1
/* Copyright (c) 1998, 1999 Thai Open Source Software Center Ltd
2
See the file COPYING for copying permission.
5
#ifdef COMPILED_FROM_DSP
7
#elif defined(MACOS_CLASSIC)
9
#elif defined(__amigaos4__)
10
#include "amigaconfig.h"
12
#ifdef HAVE_EXPAT_CONFIG_H
13
#include <expat_config.h>
15
#endif /* ndef COMPILED_FROM_DSP */
19
#include "expat_external.h"
25
#define IGNORE_SECTION_TOK_VTABLE , PREFIX(ignoreSectionTok)
27
#define IGNORE_SECTION_TOK_VTABLE /* as nothing */
31
{ PREFIX(prologTok), PREFIX(contentTok), \
32
PREFIX(cdataSectionTok) IGNORE_SECTION_TOK_VTABLE }, \
33
{ PREFIX(attributeValueTok), PREFIX(entityValueTok) }, \
35
PREFIX(nameMatchesAscii), \
39
PREFIX(charRefNumber), \
40
PREFIX(predefinedEntityName), \
41
PREFIX(updatePosition), \
44
#define VTABLE VTABLE1, PREFIX(toUtf8), PREFIX(toUtf16)
46
#define UCS2_GET_NAMING(pages, hi, lo) \
47
(namingBitmap[(pages[hi] << 3) + ((lo) >> 5)] & (1 << ((lo) & 0x1F)))
49
/* A 2 byte UTF-8 representation splits the characters 11 bits between
50
the bottom 5 and 6 bits of the bytes. We need 8 bits to index into
51
pages, 3 bits to add to that index and 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)))
59
/* A 3 byte UTF-8 representation splits the characters 16 bits between
60
the bottom 4, 6 and 6 bits of the bytes. We need 8 bits to index
61
into pages, 3 bits to add to that index and 5 bits to generate the
64
#define UTF8_GET_NAMING3(pages, byte) \
65
(namingBitmap[((pages)[((((byte)[0]) & 0xF) << 4) \
66
+ ((((byte)[1]) >> 2) & 0xF)] \
68
+ ((((byte)[1]) & 3) << 1) \
69
+ ((((byte)[2]) >> 5) & 1)] \
70
& (1 << (((byte)[2]) & 0x1F)))
72
#define UTF8_GET_NAMING(pages, p, n) \
74
? UTF8_GET_NAMING2(pages, (const unsigned char *)(p)) \
76
? UTF8_GET_NAMING3(pages, (const unsigned char *)(p)) \
79
/* Detection of invalid UTF-8 sequences is based on Table 3.1B
80
of Unicode 3.2: http://www.unicode.org/unicode/reports/tr28/
81
with the additional restriction of not allowing the Unicode
82
code points 0xFFFF and 0xFFFE (sequences EF,BF,BF and EF,BF,BE).
83
Implementation details:
84
(A & 0x80) == 0 means A < 0x80
86
(A & 0xC0) == 0xC0 means A > 0xBF
89
#define UTF8_INVALID2(p) \
90
((*p) < 0xC2 || ((p)[1] & 0x80) == 0 || ((p)[1] & 0xC0) == 0xC0)
92
#define UTF8_INVALID3(p) \
93
(((p)[2] & 0x80) == 0 \
95
((*p) == 0xEF && (p)[1] == 0xBF \
99
((p)[2] & 0xC0) == 0xC0) \
103
(p)[1] < 0xA0 || ((p)[1] & 0xC0) == 0xC0 \
105
((p)[1] & 0x80) == 0 \
107
((*p) == 0xED ? (p)[1] > 0x9F : ((p)[1] & 0xC0) == 0xC0)))
109
#define UTF8_INVALID4(p) \
110
(((p)[3] & 0x80) == 0 || ((p)[3] & 0xC0) == 0xC0 \
112
((p)[2] & 0x80) == 0 || ((p)[2] & 0xC0) == 0xC0 \
116
(p)[1] < 0x90 || ((p)[1] & 0xC0) == 0xC0 \
118
((p)[1] & 0x80) == 0 \
120
((*p) == 0xF4 ? (p)[1] > 0x8F : ((p)[1] & 0xC0) == 0xC0)))
122
static int PTRFASTCALL
123
isNever(const ENCODING *enc, const char *p)
128
static int PTRFASTCALL
129
utf8_isName2(const ENCODING *enc, const char *p)
131
return UTF8_GET_NAMING2(namePages, (const unsigned char *)p);
134
static int PTRFASTCALL
135
utf8_isName3(const ENCODING *enc, const char *p)
137
return UTF8_GET_NAMING3(namePages, (const unsigned char *)p);
140
#define utf8_isName4 isNever
142
static int PTRFASTCALL
143
utf8_isNmstrt2(const ENCODING *enc, const char *p)
145
return UTF8_GET_NAMING2(nmstrtPages, (const unsigned char *)p);
148
static int PTRFASTCALL
149
utf8_isNmstrt3(const ENCODING *enc, const char *p)
151
return UTF8_GET_NAMING3(nmstrtPages, (const unsigned char *)p);
154
#define utf8_isNmstrt4 isNever
156
static int PTRFASTCALL
157
utf8_isInvalid2(const ENCODING *enc, const char *p)
159
return UTF8_INVALID2((const unsigned char *)p);
162
static int PTRFASTCALL
163
utf8_isInvalid3(const ENCODING *enc, const char *p)
165
