emelie/tidy-html5
1
0
Fork 0
Code Issues Pull requests Releases Wiki Activity

Issue #64 - Try hard to skip '<![CDATA[ ... ]]>'

Browse source
This commit is contained in:
Geoff McLane 2015-09-24 14:21:55 +02:00
parent f4113a8643
commit 04ca419080
1 changed files with 42 additions and 2 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

42
src/lexer.c
View file

@ -1962,18 +1962,56 @@ static Bool ExpectsContent(Node *node)
* Controlled by option --skip-quotes yes|no, enum as * Controlled by option --skip-quotes yes|no, enum as
* TidySkipQuotes, off by default. * TidySkipQuotes, off by default.
\*/ \*/
#define MX_TAG_BUFF 16
static Bool IsInQuotesorComment( Lexer * lexer ) static Bool IsInQuotesorComment( Lexer * lexer )
{ {
unsigned int i; unsigned int i, cnt, off;
Bool inq, toeol, toec; Bool inq, toeol, toec;
unsigned char prev, quot, c; unsigned char prev, quot, c;
tmbchar buff[MX_TAG_BUFF];
tmbstr pnc; tmbstr pnc;
prev = quot = 0; prev = quot = 0;
inq = toeol = toec = no; inq = toeol = toec = no;
cnt = 0;
off = 0;
for ( i = lexer->txtstart; i < lexer->lexsize; i++ ) for ( i = lexer->txtstart; i < lexer->lexsize; i++ )
{ {
pnc = &lexer->lexbuf[i]; pnc = &lexer->lexbuf[i];
c = *pnc; c = *pnc;
if ( ! TY_(IsWhite)(c) ) {
if (c == '<') {
off = 0;
buff[off++] = c;
} else if (off && ((off + 1) < MX_TAG_BUFF)) {
buff[off++] = c;
if ( c == '[' ) {
buff[off] = 0;
if (TY_(tmbstrcmp)(buff,"<![CDATA[") == 0) {
/* Ignore **EVERYTHING** until "]]>" */
i++;
off = 0;
for ( ; i < lexer->lexsize; i++ )
{
pnc = &lexer->lexbuf[i];
c = *pnc;
if ( c == ']' ) {
buff[off++] = c;
} else if (off && ((off + 1) < MX_TAG_BUFF)) {
buff[off++] = c;
if ( c == '>' ) {
buff[off] = 0;
if (TY_(tmbstrcmp)(buff,"]]>") == 0) {
break;
}
off = 0;
}
}
}
}
}
}
cnt++;
}
if ( toeol ) if ( toeol )
{ {
/* continue until END OF LINE */ /* continue until END OF LINE */
@ -2009,6 +2047,8 @@ static Bool IsInQuotesorComment( Lexer * lexer )
} }
else if ( !inq && ( c == '/' ) && (prev == '/') ) else if ( !inq && ( c == '/' ) && (prev == '/') )
{ {
/* except, if we have "//<!CDATA..." or "//]]>" */
if (cnt > 2)
toeol = yes; /* set in comment, until END OF LINE */ toeol = yes; /* set in comment, until END OF LINE */
} }
else if ( !inq && ( c == '*' ) && (prev == '/')) else if ( !inq && ( c == '*' ) && (prev == '/'))