michael@0: /* Copyright (c) 1998, 1999 Thai Open Source Software Center Ltd
michael@0:    See the file COPYING for copying permission.
michael@0: */
michael@0: 
michael@0: #include <stddef.h>
michael@0: 
michael@0: #ifdef COMPILED_FROM_DSP
michael@0: #include "winconfig.h"
michael@0: #elif defined(MACOS_CLASSIC)
michael@0: #include "macconfig.h"
michael@0: #elif defined(__amigaos4__)
michael@0: #include "amigaconfig.h"
michael@0: #else
michael@0: #ifdef HAVE_EXPAT_CONFIG_H
michael@0: #include <expat_config.h>
michael@0: #endif
michael@0: #endif /* ndef COMPILED_FROM_DSP */
michael@0: 
michael@0: #include "expat_external.h"
michael@0: #include "internal.h"
michael@0: #include "xmltok.h"
michael@0: #include "nametab.h"
michael@0: 
michael@0: #ifdef XML_DTD
michael@0: #define IGNORE_SECTION_TOK_VTABLE , PREFIX(ignoreSectionTok)
michael@0: #else
michael@0: #define IGNORE_SECTION_TOK_VTABLE /* as nothing */
michael@0: #endif
michael@0: 
michael@0: #define VTABLE1 \
michael@0:   { PREFIX(prologTok), PREFIX(contentTok), \
michael@0:     PREFIX(cdataSectionTok) IGNORE_SECTION_TOK_VTABLE }, \
michael@0:   { PREFIX(attributeValueTok), PREFIX(entityValueTok) }, \
michael@0:   PREFIX(sameName), \
michael@0:   PREFIX(nameMatchesAscii), \
michael@0:   PREFIX(nameLength), \
michael@0:   PREFIX(skipS), \
michael@0:   PREFIX(getAtts), \
michael@0:   PREFIX(charRefNumber), \
michael@0:   PREFIX(predefinedEntityName), \
michael@0:   PREFIX(updatePosition), \
michael@0:   PREFIX(isPublicId)
michael@0: 
michael@0: #define VTABLE VTABLE1, PREFIX(toUtf8), PREFIX(toUtf16)
michael@0: 
michael@0: #define UCS2_GET_NAMING(pages, hi, lo) \
michael@0:    (namingBitmap[(pages[hi] << 3) + ((lo) >> 5)] & (1 << ((lo) & 0x1F)))
michael@0: 
michael@0: /* A 2 byte UTF-8 representation splits the characters 11 bits between
michael@0:    the bottom 5 and 6 bits of the bytes.  We need 8 bits to index into
michael@0:    pages, 3 bits to add to that index and 5 bits to generate the mask.
michael@0: */
michael@0: #define UTF8_GET_NAMING2(pages, byte) \
michael@0:     (namingBitmap[((pages)[(((byte)[0]) >> 2) & 7] << 3) \
michael@0:                       + ((((byte)[0]) & 3) << 1) \
michael@0:                       + ((((byte)[1]) >> 5) & 1)] \
michael@0:          & (1 << (((byte)[1]) & 0x1F)))
michael@0: 
michael@0: /* A 3 byte UTF-8 representation splits the characters 16 bits between
michael@0:    the bottom 4, 6 and 6 bits of the bytes.  We need 8 bits to index
michael@0:    into pages, 3 bits to add to that index and 5 bits to generate the
michael@0:    mask.
michael@0: */
michael@0: #define UTF8_GET_NAMING3(pages, byte) \
michael@0:   (namingBitmap[((pages)[((((byte)[0]) & 0xF) << 4) \
michael@0:                              + ((((byte)[1]) >> 2) & 0xF)] \
michael@0:                        << 3) \
michael@0:                       + ((((byte)[1]) & 3) << 1) \
michael@0:                       + ((((byte)[2]) >> 5) & 1)] \
michael@0:          & (1 << (((byte)[2]) & 0x1F)))
michael@0: 
michael@0: #define UTF8_GET_NAMING(pages, p, n) \
michael@0:   ((n) == 2 \
michael@0:   ? UTF8_GET_NAMING2(pages, (const unsigned char *)(p)) \
michael@0:   : ((n) == 3 \
michael@0:      ? UTF8_GET_NAMING3(pages, (const unsigned char *)(p)) \
michael@0:      : 0))
michael@0: 
michael@0: /* Detection of invalid UTF-8 sequences is based on Table 3.1B
michael@0:    of Unicode 3.2: http://www.unicode.org/unicode/reports/tr28/
michael@0:    with the additional restriction of not allowing the Unicode
michael@0:    code points 0xFFFF and 0xFFFE (sequences EF,BF,BF and EF,BF,BE).
michael@0:    Implementation details:
michael@0:      (A & 0x80) == 0     means A < 0x80
michael@0:    and
michael@0:      (A & 0xC0) == 0xC0  means A > 0xBF
michael@0: */
michael@0: 
michael@0: #define UTF8_INVALID2(p) \
michael@0:   ((*p) < 0xC2 || ((p)[1] & 0x80) == 0 || ((p)[1] & 0xC0) == 0xC0)
michael@0: 
michael@0: #define UTF8_INVALID3(p) \
michael@0:   (((p)[2] & 0x80) == 0 \
michael@0:   || \
michael@0:   ((*p) == 0xEF && (p)[1] == 0xBF \
michael@0:     ? \
michael@0:     (p)[2] > 0xBD \
michael@0:     : \
michael@0:     ((p)[2] & 0xC0) == 0xC0) \
michael@0:   || \
michael@0:   ((*p) == 0xE0 \
michael@0:     ? \
michael@0:     (p)[1] < 0xA0 || ((p)[1] & 0xC0) == 0xC0 \
michael@0:     : \
michael@0:     ((p)[1] & 0x80) == 0 \
michael@0:     || \
michael@0:     ((*p) == 0xED ? (p)[1] > 0x9F : ((p)[1] & 0xC0) == 0xC0)))
michael@0: 
michael@0: #define UTF8_INVALID4(p) \
michael@0:   (((p)[3] & 0x80) == 0 || ((p)[3] & 0xC0) == 0xC0 \
michael@0:   || \
michael@0:   ((p)[2] & 0x80) == 0 || ((p)[2] & 0xC0) == 0xC0 \
michael@0:   || \
michael@0:   ((*p) == 0xF0 \
michael@0:     ? \
michael@0:     (p)[1] < 0x90 || ((p)[1] & 0xC0) == 0xC0 \
michael@0:     : \
michael@0:     ((p)[1] & 0x80) == 0 \
michael@0:     || \
michael@0:     ((*p) == 0xF4 ? (p)[1] > 0x8F : ((p)[1] & 0xC0) == 0xC0)))
michael@0: 
michael@0: static int PTRFASTCALL
michael@0: isNever(const ENCODING *enc, const char *p)
michael@0: {
michael@0:   return 0;
michael@0: }
michael@0: 
michael@0: static int PTRFASTCALL
michael@0: utf8_isName2(const ENCODING *enc, const char *p)
michael@0: {
michael@0:   return UTF8_GET_NAMING2(namePages, (const unsigned char *)p);
michael@0: }
michael@0: 
michael@0: static int PTRFASTCALL
michael@0: utf8_isName3(const ENCODING *enc, const char *p)
michael@0: {
michael@0:   return UTF8_GET_NAMING3(namePages, (const unsigned char *)p);
michael@0: }
michael@0: 
michael@0: #define utf8_isName4 isNever
michael@0: 
michael@0: static int PTRFASTCALL
michael@0: utf8_isNmstrt2(const ENCODING *enc, const char *p)
michael@0: {
michael@0:   return UTF8_GET_NAMING2(nmstrtPages, (const unsigned char *)p);
michael@0: }
michael@0: 
michael@0: static int PTRFASTCALL
michael@0: utf8_isNmstrt3(const ENCODING *enc, const char *p)
michael@0: {
michael@0:   return UTF8_GET_NAMING3(nmstrtPages, (const unsigned char *)p);
michael@0: }
michael@0: 
michael@0: #define utf8_isNmstrt4 isNever
michael@0: 
michael@0: static int PTRFASTCALL
michael@0: utf8_isInvalid2(const ENCODING *enc, const char *p)
michael@0: {
michael@0:   return UTF8_INVALID2((const unsigned char *)p);
michael@0: }
michael@0: 
michael@0: static int PTRFASTCALL
michael@0: utf8_isInvalid3(const ENCODING *enc, const char *p)
michael@0: {
michael@0:   return UTF8_INVALID3((const unsigned char *)p);
michael@0: }
michael@0: 
michael@0: static int PTRFASTCALL
michael@0: utf8_isInvalid4(const ENCODING *enc, const char *p)
michael@0: {
michael@0:   return UTF8_INVALID4((const unsigned char *)p);
michael@0: }
michael@0: 
michael@0: struct normal_encoding {
michael@0:   ENCODING enc;
michael@0:   unsigned char type[256];
michael@0: #ifdef XML_MIN_SIZE
michael@0:   int (PTRFASTCALL *byteType)(const ENCODING *, const char *);
michael@0:   int (PTRFASTCALL *isNameMin)(const ENCODING *, const char *);
michael@0:   int (PTRFASTCALL *isNmstrtMin)(const ENCODING *, const char *);
michael@0:   int (PTRFASTCALL *byteToAscii)(const ENCODING *, const char *);
michael@0:   int (PTRCALL *charMatches)(const ENCODING *, const char *, int);
michael@0: #endif /* XML_MIN_SIZE */
michael@0:   int (PTRFASTCALL *isName2)(const ENCODING *, const char *);
michael@0:   int (PTRFASTCALL *isName3)(const ENCODING *, const char *);
michael@0:   int (PTRFASTCALL *isName4)(const ENCODING *, const char *);
michael@0:   int (PTRFASTCALL *isNmstrt2)(const ENCODING *, const char *);
michael@0:   int (PTRFASTCALL *isNmstrt3)(const ENCODING *, const char *);
michael@0:   int (PTRFASTCALL *isNmstrt4)(const ENCODING *, const char *);
