source: trunk/src/helpers/xmltok.c@ 97

Last change on this file since 97 was 97, checked in by umoeller, 24 years ago

XML updates.
  • Property svn:eol-style set to CRLF
  • Property svn:keywords set to Author Date Id Revision
File size: 45.7 KB
Line 
1/*
2 * Copyright (c) 1998, 1999 Thai Open Source Software Center Ltd
3 * See the file COPYING for copying permission.
4 */
5
6/* #ifdef COMPILED_FROM_DSP
7 * # include "winconfig.h"
8 * #else
9 * # include <config.h>
10 * #endif
11 */
12
13#include <memory.h>
14
15#include "expat\expat_setup.h" // V0.9.9 (2001-02-10) [umoeller]
16
17#pragma info(norea, nogen)
18 // disable "statement unreachable" and "missing break statement"
19 // this code generates those options HEAVILY
20
21#include "expat\xmltok.h"
22#include "expat\nametab.h"
23
24#ifdef XML_DTD
25#define IGNORE_SECTION_TOK_VTABLE , PREFIX(ignoreSectionTok)
26#else
27#define IGNORE_SECTION_TOK_VTABLE /* as nothing */
28#endif
29
30#define VTABLE1 \
31 { PREFIX(prologTok), PREFIX(contentTok), \
32 PREFIX(cdataSectionTok) IGNORE_SECTION_TOK_VTABLE }, \
33 { PREFIX(attributeValueTok), PREFIX(entityValueTok) }, \
34 PREFIX(sameName), \
35 PREFIX(nameMatchesAscii), \
36 PREFIX(nameLength), \
37 PREFIX(skipS), \
38 PREFIX(getAtts), \
39 PREFIX(charRefNumber), \
40 PREFIX(predefinedEntityName), \
41 PREFIX(updatePosition), \
42 PREFIX(isPublicId)
43
44#define VTABLE VTABLE1, PREFIX(toUtf8), PREFIX(toUtf16)
45
46#define UCS2_GET_NAMING(pages, hi, lo) \
47 (namingBitmap[(pages[hi] << 3) + ((lo) >> 5)] & (1 << ((lo) & 0x1F)))
48
49/* A 2 byte UTF-8 representation splits the characters 11 bits
50 * between the bottom 5 and 6 bits of the bytes.
51 * We need 8 bits to index into pages, 3 bits to add to that index and
52 * 5 bits to generate the mask. */
53#define UTF8_GET_NAMING2(pages, byte) \
54 (namingBitmap[((pages)[(((byte)[0]) >> 2) & 7] << 3) \
55 + ((((byte)[0]) & 3) << 1) \
56 + ((((byte)[1]) >> 5) & 1)] \
57 & (1 << (((byte)[1]) & 0x1F)))
58
59/* A 3 byte UTF-8 representation splits the characters 16 bits
60 * between the bottom 4, 6 and 6 bits of the bytes.
61 * We need 8 bits to index into pages, 3 bits to add to that index and
62 * 5 bits to generate the mask. */
63#define UTF8_GET_NAMING3(pages, byte) \
64 (namingBitmap[((pages)[((((byte)[0]) & 0xF) << 4) \
65 + ((((byte)[1]) >> 2) & 0xF)] \
66 << 3) \
67 + ((((byte)[1]) & 3) << 1) \
68 + ((((byte)[2]) >> 5) & 1)] \
69 & (1 << (((byte)[2]) & 0x1F)))
70
71#define UTF8_GET_NAMING(pages, p, n) \
72 ((n) == 2 \
73 ? UTF8_GET_NAMING2(pages, (const unsigned char *)(p)) \
74 : ((n) == 3 \
75 ? UTF8_GET_NAMING3(pages, (const unsigned char *)(p)) \
76 : 0))
77
78#define UTF8_INVALID3(p) \
79 ((*p) == 0xED \
80 ? (((p)[1] & 0x20) != 0) \
81 : ((*p) == 0xEF \
82 ? ((p)[1] == 0xBF && ((p)[2] == 0xBF || (p)[2] == 0xBE)) \
83 : 0))
84
85#define UTF8_INVALID4(p) ((*p) == 0xF4 && ((p)[1] & 0x30) != 0)
86
87static
88int isNever(const ENCODING * enc, const char *p)
89{
90 return 0;
91}
92
93static
94int utf8_isName2(const ENCODING * enc, const char *p)
95{
96 return UTF8_GET_NAMING2(namePages, (const unsigned char *)p);
97}
98
99static
100int utf8_isName3(const ENCODING * enc, const char *p)
101{
102 return UTF8_GET_NAMING3(namePages, (const unsigned char *)p);
103}
104
105#define utf8_isName4 isNever
106
107static
108int utf8_isNmstrt2(const ENCODING * enc, const char *p)
109{
110 return UTF8_GET_NAMING2(nmstrtPages, (const unsigned char *)p);
111}
112
113static
114int utf8_isNmstrt3(const ENCODING * enc, const char *p)
115{
116 return UTF8_GET_NAMING3(nmstrtPages, (const unsigned char *)p);
117}
118
119#define utf8_isNmstrt4 isNever
120
121#define utf8_isInvalid2 isNever
122
123static
124int utf8_isInvalid3(const ENCODING * enc, const char *p)
125{
126 return UTF8_INVALID3((const unsigned char *)p);
127}
128
129static
130int utf8_isInvalid4(const ENCODING * enc, const char *p)
131{
132 return UTF8_INVALID4((const unsigned char *)p);
133}
134
135struct normal_encoding
136{
137 ENCODING enc;
138 unsigned char type[256];
139#ifdef XML_MIN_SIZE
140 int (*byteType) (const ENCODING *, const char *);
141 int (*isNameMin) (const ENCODING *, const char *);
142 int (*isNmstrtMin) (const ENCODING *, const char *);
143 int (*byteToAscii) (const ENCODING *, const char *);
144 int (*charMatches) (const ENCODING *, const char *, int);
145#endif /* XML_MIN_SIZE */
146 int (*isName2) (const ENCODING *, const char *);
147 int (*isName3) (const ENCODING *, const char *);
148 int (*isName4) (const ENCODING *, const char *);
149 int (*isNmstrt2) (const ENCODING *, const char *);
150 int (*isNmstrt3) (const ENCODING *, const char *);
151 int (*isNmstrt4) (const ENCODING *, const char *);
152 int (*isInvalid2) (const ENCODING *, const char *);
153 int (*isInvalid3) (const ENCODING *, const char *);
154 int (*isInvalid4) (const ENCODING *, const char *);
155};
156
157#ifdef XML_MIN_SIZE
158
159#define STANDARD_VTABLE(E) \
160 E ## byteType, \
161 E ## isNameMin, \
162 E ## isNmstrtMin, \
163 E ## byteToAscii, \
164 E ## charMatches,
165
166#else
167
168#define STANDARD_VTABLE(E) /* as nothing */
169
170#endif
171
172#define NORMAL_VTABLE(E) \
173 E ## isName2, \
174 E ## isName3, \
175 E ## isName4, \
176 E ## isNmstrt2, \
177 E ## isNmstrt3, \
178 E ## isNmstrt4, \
179 E ## isInvalid2, \
180 E ## isInvalid3, \
181 E ## isInvalid4
182
183static int checkCharRefNumber(int);
184
185#include "expat\xmltok_impl.h"
186#include "expat\ascii.h"
187
188#ifdef XML_MIN_SIZE
189#define sb_isNameMin isNever
190#define sb_isNmstrtMin isNever
191#endif
192
193#ifdef XML_MIN_SIZE
