From 22973d396cebaa1a4fcecea949cd1f2ebc9e0a26 Mon Sep 17 00:00:00 2001
From: Nirav Dave <niravd@google.com>
Date: Tue, 2 Aug 2016 14:25:49 +0000
Subject: [PATCH] [MC] Fix handling of end-of-line preprocessor comments

Recommitting after fixing test.

When parsing assembly where the line comment syntax is not hash, the
lexer cannot distinguish between hash's that start a hash line comment
and one that is part of an assembly statement and must be distinguished
during parsing. Previously, this was incompletely handled by not checking
for EndOfStatement at the end of statements and interpreting hash
prefixed statements as comments.

Change EndOfStatement Parsing to check for Hash comments and reintroduce
Hash statement parsing to catch previously handled cases.

Reviewers: rnk, majnemer

Subscribers: llvm-commits

Differential Revision: https://reviews.llvm.org/D23017

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@277459 91177308-0d34-0410-b5e6-96231b3b80d8
---
 lib/MC/MCParser/AsmParser.cpp       | 34 ++++++++++++++++++++++++++++++++++
 test/MC/ARM/preserve-comments-arm.s | 10 ++++++++++
 2 files changed, 44 insertions(+)
 create mode 100644 test/MC/ARM/preserve-comments-arm.s

diff --git a/lib/MC/MCParser/AsmParser.cpp b/lib/MC/MCParser/AsmParser.cpp
index d85f1c613c7..6d3240a3e11 100644
--- a/lib/MC/MCParser/AsmParser.cpp
+++ b/lib/MC/MCParser/AsmParser.cpp
@@ -262,9 +262,23 @@ public:
     return false;
   }
 
+  bool parseEOL(const Twine &ErrMsg) {
+    if (getTok().getKind() == AsmToken::Hash) {
+      StringRef CommentStr = parseStringToEndOfStatement();
+      Lexer.Lex();
+      Lexer.UnLex(AsmToken(AsmToken::EndOfStatement, CommentStr));
+    }
+    if (getTok().getKind() != AsmToken::EndOfStatement)
+      return TokError(ErrMsg);
+    Lex();
+    return false;
+  }
+
   /// parseToken - If current token has the specified kind, eat it and
   /// return success.  Otherwise, emit the specified error and return failure.
   bool parseToken(AsmToken::TokenKind T, const Twine &ErrMsg) {
+    if (T == AsmToken::EndOfStatement)
+      return parseEOL(ErrMsg);
     if (getTok().getKind() != T)
       return TokError(ErrMsg);
     Lex();
@@ -1409,6 +1423,16 @@ bool AsmParser::parseStatement(ParseStatementInfo &Info,
     Lex();
     return false;
   }
+  if (Lexer.is(AsmToken::Hash)) {
+    // Seeing a hash here means that it was an end-of-line comment in
+    // an asm syntax where hash's are not comment and the previous
+    // statement parser did not check the end of statement. Relex as
+    // EndOfStatement.
+    StringRef CommentStr = parseStringToEndOfStatement();
+    Lexer.Lex();
+    Lexer.UnLex(AsmToken(AsmToken::EndOfStatement, CommentStr));
+    return false;
+  }
   // Statements always start with an identifier.
   AsmToken ID = getTok();
   SMLoc IDLoc = ID.getLoc();
@@ -1542,6 +1566,16 @@ bool AsmParser::parseStatement(ParseStatementInfo &Info,
     if (!Sym->isUndefined() || Sym->isVariable())
       return Error(IDLoc, "invalid symbol redefinition");
 
+    // End of Labels should be treated as end of line for lexing
+    // purposes but that information is not available to the Lexer who
+    // does not understand Labels. This may cause us to see a Hash
+    // here instead of a preprocessor line comment.
+    if (getTok().is(AsmToken::Hash)) {
+      StringRef CommentStr = parseStringToEndOfStatement();
+      Lexer.Lex();
+      Lexer.UnLex(AsmToken(AsmToken::EndOfStatement, CommentStr));
+    }
+
     // Consume any end of statement token, if present, to avoid spurious
     // AddBlankLine calls().
     if (getTok().is(AsmToken::EndOfStatement)) {
diff --git a/test/MC/ARM/preserve-comments-arm.s b/test/MC/ARM/preserve-comments-arm.s
new file mode 100644
index 00000000000..4c480096c66
--- /dev/null
+++ b/test/MC/ARM/preserve-comments-arm.s
@@ -0,0 +1,10 @@
+	@RUN: llvm-mc -preserve-comments -n -triple arm-eabi < %s > %t
+	@RUN: sed 's/#[C]omment/@Comment/g' %s > %t2
+	@RUN: diff %t %t2
+	.text
+
+	mov	r0, r0
+foo:	#Comment here
+	mov	r0, r0	@ EOL comment
+	.ident	""
+
-- 
2.11.0