AK+LibRegex: Partially implement case insensitive UTF-16 comparison

This will work for ASCII code points. Unicode case folding will be
needed for non-ASCII.
This commit is contained in:
Timothy Flynn 2021-07-21 16:38:12 -04:00 committed by Linus Groh
parent 47f6bb38a1
commit 0e6375558d
Notes: sideshowbarker 2024-07-18 08:26:28 +09:00
3 changed files with 24 additions and 0 deletions

View file

@ -4,6 +4,7 @@
* SPDX-License-Identifier: BSD-2-Clause
*/
#include <AK/CharacterTypes.h>
#include <AK/StringBuilder.h>
#include <AK/StringView.h>
#include <AK/Utf16View.h>
@ -214,6 +215,22 @@ bool Utf16View::operator==(Utf16View const& other) const
return true;
}
bool Utf16View::equals_ignoring_case(Utf16View const& other) const
{
if (length_in_code_units() == 0)
return other.length_in_code_units() == 0;
if (length_in_code_units() != other.length_in_code_units())
return false;
for (size_t i = 0; i < length_in_code_units(); ++i) {
// FIXME: Handle non-ASCII case insensitive comparisons.
if (to_ascii_lowercase(m_code_units[i]) != to_ascii_lowercase(other.m_code_units[i]))
return false;
}
return true;
}
Utf16CodePointIterator& Utf16CodePointIterator::operator++()
{
size_t code_units = length_in_code_units();

View file

@ -104,6 +104,8 @@ public:
return validate(valid_code_units);
}
bool equals_ignoring_case(Utf16View const&) const;
private:
u16 const* begin_ptr() const { return m_code_units.data(); }
u16 const* end_ptr() const { return begin_ptr() + m_code_units.size(); }

View file

@ -335,6 +335,11 @@ public:
[&](StringView other_view) { return view.equals_ignoring_case(other_view); },
[](auto&) -> bool { TODO(); });
},
[&](Utf16View view) {
return other.m_view.visit(
[&](Utf16View other_view) { return view.equals_ignoring_case(other_view); },
[](auto&) -> bool { TODO(); });
},
[](auto&) -> bool { TODO(); });
}