tf: use MSVC preprocessor version to determine which set of macros to…

… use in preprocessorUtilsLite.h

In the cases where we were previously checking whether the compiler was MSVC
when deciding whether to use an alternative set of macros versus other
toolchains, what we really wanted to know was whether MSVC's "traditional"
preprocessor was being used.

This replaces the use of ARCH_COMPILER_MSVC in preprocessorUtilsLite.h with the
newly added ARCH_PREPROCESSOR_MSVC_TRADITIONAL so that only in cases where we're
compiling with MSVC *and* using its traditional preprocessor do we use the
alternate macros. Otherwise when using MSVC's newer, more standards-conforming
preprocessor, we use the same macros as other toolchains.

pxr/base/tf/preprocessorUtilsLite.h

@@ -45,7 +45,7 @@
 /// Expand and convert the argument to a string, using a most minimal macro.
 #define TF_PP_STRINGIZE(x) TF_PP_STRINGIZE_IMPL(x)
 
-#ifdef ARCH_COMPILER_MSVC
+#ifdef ARCH_PREPROCESSOR_MSVC_TRADITIONAL
 
 /// Expand to the number of arguments passed.  For example,
 /// TF_PP_VARIADIC_SIZE(foo, bar, baz) expands to 3.  Supports up to 64
@@ -70,7 +70,7 @@
 /// and TF_PP_VARIADIC_ELEM(1, a, b, c) expands to b.
 #define TF_PP_VARIADIC_ELEM(n, ...) TF_PP_CAT(TF_PP_VAE_, n)(__VA_ARGS__,)
 
-#endif // ARCH_COMPILER_MSVC
+#endif // ARCH_PREPROCESSOR_MSVC_TRADITIONAL
 
 #define TF_PP_VARIADIC_SIZE_IMPL(a0, a1, a2, a3, a4, a5, a6, a7, a8, a9, a10, a11, a12, a13, a14, a15, a16, a17, a18, a19, a20, a21, a22, a23, a24, a25, a26, a27, a28, a29, a30, a31, a32, a33, a34, a35, a36, a37, a38, a39, a40, a41, a42, a43, a44, a45, a46, a47, a48, a49, a50, a51, a52, a53, a54, a55, a56, a57, a58, a59, a60, a61, a62, a63, size, ...) size
 
@@ -139,7 +139,7 @@
 #define TF_PP_VAE_62(a0, a1, a2, a3, a4, a5, a6, a7, a8, a9, a10, a11, a12, a13, a14, a15, a16, a17, a18, a19, a20, a21, a22, a23, a24, a25, a26, a27, a28, a29, a30, a31, a32, a33, a34, a35, a36, a37, a38, a39, a40, a41, a42, a43, a44, a45, a46, a47, a48, a49, a50, a51, a52, a53, a54, a55, a56, a57, a58, a59, a60, a61, a62, ...) a62
 #define TF_PP_VAE_63(a0, a1, a2, a3, a4, a5, a6, a7, a8, a9, a10, a11, a12, a13, a14, a15, a16, a17, a18, a19, a20, a21, a22, a23, a24, a25, a26, a27, a28, a29, a30, a31, a32, a33, a34, a35, a36, a37, a38, a39, a40, a41, a42, a43, a44, a45, a46, a47, a48, a49, a50, a51, a52, a53, a54, a55, a56, a57, a58, a59, a60, a61, a62, a63, ...) a63
 
-#ifdef ARCH_COMPILER_MSVC
+#ifdef ARCH_PREPROCESSOR_MSVC_TRADITIONAL
 
 #define TF_PP_FE_0(_macro, ...)
 #define TF_PP_FE_1(_macro, a) _macro(a)
@@ -275,7 +275,7 @@
 
 #endif
 
-#ifdef ARCH_COMPILER_MSVC
+#ifdef ARCH_PREPROCESSOR_MSVC_TRADITIONAL
 
 /// Expand the macro \p x on every variadic argument.  For example
 /// TF_PP_FOR_EACH(MACRO, foo, bar, baz) expands to MACRO(foo) MACRO(bar)
@@ -415,7 +415,7 @@
     _TF_PP_SEQ_EXPAND(_TF_PP_SEQ_DISCARD_HEAD _TF_PP_SEQ_PARTITION_HEAD(seq))
 
 #define _TF_PP_SEQ_FE_0(_macro, ...)
-#ifdef ARCH_COMPILER_MSVC
+#ifdef ARCH_PREPROCESSOR_MSVC_TRADITIONAL
 #define _TF_PP_SEQ_FE_1(_macro, data, seq) TF_PP_CAT(_macro(data, _TF_PP_SEQ_HEAD(seq)),)
 // # Generates _TF_PP_SEQ_FE_{2:229} (MSVC)
 // python3 -c 'print("\n".join(f"#define _TF_PP_SEQ_FE_{i}(_macro, data, seq) TF_PP_CAT(_macro(data, _TF_PP_SEQ_HEAD(seq)),) TF_PP_CAT(_TF_PP_SEQ_FE_{i-1}(_macro, data, _TF_PP_SEQ_TAIL(seq)),)" for i in range(2, 230)))'
@@ -889,7 +889,7 @@
 /// #undef _PRINT
 /// \endcode
 /// Limited to sequences of up to 229 elements
-#ifdef ARCH_COMPILER_MSVC
+#ifdef ARCH_PREPROCESSOR_MSVC_TRADITIONAL
 #define _TF_PP_SEQ_FOR_EACH_IMPL(_macro, size, data, seq)                   \
     TF_PP_CAT(TF_PP_CAT(_TF_PP_SEQ_FE_, size),(_macro, data, seq))
 #define TF_PP_SEQ_FOR_EACH(_macro, data, seq)                               \