194#define MINBPC(enc) ((enc)->minBytesPerChar)
195#else
196/* minimum bytes per character */
197#define MINBPC(enc) 1
198#endif
199
200#define SB_BYTE_TYPE(enc, p) \
201 (((struct normal_encoding *)(enc))->type[(unsigned char)*(p)])
202
203#ifdef XML_MIN_SIZE
204static
205int sb_byteType(const ENCODING * enc, const char *p)
206{
207 return SB_BYTE_TYPE(enc, p);
208}
209#define BYTE_TYPE(enc, p) \
210 (((const struct normal_encoding *)(enc))->byteType(enc, p))
211#else
212#define BYTE_TYPE(enc, p) SB_BYTE_TYPE(enc, p)
213#endif
214
215#ifdef XML_MIN_SIZE
216#define BYTE_TO_ASCII(enc, p) \
217 (((const struct normal_encoding *)(enc))->byteToAscii(enc, p))
218static
219int sb_byteToAscii(const ENCODING * enc, const char *p)
220{
221 return *p;
222}
223#else
224#define BYTE_TO_ASCII(enc, p) (*(p))
225#endif
226
227#define IS_NAME_CHAR(enc, p, n) \
228 (((const struct normal_encoding *)(enc))->isName ## n(enc, p))
229#define IS_NMSTRT_CHAR(enc, p, n) \
230 (((const struct normal_encoding *)(enc))->isNmstrt ## n(enc, p))
231#define IS_INVALID_CHAR(enc, p, n) \
232 (((const struct normal_encoding *)(enc))->isInvalid ## n(enc, p))
233
234#ifdef XML_MIN_SIZE
235#define IS_NAME_CHAR_MINBPC(enc, p) \
236 (((const struct normal_encoding *)(enc))->isNameMin(enc, p))
237#define IS_NMSTRT_CHAR_MINBPC(enc, p) \
238 (((const struct normal_encoding *)(enc))->isNmstrtMin(enc, p))
239#else
240#define IS_NAME_CHAR_MINBPC(enc, p) (0)
241#define IS_NMSTRT_CHAR_MINBPC(enc, p) (0)
242#endif
243
244#ifdef XML_MIN_SIZE
245#define CHAR_MATCHES(enc, p, c) \
246 (((const struct normal_encoding *)(enc))->charMatches(enc, p, c))
247static
248int sb_charMatches(const ENCODING * enc, const char *p, int c)
249{
250 return *p == c;
251}
252#else
253/* c is an ASCII character */
254#define CHAR_MATCHES(enc, p, c) (*(p) == c)
255#endif
256
257#define PREFIX(ident) normal_ ## ident
258#include "xmltok_impl.c"
259
260#undef MINBPC
261#undef BYTE_TYPE
262#undef BYTE_TO_ASCII
263#undef CHAR_MATCHES
264#undef IS_NAME_CHAR
265#undef IS_NAME_CHAR_MINBPC
266#undef IS_NMSTRT_CHAR
267#undef IS_NMSTRT_CHAR_MINBPC
268#undef IS_INVALID_CHAR
269
270enum
271{ /* UTF8_cvalN is value of masked first byte of N byte sequence */
272 UTF8_cval1 = 0x00,
273 UTF8_cval2 = 0xc0,
274 UTF8_cval3 = 0xe0,
275 UTF8_cval4 = 0xf0
276};
277
278static void utf8_toUtf8(const ENCODING * enc,
279 const char **fromP,
280 const char *fromLim,
281 char **toP,
282 const char *toLim)
283{
284 char *to;
285 const char *from;
286
287 if (fromLim - *fromP > toLim - *toP)
288 {
289 /* Avoid copying partial characters. */
290 for (fromLim = *fromP + (toLim - *toP);
291 fromLim > *fromP;
292 fromLim--)
293 if (((unsigned char)fromLim[-1] & 0xc0) != 0x80)
294 break;
295 }
296 for (to = *toP, from = *fromP;
297 from != fromLim;
298 from++, to++)
299 *to = *from;
300 *fromP = from;
301 *toP = to;
302}
303
304static void utf8_toUtf16(const ENCODING * enc,
305 const char **fromP,
306 const char *fromLim,
307 unsigned short **toP,
308 const unsigned short *toLim)
309{
310 unsigned short *to = *toP;
311 const char *from = *fromP;
312
313 while (from != fromLim && to != toLim)
314 {
315 switch (((struct normal_encoding *)enc)->type[(unsigned char)*from])
316 {
317 case BT_LEAD2:
318 *to++ = ((from[0] & 0x1f) << 6) | (from[1] & 0x3f);
319 from += 2;
320 break;
321 case BT_LEAD3:
322 *to++ = ((from[0] & 0xf) << 12) | ((from[1] & 0x3f) << 6) | (from[2] & 0x3f);
323 from += 3;
324 break;
325 case BT_LEAD4:
326 {
327 unsigned long n;
328
329 if (to + 1 == toLim)
330 break;
331 n = ((from[0] & 0x7) << 18) | ((from[1] & 0x3f) << 12) | ((from[2] & 0x3f) << 6) | (from[3] & 0x3f);
332 n -= 0x10000;
333 to[0] = (unsigned short)((n >> 10) | 0xD800);
334 to[1] = (unsigned short)((n & 0x3FF) | 0xDC00);
335 to += 2;
336 from += 4;
337 }
338 break;
339 default:
340 *to++ = *from++;
341 break;
342 }
343 }
344 *fromP = from;
345 *toP = to;
346}
347
348#ifdef XML_NS
349static const struct normal_encoding utf8_encoding_ns =
350{
351 {VTABLE1, utf8_toUtf8, utf8_toUtf16, 1, 1, 0},
352 {
353#include "expat\asciitab.h"
354#include "expat\utf8tab.h"
355 },
356 STANDARD_VTABLE(sb_) NORMAL_VTABLE(utf8_)
357};
358
359#endif
360
361static const struct normal_encoding utf8_encoding =
362{
363 {VTABLE1, utf8_toUtf8, utf8_toUtf16, 1, 1, 0},
364 {
365#define BT_COLON BT_NMSTRT
366#include "expat\asciitab.h"
367#undef BT_COLON
368#include "expat\utf8tab.h"
369 },
370 STANDARD_VTABLE(sb_) NORMAL_VTABLE(utf8_)
371};
372
373#ifdef XML_NS
374
375static const struct normal_encoding internal_utf8_encoding_ns =
376{
377 {VTABLE1, utf8_toUtf8, utf8_toUtf16, 1, 1, 0},
378 {
379#include "expat\iasciitab.h"
380#include "expat\utf8tab.h"
381 },
382 STANDARD_VTABLE(sb_) NORMAL_VTABLE(utf8_)
383};
384
385#endif
386
387static const struct normal_encoding internal_utf8_encoding =
388{
389 {VTABLE1, utf8_toUtf8, utf8_toUtf16, 1, 1, 0},
390 {
391#define BT_COLON BT_NMSTRT
392#include "expat\iasciitab.h"
393#undef BT_COLON
394#include "expat\utf8tab.h"
395 },
396 STANDARD_VTABLE(sb_) NORMAL_VTABLE(utf8_)
397};
398
399static void latin1_toUtf8(const ENCODING * enc,
400 const char **fromP,
401 const char *fromLim,
402 char **toP,
403 const char *toLim)
404{
405 for (;;)
406 {
407 unsigned char c;
408
409 if (*fromP == fromLim)
410 break;
411 c = (unsigned char)**fromP;
412 if (c & 0x80)
413 {
414 if (toLim - *toP < 2)
415 break;
416 *(*toP)++ = ((c >> 6) | UTF8_cval2);
417 *(*toP)++ = ((c & 0x3f) | 0x80);
418 (*fromP)++;
419 }
420 else
421 {
422 if (*toP == toLim)
423 break;
424 *(*toP)++ = *(*fromP)++;
425 }
426 }
427}
428
