BRNSystems/minetest
1
0
Fork 0
forked from Mirrorlandia_minetest/minetest
Code Pull Requests 1 Activity

Hand roll UTF-16 conversion in CGUITTFont (#14121)

Browse Source
This commit is contained in:
sfan5 2023-12-19 20:18:43 +01:00 committed by GitHub
parent 00d9d96e48
commit 61d0f613df
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 22 additions and 14 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

36
src/irrlicht_changes/CGUITTFont.cpp
View File

@ -32,8 +32,6 @@
#include <irrlicht.h> #include <irrlicht.h>
#include <iostream> #include <iostream>
#include <codecvt>
#include <locale>
#include "CGUITTFont.h" #include "CGUITTFont.h"
namespace irr namespace irr
@ -1251,21 +1249,31 @@ std::u32string CGUITTFont::convertWCharToU32String(const wchar_t* const charArra
{ {
static_assert(sizeof(wchar_t) == 2 || sizeof(wchar_t) == 4, "unexpected wchar size"); static_assert(sizeof(wchar_t) == 2 || sizeof(wchar_t) == 4, "unexpected wchar size");
if (sizeof(wchar_t) == 4) // Systems where wchar_t is UTF-32 if (sizeof(wchar_t) == 4) // wchar_t is UTF-32
return std::u32string(reinterpret_cast<const char32_t*>(charArray)); return std::u32string(reinterpret_cast<const char32_t*>(charArray));
// Systems where wchar_t is UTF-16: // wchar_t is UTF-16 and we need to convert.
// First, convert to UTF-8 // std::codecvt could do this for us but aside from being deprecated,
std::u16string utf16String(reinterpret_cast<const char16_t*>(charArray)); // it turns out that it's laughably slow on MSVC. Thanks Microsoft.
std::wstring_convert<std::codecvt_utf8_utf16<char16_t>, char16_t> converter1;
std::string utf8String = converter1.to_bytes(utf16String);
// Next, convert to UTF-32 std::u32string ret;
std::wstring_convert<std::codecvt_utf8<char32_t>, char32_t> converter2; ret.reserve(wcslen(charArray));
return converter2.from_bytes(utf8String); const wchar_t *p = charArray;
while (*p) {
// This is inefficient, but importantly it is _correct_, rather than a hand-rolled UTF-16 to char32_t c = *p;
// UTF-32 converter which may or may not be correct. if (c >= 0xD800 && c < 0xDC00) {
p++;
char32_t c2 = *p;
if (!c2)
break;
else if (c2 < 0xDC00 || c2 > 0xDFFF)
continue; // can't find low surrogate, skip
c = 0x10000 + ( ((c & 0x3ff) << 10) | (c2 & 0x3ff) );
}
ret.push_back(c);
p++;
}
return ret;
} }