1 /* Copyright (c) 1998, 1999 Thai Open Source Software Center Ltd
2 See the file COPYING for copying permission.
7 #ifdef COMPILED_FROM_DSP
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"
16 #ifdef HAVE_EXPAT_CONFIG_H
17 #include "expat_config.h"
19 #endif /* ndef COMPILED_FROM_DSP */
21 #include "expat_external.h"
27 #define IGNORE_SECTION_TOK_VTABLE , PREFIX(ignoreSectionTok)
29 #define IGNORE_SECTION_TOK_VTABLE /* as nothing */
33 { PREFIX(prologTok), PREFIX(contentTok), \
34 PREFIX(cdataSectionTok) IGNORE_SECTION_TOK_VTABLE }, \
35 { PREFIX(attributeValueTok), PREFIX(entityValueTok) }, \
37 PREFIX(nameMatchesAscii), \
41 PREFIX(charRefNumber), \
42 PREFIX(predefinedEntityName), \
43 PREFIX(updatePosition), \
46 #define VTABLE VTABLE1, PREFIX(toUtf8), PREFIX(toUtf16)
48 #define UCS2_GET_NAMING(pages, hi, lo) \
49 (namingBitmap[(pages[hi] << 3) + ((lo) >> 5)] & (1 << ((lo) & 0x1F)))
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.
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)))
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
66 #define UTF8_GET_NAMING3(pages, byte) \
67 (namingBitmap[((pages)[((((byte)[0]) & 0xF) << 4) \
68 + ((((byte)[1]) >> 2) & 0xF)] \
70 + ((((byte)[1]) & 3) << 1) \
71 + ((((byte)[2]) >> 5) & 1)] \
72 & (1 << (((byte)[2]) & 0x1F)))
74 #define UTF8_GET_NAMING(pages, p, n) \
76 ? UTF8_GET_NAMING2(pages, (const unsigned char *)(p)) \
78 ? UTF8_GET_NAMING3(pages, (const unsigned char *)(p)) \
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
88 (A & 0xC0) == 0xC0 means A > 0xBF
91 #define UTF8_INVALID2(p) \
92 ((*p) < 0xC2 || ((p)[1] & 0x80) == 0 || ((p)[1] & 0xC0) == 0xC0)
94 #define UTF8_INVALID3(p) \
95 (((p)[2] & 0x80) == 0 \
97 ((*p) == 0xEF && (p)[1] == 0xBF \
101 ((p)[2] & 0xC0) == 0xC0) \
105 (p)[1] < 0xA0 || ((p)[1] & 0xC0) == 0xC0 \
107 ((p)[1] & 0x80) == 0 \
109 ((*p) == 0xED ? (p)[1] > 0x9F : ((p)[1] & 0xC0) == 0xC0)))
111 #define UTF8_INVALID4(p) \
112 (((p)[3] & 0x80) == 0 || ((p)[3] & 0xC0) == 0xC0 \
114 ((p)[2] & 0x80) == 0 || ((p)[2] & 0xC0) == 0xC0 \
118 (p)[1] < 0x90 || ((p)[1] & 0xC0) == 0xC0 \
120 ((p)[1] & 0x80) == 0 \
122 ((*p) == 0xF4 ? (p)[1] > 0x8F : ((p)[1] & 0xC0) == 0xC0)))
124 static int PTRFASTCALL
125 isNever(const ENCODING *enc, const char *p)
130 static int PTRFASTCALL
131 utf8_isName2(const ENCODING *enc, const char *p)
133 return UTF8_GET_NAMING2(namePages, (const unsigned char *)p);
136 static int PTRFASTCALL
137 utf8_isName3(const ENCODING *enc, const char *p)
139 return UTF8_GET_NAMING3(namePages, (const unsigned char *)p);
142 #define utf8_isName4 isNever
144 static int PTRFASTCALL
145 utf8_isNmstrt2(const ENCODING *enc, const char *p)
147 return UTF8_GET_NAMING2(nmstrtPages, (const unsigned char *)p);
150 static int PTRFASTCALL
151 utf8_isNmstrt3(const ENCODING *enc, const char *p)