429static void latin1_toUtf16(const ENCODING * enc,
430 const char **fromP,
431 const char *fromLim,
432 unsigned short **toP,
433 const unsigned short *toLim)
434{
435 while (*fromP != fromLim && *toP != toLim)
436 *(*toP)++ = (unsigned char)*(*fromP)++;
437}
438
439#ifdef XML_NS
440
441static const struct normal_encoding latin1_encoding_ns =
442{
443 {VTABLE1, latin1_toUtf8, latin1_toUtf16, 1, 0, 0},
444 {
445#include "expat\asciitab.h"
446#include "expat\latin1tab.h"
447 },
448 STANDARD_VTABLE(sb_)
449};
450
451#endif
452
453static const struct normal_encoding latin1_encoding =
454{
455 {VTABLE1, latin1_toUtf8, latin1_toUtf16, 1, 0, 0},
456 {
457#define BT_COLON BT_NMSTRT
458#include "expat\asciitab.h"
459#undef BT_COLON
460#include "expat\latin1tab.h"
461 },
462 STANDARD_VTABLE(sb_)
463};
464
465static void ascii_toUtf8(const ENCODING * enc,
466 const char **fromP,
467 const char *fromLim,
468 char **toP,
469 const char *toLim)
470{
471 while (*fromP != fromLim && *toP != toLim)
472 *(*toP)++ = *(*fromP)++;
473}
474
475#ifdef XML_NS
476
477static const struct normal_encoding ascii_encoding_ns =
478{
479 {VTABLE1, ascii_toUtf8, latin1_toUtf16, 1, 1, 0},
480 {
481#include "expat\asciitab.h"
482/* BT_NONXML == 0 */
483 },
484 STANDARD_VTABLE(sb_)
485};
486
487#endif
488
489static const struct normal_encoding ascii_encoding =
490{
491 {VTABLE1, ascii_toUtf8, latin1_toUtf16, 1, 1, 0},
492 {
493#define BT_COLON BT_NMSTRT
494#include "expat\asciitab.h"
495#undef BT_COLON
496/* BT_NONXML == 0 */
497 },
498 STANDARD_VTABLE(sb_)
499};
500
501static int unicode_byte_type(char hi, char lo)
502{
503 switch ((unsigned char)hi)
504 {
505 case 0xD8:
506 case 0xD9:
507 case 0xDA:
508 case 0xDB:
509 return BT_LEAD4;
510 case 0xDC:
511 case 0xDD:
512 case 0xDE:
513 case 0xDF:
514 return BT_TRAIL;
515 case 0xFF:
516 switch ((unsigned char)lo)
517 {
518 case 0xFF:
519 case 0xFE:
520 return BT_NONXML;
521 }
522 break;
523 }
524 return BT_NONASCII;
525}
526
527#define DEFINE_UTF16_TO_UTF8(E) \
528static \
529void E ## toUtf8(const ENCODING *enc, \
530 const char **fromP, const char *fromLim, \
531 char **toP, const char *toLim) \
532{ \
533 const char *from; \
534 for (from = *fromP; from != fromLim; from += 2) { \
535 int plane; \
536 unsigned char lo2; \
537 unsigned char lo = GET_LO(from); \
538 unsigned char hi = GET_HI(from); \
539 switch (hi) { \
540 case 0: \
541 if (lo < 0x80) { \
542 if (*toP == toLim) { \
543 *fromP = from; \
544 return; \
545 } \
546 *(*toP)++ = lo; \
547 break; \
548 } \
549 /* fall through */ \
550 case 0x1: case 0x2: case 0x3: \
551 case 0x4: case 0x5: case 0x6: case 0x7: \
552 if (toLim - *toP < 2) { \
553 *fromP = from; \
554 return; \
555 } \
556 *(*toP)++ = ((lo >> 6) | (hi << 2) | UTF8_cval2); \
557 *(*toP)++ = ((lo & 0x3f) | 0x80); \
558 break; \
559 default: \
560 if (toLim - *toP < 3) { \
561 *fromP = from; \
562 return; \
563 } \
564 /* 16 bits divided 4, 6, 6 amongst 3 bytes */ \
565 *(*toP)++ = ((hi >> 4) | UTF8_cval3); \
566 *(*toP)++ = (((hi & 0xf) << 2) | (lo >> 6) | 0x80); \
567 *(*toP)++ = ((lo & 0x3f) | 0x80); \
568 break; \
569 case 0xD8: case 0xD9: case 0xDA: case 0xDB: \
570 if (toLim - *toP < 4) { \
571 *fromP = from; \
572 return; \
573 } \
574 plane = (((hi & 0x3) << 2) | ((lo >> 6) & 0x3)) + 1; \
575 *(*toP)++ = ((plane >> 2) | UTF8_cval4); \
576 *(*toP)++ = (((lo >> 2) & 0xF) | ((plane & 0x3) << 4) | 0x80); \
577 from += 2; \
578 lo2 = GET_LO(from); \
579 *(*toP)++ = (((lo & 0x3) << 4) \
580 | ((GET_HI(from) & 0x3) << 2) \
581 | (lo2 >> 6) \
582 | 0x80); \
583 *(*toP)++ = ((lo2 & 0x3f) | 0x80); \
584 break; \
585 } \
586 } \
587 *fromP = from; \
588}
589
590#define DEFINE_UTF16_TO_UTF16(E) \
591static \
592void E ## toUtf16(const ENCODING *enc, \
593 const char **fromP, const char *fromLim, \
594 unsigned short **toP, const unsigned short *toLim) \
595{ \
596 /* Avoid copying first half only of surrogate */ \
597 if (fromLim - *fromP > ((toLim - *toP) << 1) \
598 && (GET_HI(fromLim - 2) & 0xF8) == 0xD8) \
599 fromLim -= 2; \
600 for (; *fromP != fromLim && *toP != toLim; *fromP += 2) \
601 *(*toP)++ = (GET_HI(*fromP) << 8) | GET_LO(*fromP); \
602}
603
604#define SET2(ptr, ch) \
605 (((ptr)[0] = ((ch) & 0xff)), ((ptr)[1] = ((ch) >> 8)))
606#define GET_LO(ptr) ((unsigned char)(ptr)[0])
607#define GET_HI(ptr) ((unsigned char)(ptr)[1])
608
609DEFINE_UTF16_TO_UTF8(little2_)
610DEFINE_UTF16_TO_UTF16(little2_)
611
612#undef SET2
613#undef GET_LO
614#undef GET_HI
615
616#define SET2(ptr, ch) \
617 (((ptr)[0] = ((ch) >> 8)), ((ptr)[1] = ((ch) & 0xFF)))
618#define GET_LO(ptr) ((unsigned char)(ptr)[1])
619#define GET_HI(ptr) ((unsigned char)(ptr)[0])
620
621DEFINE_UTF16_TO_UTF8(big2_)
622DEFINE_UTF16_TO_UTF16(big2_)
623
624#undef SET2
625#undef GET_LO
626#undef GET_HI
627
628#define LITTLE2_BYTE_TYPE(enc, p) \
629 ((p)[1] == 0 \
630 ? ((struct normal_encoding *)(enc))->type[(unsigned char)*(p)] \
631 : unicode_byte_type((p)[1], (p)[0]))
632#define LITTLE2_BYTE_TO_ASCII(enc, p) ((p)[1] == 0 ? (p)[0] : -1)
633#define LITTLE2_CHAR_MATCHES(enc, p, c) ((p)[1] == 0 && (p)[0] == c)
634#define LITTLE2_IS_NAME_CHAR_MINBPC(enc, p) \
635 UCS2_GET_NAMING(namePages, (unsigned char)p[1], (unsigned char)p[0])
636#define LITTLE2_IS_NMSTRT_CHAR_MINBPC(enc, p) \
637 UCS2_GET_NAMING(nmstrtPages, (unsigned char)p[1], (unsigned char)p[0])
638
639#ifdef XML_MIN_SIZE
640
641 static
642 int little2_byteType(const ENCODING * enc, const char *p)
643{
644 return LITTLE2_BYTE_TYPE(enc, p);
645}
646
647static int little2_byteToAscii(const ENCODING * enc, const char *p)
648{
649 return LITTLE2_BYTE_TO_ASCII(enc, p);