return UTF8_INVALID3((const unsigned char *)p);
168
static int PTRFASTCALL
169
utf8_isInvalid4(const ENCODING *enc, const char *p)
171
return UTF8_INVALID4((const unsigned char *)p);
174
struct normal_encoding {
176
unsigned char type[256];
178
int (PTRFASTCALL *byteType)(const ENCODING *, const char *);
179
int (PTRFASTCALL *isNameMin)(const ENCODING *, const char *);
180
int (PTRFASTCALL *isNmstrtMin)(const ENCODING *, const char *);
181
int (PTRFASTCALL *byteToAscii)(const ENCODING *, const char *);
182
int (PTRCALL *charMatches)(const ENCODING *, const char *, int);
183
#endif /* XML_MIN_SIZE */
184
int (PTRFASTCALL *isName2)(const ENCODING *, const char *);
185
int (PTRFASTCALL *isName3)(const ENCODING *, const char *);
186
int (PTRFASTCALL *isName4)(const ENCODING *, const char *);
187
int (PTRFASTCALL *isNmstrt2)(const ENCODING *, const char *);
188
int (PTRFASTCALL *isNmstrt3)(const ENCODING *, const char *);
189
int (PTRFASTCALL *isNmstrt4)(const ENCODING *, const char *);
190
int (PTRFASTCALL *isInvalid2)(const ENCODING *, const char *);
191
int (PTRFASTCALL *isInvalid3)(const ENCODING *, const char *);
192
int (PTRFASTCALL *isInvalid4)(const ENCODING *, const char *);
195
#define AS_NORMAL_ENCODING(enc) ((const struct normal_encoding *) (enc))
199
#define STANDARD_VTABLE(E) \
208
#define STANDARD_VTABLE(E) /* as nothing */
212
#define NORMAL_VTABLE(E) \
223
static int FASTCALL checkCharRefNumber(int);
225
#include "xmltok_impl.h"
229
#define sb_isNameMin isNever
230
#define sb_isNmstrtMin isNever
234
#define MINBPC(enc) ((enc)->minBytesPerChar)
236
/* minimum bytes per character */
237
#define MINBPC(enc) 1
240
#define SB_BYTE_TYPE(enc, p) \
241
(((struct normal_encoding *)(enc))->type[(unsigned char)*(p)])
244
static int PTRFASTCALL
245
sb_byteType(const ENCODING *enc, const char *p)
247
return SB_BYTE_TYPE(enc, p);
249
#define BYTE_TYPE(enc, p) \
250
(AS_NORMAL_ENCODING(enc)->byteType(enc, p))
252
#define BYTE_TYPE(enc, p) SB_BYTE_TYPE(enc, p)
256
#define BYTE_TO_ASCII(enc, p) \
257
(AS_NORMAL_ENCODING(enc)->byteToAscii(enc, p))
258
static int PTRFASTCALL
259
sb_byteToAscii(const ENCODING *enc, const char *p)
264
#define BYTE_TO_ASCII(enc, p) (*(p))
267
#define IS_NAME_CHAR(enc, p, n) \
268
(AS_NORMAL_ENCODING(enc)->isName ## n(enc, p))
269
#define IS_NMSTRT_CHAR(enc, p, n) \
270
(AS_NORMAL_ENCODING(enc)->isNmstrt ## n(enc, p))
271
#define IS_INVALID_CHAR(enc, p, n) \
272
(AS_NORMAL_ENCODING(enc)->isInvalid ## n(enc, p))
275
#define IS_NAME_CHAR_MINBPC(enc, p) \
276
(AS_NORMAL_ENCODING(enc)->isNameMin(enc, p))
277
#define IS_NMSTRT_CHAR_MINBPC(enc, p) \
278
(AS_NORMAL_ENCODING(enc)->isNmstrtMin(enc, p))
280
#define IS_NAME_CHAR_MINBPC(enc, p) (0)
281
#define IS_NMSTRT_CHAR_MINBPC(enc, p) (0)
285
#define CHAR_MATCHES(enc, p, c) \
286
(AS_NORMAL_ENCODING(enc)->charMatches(enc, p, c))
288
sb_charMatches(const ENCODING *enc, const char *p, int c)
293
/* c is an ASCII character */
294
#define CHAR_MATCHES(enc, p, c) (*(p) == c)
297
#define PREFIX(ident) normal_ ## ident
298
#include "xmltok_impl.c"
305
#undef IS_NAME_CHAR_MINBPC
306
#undef IS_NMSTRT_CHAR
307
#undef IS_NMSTRT_CHAR_MINBPC
308
#undef IS_INVALID_CHAR
310
enum { /* UTF8_cvalN is value of masked first byte of N byte sequence */
318
utf8_toUtf8(const ENCODING *enc,
319
const char **fromP, const char *fromLim,
320
char **toP, const char *toLim)
324
if (fromLim - *fromP > toLim - *toP) {
325
/* Avoid copying partial characters. */
326
for (fromLim = *fromP + (toLim - *toP); fromLim > *fromP; fromLim--)
327
if (((unsigned char)fromLim[-1] & 0xc0) != 0x80)
330
for (to = *toP, from = *fromP; from != fromLim; from++, to++)
337
utf8_toUtf16(const ENCODING *enc,
338
const char **fromP, const char *fromLim,
339
