103 | *badPtr = ptr; |
1104 | return 0; |
1105 | } |
1106 | if (!XmlNameMatchesAscii(enc, name, nameEnd, KW_version)) { |
1107 | if (!isGeneralTextEntity) { |
1108 | *badPtr = name; |
1109 | return 0; |
1110 | } |
1111 | } |
1112 | else { |
1113 | if (versionPtr) |
1114 | *versionPtr = val; |
1115 | if (versionEndPtr) |
1116 | *versionEndPtr = ptr; |
1117 | if (!parsePseudoAttribute(enc, ptr, end, &name, &nameEnd, &val, &ptr)) { |
1118 | *badPtr = ptr; |
1119 | return 0; |
1120 | } |
1121 | if (!name) { |
1122 | if (isGeneralTextEntity) { |
1123 | /* a TextDecl must have an EncodingDecl */ |
1124 | *badPtr = ptr; |
1125 | return 0; |
1126 | } |
1127 | return 1; |
1128 | } |
1129 | } |
1130 | if (XmlNameMatchesAscii(enc, name, nameEnd, KW_encoding)) { |
1131 | int c = toAscii(enc, val, end); |
1132 | if (!(ASCII_a <= c && c <= ASCII_z) && !(ASCII_A <= c && c <= ASCII_Z)) { |
1133 | *badPtr = val; |
1134 | return 0; |
1135 | } |
1136 | if (encodingName) |
1137 | *encodingName = val; |
1138 | if (encoding) |
1139 | *encoding = encodingFinder(enc, val, ptr - enc->minBytesPerChar); |
1140 | if (!parsePseudoAttribute(enc, ptr, end, &name, &nameEnd, &val, &ptr)) { |
1141 | *badPtr = ptr; |
1142 | return 0; |
1143 | } |
1144 | if (!name) |
1145 | return 1; |
1146 | } |
1147 | if (!XmlNameMatchesAscii(enc, name, nameEnd, KW_standalone) |
1148 | || isGeneralTextEntity) { |
1149 | *badPtr = name; |
1150 | return 0; |
1151 | } |
1152 | if (XmlNameMatchesAscii(enc, val, ptr - enc->minBytesPerChar, KW_yes)) { |
1153 | if (standalone) |
1154 | *standalone = 1; |
1155 | } |
1156 | else if (XmlNameMatchesAscii(enc, val, ptr - enc->minBytesPerChar, KW_no)) { |
1157 | if (standalone) |
1158 | *standalone = 0; |
1159 | } |
1160 | else { |
1161 | *badPtr = val; |
1162 | return 0; |
1163 | } |
1164 | while (isSpace(toAscii(enc, ptr, end))) |
1165 | ptr += enc->minBytesPerChar; |
1166 | if (ptr != end) { |
1167 | *badPtr = ptr; |
1168 | return 0; |
1169 | } |
1170 | return 1; |
1171 | } |
1172 | |
1173 | static int FASTCALL |
1174 | checkCharRefNumber(int result) |
1175 | { |
1176 | switch (result >> 8) { |
1177 | case 0xD8: case 0xD9: case 0xDA: case 0xDB: |
1178 | case 0xDC: case 0xDD: case 0xDE: case 0xDF: |
1179 | return -1; |
1180 | case 0: |
1181 | if (latin1_encoding.type[result] == BT_NONXML) |
1182 | return -1; |
1183 | break; |
1184 | case 0xFF: |
1185 | if (result == 0xFFFE || result == 0xFFFF) |
1186 | return -1; |
1187 | break; |
1188 | } |
1189 | return result; |
1190 | } |
1191 | |
1192 | int FASTCALL |
1193 | XmlUtf8Encode(int c, char *buf) |
1194 | { |
1195 | enum { |
1196 | /* minN is minimum legal resulting value for N byte sequence */ |
1197 | min2 = 0x80, |
1198 | min3 = 0x800, |
1199 | min4 = 0x10000 |
1200 | }; |
1201 | |
1202 | if (c < 0) |
1203 | return 0; |
1204 | if (c < min2) { |
1205 | buf[0] = (char)(c | UTF8_cval1); |
1206 | return 1; |
1207 | } |
1208 | if (c < min3) { |