153 return UTF8_GET_NAMING3(nmstrtPages, (const unsigned char *)p);
156 #define utf8_isNmstrt4 isNever
158 static int PTRFASTCALL
159 utf8_isInvalid2(const ENCODING *enc, const char *p)
161 return UTF8_INVALID2((const unsigned char *)p);
164 static int PTRFASTCALL
165 utf8_isInvalid3(const ENCODING *enc, const char *p)
167 return UTF8_INVALID3((const unsigned char *)p);
170 static int PTRFASTCALL
171 utf8_isInvalid4(const ENCODING *enc, const char *p)
173 return UTF8_INVALID4((const unsigned char *)p);
176 struct normal_encoding {
178 unsigned char type[256];
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 *);
197 #define AS_NORMAL_ENCODING(enc) ((const struct normal_encoding *) (enc))
201 #define STANDARD_VTABLE(E) \
210 #define STANDARD_VTABLE(E) /* as nothing */
214 #define NORMAL_VTABLE(E) \
225 static int FASTCALL checkCharRefNumber(int);
227 #include "xmltok_impl.h"
231 #define sb_isNameMin isNever
232 #define sb_isNmstrtMin isNever
236 #define MINBPC(enc) ((enc)->minBytesPerChar)
238 /* minimum bytes per character */
239 #define MINBPC(enc) 1
242 #define SB_BYTE_TYPE(enc, p) \
243 (((struct normal_encoding *)(enc))->type[(unsigned char)*(p)])
246 static int PTRFASTCALL
247 sb_byteType(const ENCODING *enc, const char *p)
249 return SB_BYTE_TYPE(enc, p);
251 #define BYTE_TYPE(enc, p) \
252 (AS_NORMAL_ENCODING(enc)->byteType(enc, p))
254 #define BYTE_TYPE(enc, p) SB_BYTE_TYPE(enc, p)
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)
266 #define BYTE_TO_ASCII(enc, p) (*(p))
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))
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))
282 #define IS_NAME_CHAR_MINBPC(enc, p) (0)
283 #define IS_NMSTRT_CHAR_MINBPC(enc, p) (0)
287 #define CHAR_MATCHES(enc, p, c) \
288 (AS_NORMAL_ENCODING(enc)->charMatches(enc, p, c))
290 sb_charMatches(const ENCODING *enc, const char *p, int c)
295 /* c is an ASCII character */
296 #define CHAR_MATCHES(enc, p, c) (*(p) == c)
299 #define PREFIX(ident) normal_ ## ident
300 #define XML_TOK_IMPL_C
301 #include "xmltok_impl.c"
302 #undef XML_TOK_IMPL_C
309 #undef IS_NAME_CHAR_MINBPC
310 #undef IS_NMSTRT_CHAR
311 #undef IS_NMSTRT_CHAR_MINBPC
312 #undef IS_INVALID_CHAR
314 enum { /* UTF8_cvalN is value of masked first byte of N byte sequence */
322 utf8_toUtf8(const ENCODING *enc,
323 const char **fromP, const char *fromLim,
324 char **toP, const char *toLim)
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)
334 for (to = *toP, from = *fromP; from != fromLim; from++, to++)
341 utf8_toUtf16(const ENCODING *enc,
342 const char **fromP, const char *fromLim,
343 unsigned short **toP, const unsigned short *toLim)
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]) {
350 *to++ = (unsigned short)(((from[0] & 0x1f) << 6) | (from[1] & 0x3f));
