mirror of
https://port.numenaute.org/aleajactaest/khanat-opennel-code.git
synced 2024-11-17 13:01:42 +00:00
Ensure that html is valid utf8 and that tags come in correct(ish) order
--HG-- branch : develop
This commit is contained in:
Nimetu
parent
be95880a95
commit
c6535ee2fd
1 changed files with 138 additions and 1 deletions
|
|
@ -117,10 +117,144 @@ namespace NLGUI
|
|||
}
|
||||
}
|
||||
|
||||
// ***************************************************************************
|
||||
// http://stackoverflow.com/a/18335183
|
||||
static std::string correct_non_utf_8(const std::string &str)
|
||||
{
|
||||
int i,f_size=str.size();
|
||||
unsigned char c,c2,c3,c4;
|
||||
std::string to;
|
||||
to.reserve(f_size);
|
||||
|
||||
for(i=0 ; i<f_size ; i++){
|
||||
c=(unsigned char)(str[i]);
|
||||
if(c<32){//control char
|
||||
if(c==9 || c==10 || c==13){//allow only \t \n \r
|
||||
to.append(1,c);
|
||||
}
|
||||
continue;
|
||||
}else if(c<127){//normal ASCII
|
||||
to.append(1,c);
|
||||
continue;
|
||||
}else if(c<160){//control char (nothing should be defined here either ASCI, ISO_8859-1 or UTF8, so skipping)
|
||||
if(c==128){//fix microsoft mess, add euro
|
||||
to.append(1,226);
|
||||
to.append(1,130);
|
||||
to.append(1,172);
|
||||
}
|
||||
if(c==133){//fix IBM mess, add NEL = \n\r
|
||||
to.append(1,10);
|
||||
to.append(1,13);
|
||||
}
|
||||
continue;
|
||||
}else if(c<192){//invalid for UTF8, converting ASCII
|
||||
to.append(1,(unsigned char)194);
|
||||
to.append(1,c);
|
||||
continue;
|
||||
}else if(c<194){//invalid for UTF8, converting ASCII
|
||||
to.append(1,(unsigned char)195);
|
||||
to.append(1,c-64);
|
||||
continue;
|
||||
}else if(c<224 && i+1<f_size){//possibly 2byte UTF8
|
||||
c2=(unsigned char)(str[i+1]);
|
||||
if(c2>127 && c2<192){//valid 2byte UTF8
|
||||
if(c==194 && c2<160){//control char, skipping
|
||||
;
|
||||
}else{
|
||||
to.append(1,c);
|
||||
to.append(1,c2);
|
||||
}
|
||||
i++;
|
||||
continue;
|
||||
}
|
||||
}else if(c<240 && i+2<f_size){//possibly 3byte UTF8
|
||||
c2=(unsigned char)(str[i+1]);
|
||||
c3=(unsigned char)(str[i+2]);
|
||||
if(c2>127 && c2<192 && c3>127 && c3<192){//valid 3byte UTF8
|
||||
to.append(1,c);
|
||||
to.append(1,c2);
|
||||
to.append(1,c3);
|
||||
i+=2;
|
||||
continue;
|
||||
}
|
||||
}else if(c<245 && i+3<f_size){//possibly 4byte UTF8
|
||||
c2=(unsigned char)(str[i+1]);
|
||||
c3=(unsigned char)(str[i+2]);
|
||||
c4=(unsigned char)(str[i+3]);
|
||||
if(c2>127 && c2<192 && c3>127 && c3<192 && c4>127 && c4<192){//valid 4byte UTF8
|
||||
to.append(1,c);
|
||||
to.append(1,c2);
|
||||
to.append(1,c3);
|
||||
to.append(1,c4);
|
||||
i+=3;
|
||||
continue;
|
||||
}
|
||||
}
|
||||
//invalid UTF8, converting ASCII (c>245 || string too short for multi-byte))
|
||||
to.append(1,(unsigned char)195);
|
||||
to.append(1,c-64);
|
||||
}
|
||||
return to;
|
||||
}
|
||||
|
||||
// ***************************************************************************
|
||||
static void patchHtmlQuirks(std::string &htmlString)
|
||||
{
|
||||
size_t npos = std::string::npos;
|
||||
size_t pos;
|
||||
|
||||
// get rid of BOM (some ingame help files does not show up otherwise)
|
||||
if (htmlString.substr(0, 3) == "\xEF\xBB\xBF")
|
||||
{
|
||||
htmlString.erase(0, 3);
|
||||
}
|
||||
|
||||
// if any element is before <html>, then parser adds <html><body>
|
||||
// and original tags are ignored (their attributes not processed)
|
||||
//
|
||||
// only fix situation when there is <body> tag with attributes
|
||||
//
|
||||
// tags are considered to be lowercase
|
||||
|
||||
pos = htmlString.find("<body ");
|
||||
if (pos != npos)
|
||||
{
|
||||
size_t start = htmlString.find("<");
|
||||
// skip <!doctype html>
|
||||
if (htmlString.substr(start, 2) == "<!")
|
||||
start = htmlString.find("<", start + 1);
|
||||
|
||||
// if there is no html tag, then abort
|
||||
size_t end = htmlString.find("<html>");
|
||||
if (end != npos && start < end && end < pos)
|
||||
{
|
||||
// body tag end position
|
||||
size_t insert = htmlString.find(">", pos);
|
||||
if (insert != npos)
|
||||
{
|
||||
std::string str = htmlString.substr(start, end - start);
|
||||
htmlString.insert(insert+1, str);
|
||||
htmlString.erase(start, str.size());
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
// make sure </html> (if present) is last in document or tags coming after it are ignored
|
||||
pos = htmlString.find("</html>");
|
||||
if (pos != npos && htmlString.find("<", pos+1) > pos)
|
||||
{
|
||||
htmlString.erase(pos, 7);
|
||||
htmlString += "</html>";
|
||||
}
|
||||
|
||||
// if there is invalid utf-8 chars, then libxml will break everything after first it finds.
|
||||
htmlString = correct_non_utf_8(htmlString);
|
||||
}
|
||||
|
||||
// ***************************************************************************
|
||||
bool CGroupHTML::parseHtml(std::string htmlString)
|
||||
{
|
||||
htmlParserCtxtPtr parser = htmlCreatePushParserCtxt(NULL, NULL, NULL, 0, NULL, XML_CHAR_ENCODING_NONE);
|
||||
htmlParserCtxtPtr parser = htmlCreatePushParserCtxt(NULL, NULL, NULL, 0, NULL, XML_CHAR_ENCODING_UTF8);
|
||||
if (!parser)
|
||||
{
|
||||
nlwarning("Creating html parser context failed");
|
||||
|
|
@ -129,6 +263,9 @@ namespace NLGUI
|
|||
|
||||
htmlCtxtUseOptions(parser, HTML_PARSE_NOBLANKS | HTML_PARSE_NOERROR | HTML_PARSE_NOWARNING | HTML_PARSE_NONET);
|
||||
|
||||
// parser is little strict on tag order, so fix whats needed
|
||||
patchHtmlQuirks(htmlString);
|
||||
|
||||
htmlParseChunk(parser, htmlString.c_str(), htmlString.size(), 0);
|
||||
htmlParseChunk(parser, "", 0, 1);
|
||||
|
||||
|
|
|
|||
Loading…
Reference in a new issue