Bugfixes

2 files changed, 20 insertions, 8 deletions

diff --git a/Makefile b/Makefile
index 7ed522b..7ea075e 100644
--- a/Makefile
+++ b/Makefile
@@ -36,6 +36,7 @@ LIBS+=-fuse-ld=lld
 endif
 
 ifeq ($(COMPILER_SUITE),clang)
+# libc++ is buggy for C++20: std::hash<char8_t> not implemented. Fixed in LLVM-12
 CXXFLAGS+=-stdlib=libc++
 endif
 
@@ -55,7 +56,7 @@ LIBS+= \
 else
 LIBS+= \
 -lstdc++
-#-lstdc++fs
+-lstdc++fs
 endif
 
 SRC=\
diff --git a/src/test-unicode.cpp b/src/test-unicode.cpp
index c169fc9..5529d2c 100644
--- a/src/test-unicode.cpp
+++ b/src/test-unicode.cpp
@@ -35,13 +35,11 @@ std::vector<types_collection_type> success_sets {
 
 // Error cases: throwing upon convert to all other types
 std::vector<std::basic_string<utf8_t>> failure_strings_char8_t {
- // Note: don't encode this as u8"" since MSVC will interpret \x80 as \u0080,
-//        yet to be encoded to UTF-8 for execution encoding
- "\x80", // utf-8 continuation byte
- "\x81", // utf-8 continuation byte
- "\xc3\xc3\xa4", // initial byte of utf-8 "ä", followed by valid utf-8 "ä"
- "\xF8\x80\x80\x80\x80", // overlong encoding
- "\xF7\xBF\xBF\xBF", // valid encoding of invalid code point
+ u8"\x80", // utf-8 continuation byte
+ u8"\x81", // utf-8 continuation byte
+ u8"\xc3\xc3\xa4", // initial byte of utf-8 "ä", followed by valid utf-8 "ä"
+ u8"\xF8\x80\x80\x80\x80", // overlong encoding
+ u8"\xF7\xBF\xBF\xBF", // valid encoding of invalid code point
 };
 
 std::vector<std::basic_string<char16_t>> failure_strings_char16_t {
@@ -333,6 +331,19 @@ BOOST_AUTO_TEST_CASE(convert)
  BOOST_CHECK((unicode::convert<char32_t,char16_t>(U"abc")) == std::u16string{u"abc"});
 }
 
+BOOST_AUTO_TEST_CASE(string_u8string)
+{
+ std::string a{"\xc3\xa4"};
+
+ std::basic_string<utf8_t> b{a.begin(), a.end()};
+
+ BOOST_CHECK(b == std::basic_string<utf8_t>{u8"ä"});
+ 
+ a = std::string{b.begin(), b.end()};
+ 
+ BOOST_CHECK(a == std::string{"\xc3\xa4"});
+}
+
 // TODO:
 //
 // char8_t, char16_t, char32_t, char, wchar_t (UTF-16 on Windows, UTF-32 on Linux)