1209 | buf[0] = (char)((c >> 6) | UTF8_cval2); |
1210 | buf[1] = (char)((c & 0x3f) | 0x80); |
1211 | return 2; |
1212 | } |
1213 | if (c < min4) { |
1214 | buf[0] = (char)((c >> 12) | UTF8_cval3); |
1215 | buf[1] = (char)(((c >> 6) & 0x3f) | 0x80); |
1216 | buf[2] = (char)((c & 0x3f) | 0x80); |
1217 | return 3; |
1218 | } |
1219 | if (c < 0x110000) { |
1220 | buf[0] = (char)((c >> 18) | UTF8_cval4); |
1221 | buf[1] = (char)(((c >> 12) & 0x3f) | 0x80); |
1222 | buf[2] = (char)(((c >> 6) & 0x3f) | 0x80); |
1223 | buf[3] = (char)((c & 0x3f) | 0x80); |
1224 | return 4; |
1225 | } |
1226 | return 0; |
1227 | } |
1228 | |
1229 | int FASTCALL |
1230 | XmlUtf16Encode(int charNum, unsigned short *buf) |
1231 | { |
1232 | if (charNum < 0) |
1233 | return 0; |
1234 | if (charNum < 0x10000) { |
1235 | buf[0] = (unsigned short)charNum; |
1236 | return 1; |
1237 | } |
1238 | if (charNum < 0x110000) { |
1239 | charNum -= 0x10000; |
1240 | buf[0] = (unsigned short)((charNum >> 10) + 0xD800); |
1241 | buf[1] = (unsigned short)((charNum & 0x3FF) + 0xDC00); |
1242 | return 2; |
1243 | } |
1244 | return 0; |
1245 | } |
1246 | |
1247 | struct unknown_encoding { |
1248 | struct normal_encoding normal; |
1249 | CONVERTER convert; |
1250 | void *userData; |
1251 | unsigned short utf16[256]; |
1252 | char utf8[256][4]; |
1253 | }; |
1254 | |
1255 | #define AS_UNKNOWN_ENCODING(enc) ((const struct unknown_encoding *) (enc)) |
1256 | |
1257 | int |
1258 | XmlSizeOfUnknownEncoding(void) |
1259 | { |
1260 | return sizeof(struct unknown_encoding); |
1261 | } |
1262 | |
1263 | static int PTRFASTCALL |
1264 | unknown_isName(const ENCODING *enc, const char *p) |
1265 | { |
1266 | const struct unknown_encoding *uenc = AS_UNKNOWN_ENCODING(enc); |
1267 | int c = uenc->convert(uenc->userData, p); |
1268 | if (c & ~0xFFFF) |
1269 | return 0; |
1270 | return UCS2_GET_NAMING(namePages, c >> 8, c & 0xFF); |
1271 | } |
1272 | |
1273 | static int PTRFASTCALL |
1274 | unknown_isNmstrt(const ENCODING *enc, const char *p) |
1275 | { |
1276 | const struct unknown_encoding *uenc = AS_UNKNOWN_ENCODING(enc); |
1277 | int c = uenc->convert(uenc->userData, p); |
1278 | if (c & ~0xFFFF) |
1279 | return 0; |
1280 | return UCS2_GET_NAMING(nmstrtPages, c >> 8, c & 0xFF); |
1281 | } |
1282 | |
1283 | static int PTRFASTCALL |
1284 | unknown_isInvalid(const ENCODING *enc, const char *p) |
1285 | { |
1286 | const struct unknown_encoding *uenc = AS_UNKNOWN_ENCODING(enc); |
1287 | int c = uenc->convert(uenc->userData, p); |
1288 | return (c & ~0xFFFF) || checkCharRefNumber(c) < 0; |
1289 | } |
1290 | |
1291 | static void PTRCALL |
1292 | unknown_toUtf8(const ENCODING *enc, |
1293 | const char **fromP, const char *fromLim, |
1294 | char **toP, const char *toLim) |
1295 | { |
1296 | const struct unknown_encoding *uenc = AS_UNKNOWN_ENCODING(enc); |
1297 | char buf[XML_UTF8_ENCODE_MAX]; |
1298 | for (;;) { |
1299 | const char *utf8; |
