MC: fix layering violation introduced in r325139

[royger]

Summary

Move AsmToken from MC/MCParser/MCAsmLexer.h into MC/MCAsmMacro.h in
order to avoid MCAsmMacro definition depending on MCParser.

Note that I haven't been able to even fully build test this due to the
following issue:

llvm/tools/clang/lib/Basic/SourceManager.cpp:173:17: error: no viable conversion from 'const char [5]' to 'llvm::StringLiteral'

.StartsWith("\x00\x00\xFE\xFF", "UTF-32 (BE)")
            ^~~~~~~~~~~~~~~~~~

llvm/include/llvm/ADT/StringRef.h:857:9: note: candidate constructor (the implicit copy constructor) not viable: no known conversion from 'const char [5]' to 'const llvm::StringLiteral &' for 1st argument

class StringLiteral : public StringRef {
      ^

This is not related to my change.

[• espindola]

LGTM

This is unfortunate, but I don't think we can do better with the current split. The alternative would be to effectively revert r325139 and have clang/llc use a single parser for all the inline assembly in a file.

[labath]

llvm/tools/clang/lib/Basic/SourceManager.cpp:173:17: error: no viable conversion from 'const char [5]' to 'llvm::StringLiteral'

Are you using gcc (~4.x) by any chance? Older versions of gcc seem to have issues with the StringLiteral constructor. One solution is to wrap the argument in extra {}, so something like .StartsWith({"\x00\x00\xFE\xFF"}, "UTF-32 (BE)")

[royger]

In D44531#1039804, @labath wrote:

llvm/tools/clang/lib/Basic/SourceManager.cpp:173:17: error: no viable conversion from 'const char [5]' to 'llvm::StringLiteral'

Are you using gcc (~4.x) by any chance? Older versions of gcc seem to have issues with the StringLiteral constructor. One solution is to wrap the argument in extra {}, so something like .StartsWith({"\x00\x00\xFE\xFF"}, "UTF-32 (BE)")

No, I'm using clang 6.0:

FreeBSD clang version 6.0.0 (branches/release_60 325330) (based on LLVM 6.0.0)

TBH I'm not sure why I get this error, IIRC this is a pre-release clang 6.0, so maybe I'm missing some patches?

[dblaikie]

Ping on this patch - would be great to get the layering violation fixed.

[• espindola]

r328003

include/llvm/MC/MCAsmMacro.h

 //===- MCAsmMacro.h - Assembly Macros ---------------------------*- C++ -*-===//
 //
 //                     The LLVM Compiler Infrastructure
 //
 // This file is distributed under the University of Illinois Open Source
 // License. See LICENSE.TXT for details.
 //
 //===----------------------------------------------------------------------===//
 
 #ifndef LLVM_MC_MCASMMACRO_H
 #define LLVM_MC_MCASMMACRO_H
 
-#include "llvm/MC/MCParser/MCAsmLexer.h"
+#include "llvm/ADT/APInt.h"
+#include "llvm/ADT/StringRef.h"
 #include "llvm/Support/Debug.h"
+#include "llvm/Support/SMLoc.h"
+#include <vector>
 