354 *to++ = (unsigned short)(((from[0] & 0xf) << 12)
355 | ((from[1] & 0x3f) << 6) | (from[2] & 0x3f));
363 n = ((from[0] & 0x7) << 18) | ((from[1] & 0x3f) << 12)
364 | ((from[2] & 0x3f) << 6) | (from[3] & 0x3f);
366 to[0] = (unsigned short)((n >> 10) | 0xD800);
367 to[1] = (unsigned short)((n & 0x3FF) | 0xDC00);
383 static const struct normal_encoding utf8_encoding_ns = {
384 { VTABLE1, utf8_toUtf8, utf8_toUtf16, 1, 1, 0 },
386 #include "asciitab.h"
389 STANDARD_VTABLE(sb_) NORMAL_VTABLE(utf8_)
393 static const struct normal_encoding utf8_encoding = {
394 { VTABLE1, utf8_toUtf8, utf8_toUtf16, 1, 1, 0 },
396 #define BT_COLON BT_NMSTRT
397 #include "asciitab.h"
401 STANDARD_VTABLE(sb_) NORMAL_VTABLE(utf8_)
406 static const struct normal_encoding internal_utf8_encoding_ns = {
407 { VTABLE1, utf8_toUtf8, utf8_toUtf16, 1, 1, 0 },
409 #include "iasciitab.h"
412 STANDARD_VTABLE(sb_) NORMAL_VTABLE(utf8_)
417 static const struct normal_encoding internal_utf8_encoding = {
418 { VTABLE1, utf8_toUtf8, utf8_toUtf16, 1, 1, 0 },
420 #define BT_COLON BT_NMSTRT
421 #include "iasciitab.h"
425 STANDARD_VTABLE(sb_) NORMAL_VTABLE(utf8_)
429 latin1_toUtf8(const ENCODING *enc,
430 const char **fromP, const char *fromLim,
431 char **toP, const char *toLim)
435 if (*fromP == fromLim)
437 c = (unsigned char)**fromP;
439 if (toLim - *toP < 2)
441 *(*toP)++ = (char)((c >> 6) | UTF8_cval2);
442 *(*toP)++ = (char)((c & 0x3f) | 0x80);
448 *(*toP)++ = *(*fromP)++;
454 latin1_toUtf16(const ENCODING *enc,
455 const char **fromP, const char *fromLim,
456 unsigned short **toP, const unsigned short *toLim)
458 while (*fromP != fromLim && *toP != toLim)
459 *(*toP)++ = (unsigned char)*(*fromP)++;
464 static const struct normal_encoding latin1_encoding_ns = {
465 { VTABLE1, latin1_toUtf8, latin1_toUtf16, 1, 0, 0 },
467 #include "asciitab.h"
468 #include "latin1tab.h"
475 static const struct normal_encoding latin1_encoding = {
476 { VTABLE1, latin1_toUtf8, latin1_toUtf16, 1, 0, 0 },
478 #define BT_COLON BT_NMSTRT
479 #include "asciitab.h"
481 #include "latin1tab.h"
487 ascii_toUtf8(const ENCODING *enc,
488 const char **fromP, const char *fromLim,
489 char **toP, const char *toLim)
491 while (*fromP != fromLim && *toP != toLim)
492 *(*toP)++ = *(*fromP)++;
497 static const struct normal_encoding ascii_encoding_ns = {
498 { VTABLE1, ascii_toUtf8, latin1_toUtf16, 1, 1, 0 },
500 #include "asciitab.h"
508 static const struct normal_encoding ascii_encoding = {
509 { VTABLE1, ascii_toUtf8, latin1_toUtf16, 1, 1, 0 },
511 #define BT_COLON BT_NMSTRT
512 #include "asciitab.h"
519 static int PTRFASTCALL
520 unicode_byte_type(char hi, char lo)
522 switch ((unsigned char)hi) {
523 case 0xD8: case 0xD9: case 0xDA: case 0xDB:
525 case 0xDC: case 0xDD: case 0xDE: case 0xDF:
528 switch ((unsigned char)lo) {
