Change TokenKind's Print overload to a format_provider. (#2534)

Fundamentally this `.Print()` is wrong for debug output at present because `.fixed_spelling()` can be empty. It's also inconsistent with other enums to use it. We frequently print tokens for debugging, and it's easy to forget to specify `.name()` there.

Diagnostics use formatv, so we can provide a format_provider and address it in one spot that way. It also makes it harder to just forget to do the right thing.

toolchain/semantics/semantics_parse_tree_handler.cpp

@@ -350,7 +350,7 @@ auto SemanticsParseTreeHandler::HandleInfixOperator(ParseTree::Node parse_node)
                                      parse_node, result_type, lhs_id, rhs_id));
       break;
     default:
-      CARBON_FATAL() << "Unrecognized token kind: " << token_kind.name();
+      CARBON_FATAL() << "Unrecognized token kind: " << token_kind;
   }
 }
 
@@ -393,7 +393,7 @@ auto SemanticsParseTreeHandler::HandleLiteral(ParseTree::Node parse_node)
       break;
     }
     default:
-      CARBON_FATAL() << "Unhandled kind: " << token_kind.name();
+      CARBON_FATAL() << "Unhandled kind: " << token_kind;
   }
 }