static int PTRCALL
PREFIX(predefinedEntityName)(const ENCODING *enc, const char *ptr,
const char *end)
{
UNUSED(enc);
switch ((end - ptr)/MINBPC(enc)) {
case 2:
if (CHAR_MATCHES(enc, ptr + MINBPC(enc), ASCII_t)) {
switch (BYTE_TO_ASCII(enc, ptr)) {
case ASCII_l:
return ASCII_LT;
case ASCII_g:
return ASCII_GT;
}
}
break;
case 3:
if (CHAR_MATCHES(enc, ptr, ASCII_a)) {
ptr += MINBPC(enc);
if (CHAR_MATCHES(enc, ptr, ASCII_m)) {
ptr += MINBPC(enc);
if (CHAR_MATCHES(enc, ptr, ASCII_p))
return ASCII_AMP;
}
}
break;
case 4:
switch (BYTE_TO_ASCII(enc, ptr)) {
case ASCII_q:
ptr += MINBPC(enc);
if (CHAR_MATCHES(enc, ptr, ASCII_u)) {
ptr += MINBPC(enc);
if (CHAR_MATCHES(enc, ptr, ASCII_o)) {
ptr += MINBPC(enc);
if (CHAR_MATCHES(enc, ptr, ASCII_t))
return ASCII_QUOT;
}
}
break;
case ASCII_a:
ptr += MINBPC(enc);
if (CHAR_MATCHES(enc, ptr, ASCII_p)) {
ptr += MINBPC(enc);
if (CHAR_MATCHES(enc, ptr, ASCII_o)) {
ptr += MINBPC(enc);
if (CHAR_MATCHES(enc, ptr, ASCII_s))
return ASCII_APOS;
}
}
break;
}
}
return 0;
}
static int PTRCALL
PREFIX(isPublicId)(const ENCODING *enc, const char *ptr, const char *end,
const char **badPtr)
{
ptr += MINBPC(enc);
end -= MINBPC(enc);
for (; HAS_CHAR(enc, ptr, end); ptr += MINBPC(enc)) {
switch (BYTE_TYPE(enc, ptr)) {
case BT_DIGIT:
case BT_HEX:
case BT_MINUS:
case BT_APOS:
case BT_LPAR:
case BT_RPAR:
case BT_PLUS:
case BT_COMMA:
case BT_SOL:
case BT_EQUALS:
case BT_QUEST:
case BT_CR:
case BT_LF:
case BT_SEMI:
case BT_EXCL:
case BT_AST:
case BT_PERCNT:
case BT_NUM:
#ifdef XML_NS
case BT_COLON:
#endif
break;
case BT_S:
if (CHAR_MATCHES(enc, ptr, ASCII_TAB)) {
*badPtr = ptr;
return 0;
}
break;
case BT_NAME:
case BT_NMSTRT:
if (!(BYTE_TO_ASCII(enc, ptr) & ~0x7f))
break;
/* fall through */
default:
switch (BYTE_TO_ASCII(enc, ptr)) {
case 0x24: /* $ */
case 0x40: /* @ */
break;
default:
*badPtr = ptr;
return 0;
}
break;
}
}
return 1;
}
static
int PREFIX(predefinedEntityName)(const ENCODING *enc, const char *ptr, const char *end)
{
(void)enc;
switch (end - ptr) {
case 2 * MINBPC:
if (CHAR_MATCHES(enc, ptr + MINBPC, 't')) {
switch (BYTE_TO_ASCII(enc, ptr)) {
case 'l':
return '<';
case 'g':
return '>';
}
}
break;
case 3 * MINBPC:
if (CHAR_MATCHES(enc, ptr, 'a')) {
ptr += MINBPC;
if (CHAR_MATCHES(enc, ptr, 'm')) {
ptr += MINBPC;
if (CHAR_MATCHES(enc, ptr, 'p'))
return '&';
}
}
break;
case 4 * MINBPC:
switch (BYTE_TO_ASCII(enc, ptr)) {
case 'q':
ptr += MINBPC;
if (CHAR_MATCHES(enc, ptr, 'u')) {
ptr += MINBPC;
if (CHAR_MATCHES(enc, ptr, 'o')) {
ptr += MINBPC;
if (CHAR_MATCHES(enc, ptr, 't'))
return '"';
}
}
break;
case 'a':
ptr += MINBPC;
if (CHAR_MATCHES(enc, ptr, 'p')) {
ptr += MINBPC;
if (CHAR_MATCHES(enc, ptr, 'o')) {
ptr += MINBPC;
if (CHAR_MATCHES(enc, ptr, 's'))
return '\'';
}
}
break;
}
}
return 0;
}
static int PTRFASTCALL
PREFIX(charRefNumber)(const ENCODING *enc, const char *ptr)
{
int result = 0;
/* skip &# */
ptr += 2*MINBPC(enc);
if (CHAR_MATCHES(enc, ptr, ASCII_x)) {
for (ptr += MINBPC(enc);
!CHAR_MATCHES(enc, ptr, ASCII_SEMI);
ptr += MINBPC(enc)) {
int c = BYTE_TO_ASCII(enc, ptr);
switch (c) {
case ASCII_0: case ASCII_1: case ASCII_2: case ASCII_3: case ASCII_4:
case ASCII_5: case ASCII_6: case ASCII_7: case ASCII_8: case ASCII_9:
result <<= 4;
result |= (c - ASCII_0);
break;
case ASCII_A: case ASCII_B: case ASCII_C:
case ASCII_D: case ASCII_E: case ASCII_F:
result <<= 4;
result += 10 + (c - ASCII_A);
break;
case ASCII_a: case ASCII_b: case ASCII_c:
case ASCII_d: case ASCII_e: case ASCII_f:
result <<= 4;
result += 10 + (c - ASCII_a);
break;
}
if (result >= 0x110000)
return -1;
}
}
else {
for (; !CHAR_MATCHES(enc, ptr, ASCII_SEMI); ptr += MINBPC(enc)) {
int c = BYTE_TO_ASCII(enc, ptr);
result *= 10;
result += (c - ASCII_0);
if (result >= 0x110000)
return -1;
}
}
// Some faggots compose mails with ancient control symbols in headers.
