RGBCube/serenity
1
Fork 0
mirror of https://github.com/RGBCube/serenity synced 2025-07-24 20:07:34 +00:00
Code Issues Activity Actions

LibRegex: Don't try to consume the escaped character if at EOF

Browse source 
Fixes assert on e.g. `new RegExp("\\")`
This commit is contained in:
AnotherTest 2020-11-30 17:46:53 +03:30 • committed by Andreas Kling
parent 3bda458735
commit 6394720c87
2 changed files with 33 additions and 19 deletions
Download patch file Download diff file Expand all files Collapse all files

8
Libraries/LibRegex/RegexParser.cpp
View file

@ -1114,6 +1114,9 @@ bool ECMA262Parser::parse_atom_escape(ByteCode& stack, size_t& match_length_mini
} }
} }
if (done())
return set_error(Error::InvalidTrailingEscape);
bool negate = false; bool negate = false;
auto ch = parse_character_class_escape(negate); auto ch = parse_character_class_escape(negate);
if (!ch.has_value()) { if (!ch.has_value()) {
@ -1211,6 +1214,11 @@ bool ECMA262Parser::parse_nonempty_class_ranges(Vector<CompareTypeAndValuePair>&
} }
if (try_skip("\\")) { if (try_skip("\\")) {
if (done()) {
set_error(Error::InvalidTrailingEscape);
return {};
}
if (try_skip("f")) if (try_skip("f"))
return { { .code_point = '\f', .is_character_class = false } }; return { { .code_point = '\f', .is_character_class = false } };
if (try_skip("n")) if (try_skip("n"))

44
Libraries/LibRegex/Tests/Regex.cpp
View file

@ -474,27 +474,33 @@ TEST_CASE(simple_period_end_benchmark)
TEST_CASE(ECMA262_parse) TEST_CASE(ECMA262_parse)
{ {
constexpr const char* patterns[] { struct _test {
"^hello.$", const char* pattern;
"^(hello.)$", regex::Error expected_error { regex::Error::NoError };
"^h{0,1}ello.$",
"^hello\\W$",
"^hell\\w.$",
"^hell\\x6f1$", // ^hello1$
"^hel(?:l\\w).$",
"^hel(?<LO>l\\w).$",
"^[-a-zA-Z\\w\\s]+$",
"\\bhello\\B",
"^[\\w+/_-]+[=]{0,2}$", // #4189
"^(?:[^<]*(<[\\w\\W]+>)[^>]*$|#([\\w\\-]*)$)", // #4189
"\\/", // #4189
",/=-:", // #4243
"\\x", // Even invalid escapes are allowed if ~unicode.
}; };
for (auto& pattern : patterns) { constexpr _test tests[] {
Regex<ECMA262> re(pattern); { "^hello.$" },
EXPECT_EQ(re.parser_result.error, Error::NoError); { "^(hello.)$" },
{ "^h{0,1}ello.$" },
{ "^hello\\W$" },
{ "^hell\\w.$" },
{ "^hell\\x6f1$" }, // ^hello1$
{ "^hel(?:l\\w).$" },
{ "^hel(?<LO>l\\w).$" },
{ "^[-a-zA-Z\\w\\s]+$" },
{ "\\bhello\\B" },
{ "^[\\w+/_-]+[=]{0,2}$" }, // #4189
{ "^(?:[^<]*(<[\\w\\W]+>)[^>]*$|#([\\w\\-]*)$)" }, // #4189
{ "\\/" }, // #4189
{ ",/=-:" }, // #4243
{ "\\x" }, // Even invalid escapes are allowed if ~unicode.
{ "\\", regex::Error::InvalidTrailingEscape },
};
for (auto& test : tests) {
Regex<ECMA262> re(test.pattern);
EXPECT_EQ(re.parser_result.error, test.expected_error);
#ifdef REGEX_DEBUG #ifdef REGEX_DEBUG
dbg() << "\n"; dbg() << "\n";
RegexDebug regex_dbg(stderr); RegexDebug regex_dbg(stderr);