650}
651
652static int little2_charMatches(const ENCODING * enc, const char *p, int c)
653{
654 return LITTLE2_CHAR_MATCHES(enc, p, c);
655}
656
657static int little2_isNameMin(const ENCODING * enc, const char *p)
658{
659 return LITTLE2_IS_NAME_CHAR_MINBPC(enc, p);
660}
661
662static int little2_isNmstrtMin(const ENCODING * enc, const char *p)
663{
664 return LITTLE2_IS_NMSTRT_CHAR_MINBPC(enc, p);
665}
666
667#undef VTABLE
668#define VTABLE VTABLE1, little2_toUtf8, little2_toUtf16
669
670#else /* not XML_MIN_SIZE */
671
672#undef PREFIX
673#define PREFIX(ident) little2_ ## ident
674#define MINBPC(enc) 2
675/* CHAR_MATCHES is guaranteed to have MINBPC bytes available. */
676#define BYTE_TYPE(enc, p) LITTLE2_BYTE_TYPE(enc, p)
677#define BYTE_TO_ASCII(enc, p) LITTLE2_BYTE_TO_ASCII(enc, p)
678#define CHAR_MATCHES(enc, p, c) LITTLE2_CHAR_MATCHES(enc, p, c)
679#define IS_NAME_CHAR(enc, p, n) 0
680#define IS_NAME_CHAR_MINBPC(enc, p) LITTLE2_IS_NAME_CHAR_MINBPC(enc, p)
681#define IS_NMSTRT_CHAR(enc, p, n) (0)
682#define IS_NMSTRT_CHAR_MINBPC(enc, p) LITTLE2_IS_NMSTRT_CHAR_MINBPC(enc, p)
683
684#include "xmltok_impl.c"
685
686#undef MINBPC
687#undef BYTE_TYPE
688#undef BYTE_TO_ASCII
689#undef CHAR_MATCHES
690#undef IS_NAME_CHAR
691#undef IS_NAME_CHAR_MINBPC
692#undef IS_NMSTRT_CHAR
693#undef IS_NMSTRT_CHAR_MINBPC
694#undef IS_INVALID_CHAR
695
696#endif /* not XML_MIN_SIZE */
697
698#ifdef XML_NS
699
700 static const struct normal_encoding little2_encoding_ns =
701 {
702 {VTABLE, 2, 0,
703#if XML_BYTE_ORDER == 12
704 1
705#else
706 0
707#endif
708 },
709 {
710#include "expat\asciitab.h"
711#include "expat\latin1tab.h"
712 },
713 STANDARD_VTABLE(little2_)
714};
715
716#endif
717
718 static const struct normal_encoding little2_encoding =
719 {
720 {VTABLE, 2, 0,
721#if XML_BYTE_ORDER == 12
722 1
723#else
724 0
725#endif
726 },
727 {
728#define BT_COLON BT_NMSTRT
729#include "expat\asciitab.h"
730#undef BT_COLON
731#include "expat\latin1tab.h"
732 },
733 STANDARD_VTABLE(little2_)
734};
735
736#if XML_BYTE_ORDER != 21
737
738#ifdef XML_NS
739
740 static const struct normal_encoding internal_little2_encoding_ns =
741 {
742 {VTABLE, 2, 0, 1},
743 {
744#include "expat\iasciitab.h"
745#include "expat\latin1tab.h"
746 },
747 STANDARD_VTABLE(little2_)
748};
749
750#endif
751
752 static const struct normal_encoding internal_little2_encoding =
753 {
754 {VTABLE, 2, 0, 1},
755 {
756#define BT_COLON BT_NMSTRT
757#include "expat\iasciitab.h"
758#undef BT_COLON
759#include "expat\latin1tab.h"
760 },
761 STANDARD_VTABLE(little2_)
762};
763
764#endif
765
766
767#define BIG2_BYTE_TYPE(enc, p) \
768 ((p)[0] == 0 \
769 ? ((struct normal_encoding *)(enc))->type[(unsigned char)(p)[1]] \
770 : unicode_byte_type((p)[0], (p)[1]))
771#define BIG2_BYTE_TO_ASCII(enc, p) ((p)[0] == 0 ? (p)[1] : -1)
772#define BIG2_CHAR_MATCHES(enc, p, c) ((p)[0] == 0 && (p)[1] == c)
773#define BIG2_IS_NAME_CHAR_MINBPC(enc, p) \
774 UCS2_GET_NAMING(namePages, (unsigned char)p[0], (unsigned char)p[1])
775#define BIG2_IS_NMSTRT_CHAR_MINBPC(enc, p) \
776 UCS2_GET_NAMING(nmstrtPages, (unsigned char)p[0], (unsigned char)p[1])
777
778#ifdef XML_MIN_SIZE
779
780static int big2_byteType(const ENCODING * enc, const char *p)
781{
782 return BIG2_BYTE_TYPE(enc, p);
783}
784
785static int big2_byteToAscii(const ENCODING * enc, const char *p)
786{
787 return BIG2_BYTE_TO_ASCII(enc, p);
788}
789
790static int big2_charMatches(const ENCODING * enc, const char *p, int c)
791{
792 return BIG2_CHAR_MATCHES(enc, p, c);
793}
794
795static int big2_isNameMin(const ENCODING * enc, const char *p)
796{
797 return BIG2_IS_NAME_CHAR_MINBPC(enc, p);
798}
799
800static int big2_isNmstrtMin(const ENCODING * enc, const char *p)
801{
802 return BIG2_IS_NMSTRT_CHAR_MINBPC(enc, p);
803}
804
805#undef VTABLE
806#define VTABLE VTABLE1, big2_toUtf8, big2_toUtf16
807
808#else /* not XML_MIN_SIZE */
809
810#undef PREFIX
811#define PREFIX(ident) big2_ ## ident
812#define MINBPC(enc) 2
813/* CHAR_MATCHES is guaranteed to have MINBPC bytes available. */
814#define BYTE_TYPE(enc, p) BIG2_BYTE_TYPE(enc, p)
815#define BYTE_TO_ASCII(enc, p) BIG2_BYTE_TO_ASCII(enc, p)
816#define CHAR_MATCHES(enc, p, c) BIG2_CHAR_MATCHES(enc, p, c)
817#define IS_NAME_CHAR(enc, p, n) 0
818#define IS_NAME_CHAR_MINBPC(enc, p) BIG2_IS_NAME_CHAR_MINBPC(enc, p)
819#define IS_NMSTRT_CHAR(enc, p, n) (0)
820#define IS_NMSTRT_CHAR_MINBPC(enc, p) BIG2_IS_NMSTRT_CHAR_MINBPC(enc, p)
821
822#include "xmltok_impl.c"
823
824#undef MINBPC
825#undef BYTE_TYPE
826#undef BYTE_TO_ASCII
827#undef CHAR_MATCHES
828#undef IS_NAME_CHAR
829#undef IS_NAME_CHAR_MINBPC
830#undef IS_NMSTRT_CHAR
831#undef IS_NMSTRT_CHAR_MINBPC
832#undef IS_INVALID_CHAR
833
834#endif /* not XML_MIN_SIZE */
835
836#ifdef XML_NS
837
838 static const struct normal_encoding big2_encoding_ns =
839 {
840 {VTABLE, 2, 0,
841#if XML_BYTE_ORDER == 21
842 1
843#else
844 0
845#endif
846 },
847 {
848#include "expat\asciitab.h"
849#include "expat\latin1tab.h"
850 },
851 STANDARD_VTABLE(big2_)
852};
853
854#endif
855
856 static const struct normal_encoding big2_encoding =
857 {
858 {VTABLE, 2, 0,
859#if XML_BYTE_ORDER == 21
860 1
861#else
862 0
863#endif
864 },
865 {
866#define BT_COLON BT_NMSTRT
867#include "expat\asciitab.h"
868#undef BT_COLON
869#include "expat\latin1tab.h"
870 },
871 STANDARD_VTABLE(big2_)
872};
873
874#if XML_BYTE_ORDER != 12
875
876#ifdef XML_NS
877
878 static const struct normal_encoding internal_big2_encoding_ns =
879 {
880 {VTABLE, 2, 0, 1},
881 {
882#include "expat\iasciitab.h"
883#include "expat\latin1tab.h"
884 },
