diff --git a/six/modules/c++/six.sicd/unittests/test_CollectionInfo.cpp b/six/modules/c++/six.sicd/unittests/test_CollectionInfo.cpp
index 8b6e84847..a7db81ab7 100644
--- a/six/modules/c++/six.sicd/unittests/test_CollectionInfo.cpp
+++ b/six/modules/c++/six.sicd/unittests/test_CollectionInfo.cpp
@@ -39,16 +39,6 @@
 
 #include "TestCase.h"
 
-// It seems that a macro is better than a utility routine, see https://github.com/tahonermann/char8_t-remediation
-// C++20 changed the type of u8 to char8_t* https://en.cppreference.com/w/cpp/language/string_literal
-// Not putting this everywhere because (1) well, it's a macro, and (2) it's mostly
-// only test code that uses string literals.
-#if CODA_OSS_cpp20
-#define U8(s) u8##s
-#else
-#define U8(s) static_cast<const std::char8_t*>(static_cast<const void*>(s))
-#endif
-
 static void test_DummyData_parameters(const std::string& testName, const six::ParameterCollection& parameters,
     std::optional<bool> preserveCharacterData = std::optional<bool>())
 {
@@ -135,8 +125,10 @@ TEST_CASE(ClassificationCanada)
 {
 #ifdef _WIN32
     const std::string classificationText("NON CLASSIFI\xc9 / UNCLASSIFIED"); // ISO8859-1 "NON CLASSIFIÉ / UNCLASSIFIED"
+    const auto E_ = str::u8FromString("\xc9"); // ISO8859-1  "É"
 #else
     const std::string classificationText("NON CLASSIFI\xc3\x89 / UNCLASSIFIED"); // UTF-8 "NON CLASSIFIÉ / UNCLASSIFIED"
+    const auto E_ = str::u8FromString("\xc3\x89"); // UTF-8  "É"
 #endif
 
     auto data = createData<six::zfloat>(types::RowCol<size_t>(10, 10));
@@ -145,14 +137,14 @@ TEST_CASE(ClassificationCanada)
 
     const auto strXml = six::sicd::Utilities::toXMLString(*data, nullptr /*pSchemaPaths*/);
 
-    const auto NON_CLASSIFI = strXml.find(U8("NON CLASSIFI"));
+    const auto NON_CLASSIFI = strXml.find(str::u8FromString("NON CLASSIFI"));
     TEST_ASSERT(NON_CLASSIFI != std::string::npos);
-    const auto UNCLASSIFIED = strXml.find(U8(" / UNCLASSIFIED"));
+    const auto UNCLASSIFIED = strXml.find(str::u8FromString(" / UNCLASSIFIED"));
     TEST_ASSERT(UNCLASSIFIED != std::string::npos);
     const auto utf8 = strXml.substr(NON_CLASSIFI, UNCLASSIFIED - NON_CLASSIFI);
     TEST_ASSERT_EQ(utf8.size(), std::string("NON CLASSIFI\xc3\x89").size()); // UTF-8, "NON CLASSIFIÉ"
-    const auto E_ = utf8.find(U8("\xc3\x89")); // UTF-8,  "É"
-    TEST_ASSERT(E_ != std::string::npos);
+    const auto posE_ = utf8.find(E_);
+    TEST_ASSERT(posE_ != std::string::npos);
 
     io::U8StringStream ss;
     ss.stream() << strXml;