michael@0:   int (PTRFASTCALL *isInvalid2)(const ENCODING *, const char *);
michael@0:   int (PTRFASTCALL *isInvalid3)(const ENCODING *, const char *);
michael@0:   int (PTRFASTCALL *isInvalid4)(const ENCODING *, const char *);
michael@0: };
michael@0: 
michael@0: #define AS_NORMAL_ENCODING(enc)   ((const struct normal_encoding *) (enc))
michael@0: 
michael@0: #ifdef XML_MIN_SIZE
michael@0: 
michael@0: #define STANDARD_VTABLE(E) \
michael@0:  E ## byteType, \
michael@0:  E ## isNameMin, \
michael@0:  E ## isNmstrtMin, \
michael@0:  E ## byteToAscii, \
michael@0:  E ## charMatches,
michael@0: 
michael@0: #else
michael@0: 
michael@0: #define STANDARD_VTABLE(E) /* as nothing */
michael@0: 
michael@0: #endif
michael@0: 
michael@0: #define NORMAL_VTABLE(E) \
michael@0:  E ## isName2, \
michael@0:  E ## isName3, \
michael@0:  E ## isName4, \
michael@0:  E ## isNmstrt2, \
michael@0:  E ## isNmstrt3, \
michael@0:  E ## isNmstrt4, \
michael@0:  E ## isInvalid2, \
michael@0:  E ## isInvalid3, \
michael@0:  E ## isInvalid4
michael@0: 
michael@0: static int FASTCALL checkCharRefNumber(int);
michael@0: 
michael@0: #include "xmltok_impl.h"
michael@0: #include "ascii.h"
michael@0: 
michael@0: #ifdef XML_MIN_SIZE
michael@0: #define sb_isNameMin isNever
michael@0: #define sb_isNmstrtMin isNever
michael@0: #endif
michael@0: 
michael@0: #ifdef XML_MIN_SIZE
michael@0: #define MINBPC(enc) ((enc)->minBytesPerChar)
michael@0: #else
michael@0: /* minimum bytes per character */
michael@0: #define MINBPC(enc) 1
michael@0: #endif
michael@0: 
michael@0: #define SB_BYTE_TYPE(enc, p) \
michael@0:   (((struct normal_encoding *)(enc))->type[(unsigned char)*(p)])
michael@0: 
michael@0: #ifdef XML_MIN_SIZE
michael@0: static int PTRFASTCALL
michael@0: sb_byteType(const ENCODING *enc, const char *p)
michael@0: {
michael@0:   return SB_BYTE_TYPE(enc, p);
michael@0: }
michael@0: #define BYTE_TYPE(enc, p) \
michael@0:  (AS_NORMAL_ENCODING(enc)->byteType(enc, p))
michael@0: #else
michael@0: #define BYTE_TYPE(enc, p) SB_BYTE_TYPE(enc, p)
michael@0: #endif
michael@0: 
michael@0: #ifdef XML_MIN_SIZE
michael@0: #define BYTE_TO_ASCII(enc, p) \
michael@0:  (AS_NORMAL_ENCODING(enc)->byteToAscii(enc, p))
michael@0: static int PTRFASTCALL
michael@0: sb_byteToAscii(const ENCODING *enc, const char *p)
michael@0: {
michael@0:   return *p;
michael@0: }
michael@0: #else
michael@0: #define BYTE_TO_ASCII(enc, p) (*(p))
michael@0: #endif
michael@0: 
michael@0: #define IS_NAME_CHAR(enc, p, n) \
michael@0:  (AS_NORMAL_ENCODING(enc)->isName ## n(enc, p))
michael@0: #define IS_NMSTRT_CHAR(enc, p, n) \
michael@0:  (AS_NORMAL_ENCODING(enc)->isNmstrt ## n(enc, p))
michael@0: #define IS_INVALID_CHAR(enc, p, n) \
michael@0:  (AS_NORMAL_ENCODING(enc)->isInvalid ## n(enc, p))
michael@0: 
michael@0: #ifdef XML_MIN_SIZE
michael@0: #define IS_NAME_CHAR_MINBPC(enc, p) \
michael@0:  (AS_NORMAL_ENCODING(enc)->isNameMin(enc, p))
michael@0: #define IS_NMSTRT_CHAR_MINBPC(enc, p) \
michael@0:  (AS_NORMAL_ENCODING(enc)->isNmstrtMin(enc, p))
michael@0: #else
michael@0: #define IS_NAME_CHAR_MINBPC(enc, p) (0)
michael@0: #define IS_NMSTRT_CHAR_MINBPC(enc, p) (0)
michael@0: #endif
michael@0: 
michael@0: #ifdef XML_MIN_SIZE
michael@0: #define CHAR_MATCHES(enc, p, c) \
michael@0:  (AS_NORMAL_ENCODING(enc)->charMatches(enc, p, c))
michael@0: static int PTRCALL
michael@0: sb_charMatches(const ENCODING *enc, const char *p, int c)
michael@0: {
michael@0:   return *p == c;
michael@0: }
michael@0: #else
michael@0: /* c is an ASCII character */
michael@0: #define CHAR_MATCHES(enc, p, c) (*(p) == c)
michael@0: #endif
michael@0: 
michael@0: #define PREFIX(ident) normal_ ## ident
michael@0: #include "xmltok_impl.c"
michael@0: 
michael@0: #undef MINBPC
michael@0: #undef BYTE_TYPE
michael@0: #undef BYTE_TO_ASCII
michael@0: #undef CHAR_MATCHES
michael@0: #undef IS_NAME_CHAR
michael@0: #undef IS_NAME_CHAR_MINBPC
michael@0: #undef IS_NMSTRT_CHAR
michael@0: #undef IS_NMSTRT_CHAR_MINBPC
michael@0: #undef IS_INVALID_CHAR
michael@0: 
michael@0: enum {  /* UTF8_cvalN is value of masked first byte of N byte sequence */
michael@0:   UTF8_cval1 = 0x00,
michael@0:   UTF8_cval2 = 0xc0,
michael@0:   UTF8_cval3 = 0xe0,
michael@0:   UTF8_cval4 = 0xf0
michael@0: };
michael@0: 
michael@0: static void PTRCALL
michael@0: utf8_toUtf8(const ENCODING *enc,
michael@0:             const char **fromP, const char *fromLim,
michael@0:             char **toP, const char *toLim)
michael@0: {
michael@0:   char *to;
michael@0:   const char *from;
michael@0:   if (fromLim - *fromP > toLim - *toP) {
michael@0:     /* Avoid copying partial characters. */
michael@0:     for (fromLim = *fromP + (toLim - *toP); fromLim > *fromP; fromLim--)
michael@0:       if (((unsigned char)fromLim[-1] & 0xc0) != 0x80)
michael@0:         break;
michael@0:   }
michael@0:   for (to = *toP, from = *fromP; from != fromLim; from++, to++)
michael@0:     *to = *from;
michael@0:   *fromP = from;
michael@0:   *toP = to;
michael@0: }
michael@0: 
michael@0: static void PTRCALL
michael@0: utf8_toUtf16(const ENCODING *enc,
michael@0:              const char **fromP, const char *fromLim,
michael@0:              unsigned short **toP, const unsigned short *toLim)
michael@0: {
michael@0:   unsigned short *to = *toP;
michael@0:   const char *from = *fromP;
michael@0:   while (from != fromLim && to != toLim) {
michael@0:     switch (((struct normal_encoding *)enc)->type[(unsigned char)*from]) {
michael@0:     case BT_LEAD2:
michael@0:       *to++ = (unsigned short)(((from[0] & 0x1f) << 6) | (from[1] & 0x3f));
michael@0:       from += 2;
michael@0:       break;
michael@0:     case BT_LEAD3:
michael@0:       *to++ = (unsigned short)(((from[0] & 0xf) << 12)
michael@0:                                | ((from[1] & 0x3f) << 6) | (from[2] & 0x3f));
michael@0:       from += 3;
michael@0:       break;
michael@0:     case BT_LEAD4:
michael@0:       {
michael@0:         unsigned long n;
michael@0:         if (to + 1 == toLim)
michael@0:           goto after;
michael@0:         n = ((from[0] & 0x7) << 18) | ((from[1] & 0x3f) << 12)
michael@0:             | ((from[2] & 0x3f) << 6) | (from[3] & 0x3f);
michael@0:         n -= 0x10000;
michael@0:         to[0] = (unsigned short)((n >> 10) | 0xD800);
michael@0:         to[1] = (unsigned short)((n & 0x3FF) | 0xDC00);
michael@0:         to += 2;
michael@0:         from += 4;
michael@0:       }
michael@0:       break;
michael@0:     default:
michael@0:       *to++ = *from++;
michael@0:       break;
michael@0:     }
michael@0:   }
michael@0: after:
michael@0:   *fromP = from;
michael@0:   *toP = to;
michael@0: }
michael@0: 
michael@0: #ifdef XML_NS
michael@0: static const struct normal_encoding utf8_encoding_ns = {
michael@0:   { VTABLE1, utf8_toUtf8, utf8_toUtf16, 1, 1, 0 },
michael@0:   {
michael@0: #include "asciitab.h"
michael@0: #include "utf8tab.h"
michael@0:   },
michael@0:   STANDARD_VTABLE(sb_) NORMAL_VTABLE(utf8_)
michael@0: };
michael@0: #endif
michael@0: 
michael@0: static const struct normal_encoding utf8_encoding = {
michael@0:   { VTABLE1, utf8_toUtf8, utf8_toUtf16, 1, 1, 0 },
michael@0:   {
michael@0: #define BT_COLON BT_NMSTRT
michael@0: #include "asciitab.h"
michael@0: #undef BT_COLON
michael@0: #include "utf8tab.h"
michael@0:   },
michael@0:   STANDARD_VTABLE(sb_) NORMAL_VTABLE(utf8_)
michael@0: };
michael@0: 
michael@0: #ifdef XML_NS
michael@0: 
michael@0: static const struct normal_encoding internal_utf8_encoding_ns = {
michael@0:   { VTABLE1, utf8_toUtf8, utf8_toUtf16, 1, 1, 0 },
michael@0:   {
michael@0: #include "iasciitab.h"
michael@0: #include "utf8tab.h"
