📄 xmltok_impl.c
字号:
case BT_SEMI:
*nextTokPtr = ptr + MINBPC(enc);
return XML_TOK_CHAR_REF;
default:
*nextTokPtr = ptr;
return XML_TOK_INVALID;
}
}
}
return XML_TOK_PARTIAL;
}
/* ptr points to character following "&#" */
static int PTRCALL
PREFIX(scanCharRef)(const ENCODING *enc, const char *ptr,
const char *end, const char **nextTokPtr)
{
if (ptr != end) {
if (CHAR_MATCHES(enc, ptr, ASCII_x))
return PREFIX(scanHexCharRef)(enc, ptr + MINBPC(enc), end, nextTokPtr);
switch (BYTE_TYPE(enc, ptr)) {
case BT_DIGIT:
break;
default:
*nextTokPtr = ptr;
return XML_TOK_INVALID;
}
for (ptr += MINBPC(enc); ptr != end; ptr += MINBPC(enc)) {
switch (BYTE_TYPE(enc, ptr)) {
case BT_DIGIT:
break;
case BT_SEMI:
*nextTokPtr = ptr + MINBPC(enc);
return XML_TOK_CHAR_REF;
default:
*nextTokPtr = ptr;
return XML_TOK_INVALID;
}
}
}
return XML_TOK_PARTIAL;
}
/* ptr points to character following "&" */
static int PTRCALL
PREFIX(scanRef)(const ENCODING *enc, const char *ptr, const char *end,
const char **nextTokPtr)
{
if (ptr == end)
return XML_TOK_PARTIAL;
switch (BYTE_TYPE(enc, ptr)) {
CHECK_NMSTRT_CASES(enc, ptr, end, nextTokPtr)
case BT_NUM:
return PREFIX(scanCharRef)(enc, ptr + MINBPC(enc), end, nextTokPtr);
default:
*nextTokPtr = ptr;
return XML_TOK_INVALID;
}
while (ptr != end) {
switch (BYTE_TYPE(enc, ptr)) {
CHECK_NAME_CASES(enc, ptr, end, nextTokPtr)
case BT_SEMI:
*nextTokPtr = ptr + MINBPC(enc);
return XML_TOK_ENTITY_REF;
default:
*nextTokPtr = ptr;
return XML_TOK_INVALID;
}
}
return XML_TOK_PARTIAL;
}
/* ptr points to character following first character of attribute name */
static int PTRCALL
PREFIX(scanAtts)(const ENCODING *enc, const char *ptr, const char *end,
const char **nextTokPtr)
{
#ifdef XML_NS
int hadColon = 0;
#endif
while (ptr != end) {
switch (BYTE_TYPE(enc, ptr)) {
CHECK_NAME_CASES(enc, ptr, end, nextTokPtr)
#ifdef XML_NS
case BT_COLON:
if (hadColon) {
*nextTokPtr = ptr;
return XML_TOK_INVALID;
}
hadColon = 1;
ptr += MINBPC(enc);
if (ptr == end)
return XML_TOK_PARTIAL;
switch (BYTE_TYPE(enc, ptr)) {
CHECK_NMSTRT_CASES(enc, ptr, end, nextTokPtr)
default:
*nextTokPtr = ptr;
return XML_TOK_INVALID;
}
break;
#endif
case BT_S: case BT_CR: case BT_LF:
for (;;) {
int t;
ptr += MINBPC(enc);
if (ptr == end)
return XML_TOK_PARTIAL;
t = BYTE_TYPE(enc, ptr);
if (t == BT_EQUALS)
break;
switch (t) {
case BT_S:
case BT_LF:
case BT_CR:
break;
default:
*nextTokPtr = ptr;
return XML_TOK_INVALID;
}
}
/* fall through */
case BT_EQUALS:
{
int open;
#ifdef XML_NS
hadColon = 0;
#endif
for (;;) {
ptr += MINBPC(enc);
if (ptr == end)
return XML_TOK_PARTIAL;
open = BYTE_TYPE(enc, ptr);
if (open == BT_QUOT || open == BT_APOS)
break;
switch (open) {
case BT_S:
case BT_LF:
case BT_CR:
break;
default:
*nextTokPtr = ptr;
return XML_TOK_INVALID;
}
}
ptr += MINBPC(enc);
/* in attribute value */
for (;;) {
int t;
if (ptr == end)
return XML_TOK_PARTIAL;
t = BYTE_TYPE(enc, ptr);
if (t == open)