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) \
545 for (from = *fromP; from != fromLim; from += 2) { \
548 unsigned char lo = GET_LO(from); \
549 unsigned char hi = GET_HI(from); \
553 if (*toP == toLim) { \
561 case 0x1: case 0x2: case 0x3: \
562 case 0x4: case 0x5: case 0x6: case 0x7: \
563 if (toLim - *toP < 2) { \
567 *(*toP)++ = ((lo >> 6) | (hi << 2) | UTF8_cval2); \
568 *(*toP)++ = ((lo & 0x3f) | 0x80); \
571 if (toLim - *toP < 3) { \
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); \
580 case 0xD8: case 0xD9: case 0xDA: case 0xDB: \
581 if (toLim - *toP < 4) { \
585 plane = (((hi & 0x3) << 2) | ((lo >> 6) & 0x3)) + 1; \
586 *(*toP)++ = ((plane >> 2) | UTF8_cval4); \
587 *(*toP)++ = (((lo >> 2) & 0xF) | ((plane & 0x3) << 4) | 0x80); \
589 lo2 = GET_LO(from); \
590 *(*toP)++ = (((lo & 0x3) << 4) \
591 | ((GET_HI(from) & 0x3) << 2) \
594 *(*toP)++ = ((lo2 & 0x3f) | 0x80); \
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) \
607 /* Avoid copying first half only of surrogate */ \
608 if (fromLim - *fromP > ((toLim - *toP) << 1) \
609 && (GET_HI(fromLim - 2) & 0xF8) == 0xD8) \
611 for (; *fromP != fromLim && *toP != toLim; *fromP += 2) \
612 *(*toP)++ = (GET_HI(*fromP) << 8) | GET_LO(*fromP); \
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])
620 DEFINE_UTF16_TO_UTF8(little2_)
621 DEFINE_UTF16_TO_UTF16(little2_)
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])
632 DEFINE_UTF16_TO_UTF8(big2_)
633 DEFINE_UTF16_TO_UTF16(big2_)
639 #define LITTLE2_BYTE_TYPE(enc, p) \
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])
652 static int PTRFASTCALL
653 little2_byteType(const ENCODING *enc, const char *p)
655 return LITTLE2_BYTE_TYPE(enc, p);
658 static int PTRFASTCALL
659 little2_byteToAscii(const ENCODING *enc, const char *p)
661 return LITTLE2_BYTE_TO_ASCII(enc, p);
665 little2_charMatches(const ENCODING *enc, const char *p, int c)
667 return LITTLE2_CHAR_MATCHES(enc, p, c);
670 static int PTRFASTCALL
671 little2_isNameMin(const ENCODING *enc, const char *p)
673 return LITTLE2_IS_NAME_CHAR_MINBPC(enc, p);
676 static int PTRFASTCALL
677 little2_isNmstrtMin(const ENCODING *enc, const char *p)
679 return LITTLE2_IS_NMSTRT_CHAR_MINBPC(enc, p);
683 #define VTABLE VTABLE1, little2_toUtf8, little2_toUtf16
685 #else /* not XML_MIN_SIZE */
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)
699 #define XML_TOK_IMPL_C
700 #include "xmltok_impl.c"
701 #undef XML_TOK_IMPL_C
708 #undef IS_NAME_CHAR_MINBPC
709 #undef IS_NMSTRT_CHAR
710 #undef IS_NMSTRT_CHAR_MINBPC
711 #undef IS_INVALID_CHAR
713 #endif /* not XML_MIN_SIZE */
717 static const struct normal_encoding little2_encoding_ns = {
719 #if BYTEORDER == 1234
726 #include "asciitab.h"
727 #include "latin1tab.h"
729 STANDARD_VTABLE(little2_)