unsigned short **toP, const unsigned short *toLim)
341
unsigned short *to = *toP;
342
const char *from = *fromP;
343
while (from != fromLim && to != toLim) {
344
switch (((struct normal_encoding *)enc)->type[(unsigned char)*from]) {
346
*to++ = (unsigned short)(((from[0] & 0x1f) << 6) | (from[1] & 0x3f));
350
*to++ = (unsigned short)(((from[0] & 0xf) << 12)
351
| ((from[1] & 0x3f) << 6) | (from[2] & 0x3f));
359
n = ((from[0] & 0x7) << 18) | ((from[1] & 0x3f) << 12)
360
| ((from[2] & 0x3f) << 6) | (from[3] & 0x3f);
362
to[0] = (unsigned short)((n >> 10) | 0xD800);
363
to[1] = (unsigned short)((n & 0x3FF) | 0xDC00);
379
static const struct normal_encoding utf8_encoding_ns = {
380
{ VTABLE1, utf8_toUtf8, utf8_toUtf16, 1, 1, 0 },
382
#include "asciitab.h"
385
STANDARD_VTABLE(sb_) NORMAL_VTABLE(utf8_)
389
static const struct normal_encoding utf8_encoding = {
390
{ VTABLE1, utf8_toUtf8, utf8_toUtf16, 1, 1, 0 },
392
#define BT_COLON BT_NMSTRT
393
#include "asciitab.h"
397
STANDARD_VTABLE(sb_) NORMAL_VTABLE(utf8_)
402
static const struct normal_encoding internal_utf8_encoding_ns = {
403
{ VTABLE1, utf8_toUtf8, utf8_toUtf16, 1, 1, 0 },
405
#include "iasciitab.h"
408
STANDARD_VTABLE(sb_) NORMAL_VTABLE(utf8_)
413
static const struct normal_encoding internal_utf8_encoding = {
414
{ VTABLE1, utf8_toUtf8, utf8_toUtf16, 1, 1, 0 },
416
#define BT_COLON BT_NMSTRT
417
#include "iasciitab.h"
421
STANDARD_VTABLE(sb_) NORMAL_VTABLE(utf8_)
425
latin1_toUtf8(const ENCODING *enc,
426
const char **fromP, const char *fromLim,
427
char **toP, const char *toLim)
431
if (*fromP == fromLim)
433
c = (unsigned char)**fromP;
435
if (toLim - *toP < 2)
437
*(*toP)++ = (char)((c >> 6) | UTF8_cval2);
438
*(*toP)++ = (char)((c & 0x3f) | 0x80);
444
*(*toP)++ = *(*fromP)++;
450
latin1_toUtf16(const ENCODING *enc,
451
const char **fromP, const char *fromLim,
452
unsigned short **toP, const unsigned short *toLim)
454
while (*fromP != fromLim && *toP != toLim)
455
*(*toP)++ = (unsigned char)*(*fromP)++;
460
static const struct normal_encoding latin1_encoding_ns = {
461
{ VTABLE1, latin1_toUtf8, latin1_toUtf16, 1, 0, 0 },
463
#include "asciitab.h"
464
#include "latin1tab.h"
471
static const struct normal_encoding latin1_encoding = {
472
{ VTABLE1, latin1_toUtf8, latin1_toUtf16, 1, 0, 0 },
474
#define BT_COLON BT_NMSTRT
475
#include "asciitab.h"
477
#include "latin1tab.h"
483
ascii_toUtf8(const ENCODING *enc,
484
const char **fromP, const char *fromLim,
485
char **toP, const char *toLim)
487
while (*fromP != fromLim && *toP != toLim)
488
*(*toP)++ = *(*fromP)++;
493
static const struct normal_encoding ascii_encoding_ns = {
494
{ VTABLE1, ascii_toUtf8, latin1_toUtf16, 1, 1, 0 },
496
#include "asciitab.h"
504
static const struct normal_encoding ascii_encoding = {
505
{ VTABLE1, ascii_toUtf8, latin1_toUtf16, 1, 1, 0 },
507
#define BT_COLON BT_NMSTRT
508
#include "asciitab.h"
515
static int PTRFASTCALL
516
unicode_byte_type(char hi, char lo)
518
switch ((unsigned char)hi) {
519
case 0xD8: case 0xD9: case 0xDA: case 0xDB:
521
case 0xDC: case 0xDD: case 0xDE: case 0xDF:
524
switch ((unsigned char)lo) {
534
#define DEFINE_UTF16_TO_UTF8(E) \
535
static void PTRCALL \
536
E ## toUtf8(const ENCODING *enc, \
537
const char **fromP, const char *fromLim, \
538
char **toP, const char *toLim) \
541
for (from = *fromP; from != fromLim; from += 2) { \
544
unsigned char lo = GET_LO(from); \
545
unsigned char hi = GET_HI(from); \
549
if (*toP == toLim) { \
557
case 0x1: case 0x2: case 0x3: \
558
case 0x4: case 0x5: case 0x6: case 0x7: \
559
if (toLim - *toP < 2) { \
563
*(*toP)++ = ((lo >> 6) | (hi << 2) | UTF8_cval2); \
564
*(*toP)++ = ((lo & 0x3f) | 0x80); \
567
if (toLim - *toP < 3) { \
571
/* 16 bits divided 4, 6, 6 amongst 3 bytes */ \
572
