RGBCube/serenity
1
Fork 0
mirror of https://github.com/RGBCube/serenity synced 2025-07-25 15:57:45 +00:00
Code Issues Activity Actions

LibRegex: Support non-ASCII case-insensitive character comparisons

Browse source 
Specifically, when the Unicode flag is set, use Unicode-aware case
folding to case-insensitively compare code points.
This commit is contained in:
Timothy Flynn 2023-11-08 10:30:55 -05:00 • committed by Tim Flynn
parent 3fbf33bd37
commit e122039c99
2 changed files with 23 additions and 4 deletions
Download patch file Download diff file Expand all files Collapse all files

12
Tests/LibRegex/Regex.cpp
View file

@ -757,6 +757,18 @@ TEST_CASE(ECMA262_unicode_match)
{ "(?<\\ud835\\udcd1\\ud835\\udcfb\\ud835\\udcf8\\ud835\\udd00\\ud835\\udcf7>brown)"sv, "brown"sv, true, ECMAScriptFlags::Unicode }, { "(?<\\ud835\\udcd1\\ud835\\udcfb\\ud835\\udcf8\\ud835\\udd00\\ud835\\udcf7>brown)"sv, "brown"sv, true, ECMAScriptFlags::Unicode },
{ "^\\s+$"sv, space_and_line_terminators }, { "^\\s+$"sv, space_and_line_terminators },
{ "^\\s+$"sv, space_and_line_terminators, true, ECMAScriptFlags::Unicode }, { "^\\s+$"sv, space_and_line_terminators, true, ECMAScriptFlags::Unicode },
{ "[\\u0390]"sv, "\u1fd3"sv, false, ECMAScriptFlags::Unicode },
{ "[\\u1fd3]"sv, "\u0390"sv, false, ECMAScriptFlags::Unicode },
{ "[\\u0390]"sv, "\u1fd3"sv, true, combine_flags(ECMAScriptFlags::Unicode, ECMAScriptFlags::Insensitive) },
{ "[\\u1fd3]"sv, "\u0390"sv, true, combine_flags(ECMAScriptFlags::Unicode, ECMAScriptFlags::Insensitive) },
{ "[\\u03b0]"sv, "\u1fe3"sv, false, ECMAScriptFlags::Unicode },
{ "[\\u1fe3]"sv, "\u03b0"sv, false, ECMAScriptFlags::Unicode },
{ "[\\u03b0]"sv, "\u1fe3"sv, true, combine_flags(ECMAScriptFlags::Unicode, ECMAScriptFlags::Insensitive) },
{ "[\\u1fe3]"sv, "\u03b0"sv, true, combine_flags(ECMAScriptFlags::Unicode, ECMAScriptFlags::Insensitive) },
{ "[\\ufb05]"sv, "\ufb06"sv, false, ECMAScriptFlags::Unicode },
{ "[\\ufb06]"sv, "\ufb05"sv, false, ECMAScriptFlags::Unicode },
{ "[\\ufb05]"sv, "\ufb06"sv, true, combine_flags(ECMAScriptFlags::Unicode, ECMAScriptFlags::Insensitive) },
{ "[\\ufb06]"sv, "\ufb05"sv, true, combine_flags(ECMAScriptFlags::Unicode, ECMAScriptFlags::Insensitive) },
}; };
for (auto& test : tests) { for (auto& test : tests) {

15
Userland/Libraries/LibRegex/RegexByteCode.cpp
View file

@ -701,12 +701,19 @@ ALWAYS_INLINE void OpCode_Compare::compare_char(MatchInput const& input, MatchSt
return; return;
// FIXME: Figure out how to do this if unicode() without performing a substring split first. // FIXME: Figure out how to do this if unicode() without performing a substring split first.
auto input_view = input.view.unicode() ? input.view.substring_view(state.string_position, 1)[0] : input.view.code_unit_at(state.string_position_in_code_units); auto input_view = input.view.unicode()
? input.view.substring_view(state.string_position, 1)[0]
: input.view.code_unit_at(state.string_position_in_code_units);
bool equal; bool equal;
if (input.regex_options & AllFlags::Insensitive) if (input.regex_options & AllFlags::Insensitive) {
equal = to_ascii_lowercase(input_view) == to_ascii_lowercase(ch1); // FIXME: Implement case-insensitive matching for non-ascii characters if (input.view.unicode())
else equal = Unicode::equals_ignoring_case(Utf32View { &input_view, 1 }, Utf32View { &ch1, 1 });
else
equal = to_ascii_lowercase(input_view) == to_ascii_lowercase(ch1);
} else {
equal = input_view == ch1; equal = input_view == ch1;
}
if (equal) { if (equal) {
if (inverse) if (inverse)