734 static const struct normal_encoding little2_encoding = {
736 #if BYTEORDER == 1234
743 #define BT_COLON BT_NMSTRT
744 #include "asciitab.h"
746 #include "latin1tab.h"
748 STANDARD_VTABLE(little2_)
751 #if BYTEORDER != 4321
755 static const struct normal_encoding internal_little2_encoding_ns = {
758 #include "iasciitab.h"
759 #include "latin1tab.h"
761 STANDARD_VTABLE(little2_)
766 static const struct normal_encoding internal_little2_encoding = {
769 #define BT_COLON BT_NMSTRT
770 #include "iasciitab.h"
772 #include "latin1tab.h"
774 STANDARD_VTABLE(little2_)
780 #define BIG2_BYTE_TYPE(enc, p) \
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])
793 static int PTRFASTCALL
794 big2_byteType(const ENCODING *enc, const char *p)
796 return BIG2_BYTE_TYPE(enc, p);
799 static int PTRFASTCALL
800 big2_byteToAscii(const ENCODING *enc, const char *p)
802 return BIG2_BYTE_TO_ASCII(enc, p);
806 big2_charMatches(const ENCODING *enc, const char *p, int c)
808 return BIG2_CHAR_MATCHES(enc, p, c);
811 static int PTRFASTCALL
812 big2_isNameMin(const ENCODING *enc, const char *p)
814 return BIG2_IS_NAME_CHAR_MINBPC(enc, p);
817 static int PTRFASTCALL
818 big2_isNmstrtMin(const ENCODING *enc, const char *p)
820 return BIG2_IS_NMSTRT_CHAR_MINBPC(enc, p);
824 #define VTABLE VTABLE1, big2_toUtf8, big2_toUtf16
826 #else /* not XML_MIN_SIZE */
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)
840 #define XML_TOK_IMPL_C
841 #include "xmltok_impl.c"
842 #undef XML_TOK_IMPL_C
849 #undef IS_NAME_CHAR_MINBPC
850 #undef IS_NMSTRT_CHAR
851 #undef IS_NMSTRT_CHAR_MINBPC
852 #undef IS_INVALID_CHAR
854 #endif /* not XML_MIN_SIZE */
858 static const struct normal_encoding big2_encoding_ns = {
860 #if BYTEORDER == 4321
867 #include "asciitab.h"
868 #include "latin1tab.h"
870 STANDARD_VTABLE(big2_)
875 static const struct normal_encoding big2_encoding = {
877 #if BYTEORDER == 4321
884 #define BT_COLON BT_NMSTRT
885 #include "asciitab.h"
887 #include "latin1tab.h"
889 STANDARD_VTABLE(big2_)
892 #if BYTEORDER != 1234
896 static const struct normal_encoding internal_big2_encoding_ns = {
899 #include "iasciitab.h"
900 #include "latin1tab.h"
902 STANDARD_VTABLE(big2_)
907 static const struct normal_encoding internal_big2_encoding = {
910 #define BT_COLON BT_NMSTRT
911 #include "iasciitab.h"
913 #include "latin1tab.h"
915 STANDARD_VTABLE(big2_)
923 streqci(const char *s1, const char *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;
941 initUpdatePosition(const ENCODING *enc, const char *ptr,
942 const char *end, POSITION *pos)
944 normal_updatePosition(&utf8_encoding.enc, ptr, end, pos);
948 toAscii(const ENCODING *enc, const char *ptr, const char *end)
952 XmlUtf8Convert(enc, &ptr, end, &p, p + 1);
972 /* Return 1 if there's just optional white space or there's an S
973 followed by name=val.