885 STANDARD_VTABLE(big2_)
886};
887
888#endif
889
890 static const struct normal_encoding internal_big2_encoding =
891 {
892 {VTABLE, 2, 0, 1},
893 {
894#define BT_COLON BT_NMSTRT
895#include "expat\iasciitab.h"
896#undef BT_COLON
897#include "expat\latin1tab.h"
898 },
899 STANDARD_VTABLE(big2_)
900};
901
902#endif
903
904#undef PREFIX
905
906static int streqci(const char *s1, const char *s2)
907{
908 for (;;)
909 {
910 char c1 = *s1++;
911 char c2 = *s2++;
912
913 if (ASCII_a <= c1 && c1 <= ASCII_z)
914 c1 += ASCII_A - ASCII_a;
915 if (ASCII_a <= c2 && c2 <= ASCII_z)
916 c2 += ASCII_A - ASCII_a;
917 if (c1 != c2)
918 return 0;
919 if (!c1)
920 break;
921 }
922 return 1;
923}
924
925static void initUpdatePosition(const ENCODING * enc, const char *ptr,
926 const char *end, POSITION * pos)
927{
928 normal_updatePosition(&utf8_encoding.enc, ptr, end, pos);
929}
930
931static int toAscii(const ENCODING * enc, const char *ptr, const char *end)
932{
933 char buf[1];
934 char *p = buf;
935
936 XmlUtf8Convert(enc, &ptr, end, &p, p + 1);
937 if (p == buf)
938 return -1;
939 else
940 return buf[0];
941}
942
943static int isSpace(int c)
944{
945 switch (c)
946 {
947 case 0x20:
948 case 0xD:
949 case 0xA:
950 case 0x9:
951 return 1;
952 }
953 return 0;
954}
955
956/* Return 1 if there's just optional white space
957 * or there's an S followed by name=val. */
958static int parsePseudoAttribute(const ENCODING * enc,
959 const char *ptr,
960 const char *end,
961 const char **namePtr,
962 const char **nameEndPtr,
963 const char **valPtr,
964 const char **nextTokPtr)
965{
966 int c;
967 char open;
968
969 if (ptr == end)
970 {
971 *namePtr = 0;
972 return 1;
973 }
974 if (!isSpace(toAscii(enc, ptr, end)))
975 {
976 *nextTokPtr = ptr;
977 return 0;
978 }
979 do
980 {
981 ptr += enc->minBytesPerChar;
982 }
983 while (isSpace(toAscii(enc, ptr, end)));
984 if (ptr == end)
985 {
986 *namePtr = 0;
987 return 1;
988 }
989 *namePtr = ptr;
990 for (;;)
991 {
992 c = toAscii(enc, ptr, end);
993 if (c == -1)
994 {
995 *nextTokPtr = ptr;
996 return 0;
997 }
998 if (c == ASCII_EQUALS)
999 {
1000 *nameEndPtr = ptr;
1001 break;
1002 }
1003 if (isSpace(c))
1004 {
1005 *nameEndPtr = ptr;
1006 do
1007 {
1008 ptr += enc->minBytesPerChar;
1009 }
1010 while (isSpace(c = toAscii(enc, ptr, end)));
1011 if (c != ASCII_EQUALS)
1012 {
1013 *nextTokPtr = ptr;
1014 return 0;
1015 }
1016 break;
1017 }
1018 ptr += enc->minBytesPerChar;
1019 }
1020 if (ptr == *namePtr)
1021 {
1022 *nextTokPtr = ptr;
1023 return 0;
1024 }
1025 ptr += enc->minBytesPerChar;
1026 c = toAscii(enc, ptr, end);
1027 while (isSpace(c))
1028 {
1029 ptr += enc->minBytesPerChar;
1030 c = toAscii(enc, ptr, end);
1031 }
1032 if (c != ASCII_QUOT && c != ASCII_APOS)
1033 {
1034 *nextTokPtr = ptr;
1035 return 0;
1036 }
1037 open = c;
1038 ptr += enc->minBytesPerChar;
1039 *valPtr = ptr;
1040 for (;; ptr += enc->minBytesPerChar)
1041 {
1042 c = toAscii(enc, ptr, end);
1043 if (c == open)
1044 break;
1045 if (!(ASCII_a <= c && c <= ASCII_z)
1046 && !(ASCII_A <= c && c <= ASCII_Z)
1047 && !(ASCII_0 <= c && c <= ASCII_9)
1048 && c != ASCII_PERIOD
1049 && c != ASCII_MINUS
1050 && c != ASCII_UNDERSCORE)
1051 {
1052 *nextTokPtr = ptr;
1053 return 0;
1054 }
1055 }
1056 *nextTokPtr = ptr + enc->minBytesPerChar;
1057 return 1;
1058}
1059
1060static const char KW_version[] =
1061{
1062 ASCII_v, ASCII_e, ASCII_r, ASCII_s, ASCII_i, ASCII_o, ASCII_n, '\0'
1063};
1064
1065static const char KW_encoding[] =
1066{
1067 ASCII_e, ASCII_n, ASCII_c, ASCII_o, ASCII_d, ASCII_i, ASCII_n, ASCII_g, '\0'
1068};
1069
1070static const char KW_standalone[] =
1071{
1072 ASCII_s, ASCII_t, ASCII_a, ASCII_n, ASCII_d, ASCII_a, ASCII_l, ASCII_o, ASCII_n, ASCII_e, '\0'
1073};
1074
1075static const char KW_yes[] =
1076{
1077 ASCII_y, ASCII_e, ASCII_s, '\0'
1078};
1079
1080static const char KW_no[] =
1081{
1082 ASCII_n, ASCII_o, '\0'
1083};
1084
1085static int doParseXmlDecl(const ENCODING * (*encodingFinder) (const ENCODING *,
1086 const char *,
1087 const char *),
1088 int isGeneralTextEntity,
1089 const ENCODING * enc,
1090 const char *ptr,
1091 const char *end,
1092 const char **badPtr,
1093 const char **versionPtr,
1094 const char **versionEndPtr,
1095 const char **encodingName,
1096 const ENCODING ** encoding,
1097 int *standalone)
1098{
1099 const char *val = 0;
1100 const char *name = 0;
1101 const char *nameEnd = 0;
1102
1103 ptr += 5 * enc->minBytesPerChar;
1104 end -= 2 * enc->minBytesPerChar;
1105 if (!parsePseudoAttribute(enc, ptr, end, &name, &nameEnd, &val, &ptr) || !name)
1106 {
1107 *badPtr = ptr;
1108 return 0;
1109 }
1110 if (!XmlNameMatchesAscii(enc, name, nameEnd, KW_version))
1111 {
1112 if (!isGeneralTextEntity)
1113 {
1114 *badPtr = name;
1115 return 0;
1116 }
1117 }
1118 else
1119 {
1120 if (versionPtr)
1121 *versionPtr = val;
1122 if (versionEndPtr)
1123 *versionEndPtr = ptr;
1124 if (!parsePseudoAttribute(enc, ptr, end, &name, &nameEnd, &val, &ptr))
1125 {
1126 *badPtr = ptr;
1127 return 0;
1128 }
1129 if (!name)
1130 {
1131 if (isGeneralTextEntity)
1132 {
1133 /* a TextDecl must have an EncodingDecl */
1134 *badPtr = ptr;
1135 return 0;
1136 }
1137 return 1;
1138 }
1139 }
1140 if (XmlNameMatchesAscii(enc, name, nameEnd, KW_encoding))
1141 {
1142 int c = toAscii(enc, val, end);
1143
1144 if (!(ASCII_a <= c && c <= ASCII_z) && !(ASCII_A <= c && c <= ASCII_Z))
1145 {
1146 *badPtr = val;
1147 return 0;
1148 }
1149 if (encodingName)
1150 *encodingName = val;
1151 if (encoding)
1152 *encoding = encodingFinder(enc, val, ptr - enc->minBytesPerChar);
1153 if (!parsePseudoAttribute(enc, ptr, end, &name, &nameEnd, &val, &ptr))