michael@0:   },
michael@0:   STANDARD_VTABLE(sb_) NORMAL_VTABLE(utf8_)
michael@0: };
michael@0: 
michael@0: #endif
michael@0: 
michael@0: static const struct normal_encoding internal_utf8_encoding = {
michael@0:   { VTABLE1, utf8_toUtf8, utf8_toUtf16, 1, 1, 0 },
michael@0:   {
michael@0: #define BT_COLON BT_NMSTRT
michael@0: #include "iasciitab.h"
michael@0: #undef BT_COLON
michael@0: #include "utf8tab.h"
michael@0:   },
michael@0:   STANDARD_VTABLE(sb_) NORMAL_VTABLE(utf8_)
michael@0: };
michael@0: 
michael@0: static void PTRCALL
michael@0: latin1_toUtf8(const ENCODING *enc,
michael@0:               const char **fromP, const char *fromLim,
michael@0:               char **toP, const char *toLim)
michael@0: {
michael@0:   for (;;) {
michael@0:     unsigned char c;
michael@0:     if (*fromP == fromLim)
michael@0:       break;
michael@0:     c = (unsigned char)**fromP;
michael@0:     if (c & 0x80) {
michael@0:       if (toLim - *toP < 2)
michael@0:         break;
michael@0:       *(*toP)++ = (char)((c >> 6) | UTF8_cval2);
michael@0:       *(*toP)++ = (char)((c & 0x3f) | 0x80);
michael@0:       (*fromP)++;
michael@0:     }
michael@0:     else {
michael@0:       if (*toP == toLim)
michael@0:         break;
michael@0:       *(*toP)++ = *(*fromP)++;
michael@0:     }
michael@0:   }
michael@0: }
michael@0: 
michael@0: static void PTRCALL
michael@0: latin1_toUtf16(const ENCODING *enc,
michael@0:                const char **fromP, const char *fromLim,
michael@0:                unsigned short **toP, const unsigned short *toLim)
michael@0: {
michael@0:   while (*fromP != fromLim && *toP != toLim)
michael@0:     *(*toP)++ = (unsigned char)*(*fromP)++;
michael@0: }
michael@0: 
michael@0: #ifdef XML_NS
michael@0: 
michael@0: static const struct normal_encoding latin1_encoding_ns = {
michael@0:   { VTABLE1, latin1_toUtf8, latin1_toUtf16, 1, 0, 0 },
michael@0:   {
michael@0: #include "asciitab.h"
michael@0: #include "latin1tab.h"
michael@0:   },
michael@0:   STANDARD_VTABLE(sb_)
michael@0: };
michael@0: 
michael@0: #endif
michael@0: 
michael@0: static const struct normal_encoding latin1_encoding = {
michael@0:   { VTABLE1, latin1_toUtf8, latin1_toUtf16, 1, 0, 0 },
michael@0:   {
michael@0: #define BT_COLON BT_NMSTRT
michael@0: #include "asciitab.h"
michael@0: #undef BT_COLON
michael@0: #include "latin1tab.h"
michael@0:   },
michael@0:   STANDARD_VTABLE(sb_)
michael@0: };
michael@0: 
michael@0: static void PTRCALL
michael@0: ascii_toUtf8(const ENCODING *enc,
michael@0:              const char **fromP, const char *fromLim,
michael@0:              char **toP, const char *toLim)
michael@0: {
michael@0:   while (*fromP != fromLim && *toP != toLim)
michael@0:     *(*toP)++ = *(*fromP)++;
michael@0: }
michael@0: 
michael@0: #ifdef XML_NS
michael@0: 
michael@0: static const struct normal_encoding ascii_encoding_ns = {
michael@0:   { VTABLE1, ascii_toUtf8, latin1_toUtf16, 1, 1, 0 },
michael@0:   {
michael@0: #include "asciitab.h"
michael@0: /* BT_NONXML == 0 */
michael@0:   },
michael@0:   STANDARD_VTABLE(sb_)
michael@0: };
michael@0: 
michael@0: #endif
michael@0: 
michael@0: static const struct normal_encoding ascii_encoding = {
michael@0:   { VTABLE1, ascii_toUtf8, latin1_toUtf16, 1, 1, 0 },
michael@0:   {
michael@0: #define BT_COLON BT_NMSTRT
michael@0: #include "asciitab.h"
michael@0: #undef BT_COLON
michael@0: /* BT_NONXML == 0 */
michael@0:   },
michael@0:   STANDARD_VTABLE(sb_)
michael@0: };
michael@0: 
michael@0: static int PTRFASTCALL
michael@0: unicode_byte_type(char hi, char lo)
michael@0: {
michael@0:   switch ((unsigned char)hi) {
michael@0:   case 0xD8: case 0xD9: case 0xDA: case 0xDB:
michael@0:     return BT_LEAD4;
michael@0:   case 0xDC: case 0xDD: case 0xDE: case 0xDF:
michael@0:     return BT_TRAIL;
michael@0:   case 0xFF:
michael@0:     switch ((unsigned char)lo) {
michael@0:     case 0xFF:
michael@0:     case 0xFE:
michael@0:       return BT_NONXML;
michael@0:     }
michael@0:     break;
michael@0:   }
michael@0:   return BT_NONASCII;
michael@0: }
michael@0: 
michael@0: #define DEFINE_UTF16_TO_UTF8(E) \
michael@0: static void  PTRCALL \
michael@0: E ## toUtf8(const ENCODING *enc, \
michael@0:             const char **fromP, const char *fromLim, \
michael@0:             char **toP, const char *toLim) \
michael@0: { \
michael@0:   const char *from; \
michael@0:   for (from = *fromP; from != fromLim; from += 2) { \
michael@0:     int plane; \
michael@0:     unsigned char lo2; \
michael@0:     unsigned char lo = GET_LO(from); \
michael@0:     unsigned char hi = GET_HI(from); \
michael@0:     switch (hi) { \
michael@0:     case 0: \
michael@0:       if (lo < 0x80) { \
michael@0:         if (*toP == toLim) { \
michael@0:           *fromP = from; \
michael@0:           return; \
michael@0:         } \
michael@0:         *(*toP)++ = lo; \
michael@0:         break; \
michael@0:       } \
michael@0:       /* fall through */ \
michael@0:     case 0x1: case 0x2: case 0x3: \
michael@0:     case 0x4: case 0x5: case 0x6: case 0x7: \
michael@0:       if (toLim -  *toP < 2) { \
michael@0:         *fromP = from; \
michael@0:         return; \
michael@0:       } \
michael@0:       *(*toP)++ = ((lo >> 6) | (hi << 2) |  UTF8_cval2); \
michael@0:       *(*toP)++ = ((lo & 0x3f) | 0x80); \
michael@0:       break; \
michael@0:     default: \
michael@0:       if (toLim -  *toP < 3)  { \
michael@0:         *fromP = from; \
michael@0:         return; \
michael@0:       } \
michael@0:       /* 16 bits divided 4, 6, 6 amongst 3 bytes */ \
michael@0:       *(*toP)++ = ((hi >> 4) | UTF8_cval3); \
michael@0:       *(*toP)++ = (((hi & 0xf) << 2) | (lo >> 6) | 0x80); \
michael@0:       *(*toP)++ = ((lo & 0x3f) | 0x80); \
michael@0:       break; \
michael@0:     case 0xD8: case 0xD9: case 0xDA: case 0xDB: \
michael@0:       if (toLim -  *toP < 4) { \
michael@0:         *fromP = from; \
michael@0:         return; \
michael@0:       } \
michael@0:       plane = (((hi & 0x3) << 2) | ((lo >> 6) & 0x3)) + 1; \
michael@0:       *(*toP)++ = ((plane >> 2) | UTF8_cval4); \
michael@0:       *(*toP)++ = (((lo >> 2) & 0xF) | ((plane & 0x3) << 4) | 0x80); \
michael@0:       from += 2; \
michael@0:       lo2 = GET_LO(from); \
michael@0:       *(*toP)++ = (((lo & 0x3) << 4) \
michael@0:                    | ((GET_HI(from) & 0x3) << 2) \
michael@0:                    | (lo2 >> 6) \
michael@0:                    | 0x80); \
michael@0:       *(*toP)++ = ((lo2 & 0x3f) | 0x80); \
michael@0:       break; \
michael@0:     } \
michael@0:   } \
michael@0:   *fromP = from; \
michael@0: }
michael@0: 
michael@0: #define DEFINE_UTF16_TO_UTF16(E) \
michael@0: static void  PTRCALL \
michael@0: E ## toUtf16(const ENCODING *enc, \
michael@0:              const char **fromP, const char *fromLim, \
michael@0:              unsigned short **toP, const unsigned short *toLim) \
michael@0: { \
michael@0:   /* Avoid copying first half only of surrogate */ \
michael@0:   if (fromLim - *fromP > ((toLim - *toP) << 1) \
michael@0:       && (GET_HI(fromLim - 2) & 0xF8) == 0xD8) \
michael@0:     fromLim -= 2; \
michael@0:   for (; *fromP != fromLim && *toP != toLim; *fromP += 2) \
michael@0:     *(*toP)++ = (GET_HI(*fromP) << 8) | GET_LO(*fromP); \
michael@0: }
michael@0: 
michael@0: #define SET2(ptr, ch) \
michael@0:   (((ptr)[0] = ((ch) & 0xff)), ((ptr)[1] = ((ch) >> 8)))
michael@0: #define GET_LO(ptr) ((unsigned char)(ptr)[0])
michael@0: #define GET_HI(ptr) ((unsigned char)(ptr)[1])
michael@0: 
michael@0: DEFINE_UTF16_TO_UTF8(little2_)
michael@0: DEFINE_UTF16_TO_UTF16(little2_)
michael@0: 
michael@0: #undef SET2
