RGBCube/serenity
1
Fork 0
mirror of https://github.com/RGBCube/serenity synced 2025-05-21 17:15:07 +00:00
Code Issues Activity Actions

LibGUI: Handle utf-8 search strings in find

Browse source 
Similar to LibVT, we were iterating over needle bytes instead of code
points. This patch allows finding unicode substrings in a text document.
This commit is contained in:
Arda Cinar 2023-01-10 20:41:14 +03:00 • committed by Andreas Kling
parent 25f1e81d62
commit 1ef410eb79
1 changed files with 36 additions and 9 deletions
Download patch file Download diff file Expand all files Collapse all files

45
Userland/Libraries/LibGUI/TextDocument.cpp
View file

@ -15,6 +15,7 @@
#include <LibCore/Timer.h> #include <LibCore/Timer.h>
#include <LibGUI/TextDocument.h> #include <LibGUI/TextDocument.h>
#include <LibRegex/Regex.h> #include <LibRegex/Regex.h>
#include <LibUnicode/CharacterTypes.h>
namespace GUI { namespace GUI {
@ -499,14 +500,27 @@ TextRange TextDocument::find_next(StringView needle, TextPosition const& start,
TextPosition start_of_potential_match; TextPosition start_of_potential_match;
size_t needle_index = 0; size_t needle_index = 0;
Utf8View unicode_needle(needle);
Vector<u32> needle_code_points;
for (u32 code_point : unicode_needle)
needle_code_points.append(code_point);
do { do {
auto ch = code_point_at(position); auto ch = code_point_at(position);
// FIXME: This is not the right way to use a Unicode needle!
if (match_case ? ch == (u32)needle[needle_index] : tolower(ch) == tolower((u32)needle[needle_index])) { bool code_point_matches = false;
if (needle_index >= needle_code_points.size())
code_point_matches = false;
else if (match_case)
code_point_matches = ch == needle_code_points[needle_index];
else
code_point_matches = Unicode::to_unicode_lowercase(ch) == Unicode::to_unicode_lowercase(needle_code_points[needle_index]);
if (code_point_matches) {
if (needle_index == 0) if (needle_index == 0)
start_of_potential_match = position; start_of_potential_match = position;
++needle_index; ++needle_index;
if (needle_index >= needle.length()) if (needle_index >= needle_code_points.size())
return { start_of_potential_match, next_position_after(position, should_wrap) }; return { start_of_potential_match, next_position_after(position, should_wrap) };
} else { } else {
if (needle_index > 0) if (needle_index > 0)
@ -580,22 +594,35 @@ TextRange TextDocument::find_previous(StringView needle, TextPosition const& sta
return {}; return {};
TextPosition original_position = position; TextPosition original_position = position;
Utf8View unicode_needle(needle);
Vector<u32> needle_code_points;
for (u32 code_point : unicode_needle)
needle_code_points.append(code_point);
TextPosition end_of_potential_match; TextPosition end_of_potential_match;
size_t needle_index = needle.length() - 1; size_t needle_index = needle_code_points.size() - 1;
do { do {
auto ch = code_point_at(position); auto ch = code_point_at(position);
// FIXME: This is not the right way to use a Unicode needle!
if (match_case ? ch == (u32)needle[needle_index] : tolower(ch) == tolower((u32)needle[needle_index])) { bool code_point_matches = false;
if (needle_index == needle.length() - 1) if (needle_index >= needle_code_points.size())
code_point_matches = false;
else if (match_case)
code_point_matches = ch == needle_code_points[needle_index];
else
code_point_matches = Unicode::to_unicode_lowercase(ch) == Unicode::to_unicode_lowercase(needle_code_points[needle_index]);
if (code_point_matches) {
if (needle_index == needle_code_points.size() - 1)
end_of_potential_match = position; end_of_potential_match = position;
if (needle_index == 0) if (needle_index == 0)
return { position, next_position_after(end_of_potential_match, should_wrap) }; return { position, next_position_after(end_of_potential_match, should_wrap) };
--needle_index; --needle_index;
} else { } else {
if (needle_index < needle.length() - 1) if (needle_index < needle_code_points.size() - 1)
position = end_of_potential_match; position = end_of_potential_match;
needle_index = needle.length() - 1; needle_index = needle_code_points.size() - 1;
} }
position = previous_position_before(position, should_wrap); position = previous_position_before(position, should_wrap);
} while (position.is_valid() && position != original_position); } while (position.is_valid() && position != original_position);