From f74a4587629615ffd13bd0724868f86ba8c8f27b Mon Sep 17 00:00:00 2001
From: Eli Friedman <eli.friedman@gmail.com>
Date: Tue, 1 Nov 2011 02:14:50 +0000
Subject: [PATCH] Perform proper conversion for strings encoded in the source
 file as UTF-8.  (For now, we are assuming the source character set is always
 UTF-8; this can be easily extended if necessary.)

Tests will be coming up in a subsequent commit.

Patch by Seth Cantrell.



git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@143416 91177308-0d34-0410-b5e6-96231b3b80d8
---
 include/clang/Basic/DiagnosticLexKinds.td |  2 +
 include/clang/Lex/LiteralSupport.h        |  2 +-
 lib/Lex/LiteralSupport.cpp                | 55 ++++++++++++++++++-----
 3 files changed, 47 insertions(+), 12 deletions(-)

diff --git a/include/clang/Basic/DiagnosticLexKinds.td b/include/clang/Basic/DiagnosticLexKinds.td
index 5206fd2c27..f1e8c75d6b 100644
--- a/include/clang/Basic/DiagnosticLexKinds.td
+++ b/include/clang/Basic/DiagnosticLexKinds.td
@@ -130,6 +130,8 @@ def warn_cxx98_compat_unicode_literal : Warning<
   InGroup<CXX98Compat>, DefaultIgnore;
 def err_unsupported_string_concat : Error<
   "unsupported non-standard concatenation of string literals">;
+def err_bad_string_encoding : Error<
+  "illegal sequence in string literal">;
   
 //===----------------------------------------------------------------------===//
 // PTH Diagnostics
diff --git a/include/clang/Lex/LiteralSupport.h b/include/clang/Lex/LiteralSupport.h
index b33092c753..6f378041ae 100644
--- a/include/clang/Lex/LiteralSupport.h
+++ b/include/clang/Lex/LiteralSupport.h
@@ -197,7 +197,7 @@ public:
 
 private:
   void init(const Token *StringToks, unsigned NumStringToks);
-  void CopyStringFragment(StringRef Fragment);
+  bool CopyStringFragment(StringRef Fragment);
 };
 
 }  // end namespace clang
diff --git a/lib/Lex/LiteralSupport.cpp b/lib/Lex/LiteralSupport.cpp
index 70183fd1a0..b107531e14 100644
--- a/lib/Lex/LiteralSupport.cpp
+++ b/lib/Lex/LiteralSupport.cpp
@@ -16,6 +16,7 @@
 #include "clang/Lex/Preprocessor.h"
 #include "clang/Lex/LexDiagnostic.h"
 #include "clang/Basic/TargetInfo.h"
+#include "clang/Basic/ConvertUTF.h"
 #include "llvm/ADT/StringExtras.h"
 #include "llvm/Support/ErrorHandling.h"
 using namespace clang;
@@ -1033,7 +1034,14 @@ void StringLiteralParser::init(const Token *StringToks, unsigned NumStringToks){
         ThisTokEnd -= (ThisTokBuf - Prefix);
 
       // Copy the string over
-      CopyStringFragment(StringRef(ThisTokBuf, ThisTokEnd - ThisTokBuf));
+      if (CopyStringFragment(StringRef(ThisTokBuf,ThisTokEnd-ThisTokBuf)))
+      {
+        if (Diags)
+          Diags->Report(FullSourceLoc(StringToks[i].getLocation(), SM),
+                        diag::err_bad_string_encoding);
+        hadError = true;
+      }
+
     } else {
       assert(ThisTokBuf[0] == '"' && "Expected quote, lexer broken?");
       ++ThisTokBuf; // skip "
@@ -1060,7 +1068,13 @@ void StringLiteralParser::init(const Token *StringToks, unsigned NumStringToks){
           } while (ThisTokBuf != ThisTokEnd && ThisTokBuf[0] != '\\');
 
           // Copy the character span over.
-          CopyStringFragment(StringRef(InStart, ThisTokBuf - InStart));
+          if (CopyStringFragment(StringRef(InStart,ThisTokBuf-InStart)))
+          {
+            if (Diags)
+              Diags->Report(FullSourceLoc(StringToks[i].getLocation(), SM),
+                            diag::err_bad_string_encoding);
+            hadError = true;
+          }
           continue;
         }
         // Is this a Universal Character Name escape?
@@ -1116,20 +1130,39 @@ void StringLiteralParser::init(const Token *StringToks, unsigned NumStringToks){
 
 /// copyStringFragment - This function copies from Start to End into ResultPtr.
 /// Performs widening for multi-byte characters.
-void StringLiteralParser::CopyStringFragment(StringRef Fragment) {
+bool StringLiteralParser::CopyStringFragment(StringRef Fragment) {
+  assert(CharByteWidth==1 || CharByteWidth==2 || CharByteWidth==4);
+  ConversionResult result = conversionOK;
   // Copy the character span over.
   if (CharByteWidth == 1) {
     memcpy(ResultPtr, Fragment.data(), Fragment.size());
     ResultPtr += Fragment.size();
-  } else {
-    // Note: our internal rep of wide char tokens is always little-endian.
-    for (StringRef::iterator I=Fragment.begin(), E=Fragment.end(); I!=E; ++I) {
-      *ResultPtr++ = *I;
-      // Add zeros at the end.
-      for (unsigned i = 1, e = CharByteWidth; i != e; ++i)
-        *ResultPtr++ = 0;
-    }
+  } else if (CharByteWidth == 2) {
+    UTF8 const *sourceStart = (UTF8 const *)Fragment.data();
+    // FIXME: Make the type of the result buffer correct instead of
+    // using reinterpret_cast.
+    UTF16 *targetStart = reinterpret_cast<UTF16*>(ResultPtr);
+    ConversionFlags flags = lenientConversion;
+    result = ConvertUTF8toUTF16(
+	    &sourceStart,sourceStart + Fragment.size(),
+        &targetStart,targetStart + 2*Fragment.size(),flags);
+    if (result==conversionOK)
+      ResultPtr = reinterpret_cast<char*>(targetStart);
+  } else if (CharByteWidth == 4) {
+    UTF8 const *sourceStart = (UTF8 const *)Fragment.data();
+    // FIXME: Make the type of the result buffer correct instead of
+    // using reinterpret_cast.
+    UTF32 *targetStart = reinterpret_cast<UTF32*>(ResultPtr);
+    ConversionFlags flags = lenientConversion;
+    result = ConvertUTF8toUTF32(
+        &sourceStart,sourceStart + Fragment.size(),
+        &targetStart,targetStart + 4*Fragment.size(),flags);
+    if (result==conversionOK)
+      ResultPtr = reinterpret_cast<char*>(targetStart);
   }
+  assert((result != targetExhausted)
+         && "ConvertUTF8toUTFXX exhausted target buffer");
+  return result != conversionOK;
 }
 
 
-- 
2.40.0