michael@0: #undef GET_LO
michael@0: #undef GET_HI
michael@0: 
michael@0: #define SET2(ptr, ch) \
michael@0:   (((ptr)[0] = ((ch) >> 8)), ((ptr)[1] = ((ch) & 0xFF)))
michael@0: #define GET_LO(ptr) ((unsigned char)(ptr)[1])
michael@0: #define GET_HI(ptr) ((unsigned char)(ptr)[0])
michael@0: 
michael@0: DEFINE_UTF16_TO_UTF8(big2_)
michael@0: DEFINE_UTF16_TO_UTF16(big2_)
michael@0: 
michael@0: #undef SET2
michael@0: #undef GET_LO
michael@0: #undef GET_HI
michael@0: 
michael@0: #define LITTLE2_BYTE_TYPE(enc, p) \
michael@0:  ((p)[1] == 0 \
michael@0:   ? ((struct normal_encoding *)(enc))->type[(unsigned char)*(p)] \
michael@0:   : unicode_byte_type((p)[1], (p)[0]))
michael@0: #define LITTLE2_BYTE_TO_ASCII(enc, p) ((p)[1] == 0 ? (p)[0] : -1)
michael@0: #define LITTLE2_CHAR_MATCHES(enc, p, c) ((p)[1] == 0 && (p)[0] == c)
michael@0: #define LITTLE2_IS_NAME_CHAR_MINBPC(enc, p) \
michael@0:   UCS2_GET_NAMING(namePages, (unsigned char)p[1], (unsigned char)p[0])
michael@0: #define LITTLE2_IS_NMSTRT_CHAR_MINBPC(enc, p) \
michael@0:   UCS2_GET_NAMING(nmstrtPages, (unsigned char)p[1], (unsigned char)p[0])
michael@0: 
michael@0: #ifdef XML_MIN_SIZE
michael@0: 
michael@0: static int PTRFASTCALL
michael@0: little2_byteType(const ENCODING *enc, const char *p)
michael@0: {
michael@0:   return LITTLE2_BYTE_TYPE(enc, p);
michael@0: }
michael@0: 
michael@0: static int PTRFASTCALL
michael@0: little2_byteToAscii(const ENCODING *enc, const char *p)
michael@0: {
michael@0:   return LITTLE2_BYTE_TO_ASCII(enc, p);
michael@0: }
michael@0: 
michael@0: static int PTRCALL
michael@0: little2_charMatches(const ENCODING *enc, const char *p, int c)
michael@0: {
michael@0:   return LITTLE2_CHAR_MATCHES(enc, p, c);
michael@0: }
michael@0: 
michael@0: static int PTRFASTCALL
michael@0: little2_isNameMin(const ENCODING *enc, const char *p)
michael@0: {
michael@0:   return LITTLE2_IS_NAME_CHAR_MINBPC(enc, p);
michael@0: }
michael@0: 
michael@0: static int PTRFASTCALL
michael@0: little2_isNmstrtMin(const ENCODING *enc, const char *p)
michael@0: {
michael@0:   return LITTLE2_IS_NMSTRT_CHAR_MINBPC(enc, p);
michael@0: }
michael@0: 
michael@0: #undef VTABLE
michael@0: #define VTABLE VTABLE1, little2_toUtf8, little2_toUtf16
michael@0: 
michael@0: #else /* not XML_MIN_SIZE */
michael@0: 
michael@0: #undef PREFIX
michael@0: #define PREFIX(ident) little2_ ## ident
michael@0: #define MINBPC(enc) 2
michael@0: /* CHAR_MATCHES is guaranteed to have MINBPC bytes available. */
michael@0: #define BYTE_TYPE(enc, p) LITTLE2_BYTE_TYPE(enc, p)
michael@0: #define BYTE_TO_ASCII(enc, p) LITTLE2_BYTE_TO_ASCII(enc, p)
michael@0: #define CHAR_MATCHES(enc, p, c) LITTLE2_CHAR_MATCHES(enc, p, c)
michael@0: #define IS_NAME_CHAR(enc, p, n) 0
michael@0: #define IS_NAME_CHAR_MINBPC(enc, p) LITTLE2_IS_NAME_CHAR_MINBPC(enc, p)
michael@0: #define IS_NMSTRT_CHAR(enc, p, n) (0)
michael@0: #define IS_NMSTRT_CHAR_MINBPC(enc, p) LITTLE2_IS_NMSTRT_CHAR_MINBPC(enc, p)
michael@0: 
michael@0: #include "xmltok_impl.c"
michael@0: 
michael@0: #undef MINBPC
michael@0: #undef BYTE_TYPE
michael@0: #undef BYTE_TO_ASCII
michael@0: #undef CHAR_MATCHES
michael@0: #undef IS_NAME_CHAR
michael@0: #undef IS_NAME_CHAR_MINBPC
michael@0: #undef IS_NMSTRT_CHAR
michael@0: #undef IS_NMSTRT_CHAR_MINBPC
michael@0: #undef IS_INVALID_CHAR
michael@0: 
michael@0: #endif /* not XML_MIN_SIZE */
michael@0: 
michael@0: #ifdef XML_NS
michael@0: 
michael@0: static const struct normal_encoding little2_encoding_ns = {
michael@0:   { VTABLE, 2, 0,
michael@0: #if BYTEORDER == 1234
michael@0:     1
michael@0: #else
michael@0:     0
michael@0: #endif
michael@0:   },
michael@0:   {
michael@0: #include "asciitab.h"
michael@0: #include "latin1tab.h"
michael@0:   },
michael@0:   STANDARD_VTABLE(little2_)
michael@0: };
michael@0: 
michael@0: #endif
michael@0: 
michael@0: static const struct normal_encoding little2_encoding = {
michael@0:   { VTABLE, 2, 0,
michael@0: #if BYTEORDER == 1234
michael@0:     1
michael@0: #else
michael@0:     0
michael@0: #endif
michael@0:   },
michael@0:   {
michael@0: #define BT_COLON BT_NMSTRT
michael@0: #include "asciitab.h"
michael@0: #undef BT_COLON
michael@0: #include "latin1tab.h"
michael@0:   },
michael@0:   STANDARD_VTABLE(little2_)
michael@0: };
michael@0: 
michael@0: #if BYTEORDER != 4321
michael@0: 
michael@0: #ifdef XML_NS
michael@0: 
michael@0: static const struct normal_encoding internal_little2_encoding_ns = {
michael@0:   { VTABLE, 2, 0, 1 },
michael@0:   {
michael@0: #include "iasciitab.h"
michael@0: #include "latin1tab.h"
michael@0:   },
michael@0:   STANDARD_VTABLE(little2_)
michael@0: };
michael@0: 
michael@0: #endif
michael@0: 
michael@0: static const struct normal_encoding internal_little2_encoding = {
michael@0:   { VTABLE, 2, 0, 1 },
michael@0:   {
michael@0: #define BT_COLON BT_NMSTRT
michael@0: #include "iasciitab.h"
michael@0: #undef BT_COLON
michael@0: #include "latin1tab.h"
michael@0:   },
michael@0:   STANDARD_VTABLE(little2_)
michael@0: };
michael@0: 
michael@0: #endif
michael@0: 
michael@0: 
michael@0: #define BIG2_BYTE_TYPE(enc, p) \
michael@0:  ((p)[0] == 0 \
michael@0:   ? ((struct normal_encoding *)(enc))->type[(unsigned char)(p)[1]] \
michael@0:   : unicode_byte_type((p)[0], (p)[1]))
michael@0: #define BIG2_BYTE_TO_ASCII(enc, p) ((p)[0] == 0 ? (p)[1] : -1)
michael@0: #define BIG2_CHAR_MATCHES(enc, p, c) ((p)[0] == 0 && (p)[1] == c)
michael@0: #define BIG2_IS_NAME_CHAR_MINBPC(enc, p) \
michael@0:   UCS2_GET_NAMING(namePages, (unsigned char)p[0], (unsigned char)p[1])
michael@0: #define BIG2_IS_NMSTRT_CHAR_MINBPC(enc, p) \
michael@0:   UCS2_GET_NAMING(nmstrtPages, (unsigned char)p[0], (unsigned char)p[1])
michael@0: 
michael@0: #ifdef XML_MIN_SIZE
michael@0: 
michael@0: static int PTRFASTCALL
michael@0: big2_byteType(const ENCODING *enc, const char *p)
michael@0: {
michael@0:   return BIG2_BYTE_TYPE(enc, p);
michael@0: }
michael@0: 
michael@0: static int PTRFASTCALL
michael@0: big2_byteToAscii(const ENCODING *enc, const char *p)
michael@0: {
michael@0:   return BIG2_BYTE_TO_ASCII(enc, p);
michael@0: }
michael@0: 
michael@0: static int PTRCALL
michael@0: big2_charMatches(const ENCODING *enc, const char *p, int c)
michael@0: {
michael@0:   return BIG2_CHAR_MATCHES(enc, p, c);
michael@0: }
michael@0: 
michael@0: static int PTRFASTCALL
michael@0: big2_isNameMin(const ENCODING *enc, const char *p)
michael@0: {
michael@0:   return BIG2_IS_NAME_CHAR_MINBPC(enc, p);
michael@0: }
michael@0: 
michael@0: static int PTRFASTCALL
michael@0: big2_isNmstrtMin(const ENCODING *enc, const char *p)
michael@0: {
michael@0:   return BIG2_IS_NMSTRT_CHAR_MINBPC(enc, p);
michael@0: }
michael@0: 
michael@0: #undef VTABLE
michael@0: #define VTABLE VTABLE1, big2_toUtf8, big2_toUtf16
michael@0: 
michael@0: #else /* not XML_MIN_SIZE */
michael@0: 
michael@0: #undef PREFIX
michael@0: #define PREFIX(ident) big2_ ## ident
michael@0: #define MINBPC(enc) 2
michael@0: /* CHAR_MATCHES is guaranteed to have MINBPC bytes available. */
michael@0: #define BYTE_TYPE(enc, p) BIG2_BYTE_TYPE(enc, p)
michael@0: #define BYTE_TO_ASCII(enc, p) BIG2_BYTE_TO_ASCII(enc, p)
michael@0: #define CHAR_MATCHES(enc, p, c) BIG2_CHAR_MATCHES(enc, p, c)
michael@0: #define IS_NAME_CHAR(enc, p, n) 0
michael@0: #define IS_NAME_CHAR_MINBPC(enc, p) BIG2_IS_NAME_CHAR_MINBPC(enc, p)
michael@0: #define IS_NMSTRT_CHAR(enc, p, n) (0)
michael@0: #define IS_NMSTRT_CHAR_MINBPC(enc, p) BIG2_IS_NMSTRT_CHAR_MINBPC(enc, p)
michael@0: 
michael@0: #include "xmltok_impl.c"
michael@0: 
michael@0: #undef MINBPC
michael@0: #undef BYTE_TYPE
michael@0: #undef BYTE_TO_ASCII