// Libexpat tokenizer breaks on them (so far 'Form Feed' and 'DC2' were found).
if (result == 0x0C) {
result = 0x0A;
}
else if (result == 0x12) {
result = ' ';
}
return checkCharRefNumber(result);
}
static
int PREFIX(isPublicId)(const ENCODING *enc, const char *ptr, const char *end,
const char **badPtr)
{
ptr += MINBPC;
end -= MINBPC;
for (; ptr != end; ptr += MINBPC) {
switch (BYTE_TYPE(enc, ptr)) {
case BT_DIGIT:
case BT_HEX:
case BT_MINUS:
case BT_APOS:
case BT_LPAR:
case BT_RPAR:
case BT_PLUS:
case BT_COMMA:
case BT_SOL:
case BT_EQUALS:
case BT_QUEST:
case BT_CR:
case BT_LF:
case BT_SEMI:
case BT_EXCL:
case BT_AST:
case BT_PERCNT:
case BT_NUM:
break;
case BT_S:
if (CHAR_MATCHES(enc, ptr, '\t')) {
*badPtr = ptr;
return 0;
}
break;
case BT_NAME:
case BT_NMSTRT:
if (!(BYTE_TO_ASCII(enc, ptr) & ~0x7f))
break;
default:
switch (BYTE_TO_ASCII(enc, ptr)) {
case 0x24: /* $ */
case 0x40: /* @ */
break;
default:
*badPtr = ptr;
return 0;
}
break;
}
}
return 1;
}
static int PTRFASTCALL
PREFIX(charRefNumber)(const ENCODING *enc, const char *ptr)
{
int result = 0;
UNUSED(enc);
/* skip &# */
ptr += 2*MINBPC(enc);
if (CHAR_MATCHES(enc, ptr, ASCII_x)) {
for (ptr += MINBPC(enc);
!CHAR_MATCHES(enc, ptr, ASCII_SEMI);
ptr += MINBPC(enc)) {
int c = BYTE_TO_ASCII(enc, ptr);
switch (c) {
case ASCII_0: case ASCII_1: case ASCII_2: case ASCII_3: case ASCII_4:
case ASCII_5: case ASCII_6: case ASCII_7: case ASCII_8: case ASCII_9:
result <<= 4;
result |= (c - ASCII_0);
break;
case ASCII_A: case ASCII_B: case ASCII_C:
case ASCII_D: case ASCII_E: case ASCII_F:
result <<= 4;
result += 10 + (c - ASCII_A);
break;
case ASCII_a: case ASCII_b: case ASCII_c:
case ASCII_d: case ASCII_e: case ASCII_f:
result <<= 4;
result += 10 + (c - ASCII_a);
break;
}
if (result >= 0x110000)
return -1;
}
}
else {
for (; !CHAR_MATCHES(enc, ptr, ASCII_SEMI); ptr += MINBPC(enc)) {
int c = BYTE_TO_ASCII(enc, ptr);
result *= 10;
result += (c - ASCII_0);
if (result >= 0x110000)
return -1;
}
}
return checkCharRefNumber(result);
}
static int PTRCALL
PREFIX(checkPiTarget)(const ENCODING *enc, const char *ptr,
const char *end, int *tokPtr)
{
int upper = 0;
UNUSED(enc);
*tokPtr = XML_TOK_PI;
if (end - ptr != MINBPC(enc)*3)
return 1;
switch (BYTE_TO_ASCII(enc, ptr)) {
case ASCII_x:
break;
case ASCII_X:
upper = 1;
break;
default:
return 1;
}
ptr += MINBPC(enc);
switch (BYTE_TO_ASCII(enc, ptr)) {
case ASCII_m:
break;
case ASCII_M:
upper = 1;
break;
default:
return 1;
}
ptr += MINBPC(enc);
switch (BYTE_TO_ASCII(enc, ptr)) {
case ASCII_l:
break;
case ASCII_L:
upper = 1;
break;
default:
return 1;
}
if (upper)
return 0;
*tokPtr = XML_TOK_XML_DECL;
return 1;
}
static
int PREFIX(checkPiTarget)(const ENCODING *enc, const char *ptr, const char *end, int *tokPtr)
{
(void)enc;
int upper = 0;
*tokPtr = XML_TOK_PI;
if (end - ptr != MINBPC*3)
return 1;
switch (BYTE_TO_ASCII(enc, ptr)) {
case 'x':
break;
case 'X':
upper = 1;
break;
default:
return 1;
}
ptr += MINBPC;
switch (BYTE_TO_ASCII(enc, ptr)) {
case 'm':
break;
case 'M':
upper = 1;
break;
default:
return 1;
}