1300 | int n; |
1301 | if (*fromP == fromLim) |
1302 | break; |
1303 | utf8 = uenc->utf8[(unsigned char)**fromP]; |
1304 | n = *utf8++; |
1305 | if (n == 0) { |
1306 | int c = uenc->convert(uenc->userData, *fromP); |
1307 | n = XmlUtf8Encode(c, buf); |
1308 | if (n > toLim - *toP) |
1309 | break; |
1310 | utf8 = buf; |
1311 | *fromP += (AS_NORMAL_ENCODING(enc)->type[(unsigned char)**fromP] |
1312 | - (BT_LEAD2 - 2)); |
1313 | } |
1314 | else { |
1315 | if (n > toLim - *toP) |
1316 | break; |
1317 | (*fromP)++; |
1318 | } |
1319 | do { |
1320 | *(*toP)++ = *utf8++; |
1321 | } while (--n != 0); |
1322 | } |
1323 | } |
1324 | |
1325 | static void PTRCALL |
1326 | unknown_toUtf16(const ENCODING *enc, |
1327 | const char **fromP, const char *fromLim, |
1328 | unsigned short **toP, const unsigned short *toLim) |
1329 | { |
1330 | const struct unknown_encoding *uenc = AS_UNKNOWN_ENCODING(enc); |
1331 | while (*fromP != fromLim && *toP != toLim) { |
1332 | unsigned short c = uenc->utf16[(unsigned char)**fromP]; |
1333 | if (c == 0) { |
1334 | c = (unsigned short) |
1335 | uenc->convert(uenc->userData, *fromP); |
1336 | *fromP += (AS_NORMAL_ENCODING(enc)->type[(unsigned char)**fromP] |
1337 | - (BT_LEAD2 - 2)); |
1338 | } |
1339 | else |
1340 | (*fromP)++; |
1341 | *(*toP)++ = c; |
1342 | } |
1343 | } |
1344 | |
1345 | ENCODING * |
1346 | XmlInitUnknownEncoding(void *mem, |
1347 | int *table, |
1348 | CONVERTER convert, |
1349 | void *userData) |
1350 | { |
1351 | int i; |
1352 | struct unknown_encoding *e = (struct unknown_encoding *)mem; |
1353 | for (i = 0; i < (int)sizeof(struct normal_encoding); i++) |
1354 | ((char *)mem)[i] = ((char *)&latin1_encoding)[i]; |
1355 | for (i = 0; i < 128; i++) |
1356 | if (latin1_encoding.type[i] != BT_OTHER |
1357 | && latin1_encoding.type[i] != BT_NONXML |
1358 | && table[i] != i) |
1359 | return 0; |
1360 | for (i = 0; i < 256; i++) { |
1361 | int c = table[i]; |
1362 | if (c == -1) { |
1363 | e->normal.type[i] = BT_MALFORM; |
1364 | /* This shouldn't really get used. */ |
1365 | e->utf16[i] = 0xFFFF; |
1366 | e->utf8[i][0] = 1; |
1367 | e->utf8[i][1] = 0; |
1368 | } |
1369 | else if (c < 0) { |
1370 | if (c < -4) |
1371 | return 0; |
1372 | e->normal.type[i] = (unsigned char)(BT_LEAD2 - (c + 2)); |
1373 | e->utf8[i][0] = 0; |
1374 | e->utf16[i] = 0; |
1375 | } |
1376 | else if (c < 0x80) { |
1377 | if (latin1_encoding.type[c] != BT_OTHER |
1378 | && latin1_encoding.type[c] != BT_NONXML |
1379 | && c != i) |
1380 | return 0; |
1381 | e->normal.type[i] = latin1_encoding.type[c]; |
1382 | e->utf8[i][0] = 1; |
1383 | e->utf8[i][1] = (char)c; |
1384 | e->utf16[i] = (unsigned short)(c == 0 ? 0xFFFF : c); |
1385 | } |
1386 | else if (checkCharRefNumber(c) < 0) { |
1387 | e->normal.type[i] = BT_NONXML; |
1388 | /* This shouldn't really get used. */ |
1389 | e->utf16[i] = 0xFFFF; |
1390 | e->utf8[i][0] = 1; |
1391 | e->utf8[i][1] = 0; |
1392 | } |
1393 | else { |