michael@0: #undef CHAR_MATCHES
michael@0: #undef IS_NAME_CHAR
michael@0: #undef IS_NAME_CHAR_MINBPC
michael@0: #undef IS_NMSTRT_CHAR
michael@0: #undef IS_NMSTRT_CHAR_MINBPC
michael@0: #undef IS_INVALID_CHAR
michael@0: 
michael@0: #endif /* not XML_MIN_SIZE */
michael@0: 
michael@0: #ifdef XML_NS
michael@0: 
michael@0: static const struct normal_encoding big2_encoding_ns = {
michael@0:   { VTABLE, 2, 0,
michael@0: #if BYTEORDER == 4321
michael@0:   1
michael@0: #else
michael@0:   0
michael@0: #endif
michael@0:   },
michael@0:   {
michael@0: #include "asciitab.h"
michael@0: #include "latin1tab.h"
michael@0:   },
michael@0:   STANDARD_VTABLE(big2_)
michael@0: };
michael@0: 
michael@0: #endif
michael@0: 
michael@0: static const struct normal_encoding big2_encoding = {
michael@0:   { VTABLE, 2, 0,
michael@0: #if BYTEORDER == 4321
michael@0:   1
michael@0: #else
michael@0:   0
michael@0: #endif
michael@0:   },
michael@0:   {
michael@0: #define BT_COLON BT_NMSTRT
michael@0: #include "asciitab.h"
michael@0: #undef BT_COLON
michael@0: #include "latin1tab.h"
michael@0:   },
michael@0:   STANDARD_VTABLE(big2_)
michael@0: };
michael@0: 
michael@0: #if BYTEORDER != 1234
michael@0: 
michael@0: #ifdef XML_NS
michael@0: 
michael@0: static const struct normal_encoding internal_big2_encoding_ns = {
michael@0:   { VTABLE, 2, 0, 1 },
michael@0:   {
michael@0: #include "iasciitab.h"
michael@0: #include "latin1tab.h"
michael@0:   },
michael@0:   STANDARD_VTABLE(big2_)
michael@0: };
michael@0: 
michael@0: #endif
michael@0: 
michael@0: static const struct normal_encoding internal_big2_encoding = {
michael@0:   { VTABLE, 2, 0, 1 },
michael@0:   {
michael@0: #define BT_COLON BT_NMSTRT
michael@0: #include "iasciitab.h"
michael@0: #undef BT_COLON
michael@0: #include "latin1tab.h"
michael@0:   },
michael@0:   STANDARD_VTABLE(big2_)
michael@0: };
michael@0: 
michael@0: #endif
michael@0: 
michael@0: #undef PREFIX
michael@0: 
michael@0: static int FASTCALL
michael@0: streqci(const char *s1, const char *s2)
michael@0: {
michael@0:   for (;;) {
michael@0:     char c1 = *s1++;
michael@0:     char c2 = *s2++;
michael@0:     if (ASCII_a <= c1 && c1 <= ASCII_z)
michael@0:       c1 += ASCII_A - ASCII_a;
michael@0:     if (ASCII_a <= c2 && c2 <= ASCII_z)
michael@0:       c2 += ASCII_A - ASCII_a;
michael@0:     if (c1 != c2)
michael@0:       return 0;
michael@0:     if (!c1)
michael@0:       break;
michael@0:   }
michael@0:   return 1;
michael@0: }
michael@0: 
michael@0: static void PTRCALL
michael@0: initUpdatePosition(const ENCODING *enc, const char *ptr,
michael@0:                    const char *end, POSITION *pos)
michael@0: {
michael@0:   normal_updatePosition(&utf8_encoding.enc, ptr, end, pos);
michael@0: }
michael@0: 
michael@0: static int
michael@0: toAscii(const ENCODING *enc, const char *ptr, const char *end)
michael@0: {
michael@0:   char buf[1];
michael@0:   char *p = buf;
michael@0:   XmlUtf8Convert(enc, &ptr, end, &p, p + 1);
michael@0:   if (p == buf)
michael@0:     return -1;
michael@0:   else
michael@0:     return buf[0];
michael@0: }
michael@0: 
michael@0: static int FASTCALL
michael@0: isSpace(int c)
michael@0: {
michael@0:   switch (c) {
michael@0:   case 0x20:
michael@0:   case 0xD:
michael@0:   case 0xA:
michael@0:   case 0x9:
michael@0:     return 1;
michael@0:   }
michael@0:   return 0;
michael@0: }
michael@0: 
michael@0: /* Return 1 if there's just optional white space or there's an S
michael@0:    followed by name=val.
michael@0: */
michael@0: static int
michael@0: parsePseudoAttribute(const ENCODING *enc,
michael@0:                      const char *ptr,
michael@0:                      const char *end,
michael@0:                      const char **namePtr,
michael@0:                      const char **nameEndPtr,
michael@0:                      const char **valPtr,
michael@0:                      const char **nextTokPtr)
michael@0: {
michael@0:   int c;
michael@0:   char open;
michael@0:   if (ptr == end) {
michael@0:     *namePtr = NULL;
michael@0:     return 1;
michael@0:   }
michael@0:   if (!isSpace(toAscii(enc, ptr, end))) {
michael@0:     *nextTokPtr = ptr;
michael@0:     return 0;
michael@0:   }
michael@0:   do {
michael@0:     ptr += enc->minBytesPerChar;
michael@0:   } while (isSpace(toAscii(enc, ptr, end)));
michael@0:   if (ptr == end) {
michael@0:     *namePtr = NULL;
michael@0:     return 1;
michael@0:   }
michael@0:   *namePtr = ptr;
michael@0:   for (;;) {
michael@0:     c = toAscii(enc, ptr, end);
michael@0:     if (c == -1) {
michael@0:       *nextTokPtr = ptr;
michael@0:       return 0;
michael@0:     }
michael@0:     if (c == ASCII_EQUALS) {
michael@0:       *nameEndPtr = ptr;
michael@0:       break;
michael@0:     }
michael@0:     if (isSpace(c)) {
michael@0:       *nameEndPtr = ptr;
michael@0:       do {
michael@0:         ptr += enc->minBytesPerChar;
michael@0:       } while (isSpace(c = toAscii(enc, ptr, end)));
michael@0:       if (c != ASCII_EQUALS) {
michael@0:         *nextTokPtr = ptr;
michael@0:         return 0;
michael@0:       }
michael@0:       break;
michael@0:     }
michael@0:     ptr += enc->minBytesPerChar;
michael@0:   }
michael@0:   if (ptr == *namePtr) {
michael@0:     *nextTokPtr = ptr;
michael@0:     return 0;
michael@0:   }
michael@0:   ptr += enc->minBytesPerChar;
michael@0:   c = toAscii(enc, ptr, end);
michael@0:   while (isSpace(c)) {
michael@0:     ptr += enc->minBytesPerChar;
michael@0:     c = toAscii(enc, ptr, end);
michael@0:   }
michael@0:   if (c != ASCII_QUOT && c != ASCII_APOS) {
michael@0:     *nextTokPtr = ptr;
michael@0:     return 0;
michael@0:   }
michael@0:   open = (char)c;
michael@0:   ptr += enc->minBytesPerChar;
michael@0:   *valPtr = ptr;
michael@0:   for (;; ptr += enc->minBytesPerChar) {
michael@0:     c = toAscii(enc, ptr, end);
michael@0:     if (c == open)
michael@0:       break;
michael@0:     if (!(ASCII_a <= c && c <= ASCII_z)
michael@0:         && !(ASCII_A <= c && c <= ASCII_Z)
michael@0:         && !(ASCII_0 <= c && c <= ASCII_9)
michael@0:         && c != ASCII_PERIOD
michael@0:         && c != ASCII_MINUS
michael@0:         && c != ASCII_UNDERSCORE) {
michael@0:       *nextTokPtr = ptr;
michael@0:       return 0;
michael@0:     }
michael@0:   }
michael@0:   *nextTokPtr = ptr + enc->minBytesPerChar;
michael@0:   return 1;
michael@0: }
michael@0: 
michael@0: static const char KW_version[] = {
michael@0:   ASCII_v, ASCII_e, ASCII_r, ASCII_s, ASCII_i, ASCII_o, ASCII_n, '\0'
michael@0: };
michael@0: 
michael@0: static const char KW_encoding[] = {
michael@0:   ASCII_e, ASCII_n, ASCII_c, ASCII_o, ASCII_d, ASCII_i, ASCII_n, ASCII_g, '\0'
michael@0: };
michael@0: 
michael@0: static const char KW_standalone[] = {
michael@0:   ASCII_s, ASCII_t, ASCII_a, ASCII_n, ASCII_d, ASCII_a, ASCII_l, ASCII_o,
michael@0:   ASCII_n, ASCII_e, '\0'
michael@0: };
michael@0: 
michael@0: static const char KW_yes[] = {
michael@0:   ASCII_y, ASCII_e, ASCII_s,  '\0'
michael@0: };
michael@0: 
michael@0: static const char KW_no[] = {
michael@0:   ASCII_n, ASCII_o,  '\0'
michael@0: };
michael@0: 
michael@0: /* BEGIN MOZILLA CHANGE (http://bugzilla.mozilla.org/show_bug.cgi?id=62157) */
michael@0: static const char KW_XML_1_0[] = {
michael@0:   ASCII_1, ASCII_PERIOD, ASCII_0, '\0'
michael@0: };
michael@0: /* END MOZILLA CHANGE */
michael@0: 
michael@0: static int
michael@0: doParseXmlDecl(const ENCODING *(*encodingFinder)(const ENCODING *,
michael@0:                                                  const char *,
michael@0:                                                  const char *),
michael@0:                int isGeneralTextEntity,
