RGBCube/serenity
1
Fork 0
mirror of https://github.com/RGBCube/serenity synced 2025-07-23 12:17:36 +00:00
Code Issues Activity Actions

LibRegex: Ensure escaped code points are exactly 4 digits in length

Browse source
This commit is contained in:
Timothy Flynn 2021-08-11 09:58:08 -04:00 • committed by Linus Groh
parent e887314472
commit 2e4b6fd1ac
3 changed files with 12 additions and 7 deletions
Download patch file Download diff file Expand all files Collapse all files

2
Tests/LibRegex/Regex.cpp
View file

@ -503,6 +503,8 @@ TEST_CASE(ECMA262_parse)
{ "(?", regex::Error::InvalidCaptureGroup }, { "(?", regex::Error::InvalidCaptureGroup },
{ "\\u1234", regex::Error::NoError, regex::ECMAScriptFlags::Unicode }, { "\\u1234", regex::Error::NoError, regex::ECMAScriptFlags::Unicode },
{ "[\\u1234]", regex::Error::NoError, regex::ECMAScriptFlags::Unicode }, { "[\\u1234]", regex::Error::NoError, regex::ECMAScriptFlags::Unicode },
{ "\\u1", regex::Error::InvalidPattern, regex::ECMAScriptFlags::Unicode },
{ "[\\u1]", regex::Error::InvalidPattern, regex::ECMAScriptFlags::Unicode },
{ ",(?", regex::Error::InvalidCaptureGroup }, // #4583 { ",(?", regex::Error::InvalidCaptureGroup }, // #4583
{ "{1}", regex::Error::InvalidPattern }, { "{1}", regex::Error::InvalidPattern },
{ "{1,2}", regex::Error::InvalidPattern }, { "{1,2}", regex::Error::InvalidPattern },

13
Userland/Libraries/LibRegex/RegexParser.cpp
View file

@ -1083,7 +1083,7 @@ bool ECMA262Parser::parse_quantifiable_assertion(ByteCode& stack, size_t&, bool
return false; return false;
} }
StringView ECMA262Parser::read_digits_as_string(ReadDigitsInitialZeroState initial_zero, bool hex, int max_count) StringView ECMA262Parser::read_digits_as_string(ReadDigitsInitialZeroState initial_zero, bool hex, int max_count, int min_count)
{ {
if (!match(TokenType::Char)) if (!match(TokenType::Char))
return {}; return {};
@ -1109,12 +1109,15 @@ StringView ECMA262Parser::read_digits_as_string(ReadDigitsInitialZeroState initi
++count; ++count;
} }
if (count < min_count)
return {};
return StringView { start_token.value().characters_without_null_termination(), offset }; return StringView { start_token.value().characters_without_null_termination(), offset };
} }
Optional<unsigned> ECMA262Parser::read_digits(ECMA262Parser::ReadDigitsInitialZeroState initial_zero, bool hex, int max_count) Optional<unsigned> ECMA262Parser::read_digits(ECMA262Parser::ReadDigitsInitialZeroState initial_zero, bool hex, int max_count, int min_count)
{ {
auto str = read_digits_as_string(initial_zero, hex, max_count); auto str = read_digits_as_string(initial_zero, hex, max_count, min_count);
if (str.is_empty()) if (str.is_empty())
return {}; return {};
if (hex) if (hex)
@ -1483,7 +1486,7 @@ bool ECMA262Parser::parse_atom_escape(ByteCode& stack, size_t& match_length_mini
return false; return false;
} }
if (auto code_point = read_digits(ReadDigitsInitialZeroState::Allow, true, 4); code_point.has_value()) { if (auto code_point = read_digits(ReadDigitsInitialZeroState::Allow, true, 4, 4); code_point.has_value()) {
// In Unicode mode, we need to combine surrogate pairs into a single code point. But we also need to be // In Unicode mode, we need to combine surrogate pairs into a single code point. But we also need to be
// rather forgiving if the surrogate pairs are invalid. So if a second code unit follows this code unit, // rather forgiving if the surrogate pairs are invalid. So if a second code unit follows this code unit,
// but doesn't form a valid surrogate pair, insert bytecode for both code units individually. // but doesn't form a valid surrogate pair, insert bytecode for both code units individually.
@ -1811,7 +1814,7 @@ bool ECMA262Parser::parse_nonempty_class_ranges(Vector<CompareTypeAndValuePair>&
} }
if (try_skip("u")) { if (try_skip("u")) {
if (auto code_point = read_digits(ReadDigitsInitialZeroState::Allow, true, 4); code_point.has_value()) { if (auto code_point = read_digits(ReadDigitsInitialZeroState::Allow, true, 4, 4); code_point.has_value()) {
// FIXME: While code point ranges are supported, code point matches as "Char" are not! // FIXME: While code point ranges are supported, code point matches as "Char" are not!
return { CharClassRangeElement { .code_point = code_point.value(), .is_character_class = false } }; return { CharClassRangeElement { .code_point = code_point.value(), .is_character_class = false } };
} else if (!unicode) { } else if (!unicode) {

4
Userland/Libraries/LibRegex/RegexParser.h
View file

@ -210,8 +210,8 @@ private:
Allow, Allow,
Disallow, Disallow,
}; };
StringView read_digits_as_string(ReadDigitsInitialZeroState initial_zero = ReadDigitsInitialZeroState::Allow, bool hex = false, int max_count = -1); StringView read_digits_as_string(ReadDigitsInitialZeroState initial_zero = ReadDigitsInitialZeroState::Allow, bool hex = false, int max_count = -1, int min_count = -1);
Optional<unsigned> read_digits(ReadDigitsInitialZeroState initial_zero = ReadDigitsInitialZeroState::Allow, bool hex = false, int max_count = -1); Optional<unsigned> read_digits(ReadDigitsInitialZeroState initial_zero = ReadDigitsInitialZeroState::Allow, bool hex = false, int max_count = -1, int min_count = -1);
StringView read_capture_group_specifier(bool take_starting_angle_bracket = false); StringView read_capture_group_specifier(bool take_starting_angle_bracket = false);
struct Script { struct Script {