*(*toP)++ = ((hi >> 4) | UTF8_cval3); \
573
*(*toP)++ = (((hi & 0xf) << 2) | (lo >> 6) | 0x80); \
574
*(*toP)++ = ((lo & 0x3f) | 0x80); \
576
case 0xD8: case 0xD9: case 0xDA: case 0xDB: \
577
if (toLim - *toP < 4) { \
581
plane = (((hi & 0x3) << 2) | ((lo >> 6) & 0x3)) + 1; \
582
*(*toP)++ = ((plane >> 2) | UTF8_cval4); \
583
*(*toP)++ = (((lo >> 2) & 0xF) | ((plane & 0x3) << 4) | 0x80); \
585
lo2 = GET_LO(from); \
586
*(*toP)++ = (((lo & 0x3) << 4) \
587
| ((GET_HI(from) & 0x3) << 2) \
590
*(*toP)++ = ((lo2 & 0x3f) | 0x80); \
597
#define DEFINE_UTF16_TO_UTF16(E) \
598
static void PTRCALL \
599
E ## toUtf16(const ENCODING *enc, \
600
const char **fromP, const char *fromLim, \
601
unsigned short **toP, const unsigned short *toLim) \
603
/* Avoid copying first half only of surrogate */ \
604
if (fromLim - *fromP > ((toLim - *toP) << 1) \
605
&& (GET_HI(fromLim - 2) & 0xF8) == 0xD8) \
607
for (; *fromP != fromLim && *toP != toLim; *fromP += 2) \
608
*(*toP)++ = (GET_HI(*fromP) << 8) | GET_LO(*fromP); \
611
#define SET2(ptr, ch) \
612
(((ptr)[0] = ((ch) & 0xff)), ((ptr)[1] = ((ch) >> 8)))
613
#define GET_LO(ptr) ((unsigned char)(ptr)[0])
614
#define GET_HI(ptr) ((unsigned char)(ptr)[1])
616
DEFINE_UTF16_TO_UTF8(little2_)
617
DEFINE_UTF16_TO_UTF16(little2_)
623
#define SET2(ptr, ch) \
624
(((ptr)[0] = ((ch) >> 8)), ((ptr)[1] = ((ch) & 0xFF)))
625
#define GET_LO(ptr) ((unsigned char)(ptr)[1])
626
#define GET_HI(ptr) ((unsigned char)(ptr)[0])
628
DEFINE_UTF16_TO_UTF8(big2_)
629
DEFINE_UTF16_TO_UTF16(big2_)
635
#define LITTLE2_BYTE_TYPE(enc, p) \
637
? ((struct normal_encoding *)(enc))->type[(unsigned char)*(p)] \
638
: unicode_byte_type((p)[1], (p)[0]))
639
#define LITTLE2_BYTE_TO_ASCII(enc, p) ((p)[1] == 0 ? (p)[0] : -1)
640
#define LITTLE2_CHAR_MATCHES(enc, p, c) ((p)[1] == 0 && (p)[0] == c)
641
#define LITTLE2_IS_NAME_CHAR_MINBPC(enc, p) \
642
UCS2_GET_NAMING(namePages, (unsigned char)p[1], (unsigned char)p[0])
643
#define LITTLE2_IS_NMSTRT_CHAR_MINBPC(enc, p) \
644
UCS2_GET_NAMING(nmstrtPages, (unsigned char)p[1], (unsigned char)p[0])
648
static int PTRFASTCALL
649
little2_byteType(const ENCODING *enc, const char *p)
651
return LITTLE2_BYTE_TYPE(enc, p);
654
static int PTRFASTCALL
655
little2_byteToAscii(const ENCODING *enc, const char *p)
657
return LITTLE2_BYTE_TO_ASCII(enc, p);
661
little2_charMatches(const ENCODING *enc, const char *p, int c)
663
return LITTLE2_CHAR_MATCHES(enc, p, c);
666
static int PTRFASTCALL
667
little2_isNameMin(const ENCODING *enc, const char *p)
669
return LITTLE2_IS_NAME_CHAR_MINBPC(enc, p);
672
static int PTRFASTCALL
673
little2_isNmstrtMin(const ENCODING *enc, const char *p)
675
return LITTLE2_IS_NMSTRT_CHAR_MINBPC(enc, p);
679
#define VTABLE VTABLE1, little2_toUtf8, little2_toUtf16
681
#else /* not XML_MIN_SIZE */
684
#define PREFIX(ident) little2_ ## ident
685
#define MINBPC(enc) 2
686
/* CHAR_MATCHES is guaranteed to have MINBPC bytes available. */
687
#define BYTE_TYPE(enc, p) LITTLE2_BYTE_TYPE(enc, p)
688
#define BYTE_TO_ASCII(enc, p) LITTLE2_BYTE_TO_ASCII(enc, p)
689
#define CHAR_MATCHES(enc, p, c) LITTLE2_CHAR_MATCHES(enc, p, c)
690
#define IS_NAME_CHAR(enc, p, n) 0
691
#define IS_NAME_CHAR_MINBPC(enc, p) LITTLE2_IS_NAME_CHAR_MINBPC(enc, p)
692
#define IS_NMSTRT_CHAR(enc, p, n) (0)
693
#define IS_NMSTRT_CHAR_MINBPC(enc, p) LITTLE2_IS_NMSTRT_CHAR_MINBPC(enc, p)
695
#include "xmltok_impl.c"
702
#undef IS_NAME_CHAR_MINBPC
703
#undef IS_NMSTRT_CHAR
704
#undef IS_NMSTRT_CHAR_MINBPC
705
#undef IS_INVALID_CHAR
707
#endif /* not XML_MIN_SIZE */
711
static const struct normal_encoding little2_encoding_ns = {
713
#if BYTEORDER == 1234
720
#include "asciitab.h"
721
#include "latin1tab.h"
723
STANDARD_VTABLE(little2_)
728
static const struct normal_encoding little2_encoding = {