976 parsePseudoAttribute(const ENCODING *enc,
979 const char **namePtr,
980 const char **nameEndPtr,
982 const char **nextTokPtr)
990 if (!isSpace(toAscii(enc, ptr, end))) {
995 ptr += enc->minBytesPerChar;
996 } while (isSpace(toAscii(enc, ptr, end)));
1003 c = toAscii(enc, ptr, end);
1008 if (c == ASCII_EQUALS) {
1015 ptr += enc->minBytesPerChar;
1016 } while (isSpace(c = toAscii(enc, ptr, end)));
1017 if (c != ASCII_EQUALS) {
1023 ptr += enc->minBytesPerChar;
1025 if (ptr == *namePtr) {
1029 ptr += enc->minBytesPerChar;
1030 c = toAscii(enc, ptr, end);
1031 while (isSpace(c)) {
1032 ptr += enc->minBytesPerChar;
1033 c = toAscii(enc, ptr, end);
1035 if (c != ASCII_QUOT && c != ASCII_APOS) {
1040 ptr += enc->minBytesPerChar;
1042 for (;; ptr += enc->minBytesPerChar) {
1043 c = toAscii(enc, ptr, end);
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
1051 && c != ASCII_UNDERSCORE) {
1056 *nextTokPtr = ptr + enc->minBytesPerChar;
1060 static const char KW_version[] = {
1061 ASCII_v, ASCII_e, ASCII_r, ASCII_s, ASCII_i, ASCII_o, ASCII_n, '\0'
1064 static const char KW_encoding[] = {
1065 ASCII_e, ASCII_n, ASCII_c, ASCII_o, ASCII_d, ASCII_i, ASCII_n, ASCII_g, '\0'
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'
1073 static const char KW_yes[] = {
1074 ASCII_y, ASCII_e, ASCII_s, '\0'
1077 static const char KW_no[] = {
1078 ASCII_n, ASCII_o, '\0'
1082 doParseXmlDecl(const ENCODING *(*encodingFinder)(const ENCODING *,
1085 int isGeneralTextEntity,
1086 const ENCODING *enc,
1089 const char **badPtr,
1090 const char **versionPtr,
1091 const char **versionEndPtr,
1092 const char **encodingName,
1093 const ENCODING **encoding,
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)
1106 if (!XmlNameMatchesAscii(enc, name, nameEnd, KW_version)) {
1107 if (!isGeneralTextEntity) {
1116 *versionEndPtr = ptr;
1117 if (!parsePseudoAttribute(enc, ptr, end, &name, &nameEnd, &val, &ptr)) {
1122 if (isGeneralTextEntity) {
1123 /* a TextDecl must have an EncodingDecl */
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)) {
1137 *encodingName = val;
1139 *encoding = encodingFinder(enc, val, ptr - enc->minBytesPerChar);
1140 if (!parsePseudoAttribute(enc, ptr, end, &name, &nameEnd, &val, &ptr)) {
1147 if (!XmlNameMatchesAscii(enc, name, nameEnd, KW_standalone)
1148 || isGeneralTextEntity) {
1152 if (XmlNameMatchesAscii(enc, val, ptr - enc->minBytesPerChar, KW_yes)) {
1156 else if (XmlNameMatchesAscii(enc, val, ptr - enc->minBytesPerChar, KW_no)) {
1164 while (isSpace(toAscii(enc, ptr, end)))
1165 ptr += enc->minBytesPerChar;
1174 checkCharRefNumber(int result)
1176 switch (result >> 8) {
1177 case 0xD8: case 0xD9: case 0xDA: case 0xDB:
1178 case 0xDC: case 0xDD: case 0xDE: case 0xDF:
1181 if (latin1_encoding.type[result] == BT_NONXML)
1185 if (result == 0xFFFE || result == 0xFFFF)
1193 XmlUtf8Encode(int c, char *buf)
1196 /* minN is minimum legal resulting value for N byte sequence */
1205 buf[0] = (char)(c | UTF8_cval1);
1209 buf[0] = (char)((c >> 6) | UTF8_cval2);
1210 buf[1] = (char)((c & 0x3f) | 0x80);
1214 buf[0] = (char)((c >> 12) | UTF8_cval3);
1215 buf[1] = (char)(((c >> 6) & 0x3f) | 0x80);
1216 buf[2] = (char)((c & 0x3f) | 0x80);
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);
1230 XmlUtf16Encode(int charNum, unsigned short *buf)
1234 if (charNum < 0x10000) {
1235 buf[0] = (unsigned short)charNum;
1238 if (charNum < 0x110000) {
1240 buf[0] = (unsigned short)((charNum >> 10) + 0xD800);
1241 buf[1] = (unsigned short)((charNum & 0x3FF) + 0xDC00);
1247 struct unknown_encoding {
1248 struct normal_encoding normal;
1251 unsigned short utf16[256];
1255 #define AS_UNKNOWN_ENCODING(enc) ((const struct unknown_encoding *) (enc))