break;
switch (t) {
INVALID_CASES(ptr, nextTokPtr)
case BT_AMP:
{
int tok = PREFIX(scanRef)(enc, ptr + MINBPC(enc), end, &ptr);
if (tok <= 0) {
if (tok == XML_TOK_INVALID)
*nextTokPtr = ptr;
return tok;
}
break;
}
case BT_LT:
*nextTokPtr = ptr;
return XML_TOK_INVALID;
default:
ptr += MINBPC(enc);
break;
}
}
ptr += MINBPC(enc);
if (ptr == end)
return XML_TOK_PARTIAL;
switch (BYTE_TYPE(enc, ptr)) {
case BT_S:
case BT_CR:
case BT_LF:
break;
case BT_SOL:
goto sol;
case BT_GT:
goto gt;
default:
*nextTokPtr = ptr;
return XML_TOK_INVALID;
}
/* ptr points to closing quote */
for (;;) {
ptr += MINBPC(enc);
if (ptr == end)
return XML_TOK_PARTIAL;
switch (BYTE_TYPE(enc, ptr)) {
CHECK_NMSTRT_CASES(enc, ptr, end, nextTokPtr)
case BT_S: case BT_CR: case BT_LF:
continue;
case BT_GT:
gt:
*nextTokPtr = ptr + MINBPC(enc);
return XML_TOK_START_TAG_WITH_ATTS;
case BT_SOL:
sol:
ptr += MINBPC(enc);
if (ptr == end)
return XML_TOK_PARTIAL;
if (!CHAR_MATCHES(enc, ptr, ASCII_GT)) {
*nextTokPtr = ptr;
return XML_TOK_INVALID;
}
*nextTokPtr = ptr + MINBPC(enc);
return XML_TOK_EMPTY_ELEMENT_WITH_ATTS;
default:
*nextTokPtr = ptr;
return XML_TOK_INVALID;
}
break;
}
break;
}
default:
*nextTokPtr = ptr;
return XML_TOK_INVALID;
}
}
return XML_TOK_PARTIAL;
}
/* ptr points to character following "<" */
static int PTRCALL
PREFIX(scanLt)(const ENCODING *enc, const char *ptr, const char *end,
const char **nextTokPtr)
{
#ifdef XML_NS
int hadColon;
#endif
if (ptr == end)
return XML_TOK_PARTIAL;
switch (BYTE_TYPE(enc, ptr)) {
CHECK_NMSTRT_CASES(enc, ptr, end, nextTokPtr)
case BT_EXCL:
if ((ptr += MINBPC(enc)) == end)
return XML_TOK_PARTIAL;
switch (BYTE_TYPE(enc, ptr)) {
case BT_MINUS:
return PREFIX(scanComment)(enc, ptr + MINBPC(enc), end, nextTokPtr);
case BT_LSQB:
return PREFIX(scanCdataSection)(enc, ptr + MINBPC(enc),
end, nextTokPtr);
}
*nextTokPtr = ptr;
return XML_TOK_INVALID;
case BT_QUEST:
return PREFIX(scanPi)(enc, ptr + MINBPC(enc), end, nextTokPtr);
case BT_SOL:
return PREFIX(scanEndTag)(enc, ptr + MINBPC(enc), end, nextTokPtr);
default:
*nextTokPtr = ptr;
return XML_TOK_INVALID;
}
#ifdef XML_NS
hadColon = 0;
#endif
/* we have a start-tag */
while (ptr != end) {
switch (BYTE_TYPE(enc, ptr)) {
CHECK_NAME_CASES(enc, ptr, end, nextTokPtr)
#ifdef XML_NS
case BT_COLON:
if (hadColon) {
*nextTokPtr = ptr;
return XML_TOK_INVALID;
}
hadColon = 1;
ptr += MINBPC(enc);
if (ptr == end)
return XML_TOK_PARTIAL;
switch (BYTE_TYPE(enc, ptr)) {
CHECK_NMSTRT_CASES(enc, ptr, end, nextTokPtr)
default:
*nextTokPtr = ptr;
return XML_TOK_INVALID;
}
break;
#endif
case BT_S: case BT_CR: case BT_LF:
{
ptr += MINBPC(enc);
while (ptr != end) {
switch (BYTE_TYPE(enc, ptr)) {
CHECK_NMSTRT_CASES(enc, ptr, end, nextTokPtr)
case BT_GT:
goto gt;
case BT_SOL:
goto sol;
case BT_S: case BT_CR: case BT_LF:
ptr += MINBPC(enc);
continue;
default:
*nextTokPtr = ptr;
return XML_TOK_INVALID;
}