michael@0:                const ENCODING *enc,
michael@0:                const char *ptr,
michael@0:                const char *end,
michael@0:                const char **badPtr,
michael@0:                const char **versionPtr,
michael@0:                const char **versionEndPtr,
michael@0:                const char **encodingName,
michael@0:                const ENCODING **encoding,
michael@0:                int *standalone)
michael@0: {
michael@0:   const char *val = NULL;
michael@0:   const char *name = NULL;
michael@0:   const char *nameEnd = NULL;
michael@0:   ptr += 5 * enc->minBytesPerChar;
michael@0:   end -= 2 * enc->minBytesPerChar;
michael@0:   if (!parsePseudoAttribute(enc, ptr, end, &name, &nameEnd, &val, &ptr)
michael@0:       || !name) {
michael@0:     *badPtr = ptr;
michael@0:     return 0;
michael@0:   }
michael@0:   if (!XmlNameMatchesAscii(enc, name, nameEnd, KW_version)) {
michael@0:     if (!isGeneralTextEntity) {
michael@0:       *badPtr = name;
michael@0:       return 0;
michael@0:     }
michael@0:   }
michael@0:   else {
michael@0:     if (versionPtr)
michael@0:       *versionPtr = val;
michael@0:     if (versionEndPtr)
michael@0:       *versionEndPtr = ptr;
michael@0: /* BEGIN MOZILLA CHANGE (http://bugzilla.mozilla.org/show_bug.cgi?id=62157) */
michael@0:      /* Anything else but a version="1.0" is invalid for us, until we support later versions. */
michael@0:      if (!XmlNameMatchesAscii(enc, val, ptr - enc->minBytesPerChar, KW_XML_1_0)) {
michael@0:        *badPtr = val;
michael@0:        return 0;
michael@0:      }
michael@0: /* END MOZILLA CHANGE */
michael@0:     if (!parsePseudoAttribute(enc, ptr, end, &name, &nameEnd, &val, &ptr)) {
michael@0:       *badPtr = ptr;
michael@0:       return 0;
michael@0:     }
michael@0:     if (!name) {
michael@0:       if (isGeneralTextEntity) {
michael@0:         /* a TextDecl must have an EncodingDecl */
michael@0:         *badPtr = ptr;
michael@0:         return 0;
michael@0:       }
michael@0:       return 1;
michael@0:     }
michael@0:   }
michael@0:   if (XmlNameMatchesAscii(enc, name, nameEnd, KW_encoding)) {
michael@0:     int c = toAscii(enc, val, end);
michael@0:     if (!(ASCII_a <= c && c <= ASCII_z) && !(ASCII_A <= c && c <= ASCII_Z)) {
michael@0:       *badPtr = val;
michael@0:       return 0;
michael@0:     }
michael@0:     if (encodingName)
michael@0:       *encodingName = val;
michael@0:     if (encoding)
michael@0:       *encoding = encodingFinder(enc, val, ptr - enc->minBytesPerChar);
michael@0:     if (!parsePseudoAttribute(enc, ptr, end, &name, &nameEnd, &val, &ptr)) {
michael@0:       *badPtr = ptr;
michael@0:       return 0;
michael@0:     }
michael@0:     if (!name)
michael@0:       return 1;
michael@0:   }
michael@0:   if (!XmlNameMatchesAscii(enc, name, nameEnd, KW_standalone)
michael@0:       || isGeneralTextEntity) {
michael@0:     *badPtr = name;
michael@0:     return 0;
michael@0:   }
michael@0:   if (XmlNameMatchesAscii(enc, val, ptr - enc->minBytesPerChar, KW_yes)) {
michael@0:     if (standalone)
michael@0:       *standalone = 1;
michael@0:   }
michael@0:   else if (XmlNameMatchesAscii(enc, val, ptr - enc->minBytesPerChar, KW_no)) {
michael@0:     if (standalone)
michael@0:       *standalone = 0;
michael@0:   }
michael@0:   else {
michael@0:     *badPtr = val;
michael@0:     return 0;
michael@0:   }
michael@0:   while (isSpace(toAscii(enc, ptr, end)))
michael@0:     ptr += enc->minBytesPerChar;
michael@0:   if (ptr != end) {
michael@0:     *badPtr = ptr;
michael@0:     return 0;
michael@0:   }
michael@0:   return 1;
michael@0: }
michael@0: 
michael@0: static int FASTCALL
michael@0: checkCharRefNumber(int result)
michael@0: {
michael@0:   switch (result >> 8) {
michael@0:   case 0xD8: case 0xD9: case 0xDA: case 0xDB:
michael@0:   case 0xDC: case 0xDD: case 0xDE: case 0xDF:
michael@0:     return -1;
michael@0:   case 0:
michael@0:     if (latin1_encoding.type[result] == BT_NONXML)
michael@0:       return -1;
michael@0:     break;
michael@0:   case 0xFF:
michael@0:     if (result == 0xFFFE || result == 0xFFFF)
michael@0:       return -1;
michael@0:     break;
michael@0:   }
michael@0:   return result;
michael@0: }
michael@0: 
michael@0: int FASTCALL
michael@0: XmlUtf8Encode(int c, char *buf)
michael@0: {
michael@0:   enum {
michael@0:     /* minN is minimum legal resulting value for N byte sequence */
michael@0:     min2 = 0x80,
michael@0:     min3 = 0x800,
michael@0:     min4 = 0x10000
michael@0:   };
michael@0: 
michael@0:   if (c < 0)
michael@0:     return 0;
michael@0:   if (c < min2) {
michael@0:     buf[0] = (char)(c | UTF8_cval1);
michael@0:     return 1;
michael@0:   }
michael@0:   if (c < min3) {
michael@0:     buf[0] = (char)((c >> 6) | UTF8_cval2);
michael@0:     buf[1] = (char)((c & 0x3f) | 0x80);
michael@0:     return 2;
michael@0:   }
michael@0:   if (c < min4) {
michael@0:     buf[0] = (char)((c >> 12) | UTF8_cval3);
michael@0:     buf[1] = (char)(((c >> 6) & 0x3f) | 0x80);
michael@0:     buf[2] = (char)((c & 0x3f) | 0x80);
michael@0:     return 3;
michael@0:   }
michael@0:   if (c < 0x110000) {
michael@0:     buf[0] = (char)((c >> 18) | UTF8_cval4);
michael@0:     buf[1] = (char)(((c >> 12) & 0x3f) | 0x80);
michael@0:     buf[2] = (char)(((c >> 6) & 0x3f) | 0x80);
michael@0:     buf[3] = (char)((c & 0x3f) | 0x80);
michael@0:     return 4;
michael@0:   }
michael@0:   return 0;
michael@0: }
michael@0: 
michael@0: int FASTCALL
michael@0: XmlUtf16Encode(int charNum, unsigned short *buf)
michael@0: {
michael@0:   if (charNum < 0)
michael@0:     return 0;
michael@0:   if (charNum < 0x10000) {
michael@0:     buf[0] = (unsigned short)charNum;
michael@0:     return 1;
michael@0:   }
michael@0:   if (charNum < 0x110000) {
michael@0:     charNum -= 0x10000;
michael@0:     buf[0] = (unsigned short)((charNum >> 10) + 0xD800);
michael@0:     buf[1] = (unsigned short)((charNum & 0x3FF) + 0xDC00);
michael@0:     return 2;
michael@0:   }
michael@0:   return 0;
michael@0: }
michael@0: 
michael@0: struct unknown_encoding {
michael@0:   struct normal_encoding normal;
michael@0:   CONVERTER convert;
michael@0:   void *userData;
michael@0:   unsigned short utf16[256];
michael@0:   char utf8[256][4];
michael@0: };
michael@0: 
michael@0: #define AS_UNKNOWN_ENCODING(enc)  ((const struct unknown_encoding *) (enc))
michael@0: 
michael@0: int
michael@0: XmlSizeOfUnknownEncoding(void)
michael@0: {
michael@0:   return sizeof(struct unknown_encoding);
michael@0: }
michael@0: 
michael@0: static int PTRFASTCALL
michael@0: unknown_isName(const ENCODING *enc, const char *p)
michael@0: {
michael@0:   const struct unknown_encoding *uenc = AS_UNKNOWN_ENCODING(enc);
michael@0:   int c = uenc->convert(uenc->userData, p);
michael@0:   if (c & ~0xFFFF)
michael@0:     return 0;
michael@0:   return UCS2_GET_NAMING(namePages, c >> 8, c & 0xFF);
michael@0: }
michael@0: 
michael@0: static int PTRFASTCALL
michael@0: unknown_isNmstrt(const ENCODING *enc, const char *p)
michael@0: {
michael@0:   const struct unknown_encoding *uenc = AS_UNKNOWN_ENCODING(enc);
michael@0:   int c = uenc->convert(uenc->userData, p);
michael@0:   if (c & ~0xFFFF)
michael@0:     return 0;
michael@0:   return UCS2_GET_NAMING(nmstrtPages, c >> 8, c & 0xFF);
michael@0: }
michael@0: 
michael@0: static int PTRFASTCALL
michael@0: unknown_isInvalid(const ENCODING *enc, const char *p)
michael@0: {
michael@0:   const struct unknown_encoding *uenc = AS_UNKNOWN_ENCODING(enc);
michael@0:   int c = uenc->convert(uenc->userData, p);
michael@0:   return (c & ~0xFFFF) || checkCharRefNumber(c) < 0;