 namespace llvm {
 
+/// Target independent representation for an assembler token.
+class AsmToken {
+public:
+  enum TokenKind {
+    // Markers
+    Eof, Error,
+
+    // String values.
+    Identifier,
+    String,
+
+    // Integer values.
+    Integer,
+    BigNum, // larger than 64 bits
+
+    // Real values.
+    Real,
+
+    // Comments
+    Comment,
+    HashDirective,
+    // No-value.
+    EndOfStatement,
+    Colon,
+    Space,
+    Plus, Minus, Tilde,
+    Slash,     // '/'
+    BackSlash, // '\'
+    LParen, RParen, LBrac, RBrac, LCurly, RCurly,
+    Star, Dot, Comma, Dollar, Equal, EqualEqual,
+
+    Pipe, PipePipe, Caret,
+    Amp, AmpAmp, Exclaim, ExclaimEqual, Percent, Hash,
+    Less, LessEqual, LessLess, LessGreater,
+    Greater, GreaterEqual, GreaterGreater, At,
+
+    // MIPS unary expression operators such as %neg.
+    PercentCall16, PercentCall_Hi, PercentCall_Lo, PercentDtprel_Hi,
+    PercentDtprel_Lo, PercentGot, PercentGot_Disp, PercentGot_Hi, PercentGot_Lo,
+    PercentGot_Ofst, PercentGot_Page, PercentGottprel, PercentGp_Rel, PercentHi,
+    PercentHigher, PercentHighest, PercentLo, PercentNeg, PercentPcrel_Hi,
+    PercentPcrel_Lo, PercentTlsgd, PercentTlsldm, PercentTprel_Hi,
+    PercentTprel_Lo
+  };
+
+private:
+  TokenKind Kind;
+
+  /// A reference to the entire token contents; this is always a pointer into
+  /// a memory buffer owned by the source manager.
+  StringRef Str;
+
+  APInt IntVal;
+
+public:
+  AsmToken() = default;
+  AsmToken(TokenKind Kind, StringRef Str, APInt IntVal)
+      : Kind(Kind), Str(Str), IntVal(std::move(IntVal)) {}
+  AsmToken(TokenKind Kind, StringRef Str, int64_t IntVal = 0)
+      : Kind(Kind), Str(Str), IntVal(64, IntVal, true) {}
+
+  TokenKind getKind() const { return Kind; }
+  bool is(TokenKind K) const { return Kind == K; }
+  bool isNot(TokenKind K) const { return Kind != K; }
+
+  SMLoc getLoc() const;
+  SMLoc getEndLoc() const;
+  SMRange getLocRange() const;
+
+  /// Get the contents of a string token (without quotes).
+  StringRef getStringContents() const {
+    assert(Kind == String && "This token isn't a string!");
+    return Str.slice(1, Str.size() - 1);
+  }
+
+  /// Get the identifier string for the current token, which should be an
+  /// identifier or a string. This gets the portion of the string which should
+  /// be used as the identifier, e.g., it does not include the quotes on
+  /// strings.
+  StringRef getIdentifier() const {
+    if (Kind == Identifier)
+      return getString();
+    return getStringContents();
+  }
+
+  /// Get the string for the current token, this includes all characters (for
+  /// example, the quotes on strings) in the token.
+  ///
+  /// The returned StringRef points into the source manager's memory buffer, and
+  /// is safe to store across calls to Lex().
+  StringRef getString() const { return Str; }
+
+  // FIXME: Don't compute this in advance, it makes every token larger, and is
+  // also not generally what we want (it is nicer for recovery etc. to lex 123br
+  // as a single token, then diagnose as an invalid number).
+  int64_t getIntVal() const {
+    assert(Kind == Integer && "This token isn't an integer!");
+    return IntVal.getZExtValue();
+  }
+
+  APInt getAPIntVal() const {
+    assert((Kind == Integer || Kind == BigNum) &&
+           "This token isn't an integer!");
+    return IntVal;
+  }
+
+  void dump(raw_ostream &OS) const;
+  void dump() const { dump(dbgs()); }
+};
+
 struct MCAsmMacroParameter {
   StringRef Name;
   std::vector<AsmToken> Value;
   bool Required = false;
   bool Vararg = false;
 
   MCAsmMacroParameter() = default;
 

include/llvm/MC/MCParser/MCAsmLexer.h

 //===- llvm/MC/MCAsmLexer.h - Abstract Asm Lexer Interface ------*- C++ -*-===//
 //
 //                     The LLVM Compiler Infrastructure
 //
 // This file is distributed under the University of Illinois Open Source
 // License. See LICENSE.TXT for details.
 //
 //===----------------------------------------------------------------------===//
 
 #ifndef LLVM_MC_MCPARSER_MCASMLEXER_H
 #define LLVM_MC_MCPARSER_MCASMLEXER_H
 
-#include "llvm/ADT/APInt.h"
 #include "llvm/ADT/ArrayRef.h"
 #include "llvm/ADT/SmallVector.h"
-#include "llvm/ADT/StringRef.h"
-#include "llvm/Support/Debug.h"
-#include "llvm/Support/SMLoc.h"
+#include "llvm/MC/MCAsmMacro.h"
 #include <algorithm>
 #include <cassert>
 #include <cstddef>
 #include <cstdint>
 #include <string>
 