1154 {
1155 *badPtr = ptr;
1156 return 0;
1157 }
1158 if (!name)
1159 return 1;
1160 }
1161 if (!XmlNameMatchesAscii(enc, name, nameEnd, KW_standalone) || isGeneralTextEntity)
1162 {
1163 *badPtr = name;
1164 return 0;
1165 }
1166 if (XmlNameMatchesAscii(enc, val, ptr - enc->minBytesPerChar, KW_yes))
1167 {
1168 if (standalone)
1169 *standalone = 1;
1170 }
1171 else if (XmlNameMatchesAscii(enc, val, ptr - enc->minBytesPerChar, KW_no))
1172 {
1173 if (standalone)
1174 *standalone = 0;
1175 }
1176 else
1177 {
1178 *badPtr = val;
1179 return 0;
1180 }
1181 while (isSpace(toAscii(enc, ptr, end)))
1182 ptr += enc->minBytesPerChar;
1183 if (ptr != end)
1184 {
1185 *badPtr = ptr;
1186 return 0;
1187 }
1188 return 1;
1189}
1190
1191static int checkCharRefNumber(int result)
1192{
1193 switch (result >> 8)
1194 {
1195 case 0xD8:
1196 case 0xD9:
1197 case 0xDA:
1198 case 0xDB:
1199 case 0xDC:
1200 case 0xDD:
1201 case 0xDE:
1202 case 0xDF:
1203 return -1;
1204 case 0:
1205 if (latin1_encoding.type[result] == BT_NONXML)
1206 return -1;
1207 break;
1208 case 0xFF:
1209 if (result == 0xFFFE || result == 0xFFFF)
1210 return -1;
1211 break;
1212 }
1213 return result;
1214}
1215
1216int XmlUtf8Encode(int c, char *buf)
1217{
1218 enum
1219 {
1220 /* minN is minimum legal resulting value for N byte sequence */
1221 min2 = 0x80,
1222 min3 = 0x800,
1223 min4 = 0x10000
1224 };
1225
1226 if (c < 0)
1227 return 0;
1228 if (c < min2)
1229 {
1230 buf[0] = (c | UTF8_cval1);
1231 return 1;
1232 }
1233 if (c < min3)
1234 {
1235 buf[0] = ((c >> 6) | UTF8_cval2);
1236 buf[1] = ((c & 0x3f) | 0x80);
1237 return 2;
1238 }
1239 if (c < min4)
1240 {
1241 buf[0] = ((c >> 12) | UTF8_cval3);
1242 buf[1] = (((c >> 6) & 0x3f) | 0x80);
1243 buf[2] = ((c & 0x3f) | 0x80);
1244 return 3;
1245 }
1246 if (c < 0x110000)
1247 {
1248 buf[0] = ((c >> 18) | UTF8_cval4);
1249 buf[1] = (((c >> 12) & 0x3f) | 0x80);
1250 buf[2] = (((c >> 6) & 0x3f) | 0x80);
1251 buf[3] = ((c & 0x3f) | 0x80);
1252 return 4;
1253 }
1254 return 0;
1255}
1256
1257int XmlUtf16Encode(int charNum, unsigned short *buf)
1258{
1259 if (charNum < 0)
1260 return 0;
1261 if (charNum < 0x10000)
1262 {
1263 buf[0] = charNum;
1264 return 1;
1265 }
1266 if (charNum < 0x110000)
1267 {
1268 charNum -= 0x10000;
1269 buf[0] = (charNum >> 10) + 0xD800;
1270 buf[1] = (charNum & 0x3FF) + 0xDC00;
1271 return 2;
1272 }
1273 return 0;
1274}
1275
1276struct unknown_encoding
1277{
1278 struct normal_encoding normal;
1279 int (*convert) (void *userData, const char *p);
1280 void *userData;
1281 unsigned short utf16[256];
1282 char utf8[256][4];
1283};
1284
1285int XmlSizeOfUnknownEncoding(void)
1286{
1287 return sizeof(struct unknown_encoding);
1288}
1289
1290static int unknown_isName(const ENCODING * enc, const char *p)
1291{
1292 int c = ((const struct unknown_encoding *)enc)
1293 ->convert(((const struct unknown_encoding *)enc)->userData, p);
1294
1295 if (c & ~0xFFFF)
1296 return 0;
1297 return UCS2_GET_NAMING(namePages, c >> 8, c & 0xFF);
1298}
1299
1300static int unknown_isNmstrt(const ENCODING * enc, const char *p)
1301{
1302 int c = ((const struct unknown_encoding *)enc)
1303 ->convert(((const struct unknown_encoding *)enc)->userData, p);
1304
1305 if (c & ~0xFFFF)
1306 return 0;
1307 return UCS2_GET_NAMING(nmstrtPages, c >> 8, c & 0xFF);
1308}
1309
1310static int unknown_isInvalid(const ENCODING * enc, const char *p)
1311{
1312 int c = ((const struct unknown_encoding *)enc)
1313 ->convert(((const struct unknown_encoding *)enc)->userData, p);
1314
1315 return (c & ~0xFFFF) || checkCharRefNumber(c) < 0;
1316}
1317
1318static void unknown_toUtf8(const ENCODING * enc,
1319 const char **fromP,
1320 const char *fromLim,
1321 char **toP,
1322 const char *toLim)
1323{
1324 char buf[XML_UTF8_ENCODE_MAX];
1325
1326 for (;;)
1327 {
1328 const char *utf8;
1329 int n;
1330
1331 if (*fromP == fromLim)
1332 break;
1333 utf8 = ((const struct unknown_encoding *)enc)->utf8[(unsigned char)**fromP];
1334 n = *utf8++;
1335 if (n == 0)
1336 {
1337 int c = ((const struct unknown_encoding *)enc)
1338 ->convert(((const struct unknown_encoding *)enc)->userData, *fromP);
1339
1340 n = XmlUtf8Encode(c, buf);
1341 if (n > toLim - *toP)
1342 break;
1343 utf8 = buf;
1344 *fromP += ((const struct normal_encoding *)enc)->type[(unsigned char)**fromP]
1345 - (BT_LEAD2 - 2);
1346 }
1347 else
1348 {
1349 if (n > toLim - *toP)
1350 break;
1351 (*fromP)++;
1352 }
1353 do
1354 {
1355 *(*toP)++ = *utf8++;
1356 }
1357 while (--n != 0);
1358 }
1359}
1360
1361static void unknown_toUtf16(const ENCODING * enc,
1362 const char **fromP,
1363 const char *fromLim,
1364 unsigned short **toP,
1365 const unsigned short *toLim)
1366{
1367 while (*fromP != fromLim && *toP != toLim)
1368 {
1369 unsigned short c
1370 = ((const struct unknown_encoding *)enc)->utf16[(unsigned char)**fromP];
1371
1372 if (c == 0)
1373 {
1374 c = (unsigned short)((const struct unknown_encoding *)enc)
1375 ->convert(((const struct unknown_encoding *)enc)->userData, *fromP);
1376 *fromP += ((const struct normal_encoding *)enc)->type[(unsigned char)**fromP]
1377 - (BT_LEAD2 - 2);
1378 }
1379 else
1380 (*fromP)++;
1381 *(*toP)++ = c;
1382 }
1383}
1384
1385/*
1386 *@@ XmlInitUnknownEncoding:
1387 *
1388 *@@changed V0.9.14 (2001-08-09) [umoeller]: couple of performance hacks
1389 */
1390
1391ENCODING* XmlInitUnknownEncoding(void *mem,
1392 int *table,
1393 int (*convert) (void *userData, const char *p),
1394 void *userData)
1395{
1396 int i;
1397 struct unknown_encoding *e = mem;
1398
1399 // gee, isn't this a regular memcpy?!?
