From: John McCall <rjmccall@apple.com>
Date: Thu, 15 Jul 2010 21:53:03 +0000 (+0000)
Subject: After some discussion, the ABI list settled on a much more sensible mangling
X-Git-Url: https://granicus.if.org/sourcecode?a=commitdiff_plain;h=1658c39227fe0dc53e6365799a579eeb22b478d2;p=clang

After some discussion, the ABI list settled on a much more sensible mangling
for string literals.



git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@108464 91177308-0d34-0410-b5e6-96231b3b80d8
---

diff --git a/lib/CodeGen/Mangle.cpp b/lib/CodeGen/Mangle.cpp
index 30ee541c00..58b0246e36 100644
--- a/lib/CodeGen/Mangle.cpp
+++ b/lib/CodeGen/Mangle.cpp
@@ -1974,13 +1974,10 @@ void CXXNameMangler::mangleExpression(const Expr *E) {
   }
 
   case Expr::StringLiteralClass: {
-    // Proposal from David Vandervoorde, 2010.06.30.
-    // I've sent a comment off asking whether this needs to also
-    // represent the length of the string.
+    // Revised proposal from David Vandervoorde, 2010.07.15.
     Out << 'L';
-    const ConstantArrayType *T = cast<ConstantArrayType>(E->getType());
-    QualType CharTy = T->getElementType().getUnqualifiedType();
-    mangleType(CharTy);
+    assert(isa<ConstantArrayType>(E->getType()));
+    mangleType(E->getType());
     Out << 'E';
     break;
   }