730
#if BYTEORDER == 1234
737
#define BT_COLON BT_NMSTRT
738
#include "asciitab.h"
740
#include "latin1tab.h"
742
STANDARD_VTABLE(little2_)
745
#if BYTEORDER != 4321
749
static const struct normal_encoding internal_little2_encoding_ns = {
752
#include "iasciitab.h"
753
#include "latin1tab.h"
755
STANDARD_VTABLE(little2_)
760
static const struct normal_encoding internal_little2_encoding = {
763
#define BT_COLON BT_NMSTRT
764
#include "iasciitab.h"
766
#include "latin1tab.h"
768
STANDARD_VTABLE(little2_)
774
#define BIG2_BYTE_TYPE(enc, p) \
776
? ((struct normal_encoding *)(enc))->type[(unsigned char)(p)[1]] \
777
: unicode_byte_type((p)[0], (p)[1]))
778
#define BIG2_BYTE_TO_ASCII(enc, p) ((p)[0] == 0 ? (p)[1] : -1)
779
#define BIG2_CHAR_MATCHES(enc, p, c) ((p)[0] == 0 && (p)[1] == c)
780
#define BIG2_IS_NAME_CHAR_MINBPC(enc, p) \
781
UCS2_GET_NAMING(namePages, (unsigned char)p[0], (unsigned char)p[1])
782
#define BIG2_IS_NMSTRT_CHAR_MINBPC(enc, p) \
783
UCS2_GET_NAMING(nmstrtPages, (unsigned char)p[0], (unsigned char)p[1])
787
static int PTRFASTCALL
788
big2_byteType(const ENCODING *enc, const char *p)
790
return BIG2_BYTE_TYPE(enc, p);
793
static int PTRFASTCALL
794
big2_byteToAscii(const ENCODING *enc, const char *p)
796
return BIG2_BYTE_TO_ASCII(enc, p);
800
big2_charMatches(const ENCODING *enc, const char *p, int c)
802
return BIG2_CHAR_MATCHES(enc, p, c);
805
static int PTRFASTCALL
806
big2_isNameMin(const ENCODING *enc, const char *p)
808
return BIG2_IS_NAME_CHAR_MINBPC(enc, p);
811
static int PTRFASTCALL
812
big2_isNmstrtMin(const ENCODING *enc, const char *p)
814
return BIG2_IS_NMSTRT_CHAR_MINBPC(enc, p);
818
#define VTABLE VTABLE1, big2_toUtf8, big2_toUtf16
820
#else /* not XML_MIN_SIZE */
823
#define PREFIX(ident) big2_ ## ident
824
#define MINBPC(enc) 2
825
/* CHAR_MATCHES is guaranteed to have MINBPC bytes available. */
826
#define BYTE_TYPE(enc, p) BIG2_BYTE_TYPE(enc, p)
827
#define BYTE_TO_ASCII(enc, p) BIG2_BYTE_TO_ASCII(enc, p)
828
#define CHAR_MATCHES(enc, p, c) BIG2_CHAR_MATCHES(enc, p, c)
829
#define IS_NAME_CHAR(enc, p, n) 0
830
#define IS_NAME_CHAR_MINBPC(enc, p) BIG2_IS_NAME_CHAR_MINBPC(enc, p)
831
#define IS_NMSTRT_CHAR(enc, p, n) (0)
832
#define IS_NMSTRT_CHAR_MINBPC(enc, p) BIG2_IS_NMSTRT_CHAR_MINBPC(enc, p)
834
#include "xmltok_impl.c"
841
#undef IS_NAME_CHAR_MINBPC
842
#undef IS_NMSTRT_CHAR
843
#undef IS_NMSTRT_CHAR_MINBPC
844
#undef IS_INVALID_CHAR
846
#endif /* not XML_MIN_SIZE */
850
static const struct normal_encoding big2_encoding_ns = {
852
#if BYTEORDER == 4321
859
#include "asciitab.h"
860
#include "latin1tab.h"
862
STANDARD_VTABLE(big2_)
867
static const struct normal_encoding big2_encoding = {
869
#if BYTEORDER == 4321
876
#define BT_COLON BT_NMSTRT
877
#include "asciitab.h"
879
#include "latin1tab.h"
881
STANDARD_VTABLE(big2_)
884
#if BYTEORDER != 1234
888
static const struct normal_encoding internal_big2_encoding_ns = {
891
#include "iasciitab.h"
892
#include "latin1tab.h"
894
STANDARD_VTABLE(big2_)
899
static const struct normal_encoding internal_big2_encoding = {
902
#define BT_COLON BT_NMSTRT
903
#include "iasciitab.h"
905
#include "latin1tab.h"
907
STANDARD_VTABLE(big2_)
915
streqci(const char *s1, const char *s2)
920
if (ASCII_a <= c1 && c1 <= ASCII_z)
921
c1 += ASCII_A - ASCII_a;
922
if (ASCII_a <= c2 && c2 <= ASCII_z)
923
c2 += ASCII_A - ASCII_a;
933
initUpdatePosition(const ENCODING *enc, const char *ptr,
934
const char *end, POSITION *pos)
936
normal_updatePosition(&utf8_encoding.enc, ptr, end, pos);
940
toAscii(const ENCODING *enc, const char *ptr, const char *end)
944
XmlUtf8Convert(enc, &ptr, end, &p, p + 1);
964
/* Return 1 if there's just optional white space or there's an S
965
followed by name=val.