return PREFIX(scanAtts)(enc, ptr, end, nextTokPtr);
}
return XML_TOK_PARTIAL;
}
case BT_GT:
gt:
*nextTokPtr = ptr + MINBPC(enc);
return XML_TOK_START_TAG_NO_ATTS;
case BT_SOL:
sol:
ptr += MINBPC(enc);
if (ptr == end)
return XML_TOK_PARTIAL;
if (!CHAR_MATCHES(enc, ptr, ASCII_GT)) {
*nextTokPtr = ptr;
return XML_TOK_INVALID;
}
*nextTokPtr = ptr + MINBPC(enc);
return XML_TOK_EMPTY_ELEMENT_NO_ATTS;
default:
*nextTokPtr = ptr;
return XML_TOK_INVALID;
}
}
return XML_TOK_PARTIAL;
}
static int PTRCALL
PREFIX(contentTok)(const ENCODING *enc, const char *ptr, const char *end,
const char **nextTokPtr)
{
if (ptr == end)
return XML_TOK_NONE;
if (MINBPC(enc) > 1) {
size_t n = end - ptr;
if (n & (MINBPC(enc) - 1)) {
n &= ~(MINBPC(enc) - 1);
if (n == 0)
return XML_TOK_PARTIAL;
end = ptr + n;
}
}
switch (BYTE_TYPE(enc, ptr)) {
case BT_LT:
return PREFIX(scanLt)(enc, ptr + MINBPC(enc), end, nextTokPtr);
case BT_AMP:
return PREFIX(scanRef)(enc, ptr + MINBPC(enc), end, nextTokPtr);
case BT_CR:
ptr += MINBPC(enc);
if (ptr == end)
return XML_TOK_TRAILING_CR;
if (BYTE_TYPE(enc, ptr) == BT_LF)
ptr += MINBPC(enc);
*nextTokPtr = ptr;
return XML_TOK_DATA_NEWLINE;
case BT_LF:
*nextTokPtr = ptr + MINBPC(enc);
return XML_TOK_DATA_NEWLINE;
case BT_RSQB:
ptr += MINBPC(enc);
if (ptr == end)
return XML_TOK_TRAILING_RSQB;
if (!CHAR_MATCHES(enc, ptr, ASCII_RSQB))
break;
ptr += MINBPC(enc);
if (ptr == end)
return XML_TOK_TRAILING_RSQB;
if (!CHAR_MATCHES(enc, ptr, ASCII_GT)) {
ptr -= MINBPC(enc);
break;
}
*nextTokPtr = ptr;
return XML_TOK_INVALID;
INVALID_CASES(ptr, nextTokPtr)
default:
ptr += MINBPC(enc);
break;
}
while (ptr != end) {
switch (BYTE_TYPE(enc, ptr)) {
#define LEAD_CASE(n) \
case BT_LEAD ## n: \
if (end - ptr < n || IS_INVALID_CHAR(enc, ptr, n)) { \
*nextTokPtr = ptr; \
return XML_TOK_DATA_CHARS; \
} \
ptr += n; \
break;
LEAD_CASE(2) LEAD_CASE(3) LEAD_CASE(4)
#undef LEAD_CASE
case BT_RSQB:
if (ptr + MINBPC(enc) != end) {
if (!CHAR_MATCHES(enc, ptr + MINBPC(enc), ASCII_RSQB)) {
ptr += MINBPC(enc);
break;
}
if (ptr + 2*MINBPC(enc) != end) {
if (!CHAR_MATCHES(enc, ptr + 2*MINBPC(enc), ASCII_GT)) {
ptr += MINBPC(enc);
break;
}
*nextTokPtr = ptr + 2*MINBPC(enc);
return XML_TOK_INVALID;
}
}
/* fall through */
case BT_AMP:
case BT_LT:
case BT_NONXML:
case BT_MALFORM:
case BT_TRAIL:
case BT_CR:
case BT_LF:
*nextTokPtr = ptr;
return XML_TOK_DATA_CHARS;
default:
ptr += MINBPC(enc);
break;
}
}
*nextTokPtr = ptr;
return XML_TOK_DATA_CHARS;
}
/* ptr points to character following "%" */
static int PTRCALL
PREFIX(scanPercent)(const ENCODING *enc, const char *ptr, const char *end,
const char **nextTokPtr)
{
if (ptr == end)
return XML_TOK_PARTIAL;
switch (BYTE_TYPE(enc, ptr)) {
CHECK_NMSTRT_CASES(enc, ptr, end, nextTokPtr)
case BT_S: case BT_LF: case BT_CR: case BT_PERCNT:
*nextTokPtr = ptr;
return XML_TOK_PERCENT;
⌨️ 快捷键说明
复制代码
Ctrl + C
搜索代码
Ctrl + F
全屏模式
F11
切换主题
Ctrl + Shift + D
显示快捷键
?
增大字号
Ctrl + =
减小字号
Ctrl + -