 namespace llvm {
 
-/// Target independent representation for an assembler token.
-class AsmToken {
-public:
-  enum TokenKind {
-    // Markers
-    Eof, Error,
-
-    // String values.
-    Identifier,
-    String,
-
-    // Integer values.
-    Integer,
-    BigNum, // larger than 64 bits
-
-    // Real values.
-    Real,
-
-    // Comments
-    Comment,
-    HashDirective,
-    // No-value.
-    EndOfStatement,
-    Colon,
-    Space,
-    Plus, Minus, Tilde,
-    Slash,     // '/'
-    BackSlash, // '\'
-    LParen, RParen, LBrac, RBrac, LCurly, RCurly,
-    Star, Dot, Comma, Dollar, Equal, EqualEqual,
-
-    Pipe, PipePipe, Caret,
-    Amp, AmpAmp, Exclaim, ExclaimEqual, Percent, Hash,
-    Less, LessEqual, LessLess, LessGreater,
-    Greater, GreaterEqual, GreaterGreater, At,
-
-    // MIPS unary expression operators such as %neg.
-    PercentCall16, PercentCall_Hi, PercentCall_Lo, PercentDtprel_Hi,
-    PercentDtprel_Lo, PercentGot, PercentGot_Disp, PercentGot_Hi, PercentGot_Lo,
-    PercentGot_Ofst, PercentGot_Page, PercentGottprel, PercentGp_Rel, PercentHi,
-    PercentHigher, PercentHighest, PercentLo, PercentNeg, PercentPcrel_Hi,
-    PercentPcrel_Lo, PercentTlsgd, PercentTlsldm, PercentTprel_Hi,
-    PercentTprel_Lo
-  };
-
-private:
-  TokenKind Kind;
-
-  /// A reference to the entire token contents; this is always a pointer into
-  /// a memory buffer owned by the source manager.
-  StringRef Str;
-
-  APInt IntVal;
-
-public:
-  AsmToken() = default;
-  AsmToken(TokenKind Kind, StringRef Str, APInt IntVal)
-      : Kind(Kind), Str(Str), IntVal(std::move(IntVal)) {}
-  AsmToken(TokenKind Kind, StringRef Str, int64_t IntVal = 0)
-      : Kind(Kind), Str(Str), IntVal(64, IntVal, true) {}
-
-  TokenKind getKind() const { return Kind; }
-  bool is(TokenKind K) const { return Kind == K; }
-  bool isNot(TokenKind K) const { return Kind != K; }
-
-  SMLoc getLoc() const;
-  SMLoc getEndLoc() const;
-  SMRange getLocRange() const;
-
-  /// Get the contents of a string token (without quotes).
-  StringRef getStringContents() const {
-    assert(Kind == String && "This token isn't a string!");
-    return Str.slice(1, Str.size() - 1);
-  }
-
-  /// Get the identifier string for the current token, which should be an
-  /// identifier or a string. This gets the portion of the string which should
-  /// be used as the identifier, e.g., it does not include the quotes on
-  /// strings.
-  StringRef getIdentifier() const {
-    if (Kind == Identifier)
-      return getString();
-    return getStringContents();
-  }
-
-  /// Get the string for the current token, this includes all characters (for
-  /// example, the quotes on strings) in the token.
-  ///
-  /// The returned StringRef points into the source manager's memory buffer, and
-  /// is safe to store across calls to Lex().
-  StringRef getString() const { return Str; }
-
-  // FIXME: Don't compute this in advance, it makes every token larger, and is
-  // also not generally what we want (it is nicer for recovery etc. to lex 123br
-  // as a single token, then diagnose as an invalid number).
-  int64_t getIntVal() const {
-    assert(Kind == Integer && "This token isn't an integer!");
-    return IntVal.getZExtValue();
-  }
-
-  APInt getAPIntVal() const {
-    assert((Kind == Integer || Kind == BigNum) &&
-           "This token isn't an integer!");
-    return IntVal;
-  }
-
-  void dump(raw_ostream &OS) const;
-  void dump() const { dump(dbgs()); }
-};
-
 /// A callback class which is notified of each comment in an assembly file as
 /// it is lexed.
 class AsmCommentConsumer {
 public:
   virtual ~AsmCommentConsumer() = default;
 
   /// Callback function for when a comment is lexed. Loc is the start of the
   /// comment text (excluding the comment-start marker). CommentText is the text