Avoid non-ascii characters in catch.hpp

[autocompile]

Describe the bug
When compiling with Visual Studio (using compound single-file catch.hpp) with /W4, Visual Studio warns about non-ascii character found in catch.hpp.
NOTE: I'm NOT using an English version of Windows.

Expected behavior
Don't use non-ascii characters in Catch source files... as they are not necessary.

Reproduction steps
Just

#define CATCH_CONFIG_MAIN
#include "./catch.hpp"

is enough to trigger the VS warning.

Platform information:

• OS: Windows 10
• Compiler+version: Visual Studio 2017 15.9.7
• Catch version: v2.6.1

Additional context
Non-ascii characters exists in 2 lines of compound single-file catch.hpp (v2.6.1):

@@ -3815,7 +3815,7 @@ namespace Generators {
     template<typename L>
     // Note: The type after -> is weird, because VS2015 cannot parse
     //       the expression used in the typedef inside, when it is in
-    //       return type. Yeah, ¯\_(ツ)_/¯
+    //       return type. Yeah.
     auto generate( SourceLineInfo const& lineInfo, L const& generatorExpression ) -> decltype(std::declval<decltype(generatorExpression())>().get()) {
         using UnderlyingType = typename decltype(generatorExpression())::type;

@@ -13315,7 +13315,7 @@ public:
         case Unit::Nanoseconds:
             return "ns";
         case Unit::Microseconds:
-            return "µs";
+            return "us";
         case Unit::Milliseconds:
             return "ms";
         case Unit::Seconds:

For the first occurrence, just remove the emoticon in the comment.
For the second occurrence, consider two ways:

• 1. As the diff above shows, use English letter u instead of Greek letter µ. It's clear for users to understand us refers to microseconds.
• 2. Use an escape sequence for Greek letter µ, like return "\x????s";. But you have to choose an encoding first.

Personally, I strongly suggest the first approach, since the bytes presentation of µ may vary among encodings.

[horenmar]

Oh well.