1258 XmlSizeOfUnknownEncoding(void)
1260 return sizeof(struct unknown_encoding);
1263 static int PTRFASTCALL
1264 unknown_isName(const ENCODING *enc, const char *p)
1266 const struct unknown_encoding *uenc = AS_UNKNOWN_ENCODING(enc);
1267 int c = uenc->convert(uenc->userData, p);
1270 return UCS2_GET_NAMING(namePages, c >> 8, c & 0xFF);
1273 static int PTRFASTCALL
1274 unknown_isNmstrt(const ENCODING *enc, const char *p)
1276 const struct unknown_encoding *uenc = AS_UNKNOWN_ENCODING(enc);
1277 int c = uenc->convert(uenc->userData, p);
1280 return UCS2_GET_NAMING(nmstrtPages, c >> 8, c & 0xFF);
1283 static int PTRFASTCALL
1284 unknown_isInvalid(const ENCODING *enc, const char *p)
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;
1292 unknown_toUtf8(const ENCODING *enc,
1293 const char **fromP, const char *fromLim,
1294 char **toP, const char *toLim)
1296 const struct unknown_encoding *uenc = AS_UNKNOWN_ENCODING(enc);
1297 char buf[XML_UTF8_ENCODE_MAX];
1301 if (*fromP == fromLim)
1303 utf8 = uenc->utf8[(unsigned char)**fromP];
1306 int c = uenc->convert(uenc->userData, *fromP);
1307 n = XmlUtf8Encode(c, buf);
1308 if (n > toLim - *toP)
1311 *fromP += (AS_NORMAL_ENCODING(enc)->type[(unsigned char)**fromP]
1315 if (n > toLim - *toP)
1320 *(*toP)++ = *utf8++;
1326 unknown_toUtf16(const ENCODING *enc,
1327 const char **fromP, const char *fromLim,
1328 unsigned short **toP, const unsigned short *toLim)
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];
1334 c = (unsigned short)
1335 uenc->convert(uenc->userData, *fromP);
1336 *fromP += (AS_NORMAL_ENCODING(enc)->type[(unsigned char)**fromP]
1346 XmlInitUnknownEncoding(void *mem,
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
1360 for (i = 0; i < 256; i++) {
1363 e->normal.type[i] = BT_MALFORM;
1364 /* This shouldn't really get used. */
1365 e->utf16[i] = 0xFFFF;
1372 e->normal.type[i] = (unsigned char)(BT_LEAD2 - (c + 2));
1376 else if (c < 0x80) {
1377 if (latin1_encoding.type[c] != BT_OTHER
1378 && latin1_encoding.type[c] != BT_NONXML
1381 e->normal.type[i] = latin1_encoding.type[c];
1383 e->utf8[i][1] = (char)c;
1384 e->utf16[i] = (unsigned short)(c == 0 ? 0xFFFF : c);
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;
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;
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;
1406 e->userData = userData;
1407 e->convert = 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;
1419 e->normal.enc.utf8Convert = unknown_toUtf8;
1420 e->normal.enc.utf16Convert = unknown_toUtf16;
1421 return &(e->normal.enc);
1424 /* If this enumeration is changed, getEncodingIndex and encodings
1425 must also be changed. */
1434 /* must match encodingNames up to here */
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'
1442 static const char KW_US_ASCII[] = {
1443 ASCII_U, ASCII_S, ASCII_MINUS, ASCII_A, ASCII_S, ASCII_C, ASCII_I, ASCII_I,
1446 static const char KW_UTF_8[] = {
1447 ASCII_U, ASCII_T, ASCII_F, ASCII_MINUS, ASCII_8, '\0'
1449 static const char KW_UTF_16[] = {
1450 ASCII_U, ASCII_T, ASCII_F, ASCII_MINUS, ASCII_1, ASCII_6, '\0'
1452 static const char KW_UTF_16BE[] = {
1453 ASCII_U, ASCII_T, ASCII_F, ASCII_MINUS, ASCII_1, ASCII_6, ASCII_B, ASCII_E,
1456 static const char KW_UTF_16LE[] = {
1457 ASCII_U, ASCII_T, ASCII_F, ASCII_MINUS, ASCII_1, ASCII_6, ASCII_L, ASCII_E,
1462 getEncodingIndex(const char *name)
1464 static const char * const encodingNames[] = {
1475 for (i = 0; i < (int)(sizeof(encodingNames)/sizeof(encodingNames[0])); i++)
1476 if (streqci(name, encodingNames[i]))
1481 /* For binary compatibility, we store the index of the encoding
1482 specified at initialization in the isUtf16 member.