1400 /* for (i = 0;
1401 i < (int)sizeof(struct normal_encoding);
1402 i++)
1403 ((char *)mem)[i] = ((char *)&latin1_encoding)[i]; */
1404
1405 // replaced the above with this V0.9.14 (2001-08-09) [umoeller]
1406 memcpy(mem, &latin1_encoding, sizeof(struct normal_encoding));
1407
1408 for (i = 0; i < 128; i++)
1409 if ( latin1_encoding.type[i] != BT_OTHER
1410 && latin1_encoding.type[i] != BT_NONXML
1411 && table[i] != i
1412 )
1413 return 0;
1414
1415 for (i = 0; i < 256; i++)
1416 {
1417 int c = table[i];
1418
1419 if (c == -1)
1420 {
1421 e->normal.type[i] = BT_MALFORM;
1422 /* This shouldn't really get used. */
1423 e->utf16[i] = 0xFFFF;
1424 e->utf8[i][0] = 1;
1425 e->utf8[i][1] = 0;
1426 }
1427 else if (c < 0)
1428 {
1429 if (c < -4)
1430 return 0;
1431 e->normal.type[i] = BT_LEAD2 - (c + 2);
1432 e->utf8[i][0] = 0;
1433 e->utf16[i] = 0;
1434 }
1435 else if (c < 0x80)
1436 {
1437 if ( latin1_encoding.type[c] != BT_OTHER
1438 && latin1_encoding.type[c] != BT_NONXML
1439 && c != i
1440 )
1441 return 0;
1442 e->normal.type[i] = latin1_encoding.type[c];
1443 e->utf8[i][0] = 1;
1444 e->utf8[i][1] = (char)c;
1445 e->utf16[i] = c == 0 ? 0xFFFF : c;
1446 }
1447 else if (checkCharRefNumber(c) < 0)
1448 {
1449 e->normal.type[i] = BT_NONXML;
1450 /* This shouldn't really get used. */
1451 e->utf16[i] = 0xFFFF;
1452 e->utf8[i][0] = 1;
1453 e->utf8[i][1] = 0;
1454 }
1455 else
1456 {
1457 if (c > 0xFFFF)
1458 return 0;
1459 if (UCS2_GET_NAMING(nmstrtPages, c >> 8, c & 0xff))
1460 e->normal.type[i] = BT_NMSTRT;
1461 else if (UCS2_GET_NAMING(namePages, c >> 8, c & 0xff))
1462 e->normal.type[i] = BT_NAME;
1463 else
1464 e->normal.type[i] = BT_OTHER;
1465 e->utf8[i][0] = (char)XmlUtf8Encode(c, e->utf8[i] + 1);
1466 e->utf16[i] = c;
1467 }
1468 }
1469 e->userData = userData;
1470 e->convert = convert;
1471 if (convert)
1472 {
1473 e->normal.isName2 = unknown_isName;
1474 e->normal.isName3 = unknown_isName;
1475 e->normal.isName4 = unknown_isName;
1476 e->normal.isNmstrt2 = unknown_isNmstrt;
1477 e->normal.isNmstrt3 = unknown_isNmstrt;
1478 e->normal.isNmstrt4 = unknown_isNmstrt;
1479 e->normal.isInvalid2 = unknown_isInvalid;
1480 e->normal.isInvalid3 = unknown_isInvalid;
1481 e->normal.isInvalid4 = unknown_isInvalid;
1482 }
1483 e->normal.enc.utf8Convert = unknown_toUtf8;
1484 e->normal.enc.utf16Convert = unknown_toUtf16;
1485 return &(e->normal.enc);
1486}
1487
1488/* If this enumeration is changed, getEncodingIndex and encodings
1489 * must also be changed. */
1490enum
1491{
1492 UNKNOWN_ENC = -1,
1493 ISO_8859_1_ENC = 0,
1494 US_ASCII_ENC,
1495 UTF_8_ENC,
1496 UTF_16_ENC,
1497 UTF_16BE_ENC,
1498 UTF_16LE_ENC,
1499 /* must match encodingNames up to here */
1500 NO_ENC
1501};
1502
1503static const char KW_ISO_8859_1[] =
1504{
1505 ASCII_I, ASCII_S, ASCII_O, ASCII_MINUS, ASCII_8, ASCII_8, ASCII_5, ASCII_9, ASCII_MINUS, ASCII_1, '\0'
1506};
1507static const char KW_US_ASCII[] =
1508{
1509 ASCII_U, ASCII_S, ASCII_MINUS, ASCII_A, ASCII_S, ASCII_C, ASCII_I, ASCII_I, '\0'
1510};
1511static const char KW_UTF_8[] =
1512{
1513 ASCII_U, ASCII_T, ASCII_F, ASCII_MINUS, ASCII_8, '\0'
1514};
1515static const char KW_UTF_16[] =
1516{
1517 ASCII_U, ASCII_T, ASCII_F, ASCII_MINUS, ASCII_1, ASCII_6, '\0'
1518};
1519static const char KW_UTF_16BE[] =
1520{
1521 ASCII_U, ASCII_T, ASCII_F, ASCII_MINUS, ASCII_1, ASCII_6, ASCII_B, ASCII_E, '\0'
1522};
1523static const char KW_UTF_16LE[] =
1524{
1525 ASCII_U, ASCII_T, ASCII_F, ASCII_MINUS, ASCII_1, ASCII_6, ASCII_L, ASCII_E, '\0'
1526};
1527
1528static int getEncodingIndex(const char *name)
1529{
1530 static const char *encodingNames[] =
1531 {
1532 KW_ISO_8859_1,
1533 KW_US_ASCII,
1534 KW_UTF_8,
1535 KW_UTF_16,
1536 KW_UTF_16BE,
1537 KW_UTF_16LE,
1538 };
1539 int i;
1540
1541 if (name == 0)
1542 return NO_ENC;
1543 for (i = 0;
1544 i < (int)(sizeof(encodingNames) / sizeof(encodingNames[0]));
1545 i++)
1546 if (streqci(name, encodingNames[i]))
1547 return i;
1548 return UNKNOWN_ENC;
1549}
1550
1551/* For binary compatibility, we store the index of the encoding specified
1552 * at initialization in the isUtf16 member. */
1553
1554#define INIT_ENC_INDEX(enc) ((int)(enc)->initEnc.isUtf16)
1555#define SET_INIT_ENC_INDEX(enc, i) ((enc)->initEnc.isUtf16 = (char)i)
1556
1557/* This is what detects the encoding.