1394 | if (c > 0xFFFF) |
1395 | return 0; |
1396 | if (UCS2_GET_NAMING(nmstrtPages, c >> 8, c & 0xff)) |
1397 | e->normal.type[i] = BT_NMSTRT; |
1398 | else if (UCS2_GET_NAMING(namePages, c >> 8, c & 0xff)) |
1399 | e->normal.type[i] = BT_NAME; |
1400 | else |
1401 | e->normal.type[i] = BT_OTHER; |
1402 | e->utf8[i][0] = (char)XmlUtf8Encode(c, e->utf8[i] + 1); |
1403 | e->utf16[i] = (unsigned short)c; |
1404 | } |
1405 | } |
1406 | e->userData = userData; |
1407 | e->convert = convert; |
1408 | if (convert) { |
1409 | e->normal.isName2 = unknown_isName; |
1410 | e->normal.isName3 = unknown_isName; |
1411 | e->normal.isName4 = unknown_isName; |
1412 | e->normal.isNmstrt2 = unknown_isNmstrt; |
1413 | e->normal.isNmstrt3 = unknown_isNmstrt; |
1414 | e->normal.isNmstrt4 = unknown_isNmstrt; |
1415 | e->normal.isInvalid2 = unknown_isInvalid; |
1416 | e->normal.isInvalid3 = unknown_isInvalid; |
1417 | e->normal.isInvalid4 = unknown_isInvalid; |
1418 | } |
1419 | e->normal.enc.utf8Convert = unknown_toUtf8; |
1420 | e->normal.enc.utf16Convert = unknown_toUtf16; |
1421 | return &(e->normal.enc); |
1422 | } |
1423 | |
1424 | /* If this enumeration is changed, getEncodingIndex and encodings |
1425 | must also be changed. */ |
1426 | enum { |
1427 | UNKNOWN_ENC = -1, |
1428 | ISO_8859_1_ENC = 0, |
1429 | US_ASCII_ENC, |
1430 | UTF_8_ENC, |
1431 | UTF_16_ENC, |
1432 | UTF_16BE_ENC, |
1433 | UTF_16LE_ENC, |
1434 | /* must match encodingNames up to here */ |
1435 | NO_ENC |
1436 | }; |
1437 | |
1438 | static const char KW_ISO_8859_1[] = { |
1439 | ASCII_I, ASCII_S, ASCII_O, ASCII_MINUS, ASCII_8, ASCII_8, ASCII_5, ASCII_9, |
1440 | ASCII_MINUS, ASCII_1, '\0' |
1441 | }; |
1442 | static const char KW_US_ASCII[] = { |
1443 | ASCII_U, ASCII_S, ASCII_MINUS, ASCII_A, ASCII_S, ASCII_C, ASCII_I, ASCII_I, |
1444 | '\0' |
1445 | }; |
1446 | static const char KW_UTF_8[] = { |
1447 | ASCII_U, ASCII_T, ASCII_F, ASCII_MINUS, ASCII_8, '\0' |
1448 | }; |
1449 | static const char KW_UTF_16[] = { |
1450 | ASCII_U, ASCII_T, ASCII_F, ASCII_MINUS, ASCII_1, ASCII_6, '\0' |
1451 | }; |
1452 | static const char KW_UTF_16BE[] = { |
1453 | ASCII_U, ASCII_T, ASCII_F, ASCII_MINUS, ASCII_1, ASCII_6, ASCII_B, ASCII_E, |
1454 | '\0' |
1455 | }; |
1456 | static const char KW_UTF_16LE[] = { |
1457 | ASCII_U, ASCII_T, ASCII_F, ASCII_MINUS, ASCII_1, ASCII_6, ASCII_L, ASCII_E, |
1458 | '\0' |
1459 | }; |
1460 | |
1461 | static int FASTCALL |
1462 | getEncodingIndex(const char *name) |
1463 | { |
1464 | static const char * const encodingNames[] = { |
1465 | KW_ISO_8859_1, |
1466 | KW_US_ASCII, |
1467 | KW_UTF_8, |
1468 | KW_UTF_16, |
1469 | KW_UTF_16BE, |
1470 | KW_UTF_16LE, |
1471 | }; |
1472 | int i; |
1473 | if (name == NULL) |
1474 | return NO_ENC; |
1475 | for (i = 0; i < (int)(sizeof(encodingNames)/sizeof(encodingNames[0])); i++) |
1476 | if (streqci(name, encodingNames[i])) |
1477 | return i; |
1478 | return UNKNOWN_ENC; |
