void TxtReaderCoreUtf16::readDocument(ZLInputStream &stream) {
const size_t BUFSIZE = 2048;
char *buffer = new char[BUFSIZE];
std::string str;
size_t length;
do {
length = stream.read(buffer, BUFSIZE);
char *start = buffer;
const char *end = buffer + length;
for (char *ptr = start; ptr < end; ptr += 2) {
const char chr = getAscii(ptr);
if (chr == '\n' || chr == '\r') {
bool skipNewLine = false;
if (chr == '\r' && ptr + 2 != end && getAscii(ptr + 2) == '\n') {
skipNewLine = true;
setAscii(ptr, '\n');
}
if (start != ptr) {
str.erase();
myReader.myConverter->convert(str, start, ptr + 2);
myReader.characterDataHandler(str);
}
if (skipNewLine) {
ptr += 2;
}
start = ptr + 2;
myReader.newLineHandler();
} else if (chr != 0 && isspace(chr)) {
if (chr != '\t') {
setAscii(ptr, ' ');
}
}
}
if (start != end) {
str.erase();
myReader.myConverter->convert(str, start, end);
myReader.characterDataHandler(str);
}
} while (length == BUFSIZE);
delete[] buffer;
}
void KURLGooglePrivate::replaceComponents(const Replacements& replacements)
{
url_canon::RawCanonOutputT<char> output;
url_parse::Parsed newParsed;
m_isValid = url_util::ReplaceComponents(utf8String().data(),
utf8String().length(), m_parsed, replacements, 0, &output, &newParsed);
m_parsed = newParsed;
if (m_parsed.ref.is_nonempty())
setUtf8(CString(output.data(), output.length()));
else
setAscii(CString(output.data(), output.length()));
}
// Note: code mostly duplicated below.
void KURLGooglePrivate::init(const KURL& base, const char* rel, int relLength,
const TextEncoding* queryEncoding)
{
// As a performance optimization, we do not use the charset converter if
// encoding is UTF-8 or other Unicode encodings. Note that this is
// per HTML5 2.5.3 (resolving URL). The URL canonicalizer will be
// more efficient with no charset converter object because it
// can do UTF-8 internally with no extra copies.
// We feel free to make the charset converter object every time since it's
// just a wrapper around a reference.
KURLCharsetConverter charsetConverterObject(queryEncoding);
KURLCharsetConverter* charsetConverter =
(!queryEncoding || isUnicodeEncoding(queryEncoding)) ? 0 :
&charsetConverterObject;
url_canon::RawCanonOutputT<char> output;
const CString& baseStr = base.m_url.utf8String();
m_isValid = url_util::ResolveRelative(baseStr.data(), baseStr.length(),
base.m_url.m_parsed, rel, relLength,
charsetConverter,
&output, &m_parsed);
// See FIXME in KURLGooglePrivate in the header. If canonicalization has not
// changed the string, we can avoid an extra allocation by using assignment.
//
// When KURL encounters an error such that the URL is invalid and empty
// (for example, resolving a relative URL on a non-hierarchical base), it
// will produce an isNull URL, and calling setUtf8 will produce an empty
// non-null URL. This is unlikely to affect anything, but we preserve this
// just in case.
if (m_isValid || output.length()) {
// Without ref, the whole url is guaranteed to be ASCII-only.
if (m_parsed.ref.is_nonempty())
setUtf8(CString(output.data(), output.length()));
else
setAscii(CString(output.data(), output.length()));
} else {
// WebCore expects resolved URLs to be empty rather than NULL.
setUtf8(CString("", 0));
}
}
// Note: code mostly duplicated above. See FIXMEs and comments there.
void KURLGooglePrivate::init(const KURL& base, const UChar* rel, int relLength,
const TextEncoding* queryEncoding)
{
KURLCharsetConverter charsetConverterObject(queryEncoding);
KURLCharsetConverter* charsetConverter =
(!queryEncoding || isUnicodeEncoding(queryEncoding)) ? 0 :
&charsetConverterObject;
url_canon::RawCanonOutputT<char> output;
const CString& baseStr = base.m_url.utf8String();
m_isValid = url_util::ResolveRelative(baseStr.data(), baseStr.length(),
base.m_url.m_parsed, rel, relLength,
charsetConverter,
&output, &m_parsed);
if (m_isValid || output.length()) {
if (m_parsed.ref.is_nonempty())
setUtf8(CString(output.data(), output.length()));
else
setAscii(CString(output.data(), output.length()));
} else
setUtf8(CString("", 0));
}
