[libSyntax] Remove the C++ SyntaxClassifier

The recommended way forward is to use the SyntaxClassifier on the Swift
side.

By removing the C++ SyntaxClassifier, we can also eliminate the
-force-libsyntax-based-processing option that was used to bootstrap
incremental parsing and would generate the syntax map from a syntax
tree.

unittests/SourceKit/SwiftLang/CursorInfoTest.cpp

@@ -36,6 +36,8 @@ class NullEditorConsumer : public EditorConsumer {
     llvm_unreachable("unexpected error");
   }
 
+  bool syntaxMapEnabled() override { return true; }
+
   bool handleSyntaxMap(unsigned Offset, unsigned Length, UIdent Kind) override {
     return false;
   }
@@ -44,6 +46,8 @@ class NullEditorConsumer : public EditorConsumer {
                                 UIdent Kind, bool isSystem) override {
     return false;
   }
+  
+  bool documentStructureEnabled() override { return false; }
 
   bool beginDocumentSubStructure(unsigned Offset, unsigned Length,
                                  UIdent Kind, UIdent AccessLevel,
@@ -100,8 +104,6 @@ class NullEditorConsumer : public EditorConsumer {
       std::vector<SourceFileRange> ReuseRegions) override {
     return false;
   }
-
-  bool forceLibSyntaxBasedProcessing() override { return false; }
 public:
   bool needsSema = false;
 };