michael@0: }
michael@0: 
michael@0: static void PTRCALL
michael@0: unknown_toUtf8(const ENCODING *enc,
michael@0:                const char **fromP, const char *fromLim,
michael@0:                char **toP, const char *toLim)
michael@0: {
michael@0:   const struct unknown_encoding *uenc = AS_UNKNOWN_ENCODING(enc);
michael@0:   char buf[XML_UTF8_ENCODE_MAX];
michael@0:   for (;;) {
michael@0:     const char *utf8;
michael@0:     int n;
michael@0:     if (*fromP == fromLim)
michael@0:       break;
michael@0:     utf8 = uenc->utf8[(unsigned char)**fromP];
michael@0:     n = *utf8++;
michael@0:     if (n == 0) {
michael@0:       int c = uenc->convert(uenc->userData, *fromP);
michael@0:       n = XmlUtf8Encode(c, buf);
michael@0:       if (n > toLim - *toP)
michael@0:         break;
michael@0:       utf8 = buf;
michael@0:       *fromP += (AS_NORMAL_ENCODING(enc)->type[(unsigned char)**fromP]
michael@0:                  - (BT_LEAD2 - 2));
michael@0:     }
michael@0:     else {
michael@0:       if (n > toLim - *toP)
michael@0:         break;
michael@0:       (*fromP)++;
michael@0:     }
michael@0:     do {
michael@0:       *(*toP)++ = *utf8++;
michael@0:     } while (--n != 0);
michael@0:   }
michael@0: }
michael@0: 
michael@0: static void PTRCALL
michael@0: unknown_toUtf16(const ENCODING *enc,
michael@0:                 const char **fromP, const char *fromLim,
michael@0:                 unsigned short **toP, const unsigned short *toLim)
michael@0: {
michael@0:   const struct unknown_encoding *uenc = AS_UNKNOWN_ENCODING(enc);
michael@0:   while (*fromP != fromLim && *toP != toLim) {
michael@0:     unsigned short c = uenc->utf16[(unsigned char)**fromP];
michael@0:     if (c == 0) {
michael@0:       c = (unsigned short)
michael@0:           uenc->convert(uenc->userData, *fromP);
michael@0:       *fromP += (AS_NORMAL_ENCODING(enc)->type[(unsigned char)**fromP]
michael@0:                  - (BT_LEAD2 - 2));
michael@0:     }
michael@0:     else
michael@0:       (*fromP)++;
michael@0:     *(*toP)++ = c;
michael@0:   }
michael@0: }
michael@0: 
michael@0: ENCODING *
michael@0: XmlInitUnknownEncoding(void *mem,
michael@0:                        int *table,
michael@0:                        CONVERTER convert, 
michael@0:                        void *userData)
michael@0: {
michael@0:   int i;
michael@0:   struct unknown_encoding *e = (struct unknown_encoding *)mem;
michael@0:   for (i = 0; i < (int)sizeof(struct normal_encoding); i++)
michael@0:     ((char *)mem)[i] = ((char *)&latin1_encoding)[i];
michael@0:   for (i = 0; i < 128; i++)
michael@0:     if (latin1_encoding.type[i] != BT_OTHER
michael@0:         && latin1_encoding.type[i] != BT_NONXML
michael@0:         && table[i] != i)
michael@0:       return 0;
michael@0:   for (i = 0; i < 256; i++) {
michael@0:     int c = table[i];
michael@0:     if (c == -1) {
michael@0:       e->normal.type[i] = BT_MALFORM;
michael@0:       /* This shouldn't really get used. */
michael@0:       e->utf16[i] = 0xFFFF;
michael@0:       e->utf8[i][0] = 1;
michael@0:       e->utf8[i][1] = 0;
michael@0:     }
michael@0:     else if (c < 0) {
michael@0:       if (c < -4)
michael@0:         return 0;
michael@0:       e->normal.type[i] = (unsigned char)(BT_LEAD2 - (c + 2));
michael@0:       e->utf8[i][0] = 0;
michael@0:       e->utf16[i] = 0;
michael@0:     }
michael@0:     else if (c < 0x80) {
michael@0:       if (latin1_encoding.type[c] != BT_OTHER
michael@0:           && latin1_encoding.type[c] != BT_NONXML
michael@0:           && c != i)
michael@0:         return 0;
michael@0:       e->normal.type[i] = latin1_encoding.type[c];
michael@0:       e->utf8[i][0] = 1;
michael@0:       e->utf8[i][1] = (char)c;
michael@0:       e->utf16[i] = (unsigned short)(c == 0 ? 0xFFFF : c);
michael@0:     }
michael@0:     else if (checkCharRefNumber(c) < 0) {
michael@0:       e->normal.type[i] = BT_NONXML;
michael@0:       /* This shouldn't really get used. */
michael@0:       e->utf16[i] = 0xFFFF;
michael@0:       e->utf8[i][0] = 1;
michael@0:       e->utf8[i][1] = 0;
michael@0:     }
michael@0:     else {
michael@0:       if (c > 0xFFFF)
michael@0:         return 0;
michael@0:       if (UCS2_GET_NAMING(nmstrtPages, c >> 8, c & 0xff))
michael@0:         e->normal.type[i] = BT_NMSTRT;
michael@0:       else if (UCS2_GET_NAMING(namePages, c >> 8, c & 0xff))
michael@0:         e->normal.type[i] = BT_NAME;
michael@0:       else
michael@0:         e->normal.type[i] = BT_OTHER;
michael@0:       e->utf8[i][0] = (char)XmlUtf8Encode(c, e->utf8[i] + 1);
michael@0:       e->utf16[i] = (unsigned short)c;
michael@0:     }
michael@0:   }
michael@0:   e->userData = userData;
michael@0:   e->convert = convert;
michael@0:   if (convert) {
michael@0:     e->normal.isName2 = unknown_isName;
michael@0:     e->normal.isName3 = unknown_isName;
michael@0:     e->normal.isName4 = unknown_isName;
michael@0:     e->normal.isNmstrt2 = unknown_isNmstrt;
michael@0:     e->normal.isNmstrt3 = unknown_isNmstrt;
michael@0:     e->normal.isNmstrt4 = unknown_isNmstrt;
michael@0:     e->normal.isInvalid2 = unknown_isInvalid;
michael@0:     e->normal.isInvalid3 = unknown_isInvalid;
michael@0:     e->normal.isInvalid4 = unknown_isInvalid;
michael@0:   }
michael@0:   e->normal.enc.utf8Convert = unknown_toUtf8;
michael@0:   e->normal.enc.utf16Convert = unknown_toUtf16;
michael@0:   return &(e->normal.enc);
michael@0: }
michael@0: 
michael@0: /* If this enumeration is changed, getEncodingIndex and encodings
michael@0: must also be changed. */
michael@0: enum {
michael@0:   UNKNOWN_ENC = -1,
michael@0:   ISO_8859_1_ENC = 0,
michael@0:   US_ASCII_ENC,
michael@0:   UTF_8_ENC,
michael@0:   UTF_16_ENC,
michael@0:   UTF_16BE_ENC,
michael@0:   UTF_16LE_ENC,
michael@0:   /* must match encodingNames up to here */
michael@0:   NO_ENC
michael@0: };
michael@0: 
michael@0: static const char KW_ISO_8859_1[] = {
michael@0:   ASCII_I, ASCII_S, ASCII_O, ASCII_MINUS, ASCII_8, ASCII_8, ASCII_5, ASCII_9,
michael@0:   ASCII_MINUS, ASCII_1, '\0'
michael@0: };
michael@0: static const char KW_US_ASCII[] = {
michael@0:   ASCII_U, ASCII_S, ASCII_MINUS, ASCII_A, ASCII_S, ASCII_C, ASCII_I, ASCII_I,
michael@0:   '\0'
michael@0: };
michael@0: static const char KW_UTF_8[] =  {
michael@0:   ASCII_U, ASCII_T, ASCII_F, ASCII_MINUS, ASCII_8, '\0'
michael@0: };
michael@0: static const char KW_UTF_16[] = {
michael@0:   ASCII_U, ASCII_T, ASCII_F, ASCII_MINUS, ASCII_1, ASCII_6, '\0'
michael@0: };
michael@0: static const char KW_UTF_16BE[] = {
michael@0:   ASCII_U, ASCII_T, ASCII_F, ASCII_MINUS, ASCII_1, ASCII_6, ASCII_B, ASCII_E,
michael@0:   '\0'
michael@0: };
michael@0: static const char KW_UTF_16LE[] = {
michael@0:   ASCII_U, ASCII_T, ASCII_F, ASCII_MINUS, ASCII_1, ASCII_6, ASCII_L, ASCII_E,
michael@0:   '\0'
michael@0: };
michael@0: 
michael@0: static int FASTCALL
michael@0: getEncodingIndex(const char *name)
michael@0: {
michael@0:   static const char * const encodingNames[] = {
michael@0:     KW_ISO_8859_1,
michael@0:     KW_US_ASCII,
michael@0:     KW_UTF_8,
michael@0:     KW_UTF_16,
michael@0:     KW_UTF_16BE,
michael@0:     KW_UTF_16LE,
michael@0:   };
michael@0:   int i;
michael@0:   if (name == NULL)
michael@0:     return NO_ENC;
michael@0:   for (i = 0; i < (int)(sizeof(encodingNames)/sizeof(encodingNames[0])); i++)
michael@0:     if (streqci(name, encodingNames[i]))
michael@0:       return i;
michael@0:   return UNKNOWN_ENC;
michael@0: }
michael@0: 
michael@0: /* For binary compatibility, we store the index of the encoding
michael@0:    specified at initialization in the isUtf16 member.
