From 692fe84331894a55afe679f47259329098c45682 Mon Sep 17 00:00:00 2001
From: Jan Wojciechowski <janwoj0@gmail.com>
Date: Sun, 5 Mar 2023 09:40:34 +0100
Subject: [PATCH] Fix condition for trailing bytes count in UTF-8 decoder. Test
 added to check if a replacement characters is added to output.

---
 include/SFML/System/Utf.inl |  2 +-
 test/System/String.test.cpp | 36 ++++++++++++++++++++++++------------
 2 files changed, 25 insertions(+), 13 deletions(-)

diff --git a/include/SFML/System/Utf.inl b/include/SFML/System/Utf.inl
index 02211135..d72931af 100644
--- a/include/SFML/System/Utf.inl
+++ b/include/SFML/System/Utf.inl
@@ -69,7 +69,7 @@ In Utf<8>::decode(In begin, In end, std::uint32_t& output, std::uint32_t replace
 
     // decode the character
     int trailingBytes = trailing[static_cast<std::uint8_t>(*begin)];
-    if (begin + trailingBytes < end)
+    if (trailingBytes < std::distance(begin, end))
     {
         output = 0;
 
diff --git a/test/System/String.test.cpp b/test/System/String.test.cpp
index 20cde7f7..a0f96661 100644
--- a/test/System/String.test.cpp
+++ b/test/System/String.test.cpp
@@ -230,18 +230,30 @@ TEST_CASE("[System] sf::String")
 
     SUBCASE("fromUtf8()")
     {
-        constexpr std::array<std::uint8_t, 4> characters{'w', 'x', 'y', 'z'};
-        const sf::String                      string = sf::String::fromUtf8(characters.begin(), characters.end());
-        CHECK(std::string(string) == "wxyz"s);
-        CHECK(std::wstring(string) == L"wxyz"s);
-        CHECK(string.toAnsiString() == "wxyz"s);
-        CHECK(string.toWideString() == L"wxyz"s);
-        CHECK(string.toUtf8() == std::basic_string<std::uint8_t>{'w', 'x', 'y', 'z'});
-        CHECK(string.toUtf16() == u"wxyz"s);
-        CHECK(string.toUtf32() == U"wxyz"s);
-        CHECK(string.getSize() == 4);
-        CHECK(!string.isEmpty());
-        CHECK(string.getData() != nullptr);
+        SUBCASE("Nominal")
+        {
+            constexpr std::array<std::uint8_t, 4> characters{'w', 'x', 'y', 'z'};
+            const sf::String                      string = sf::String::fromUtf8(characters.begin(), characters.end());
+            CHECK(std::string(string) == "wxyz"s);
+            CHECK(std::wstring(string) == L"wxyz"s);
+            CHECK(string.toAnsiString() == "wxyz"s);
+            CHECK(string.toWideString() == L"wxyz"s);
+            CHECK(string.toUtf8() == std::basic_string<std::uint8_t>{'w', 'x', 'y', 'z'});
+            CHECK(string.toUtf16() == u"wxyz"s);
+            CHECK(string.toUtf32() == U"wxyz"s);
+            CHECK(string.getSize() == 4);
+            CHECK(!string.isEmpty());
+            CHECK(string.getData() != nullptr);
+        }
+
+        SUBCASE("Insufficient input")
+        {
+            constexpr std::array<std::uint8_t, 1> characters{251};
+            const sf::String                      string = sf::String::fromUtf8(characters.begin(), characters.end());
+            constexpr char32_t                    defaultReplacementCharacter = 0;
+            CHECK(string.getSize() == 1);
+            CHECK(string[0] == defaultReplacementCharacter);
+        }
     }
 
     SUBCASE("fromUtf16()")