diff options
| author | Reid Kleckner <rnk@google.com> | 2017-10-16 22:20:03 +0000 |
|---|---|---|
| committer | Reid Kleckner <rnk@google.com> | 2017-10-16 22:20:03 +0000 |
| commit | 983f9af008ef293025c0e0198f0726ee64ffc793 (patch) | |
| tree | ad5fe4cf17566f213b2a67d8b0631cb790b95372 | |
| parent | 7b58f472aaaca5e1195d0c6a47e6a2657fb74ff2 (diff) | |
| download | llvm-983f9af008ef293025c0e0198f0726ee64ffc793.tar.gz | |
[MC] Lex CRLF as one token
This will prevent doubling of line endings when parsing assembly and
emitting assembly.
Otherwise we'd parse the directive, consume the end of statement, hit
the next end of statement, and emit a fresh newline.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@315943 91177308-0d34-0410-b5e6-96231b3b80d8
| -rw-r--r-- | lib/MC/MCParser/AsmLexer.cpp | 10 | ||||
| -rw-r--r-- | test/MC/X86/crlf.test | 5 |
2 files changed, 14 insertions, 1 deletions
diff --git a/lib/MC/MCParser/AsmLexer.cpp b/lib/MC/MCParser/AsmLexer.cpp index e9123b9d714..b83b6d3dcf6 100644 --- a/lib/MC/MCParser/AsmLexer.cpp +++ b/lib/MC/MCParser/AsmLexer.cpp @@ -606,8 +606,16 @@ AsmToken AsmLexer::LexToken() { return LexToken(); // Ignore whitespace. else return AsmToken(AsmToken::Space, StringRef(TokStart, CurPtr - TokStart)); + case '\r': { + IsAtStartOfLine = true; + IsAtStartOfStatement = true; + // If this is a CR followed by LF, treat that as one token. + if (CurPtr != CurBuf.end() && *CurPtr == '\n') + ++CurPtr; + return AsmToken(AsmToken::EndOfStatement, + StringRef(TokStart, CurPtr - TokStart)); + } case '\n': - case '\r': IsAtStartOfLine = true; IsAtStartOfStatement = true; return AsmToken(AsmToken::EndOfStatement, StringRef(TokStart, 1)); diff --git a/test/MC/X86/crlf.test b/test/MC/X86/crlf.test new file mode 100644 index 00000000000..66014456138 --- /dev/null +++ b/test/MC/X86/crlf.test @@ -0,0 +1,5 @@ +RUN: printf '\r\n\r\n' | llvm-mc -as-lex | FileCheck %s +There should only be two end of statements. +CHECK: EndOfStatement +CHECK: EndOfStatement +CHECK-NOT: EndOfStatement |