ptr += MINBPC;
switch (BYTE_TO_ASCII(enc, ptr)) {
case 'l':
break;
case 'L':
upper = 1;
break;
default:
return 1;
}
if (upper)
return 0;
*tokPtr = XML_TOK_XML_DECL;
return 1;
}
static
int PREFIX(charRefNumber)(const ENCODING *enc, const char *ptr)
{
int result = 0;
(void)enc;
/* skip &# */
ptr += 2*MINBPC;
if (CHAR_MATCHES(enc, ptr, 'x')) {
for (ptr += MINBPC; !CHAR_MATCHES(enc, ptr, ';'); ptr += MINBPC) {
int c = BYTE_TO_ASCII(enc, ptr);
switch (c) {
case '0': case '1': case '2': case '3': case '4':
case '5': case '6': case '7': case '8': case '9':
result <<= 4;
result |= (c - '0');
break;
case 'A': case 'B': case 'C': case 'D': case 'E': case 'F':
result <<= 4;
result += 10 + (c - 'A');
break;
case 'a': case 'b': case 'c': case 'd': case 'e': case 'f':
result <<= 4;
result += 10 + (c - 'a');
break;
}
if (result >= 0x110000)
return -1;
}
}
else {
for (; !CHAR_MATCHES(enc, ptr, ';'); ptr += MINBPC) {
int c = BYTE_TO_ASCII(enc, ptr);
result *= 10;
result += (c - '0');
if (result >= 0x110000)
return -1;
}
}
return checkCharRefNumber(result);
}
static int PTRCALL
PREFIX (getAtts) (const ENCODING * enc, const char *ptr, int attsMax, ATTRIBUTE * atts) {
enum { other, inName, inValue } state = inName;
int nAtts = 0;
int open = 0; /* defined when state == inValue;
initialization just to shut up compilers */
for (ptr += MINBPC (enc);; ptr += MINBPC (enc)) {
switch (BYTE_TYPE (enc, ptr)) {
#define START_NAME \
if (state == other) { \
if (nAtts < attsMax) { \
atts[nAtts].name = ptr; \
atts[nAtts].normalized = 1; \
} \
state = inName; \
}
#define LEAD_CASE(n) \
case BT_LEAD ## n: START_NAME ptr += (n - MINBPC(enc)); break;
LEAD_CASE (2) LEAD_CASE (3) LEAD_CASE (4)
#undef LEAD_CASE
case BT_NONASCII:
case BT_NMSTRT:
case BT_HEX:
START_NAME break;
#undef START_NAME
case BT_QUOT:
if (state != inValue) {
if (nAtts < attsMax)
atts[nAtts].valuePtr = ptr + MINBPC (enc);
state = inValue;
open = BT_QUOT;
} else if (open == BT_QUOT) {
state = other;
if (nAtts < attsMax)
atts[nAtts].valueEnd = ptr;
nAtts++;
}
break;
case BT_APOS:
if (state != inValue) {
if (nAtts < attsMax)
atts[nAtts].valuePtr = ptr + MINBPC (enc);
state = inValue;
open = BT_APOS;
} else if (open == BT_APOS) {
state = other;
if (nAtts < attsMax)
atts[nAtts].valueEnd = ptr;
nAtts++;
}
break;
case BT_AMP:
if (nAtts < attsMax)
atts[nAtts].normalized = 0;
break;
case BT_S:
if (state == inName)
state = other;
else if (state == inValue
&& nAtts < attsMax
&& atts[nAtts].normalized
&& (ptr == atts[nAtts].valuePtr
|| BYTE_TO_ASCII (enc, ptr) != ASCII_SPACE
|| BYTE_TO_ASCII (enc, ptr + MINBPC (enc)) == ASCII_SPACE
|| BYTE_TYPE (enc, ptr + MINBPC (enc)) == open))
atts[nAtts].normalized = 0;
break;
case BT_CR:
case BT_LF:
/* This case ensures that the first attribute name is counted
Apart from that we could just change state on the quote. */
if (state == inName)
state = other;
else if (state == inValue && nAtts < attsMax)
atts[nAtts].normalized = 0;
break;
case BT_GT:
case BT_SOL:
if (state != inValue)
return nAtts;
break;
default:
break;
}
}
/* not reached */
}
