fix unicode string literal escape

This commit is contained in:
MihailRis
2025-09-23 00:08:05 +03:00
parent 8b38d57966
commit 28c821006a
5 changed files with 63 additions and 6 deletions
+1 -1
View File
@@ -32,7 +32,7 @@ protected:
void goBack(size_t count = 1);
void reset();
int64_t parseSimpleInt(int base);
int64_t parseSimpleInt(int base, size_t maxLength = 0xFFFFFFFF);
dv::value parseNumber(int sign);
dv::value parseNumber();
StringT parseString(CharT chr, bool closeRequired = true);
+6 -3
View File
@@ -349,7 +349,10 @@ std::basic_string<CharT> BasicParser<CharT>::parseXmlName() {
}
template <typename CharT>
int64_t BasicParser<CharT>::parseSimpleInt(int base) {
int64_t BasicParser<CharT>::parseSimpleInt(int base, size_t maxLength) {
if (maxLength == 0) return 0;
size_t start = pos;
CharT c = peek();
int index = hexchar2int(c);
if (index == -1 || index >= base) {
@@ -357,7 +360,7 @@ int64_t BasicParser<CharT>::parseSimpleInt(int base) {
}
int64_t value = index;
pos++;
while (hasNext()) {
while (hasNext() && pos - start < maxLength) {
c = source[pos];
while (c == '_') {
c = source[++pos];
@@ -476,7 +479,7 @@ std::basic_string<CharT> BasicParser<CharT>::parseString(
continue;
}
if (c == 'u' || c == 'x') {
int codepoint = parseSimpleInt(16);
int codepoint = parseSimpleInt(16, c == 'u' ? 4 : 2);
ubyte bytes[4];
int size = util::encode_utf8(codepoint, bytes);
CharT chars[4];
+1 -1
View File
@@ -22,7 +22,7 @@ namespace markdown {
Result<wchar_t> process(std::wstring_view source, bool eraseMarkdown);
template <typename CharT>
inline std::basic_string<CharT> escape(std::string_view source) {
inline std::basic_string<CharT> escape(std::basic_string_view<CharT> source) {
std::basic_stringstream<CharT> ss;
int pos = 0;
while (pos < source.size()) {
+1 -1
View File
@@ -40,7 +40,7 @@ std::string util::escape(std::string_view s, bool escapeUnicode) {
uint cpsize;
int codepoint = decode_utf8(cpsize, s.data() + pos);
if (escapeUnicode) {
ss << "\\u" << std::hex << codepoint;
ss << "\\u" << std::setw(4) << std::setfill('0') << std::hex << codepoint;
} else {
ss << std::string(s.data() + pos, cpsize);
}