968
parsePseudoAttribute(const ENCODING *enc,
971
const char **namePtr,
972
const char **nameEndPtr,
974
const char **nextTokPtr)
982
if (!isSpace(toAscii(enc, ptr, end))) {
987
ptr += enc->minBytesPerChar;
988
} while (isSpace(toAscii(enc, ptr, end)));
995
c = toAscii(enc, ptr, end);
1000
if (c == ASCII_EQUALS) {
1007
ptr += enc->minBytesPerChar;
1008
} while (isSpace(c = toAscii(enc, ptr, end)));
1009
if (c != ASCII_EQUALS) {
1015
ptr += enc->minBytesPerChar;
1017
if (ptr == *namePtr) {
1021
ptr += enc->minBytesPerChar;
1022
c = toAscii(enc, ptr, end);
1023
while (isSpace(c)) {
1024
ptr += enc->minBytesPerChar;
1025
c = toAscii(enc, ptr, end);
1027
if (c != ASCII_QUOT && c != ASCII_APOS) {
1032
ptr += enc->minBytesPerChar;
1034
for (;; ptr += enc->minBytesPerChar) {
1035
c = toAscii(enc, ptr, end);
1038
if (!(ASCII_a <= c && c <= ASCII_z)
1039
&& !(ASCII_A <= c && c <= ASCII_Z)
1040
&& !(ASCII_0 <= c && c <= ASCII_9)
1041
&& c != ASCII_PERIOD
1043
&& c != ASCII_UNDERSCORE) {
1048
*nextTokPtr = ptr + enc->minBytesPerChar;
1052
static const char KW_version[] = {
1053
ASCII_v, ASCII_e, ASCII_r, ASCII_s, ASCII_i, ASCII_o, ASCII_n, '\0'
1056
static const char KW_encoding[] = {
1057
ASCII_e, ASCII_n, ASCII_c, ASCII_o, ASCII_d, ASCII_i, ASCII_n, ASCII_g, '\0'
1060
static const char KW_standalone[] = {
1061
ASCII_s, ASCII_t, ASCII_a, ASCII_n, ASCII_d, ASCII_a, ASCII_l, ASCII_o,
1062
ASCII_n, ASCII_e, '\0'
1065
static const char KW_yes[] = {
1066
ASCII_y, ASCII_e, ASCII_s, '\0'
1069
static const char KW_no[] = {
1070
ASCII_n, ASCII_o, '\0'
1074
doParseXmlDecl(const ENCODING *(*encodingFinder)(const ENCODING *,
1077
int isGeneralTextEntity,
1078
const ENCODING *enc,
1081
const char **badPtr,
1082
const char **versionPtr,
1083
const char **versionEndPtr,
1084
const char **encodingName,
1085
const ENCODING **encoding,
1088
const char *val = NULL;
1089
const char *name = NULL;
1090
const char *nameEnd = NULL;
1091
ptr += 5 * enc->minBytesPerChar;
1092
end -= 2 * enc->minBytesPerChar;
1093
if (!parsePseudoAttribute(enc, ptr, end, &name, &nameEnd, &val, &ptr)
1098
if (!XmlNameMatchesAscii(enc, name, nameEnd, KW_version)) {
1099
if (!isGeneralTextEntity) {
1108
*versionEndPtr = ptr;
1109
if (!parsePseudoAttribute(enc, ptr, end, &name, &nameEnd, &val, &ptr)) {
1114
if (isGeneralTextEntity) {
1115
/* a TextDecl must have an EncodingDecl */
1122
if (XmlNameMatchesAscii(enc, name, nameEnd, KW_encoding)) {
1123
int c = toAscii(enc, val, end);
1124
if (!(ASCII_a <= c && c <= ASCII_z) && !(ASCII_A <= c && c <= ASCII_Z)) {
1129
*encodingName = val;
1131
*encoding = encodingFinder(enc, val, ptr - enc->minBytesPerChar);
1132
if (!parsePseudoAttribute(enc, ptr, end, &name, &nameEnd, &val, &ptr)) {
1139
if (!XmlNameMatchesAscii(enc, name, nameEnd, KW_standalone)
1140
|| isGeneralTextEntity) {
1144
if (XmlNameMatchesAscii(enc, val, ptr - enc->minBytesPerChar, KW_yes)) {
1148
else if (XmlNameMatchesAscii(enc, val, ptr - enc->minBytesPerChar, KW_no)) {
1156
while (isSpace(toAscii(enc, ptr, end)))
1157
ptr += enc->minBytesPerChar;
1166
checkCharRefNumber(int result)
1168
switch (result >> 8) {
1169
case 0xD8: case 0xD9: case 0xDA: case 0xDB:
1170
case 0xDC: case 0xDD: case 0xDE: case 0xDF:
1173
if (latin1_encoding.type[result] == BT_NONXML)
1177
if (result == 0xFFFE || result == 0xFFFF)
1185
XmlUtf8Encode(int c, char *buf)
1188
/* minN is minimum legal resulting value for N byte sequence */
1197
buf[0] = (char)(c | UTF8_cval1);
1201
buf[0] = (char)((c >> 6) | UTF8_cval2);
1202
buf[1] = (char)((c & 0x3f) | 0x80);
1206
buf[0] = (char)((c >> 12) | UTF8_cval3);
1207
buf[1] = (char)(((c >> 6) & 0x3f) | 0x80);
1208
buf[2] = (char)((c & 0x3f) | 0x80);
1212
buf[0] = (char)((c >> 18) | UTF8_cval4);
1213
buf[1] = (char)(((c >> 12) & 0x3f) | 0x80);
1214
buf[2] = (char)(((c >> 6) & 0x3f) | 0x80);
1215
buf[3] = (char)((c & 0x3f) | 0x80);
1222
XmlUtf16Encode(int charNum, unsigned short *buf)
1226
if (charNum < 0x10000) {
1227
buf[0] = (unsigned short)charNum;
1230
if (charNum < 0x110000) {
1232
buf[0] = (unsigned short)((charNum >> 10) + 0xD800);
1233
buf[1] = (unsigned short)((charNum & 0x3FF) + 0xDC00);
1239
struct unknown_encoding {
1240
struct normal_encoding normal;
1243
unsigned short utf16[256];
1247
#define AS_UNKNOWN_ENCODING(enc) ((const struct unknown_encoding *) (enc))
1250
XmlSizeOfUnknownEncoding(void)
1252
return sizeof(struct unknown_encoding);
1255
static int PTRFASTCALL
1256
unknown_isName(const ENCODING *enc, const char *p)
1258
const struct unknown_encoding *uenc = AS_UNKNOWN_ENCODING(enc);
1259
int c = uenc->convert(uenc->userData, p);
1262
