Rewrite the tblgen parser in a recursive descent style, eliminating the bison parser.

This makes the parser much easier to understand, eliminates a ton of global variables,
and gives tblgen nice caret diagnostics.  It is also faster, but tblgen probably doesn't
care about performance.

There are a couple of FIXMEs which I will take care of next.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@44274 91177308-0d34-0410-b5e6-96231b3b80d8

1 files changed, 66 insertions, 13 deletions

diff --git a/utils/TableGen/TGLexer.h b/utils/TableGen/TGLexer.h
index ef786b6f28..9f1b535cd4 100644
--- a/utils/TableGen/TGLexer.h
+++ b/utils/TableGen/TGLexer.h
@@ -20,12 +20,49 @@
 
 namespace llvm {
 class MemoryBuffer;
+  
+namespace tgtok {
+  enum TokKind {
+    // Markers
+    Eof, Error,
+    
+    // Tokens with no info.
+    minus, plus,        // - +
+    l_square, r_square, // [ ]
+    l_brace, r_brace,   // { }
+    l_paren, r_paren,   // ( )
+    less, greater,      // < >
+    colon, semi,        // ; :
+    comma, period,      // , .
+    equal, question,    // = ?
+    
+    // Keywords.
+    Bit, Bits, Class, Code, Dag, Def, Defm, Field, In, Int, Let, List,
+    MultiClass, String,
+    
+    // !keywords.
+    XConcat, XSRA, XSRL, XSHL, XStrConcat,
+    
+    // Integer value.
+    IntVal,
+    
+    // String valued tokens.
+    Id, StrVal, VarName, CodeFragment
+  };
+}
 
+/// TGLexer - TableGen Lexer class.
 class TGLexer {
   const char *CurPtr;
   unsigned CurLineNo;
   MemoryBuffer *CurBuf;
 
+  // Information about the current token.
+  const char *TokStart;
+  tgtok::TokKind CurCode;
+  std::string CurStrVal;  // This is valid for ID, STRVAL, VARNAME, CODEFRAGMENT
+  int CurIntVal;          // This is valid for INTVAL.
+  
   /// IncludeRec / IncludeStack - This captures the current set of include
   /// directives we are nested within.
   struct IncludeRec {
@@ -40,7 +77,6 @@ class TGLexer {
   // IncludeDirectories - This is the list of directories we should search for
   // include files in.
   std::vector<std::string> IncludeDirectories;
-  const char *TokStart;
 public:
   TGLexer(MemoryBuffer *StartBuf);
   ~TGLexer();
@@ -49,29 +85,46 @@ public:
     IncludeDirectories = Dirs;
   }
   
-  int LexToken();
+  tgtok::TokKind Lex() {
+    return CurCode = LexToken();
+  }
+  
+  tgtok::TokKind getCode() const { return CurCode; }
 
-  typedef const char* LocationTy;
-  LocationTy getTokenStart() const { return TokStart; }
+  const std::string &getCurStrVal() const {
+    assert((CurCode == tgtok::Id || CurCode == tgtok::StrVal || 
+            CurCode == tgtok::VarName || CurCode == tgtok::CodeFragment) &&
+           "This token doesn't have a string value");
+    return CurStrVal;
+  }
+  int getCurIntVal() const {
+    assert(CurCode == tgtok::IntVal && "This token isn't an integer");
+    return CurIntVal;
+  }
 
-  void PrintError(LocationTy Loc, const std::string &Msg) const;
+  typedef const char* LocTy;
+  LocTy getLoc() const { return TokStart; }
+
+  void PrintError(LocTy Loc, const std::string &Msg) const;
   
-  std::ostream &err() const;
   void PrintIncludeStack(std::ostream &OS) const;
   
 private:
-  int ReturnError(const char *Loc, const std::string &Msg);
+  /// LexToken - Read the next token and return its code.
+  tgtok::TokKind LexToken();
+  
+  tgtok::TokKind ReturnError(const char *Loc, const std::string &Msg);
   
   int getNextChar();
   void SkipBCPLComment();
   bool SkipCComment();
-  int LexIdentifier();
+  tgtok::TokKind LexIdentifier();
   bool LexInclude();
-  int LexString();
-  int LexVarName();
-  int LexNumber();
-  int LexBracket();
-  int LexExclaim();
+  tgtok::TokKind LexString();
+  tgtok::TokKind LexVarName();
+  tgtok::TokKind LexNumber();
+  tgtok::TokKind LexBracket();
+  tgtok::TokKind LexExclaim();
 };
   
 } // end namespace llvm