1479 | } |
1480 | |
1481 | /* For binary compatibility, we store the index of the encoding |
1482 | specified at initialization in the isUtf16 member. |
1483 | */ |
1484 | |
1485 | #define INIT_ENC_INDEX(enc) ((int)(enc)->initEnc.isUtf16) |
1486 | #define SET_INIT_ENC_INDEX(enc, i) ((enc)->initEnc.isUtf16 = (char)i) |
1487 | |
1488 | /* This is what detects the encoding. encodingTable maps from |
1489 | encoding indices to encodings; INIT_ENC_INDEX(enc) is the index of |
1490 | the external (protocol) specified encoding; state is |
1491 | XML_CONTENT_STATE if we're parsing an external text entity, and |
1492 | XML_PROLOG_STATE otherwise. |
1493 | */ |
1494 | |
1495 | |
1496 | static int |
1497 | initScan(const ENCODING * const *encodingTable, |
1498 | const INIT_ENCODING *enc, |
1499 | int state, |
1500 | const char *ptr, |
1501 | const char *end, |
1502 | const char **nextTokPtr) |
1503 | { |
1504 | const ENCODING **encPtr; |
1505 | |
1506 | if (ptr == end) |
1507 | return XML_TOK_NONE; |
1508 | encPtr = enc->encPtr; |
1509 | if (ptr + 1 == end) { |
1510 | /* only a single byte available for auto-detection */ |
1511 | #ifndef XML_DTD /* FIXME */ |
1512 | /* a well-formed document entity must have more than one byte */ |
1513 | if (state != XML_CONTENT_STATE) |
1514 | return XML_TOK_PARTIAL; |
1515 | #endif |
1516 | /* so we're parsing an external text entity... */ |
1517 | /* if UTF-16 was externally specified, then we need at least 2 bytes */ |
1518 | switch (INIT_ENC_INDEX(enc)) { |
1519 | case UTF_16_ENC: |
1520 | case UTF_16LE_ENC: |
1521 | case UTF_16BE_ENC: |
1522 | return XML_TOK_PARTIAL; |
1523 | } |
1524 | switch ((unsigned char)*ptr) { |
1525 | case 0xFE: |
1526 | case 0xFF: |
1527 | case 0xEF: /* possibly first byte of UTF-8 BOM */ |
1528 | if (INIT_ENC_INDEX(enc) == ISO_8859_1_ENC |
1529 | && state == XML_CONTENT_STATE) |
1530 | break; |
1531 | /* fall through */ |
1532 | case 0x00: |
1533 | case 0x3C: |
1534 | return XML_TOK_PARTIAL; |
1535 | } |
1536 | } |
1537 | else { |
1538 | switch (((unsigned char)ptr[0] << 8) | (unsigned char)ptr[1]) { |
1539 | case 0xFEFF: |
1540 | if (INIT_ENC_INDEX(enc) == ISO_8859_1_ENC |
1541 | && state == XML_CONTENT_STATE) |
1542 | break; |
1543 | *nextTokPtr = ptr + 2; |
1544 | *encPtr = encodingTable[UTF_16BE_ENC]; |
1545 | return XML_TOK_BOM; |
1546 | /* 00 3C is handled in the default case */ |
1547 | case 0x3C00: |
1548 | if ((INIT_ENC_INDEX(enc) == UTF_16BE_ENC |
1549 | || INIT_ENC_INDEX(enc) == UTF_16_ENC) |
1550 | && state == XML_CONTENT_STATE) |
1551 | break; |
1552 | *encPtr = encodingTable[UTF_16LE_ENC]; |
1553 | return XmlTok(*encPtr, state, ptr, end, nextTokPtr); |
1554 | case 0xFFFE: |
1555 | if (INIT_ENC_INDEX(enc) == ISO_8859_1_ENC |
1556 | && state == XML_CONTENT_STATE) |
1557 | break; |
1558 | *nextTokPtr = ptr + 2; |
1559 | *encPtr = encodingTable[UTF_16LE_ENC]; |
1560 | return XML_TOK_BOM; |
1561 | case 0xEFBB: |