return UCS2_GET_NAMING(namePages, c >> 8, c & 0xFF);
1265
static int PTRFASTCALL
1266
unknown_isNmstrt(const ENCODING *enc, const char *p)
1268
const struct unknown_encoding *uenc = AS_UNKNOWN_ENCODING(enc);
1269
int c = uenc->convert(uenc->userData, p);
1272
return UCS2_GET_NAMING(nmstrtPages, c >> 8, c & 0xFF);
1275
static int PTRFASTCALL
1276
unknown_isInvalid(const ENCODING *enc, const char *p)
1278
const struct unknown_encoding *uenc = AS_UNKNOWN_ENCODING(enc);
1279
int c = uenc->convert(uenc->userData, p);
1280
return (c & ~0xFFFF) || checkCharRefNumber(c) < 0;
1284
unknown_toUtf8(const ENCODING *enc,
1285
const char **fromP, const char *fromLim,
1286
char **toP, const char *toLim)
1288
const struct unknown_encoding *uenc = AS_UNKNOWN_ENCODING(enc);
1289
char buf[XML_UTF8_ENCODE_MAX];
1293
if (*fromP == fromLim)
1295
utf8 = uenc->utf8[(unsigned char)**fromP];
1298
int c = uenc->convert(uenc->userData, *fromP);
1299
n = XmlUtf8Encode(c, buf);
1300
if (n > toLim - *toP)
1303
*fromP += (AS_NORMAL_ENCODING(enc)->type[(unsigned char)**fromP]
1307
if (n > toLim - *toP)
1312
*(*toP)++ = *utf8++;
1318
unknown_toUtf16(const ENCODING *enc,
1319
const char **fromP, const char *fromLim,
1320
unsigned short **toP, const unsigned short *toLim)
1322
const struct unknown_encoding *uenc = AS_UNKNOWN_ENCODING(enc);
1323
while (*fromP != fromLim && *toP != toLim) {
1324
unsigned short c = uenc->utf16[(unsigned char)**fromP];
1326
c = (unsigned short)
1327
uenc->convert(uenc->userData, *fromP);
1328
*fromP += (AS_NORMAL_ENCODING(enc)->type[(unsigned char)**fromP]
1338
XmlInitUnknownEncoding(void *mem,
1344
struct unknown_encoding *e = (struct unknown_encoding *)mem;
1345
for (i = 0; i < (int)sizeof(struct normal_encoding); i++)
1346
((char *)mem)[i] = ((char *)&latin1_encoding)[i];
1347
for (i = 0; i < 128; i++)
1348
if (latin1_encoding.type[i] != BT_OTHER
1349
&& latin1_encoding.type[i] != BT_NONXML
1352
for (i = 0; i < 256; i++) {
1355
e->normal.type[i] = BT_MALFORM;
1356
/* This shouldn't really get used. */
1357
e->utf16[i] = 0xFFFF;
1364
e->normal.type[i] = (unsigned char)(BT_LEAD2 - (c + 2));
1368
else if (c < 0x80) {
1369
if (latin1_encoding.type[c] != BT_OTHER
1370
&& latin1_encoding.type[c] != BT_NONXML
1373
e->normal.type[i] = latin1_encoding.type[c];
1375
e->utf8[i][1] = (char)c;
1376
e->utf16[i] = (unsigned short)(c == 0 ? 0xFFFF : c);
1378
else if (checkCharRefNumber(c) < 0) {
1379
e->normal.type[i] = BT_NONXML;
1380
/* This shouldn't really get used. */
1381
e->utf16[i] = 0xFFFF;
1388
if (UCS2_GET_NAMING(nmstrtPages, c >> 8, c & 0xff))
1389
e->normal.type[i] = BT_NMSTRT;
1390
else if (UCS2_GET_NAMING(namePages, c >> 8, c & 0xff))
1391
e->normal.type[i] = BT_NAME;
1393
e->normal.type[i] = BT_OTHER;
1394
e->utf8[i][0] = (char)XmlUtf8Encode(c, e->utf8[i] + 1);
1395
e->utf16[i] = (unsigned short)c;
1398
e->userData = userData;
1399
e->convert = convert;
1401
e->normal.isName2 = unknown_isName;
1402
e->normal.isName3 = unknown_isName;
1403
e->normal.isName4 = unknown_isName;
1404
e->normal.isNmstrt2 = unknown_isNmstrt;
1405
e->normal.isNmstrt3 = unknown_isNmstrt;
1406
e->normal.isNmstrt4 = unknown_isNmstrt;
1407
e->normal.isInvalid2 = unknown_isInvalid;
1408
e->normal.isInvalid3 = unknown_isInvalid;
1409
e->normal.isInvalid4 = unknown_isInvalid;
1411
e->normal.enc.utf8Convert = unknown_toUtf8;
1412
e->normal.enc.utf16Convert = unknown_toUtf16;
1413
return &(e->normal.enc);
1416
/* If this enumeration is changed, getEncodingIndex and encodings
1417
must also be changed. */
1426
/* must match encodingNames up to here */
1430
static const char KW_ISO_8859_1[] = {
1431
ASCII_I, ASCII_S, ASCII_O, ASCII_MINUS, ASCII_8, ASCII_8, ASCII_5, ASCII_9,
1432
ASCII_MINUS, ASCII_1, '\0'
1434
static const char KW_US_ASCII[] = {
1435
ASCII_U, ASCII_S, ASCII_MINUS, ASCII_A, ASCII_S, ASCII_C, ASCII_I, ASCII_I,
1438
static const char KW_UTF_8[] = {
1439
ASCII_U, ASCII_T, ASCII_F, ASCII_MINUS, ASCII_8, '\0'
1441
static const char KW_UTF_16[] = {
1442
ASCII_U, ASCII_T, ASCII_F, ASCII_MINUS, ASCII_1, ASCII_6, '\0'
1444
static const char KW_UTF_16BE[] = {
1445
ASCII_U, ASCII_T, ASCII_F, ASCII_MINUS, ASCII_1, ASCII_6, ASCII_B, ASCII_E,
1448
static const char KW_UTF_16LE[] = {
1449
ASCII_U, ASCII_T, ASCII_F, ASCII_MINUS, ASCII_1, ASCII_6, ASCII_L, ASCII_E,
1454
getEncodingIndex(const char *name)
1456
static const char * const encodingNames[] = {
1467
for (i = 0; i < (int)(sizeof(encodingNames)/sizeof(encodingNames[0])); i++)
1468
if (streqci(name, encodingNames[i]))
1473
/* For binary compatibility, we store the index of the encoding
1474
specified at initialization in the isUtf16 member.