1558 * encodingTable maps from encoding indices to encodings;
1559 * INIT_ENC_INDEX(enc) is the index of the external (protocol) specified encoding;
1560 * state is XML_CONTENT_STATE if we're parsing an external text entity,
1561 * and XML_PROLOG_STATE otherwise.
1562 */
1563
1564
1565static int initScan(const ENCODING ** encodingTable,
1566 const INIT_ENCODING * enc,
1567 int state,
1568 const char *ptr,
1569 const char *end,
1570 const char **nextTokPtr)
1571{
1572 const ENCODING **encPtr;
1573
1574 if (ptr == end)
1575 return XML_TOK_NONE;
1576 encPtr = enc->encPtr;
1577 if (ptr + 1 == end)
1578 {
1579 /* only a single byte available for auto-detection */
1580#ifndef XML_DTD /* FIXME */
1581 /* a well-formed document entity must have more than one byte */
1582 if (state != XML_CONTENT_STATE)
1583 return XML_TOK_PARTIAL;
1584#endif
1585 /* so we're parsing an external text entity... */
1586 /* if UTF-16 was externally specified, then we need at least 2 bytes */
1587 switch (INIT_ENC_INDEX(enc))
1588 {
1589 case UTF_16_ENC:
1590 case UTF_16LE_ENC:
1591 case UTF_16BE_ENC:
1592 return XML_TOK_PARTIAL;
1593 }
1594 switch ((unsigned char)*ptr)
1595 {
1596 case 0xFE:
1597 case 0xFF:
1598 case 0xEF: /* possibly first byte of UTF-8 BOM */
1599 if (INIT_ENC_INDEX(enc) == ISO_8859_1_ENC
1600 && state == XML_CONTENT_STATE)
1601 break;
1602 /* fall through */
1603 case 0x00:
1604 case 0x3C:
1605 return XML_TOK_PARTIAL;
1606 }
1607 }
1608 else
1609 {
1610 switch (((unsigned char)ptr[0] << 8) | (unsigned char)ptr[1])
1611 {
1612 case 0xFEFF:
1613 if (INIT_ENC_INDEX(enc) == ISO_8859_1_ENC
1614 && state == XML_CONTENT_STATE)
1615 break;
1616 *nextTokPtr = ptr + 2;
1617 *encPtr = encodingTable[UTF_16BE_ENC];
1618 return XML_TOK_BOM;
1619 /* 00 3C is handled in the default case */
1620 case 0x3C00:
1621 if ((INIT_ENC_INDEX(enc) == UTF_16BE_ENC
1622 || INIT_ENC_INDEX(enc) == UTF_16_ENC)
1623 && state == XML_CONTENT_STATE)
1624 break;
1625 *encPtr = encodingTable[UTF_16LE_ENC];
1626 return XmlTok(*encPtr, state, ptr, end, nextTokPtr);
1627 case 0xFFFE:
1628 if (INIT_ENC_INDEX(enc) == ISO_8859_1_ENC
1629 && state == XML_CONTENT_STATE)
1630 break;
1631 *nextTokPtr = ptr + 2;
1632 *encPtr = encodingTable[UTF_16LE_ENC];
1633 return XML_TOK_BOM;
1634 case 0xEFBB:
1635 /* Maybe a UTF-8 BOM (EF BB BF) */
1636 /* If there's an explicitly specified (external) encoding
1637 * of ISO-8859-1 or some flavour of UTF-16
1638 * and this is an external text entity,
1639 * don't look for the BOM,
1640 * because it might be a legal data. */
1641 if (state == XML_CONTENT_STATE)
1642 {
1643 int e = INIT_ENC_INDEX(enc);
1644
1645 if (e == ISO_8859_1_ENC || e == UTF_16BE_ENC || e == UTF_16LE_ENC || e == UTF_16_ENC)
1646 break;
1647 }
1648 if (ptr + 2 == end)
1649 return XML_TOK_PARTIAL;
1650 if ((unsigned char)ptr[2] == 0xBF)
1651 {
1652 *nextTokPtr = ptr + 3;
1653 *encPtr = encodingTable[UTF_8_ENC];
1654 return XML_TOK_BOM;
1655 }
1656 break;
1657 default:
1658 if (ptr[0] == '\0')
1659 {
1660 /* 0 isn't a legal data character. Furthermore a document entity can only
1661 * start with ASCII characters. So the only way this can fail to be big-endian
1662 * UTF-16 if it it's an external parsed general entity that's labelled as
1663 * UTF-16LE. */
1664 if (state == XML_CONTENT_STATE && INIT_ENC_INDEX(enc) == UTF_16LE_ENC)
1665 break;
1666 *encPtr = encodingTable[UTF_16BE_ENC];
1667 return XmlTok(*encPtr, state, ptr, end, nextTokPtr);
1668 }
1669 else if (ptr[1] == '\0')
1670 {
1671 /* We could recover here in the case:
1672 * - parsing an external entity
1673 * - second byte is 0
1674 * - no externally specified encoding
1675 * - no encoding declaration
1676 * by assuming UTF-16LE. But we don't, because this would mean when
1677 * presented just with a single byte, we couldn't reliably determine
1678 * whether we needed further bytes. */
1679 if (state == XML_CONTENT_STATE)
1680 break;
1681 *encPtr = encodingTable[UTF_16LE_ENC];
1682 return XmlTok(*encPtr, state, ptr, end, nextTokPtr);
1683 }
1684 break;
1685 }
1686 }
1687 *encPtr = encodingTable[INIT_ENC_INDEX(enc)];
1688 return XmlTok(*encPtr, state, ptr, end, nextTokPtr);
1689}
1690
1691
1692#define NS(x) x
1693#define ns(x) x
1694#include "xmltok_ns.c"
1695#undef NS
1696#undef ns
1697
1698#ifdef XML_NS
1699
1700#define NS(x) x ## NS
1701#define ns(x) x ## _ns
1702
1703#include "xmltok_ns.c"
1704
1705#undef NS
1706#undef ns
1707
1708ENCODING * XmlInitUnknownEncodingNS(void *mem,
1709 int *table,
1710 int (*convert) (void *userData, const char *p),
1711 void *userData)
1712{
1713 ENCODING *enc = XmlInitUnknownEncoding(mem, table, convert, userData);
1714
1715 if (enc)
1716 ((struct normal_encoding *)enc)->type[ASCII_COLON] = BT_COLON;
1717 return enc;
1718}
1719
1720#endif /* XML_NS */
Note: See TracBrowser for help on using the repository browser.