1562 | /* Maybe a UTF-8 BOM (EF BB BF) */ |
1563 | /* If there's an explicitly specified (external) encoding |
1564 | of ISO-8859-1 or some flavour of UTF-16 |
1565 | and this is an external text entity, |
1566 | don't look for the BOM, |
1567 | because it might be a legal data. |
1568 | */ |
1569 | if (state == XML_CONTENT_STATE) { |
1570 | int e = INIT_ENC_INDEX(enc); |
1571 | if (e == ISO_8859_1_ENC || e == UTF_16BE_ENC |
1572 | || e == UTF_16LE_ENC || e == UTF_16_ENC) |
1573 | break; |
1574 | } |
1575 | if (ptr + 2 == end) |
1576 | return XML_TOK_PARTIAL; |
1577 | if ((unsigned char)ptr[2] == 0xBF) { |
1578 | *nextTokPtr = ptr + 3; |
1579 | *encPtr = encodingTable[UTF_8_ENC]; |
1580 | return XML_TOK_BOM; |
1581 | } |
1582 | break; |
1583 | default: |
1584 | if (ptr[0] == '\0') { |
1585 | /* 0 isn't a legal data character. Furthermore a document |
1586 | entity can only start with ASCII characters. So the only |
1587 | way this can fail to be big-endian UTF-16 if it it's an |
1588 | external parsed general entity that's labelled as |
1589 | UTF-16LE. |
1590 | */ |
1591 | if (state == XML_CONTENT_STATE && INIT_ENC_INDEX(enc) == UTF_16LE_ENC) |
1592 | break; |
1593 | *encPtr = encodingTable[UTF_16BE_ENC]; |
1594 | return XmlTok(*encPtr, state, ptr, end, nextTokPtr); |
1595 | } |
1596 | else if (ptr[1] == '\0') { |
1597 | /* We could recover here in the case: |
1598 | - parsing an external entity |
1599 | - second byte is 0 |
1600 | - no externally specified encoding |
1601 | - no encoding declaration |
1602 | by assuming UTF-16LE. But we don't, because this would mean when |
1603 | presented just with a single byte, we couldn't reliably determine |
1604 | whether we needed further bytes. |
1605 | */ |
1606 | if (state == XML_CONTENT_STATE) |
1607 | break; |
1608 | *encPtr = encodingTable[UTF_16LE_ENC]; |
1609 | return XmlTok(*encPtr, state, ptr, end, nextTokPtr); |
1610 | } |
1611 | break; |
1612 | } |
1613 | } |
1614 | *encPtr = encodingTable[INIT_ENC_INDEX(enc)]; |
1615 | return XmlTok(*encPtr, state, ptr, end, nextTokPtr); |
1616 | } |
1617 | |
1618 | |
1619 | #define NS(x) x |
1620 | #define ns(x) x |
1621 | #define XML_TOK_NS_C |
1622 | #include "xmltok_ns.c" |
1623 | #undef XML_TOK_NS_C |
1624 | #undef NS |
1625 | #undef ns |
1626 | |
1627 | #ifdef XML_NS |
1628 | |
1629 | #define NS(x) x ## NS |
1630 | #define ns(x) x ## _ns |
1631 | |
1632 | #define XML_TOK_NS_C |
1633 | #include "xmltok_ns.c" |
1634 | #undef XML_TOK_NS_C |
1635 | |
1636 | #undef NS |
1637 | #undef ns |
1638 | |
1639 | ENCODING * |
1640 | XmlInitUnknownEncodingNS(void *mem, |
1641 | int *table, |
1642 | CONVERTER convert, |
1643 | void *userData) |
1644 | { |
1645 | ENCODING *enc = XmlInitUnknownEncoding(mem, table, convert, userData); |
1646 | if (enc) |
1647 | ((struct normal_encoding *)enc)->type[ASCII_COLON] = BT_COLON; |
1648 | return enc; |
1649 | } |
1650 | |
1651 | #endif /* XML_NS */ |
1652 | |