Visual Studio 16.8 has different floating point precision

GitHub action runners now feature Visual Studio 16.8.1 (previously:
16.7). According to its STL changelog, this update includes the
following change:

microsoft/STL#1173 (review)

Therefore update the unit test to expect 13 trailing zeros for MSVC
version >= 19.28.

Author: tautschnig

jbmc/unit/java_bytecode/expr2java.cpp

@@ -99,6 +99,10 @@ TEST_CASE(
 #ifndef _MSC_VER
     REQUIRE(
       floating_point_to_java_string(value) == "0x1p+37f /* 1.37439e+11 */");
+#elif _MSC_VER >= 1928
+    REQUIRE(
+      floating_point_to_java_string(value) ==
+      "0x1.0000000000000p+37f /* 1.37439e+11 */");
 #else
     REQUIRE(
       floating_point_to_java_string(value) ==
@@ -112,6 +116,10 @@ TEST_CASE(
 #ifndef _MSC_VER
     REQUIRE(
       floating_point_to_java_string(value) == "0x1p+37 /* 1.37439e+11 */");
+#elif _MSC_VER >= 1928
+    REQUIRE(
+      floating_point_to_java_string(value) ==
+      "0x1.0000000000000p+37 /* 1.37439e+11 */");
 #else
     REQUIRE(
       floating_point_to_java_string(value) ==