michael@0: */
michael@0: 
michael@0: #define INIT_ENC_INDEX(enc) ((int)(enc)->initEnc.isUtf16)
michael@0: #define SET_INIT_ENC_INDEX(enc, i) ((enc)->initEnc.isUtf16 = (char)i)
michael@0: 
michael@0: /* This is what detects the encoding.  encodingTable maps from
michael@0:    encoding indices to encodings; INIT_ENC_INDEX(enc) is the index of
michael@0:    the external (protocol) specified encoding; state is
michael@0:    XML_CONTENT_STATE if we're parsing an external text entity, and
michael@0:    XML_PROLOG_STATE otherwise.
michael@0: */
michael@0: 
michael@0: 
michael@0: static int
michael@0: initScan(const ENCODING * const *encodingTable,
michael@0:          const INIT_ENCODING *enc,
michael@0:          int state,
michael@0:          const char *ptr,
michael@0:          const char *end,
michael@0:          const char **nextTokPtr)
michael@0: {
michael@0:   const ENCODING **encPtr;
michael@0: 
michael@0:   if (ptr == end)
michael@0:     return XML_TOK_NONE;
michael@0:   encPtr = enc->encPtr;
michael@0:   if (ptr + 1 == end) {
michael@0:     /* only a single byte available for auto-detection */
michael@0: #ifndef XML_DTD /* FIXME */
michael@0:     /* a well-formed document entity must have more than one byte */
michael@0:     if (state != XML_CONTENT_STATE)
michael@0:       return XML_TOK_PARTIAL;
michael@0: #endif
michael@0:     /* so we're parsing an external text entity... */
michael@0:     /* if UTF-16 was externally specified, then we need at least 2 bytes */
michael@0:     switch (INIT_ENC_INDEX(enc)) {
michael@0:     case UTF_16_ENC:
michael@0:     case UTF_16LE_ENC:
michael@0:     case UTF_16BE_ENC:
michael@0:       return XML_TOK_PARTIAL;
michael@0:     }
michael@0:     switch ((unsigned char)*ptr) {
michael@0:     case 0xFE:
michael@0:     case 0xFF:
michael@0:     case 0xEF: /* possibly first byte of UTF-8 BOM */
michael@0:       if (INIT_ENC_INDEX(enc) == ISO_8859_1_ENC
michael@0:           && state == XML_CONTENT_STATE)
michael@0:         break;
michael@0:       /* fall through */
michael@0:     case 0x00:
michael@0:     case 0x3C:
michael@0:       return XML_TOK_PARTIAL;
michael@0:     }
michael@0:   }
michael@0:   else {
michael@0:     switch (((unsigned char)ptr[0] << 8) | (unsigned char)ptr[1]) {
michael@0:     case 0xFEFF:
michael@0:       if (INIT_ENC_INDEX(enc) == ISO_8859_1_ENC
michael@0:           && state == XML_CONTENT_STATE)
michael@0:         break;
michael@0:       *nextTokPtr = ptr + 2;
michael@0:       *encPtr = encodingTable[UTF_16BE_ENC];
michael@0:       return XML_TOK_BOM;
michael@0:     /* 00 3C is handled in the default case */
michael@0:     case 0x3C00:
michael@0:       if ((INIT_ENC_INDEX(enc) == UTF_16BE_ENC
michael@0:            || INIT_ENC_INDEX(enc) == UTF_16_ENC)
michael@0:           && state == XML_CONTENT_STATE)
michael@0:         break;
michael@0:       *encPtr = encodingTable[UTF_16LE_ENC];
michael@0:       return XmlTok(*encPtr, state, ptr, end, nextTokPtr);
michael@0:     case 0xFFFE:
michael@0:       if (INIT_ENC_INDEX(enc) == ISO_8859_1_ENC
michael@0:           && state == XML_CONTENT_STATE)
michael@0:         break;
michael@0:       *nextTokPtr = ptr + 2;
michael@0:       *encPtr = encodingTable[UTF_16LE_ENC];
michael@0:       return XML_TOK_BOM;
michael@0:     case 0xEFBB:
michael@0:       /* Maybe a UTF-8 BOM (EF BB BF) */
michael@0:       /* If there's an explicitly specified (external) encoding
michael@0:          of ISO-8859-1 or some flavour of UTF-16
michael@0:          and this is an external text entity,
michael@0:          don't look for the BOM,
michael@0:          because it might be a legal data.
michael@0:       */
michael@0:       if (state == XML_CONTENT_STATE) {
michael@0:         int e = INIT_ENC_INDEX(enc);
michael@0:         if (e == ISO_8859_1_ENC || e == UTF_16BE_ENC
michael@0:             || e == UTF_16LE_ENC || e == UTF_16_ENC)
michael@0:           break;
michael@0:       }
michael@0:       if (ptr + 2 == end)
michael@0:         return XML_TOK_PARTIAL;
michael@0:       if ((unsigned char)ptr[2] == 0xBF) {
michael@0:         *nextTokPtr = ptr + 3;
michael@0:         *encPtr = encodingTable[UTF_8_ENC];
michael@0:         return XML_TOK_BOM;
michael@0:       }
michael@0:       break;
michael@0:     default:
michael@0:       if (ptr[0] == '\0') {
michael@0:         /* 0 isn't a legal data character. Furthermore a document
michael@0:            entity can only start with ASCII characters.  So the only
michael@0:            way this can fail to be big-endian UTF-16 if it it's an
michael@0:            external parsed general entity that's labelled as
michael@0:            UTF-16LE.
michael@0:         */
michael@0:         if (state == XML_CONTENT_STATE && INIT_ENC_INDEX(enc) == UTF_16LE_ENC)
michael@0:           break;
michael@0:         *encPtr = encodingTable[UTF_16BE_ENC];
michael@0:         return XmlTok(*encPtr, state, ptr, end, nextTokPtr);
michael@0:       }
michael@0:       else if (ptr[1] == '\0') {
michael@0:         /* We could recover here in the case:
michael@0:             - parsing an external entity
michael@0:             - second byte is 0
michael@0:             - no externally specified encoding
michael@0:             - no encoding declaration
michael@0:            by assuming UTF-16LE.  But we don't, because this would mean when
michael@0:            presented just with a single byte, we couldn't reliably determine
michael@0:            whether we needed further bytes.
michael@0:         */
michael@0:         if (state == XML_CONTENT_STATE)
michael@0:           break;
michael@0:         *encPtr = encodingTable[UTF_16LE_ENC];
michael@0:         return XmlTok(*encPtr, state, ptr, end, nextTokPtr);
michael@0:       }
michael@0:       break;
michael@0:     }
michael@0:   }
michael@0:   *encPtr = encodingTable[INIT_ENC_INDEX(enc)];
michael@0:   return XmlTok(*encPtr, state, ptr, end, nextTokPtr);
michael@0: }
michael@0: 
michael@0: 
michael@0: #define NS(x) x
michael@0: #define ns(x) x
michael@0: #include "xmltok_ns.c"
michael@0: #undef NS
michael@0: #undef ns
michael@0: 
michael@0: #ifdef XML_NS
michael@0: 
michael@0: #define NS(x) x ## NS
michael@0: #define ns(x) x ## _ns
michael@0: 
michael@0: #include "xmltok_ns.c"
michael@0: 
michael@0: #undef NS
michael@0: #undef ns
michael@0: 
michael@0: ENCODING *
michael@0: XmlInitUnknownEncodingNS(void *mem,
michael@0:                          int *table,
michael@0:                          CONVERTER convert, 
michael@0:                          void *userData)
michael@0: {
michael@0:   ENCODING *enc = XmlInitUnknownEncoding(mem, table, convert, userData);
michael@0:   if (enc)
michael@0:     ((struct normal_encoding *)enc)->type[ASCII_COLON] = BT_COLON;
michael@0:   return enc;
michael@0: }
michael@0: 
michael@0: #endif /* XML_NS */
michael@0: 
michael@0: /* BEGIN MOZILLA CHANGE (Mozilla extensions for QName checking) */
michael@0: #ifdef MOZILLA_CLIENT
michael@0: #include "moz_extensions.c"
michael@0: #endif /* MOZILLA_CLIENT */
michael@0: /* END MOZILLA CHANGE */