1477
#define INIT_ENC_INDEX(enc) ((int)(enc)->initEnc.isUtf16)
1478
#define SET_INIT_ENC_INDEX(enc, i) ((enc)->initEnc.isUtf16 = (char)i)
1480
/* This is what detects the encoding. encodingTable maps from
1481
encoding indices to encodings; INIT_ENC_INDEX(enc) is the index of
1482
the external (protocol) specified encoding; state is
1483
XML_CONTENT_STATE if we're parsing an external text entity, and
1484
XML_PROLOG_STATE otherwise.
1489
initScan(const ENCODING * const *encodingTable,
1490
const INIT_ENCODING *enc,
1494
const char **nextTokPtr)
1496
const ENCODING **encPtr;
1499
return XML_TOK_NONE;
1500
encPtr = enc->encPtr;
1501
if (ptr + 1 == end) {
1502
/* only a single byte available for auto-detection */
1503
#ifndef XML_DTD /* FIXME */
1504
/* a well-formed document entity must have more than one byte */
1505
if (state != XML_CONTENT_STATE)
1506
return XML_TOK_PARTIAL;
1508
/* so we're parsing an external text entity... */
1509
/* if UTF-16 was externally specified, then we need at least 2 bytes */
1510
switch (INIT_ENC_INDEX(enc)) {
1514
return XML_TOK_PARTIAL;
1516
switch ((unsigned char)*ptr) {
1519
case 0xEF: /* possibly first byte of UTF-8 BOM */
1520
if (INIT_ENC_INDEX(enc) == ISO_8859_1_ENC
1521
&& state == XML_CONTENT_STATE)
1526
return XML_TOK_PARTIAL;
1530
switch (((unsigned char)ptr[0] << 8) | (unsigned char)ptr[1]) {
1532
if (INIT_ENC_INDEX(enc) == ISO_8859_1_ENC
1533
&& state == XML_CONTENT_STATE)
1535
*nextTokPtr = ptr + 2;
1536
*encPtr = encodingTable[UTF_16BE_ENC];
1538
/* 00 3C is handled in the default case */
1540
if ((INIT_ENC_INDEX(enc) == UTF_16BE_ENC
1541
|| INIT_ENC_INDEX(enc) == UTF_16_ENC)
1542
&& state == XML_CONTENT_STATE)
1544
*encPtr = encodingTable[UTF_16LE_ENC];
1545
return XmlTok(*encPtr, state, ptr, end, nextTokPtr);
1547
if (INIT_ENC_INDEX(enc) == ISO_8859_1_ENC
1548
&& state == XML_CONTENT_STATE)
1550
*nextTokPtr = ptr + 2;
1551
*encPtr = encodingTable[UTF_16LE_ENC];
1554
/* Maybe a UTF-8 BOM (EF BB BF) */
1555
/* If there's an explicitly specified (external) encoding
1556
of ISO-8859-1 or some flavour of UTF-16
1557
and this is an external text entity,
1558
don't look for the BOM,
1559
because it might be a legal data.
1561
if (state == XML_CONTENT_STATE) {
1562
int e = INIT_ENC_INDEX(enc);
1563
if (e == ISO_8859_1_ENC || e == UTF_16BE_ENC
1564
|| e == UTF_16LE_ENC || e == UTF_16_ENC)
1568
return XML_TOK_PARTIAL;
1569
if ((unsigned char)ptr[2] == 0xBF) {
1570
*nextTokPtr = ptr + 3;
1571
*encPtr = encodingTable[UTF_8_ENC];
1576
if (ptr[0] == '\0') {
1577
/* 0 isn't a legal data character. Furthermore a document
1578
entity can only start with ASCII characters. So the only
1579
way this can fail to be big-endian UTF-16 if it it's an
1580
external parsed general entity that's labelled as
1583
if (state == XML_CONTENT_STATE && INIT_ENC_INDEX(enc) == UTF_16LE_ENC)
1585
*encPtr = encodingTable[UTF_16BE_ENC];
1586
return XmlTok(*encPtr, state, ptr, end, nextTokPtr);
1588
else if (ptr[1] == '\0') {
1589
/* We could recover here in the case:
1590
- parsing an external entity
1592
- no externally specified encoding
1593
- no encoding declaration
1594
by assuming UTF-16LE. But we don't, because this would mean when
1595
presented just with a single byte, we couldn't reliably determine
1596
whether we needed further bytes.
1598
if (state == XML_CONTENT_STATE)
1600
*encPtr = encodingTable[UTF_16LE_ENC];
1601
return XmlTok(*encPtr, state, ptr, end, nextTokPtr);
1606
*encPtr = encodingTable[INIT_ENC_INDEX(enc)];
1607
return XmlTok(*encPtr, state, ptr, end, nextTokPtr);
1613
#include "xmltok_ns.c"
1619
#define NS(x) x ## NS
1620
#define ns(x) x ## _ns
1622
#include "xmltok_ns.c"
1628
XmlInitUnknownEncodingNS(void *mem,
1633
ENCODING *enc = XmlInitUnknownEncoding(mem, table, convert, userData);
1635
((struct normal_encoding *)enc)->type[ASCII_COLON] = BT_COLON;