1485 #define INIT_ENC_INDEX(enc) ((int)(enc)->initEnc.isUtf16)
1486 #define SET_INIT_ENC_INDEX(enc, i) ((enc)->initEnc.isUtf16 = (char)i)
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.
1497 initScan(const ENCODING * const *encodingTable,
1498 const INIT_ENCODING *enc,
1502 const char **nextTokPtr)
1504 const ENCODING **encPtr;
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;
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)) {
1522 return XML_TOK_PARTIAL;
1524 switch ((unsigned char)*ptr) {
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)
1534 return XML_TOK_PARTIAL;
1538 switch (((unsigned char)ptr[0] << 8) | (unsigned char)ptr[1]) {
1540 if (INIT_ENC_INDEX(enc) == ISO_8859_1_ENC
1541 && state == XML_CONTENT_STATE)
1543 *nextTokPtr = ptr + 2;
1544 *encPtr = encodingTable[UTF_16BE_ENC];
1546 /* 00 3C is handled in the default case */
1548 if ((INIT_ENC_INDEX(enc) == UTF_16BE_ENC
1549 || INIT_ENC_INDEX(enc) == UTF_16_ENC)
1550 && state == XML_CONTENT_STATE)
1552 *encPtr = encodingTable[UTF_16LE_ENC];
1553 return XmlTok(*encPtr, state, ptr, end, nextTokPtr);
1555 if (INIT_ENC_INDEX(enc) == ISO_8859_1_ENC
1556 && state == XML_CONTENT_STATE)
1558 *nextTokPtr = ptr + 2;
1559 *encPtr = encodingTable[UTF_16LE_ENC];
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.
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)
1576 return XML_TOK_PARTIAL;
1577 if ((unsigned char)ptr[2] == 0xBF) {
1578 *nextTokPtr = ptr + 3;
1579 *encPtr = encodingTable[UTF_8_ENC];
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
1591 if (state == XML_CONTENT_STATE && INIT_ENC_INDEX(enc) == UTF_16LE_ENC)
1593 *encPtr = encodingTable[UTF_16BE_ENC];
1594 return XmlTok(*encPtr, state, ptr, end, nextTokPtr);
1596 else if (ptr[1] == '\0') {
1597 /* We could recover here in the case:
1598 - parsing an external entity
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.
1606 if (state == XML_CONTENT_STATE)
1608 *encPtr = encodingTable[UTF_16LE_ENC];
1609 return XmlTok(*encPtr, state, ptr, end, nextTokPtr);
1614 *encPtr = encodingTable[INIT_ENC_INDEX(enc)];
1615 return XmlTok(*encPtr, state, ptr, end, nextTokPtr);
1621 #define XML_TOK_NS_C
1622 #include "xmltok_ns.c"
1629 #define NS(x) x ## NS
1630 #define ns(x) x ## _ns
1632 #define XML_TOK_NS_C
1633 #include "xmltok_ns.c"
1640 XmlInitUnknownEncodingNS(void *mem,
1645 ENCODING *enc = XmlInitUnknownEncoding(mem, table, convert, userData);
1647 ((struct normal_encoding *)enc)->type[ASCII_COLON] = BT_COLON;