Home | History | Annotate | Download | only in Basic
      1 //===-- ConvertUTFWrapper.cpp - Wrap ConvertUTF.h with clang data types -----===
      2 //
      3 //                     The LLVM Compiler Infrastructure
      4 //
      5 // This file is distributed under the University of Illinois Open Source
      6 // License. See LICENSE.TXT for details.
      7 //
      8 //===----------------------------------------------------------------------===//
      9 
     10 #include "clang/Basic/ConvertUTF.h"
     11 #include "clang/Basic/LLVM.h"
     12 
     13 namespace clang {
     14 
     15 bool ConvertUTF8toWide(unsigned WideCharWidth, llvm::StringRef Source,
     16                        char *&ResultPtr, const UTF8 *&ErrorPtr) {
     17   assert(WideCharWidth == 1 || WideCharWidth == 2 || WideCharWidth == 4);
     18   ConversionResult result = conversionOK;
     19   // Copy the character span over.
     20   if (WideCharWidth == 1) {
     21     const UTF8 *Pos = reinterpret_cast<const UTF8*>(Source.begin());
     22     if (!isLegalUTF8String(&Pos, reinterpret_cast<const UTF8*>(Source.end()))) {
     23       result = sourceIllegal;
     24       ErrorPtr = Pos;
     25     } else {
     26       memcpy(ResultPtr, Source.data(), Source.size());
     27       ResultPtr += Source.size();
     28     }
     29   } else if (WideCharWidth == 2) {
     30     const UTF8 *sourceStart = (const UTF8*)Source.data();
     31     // FIXME: Make the type of the result buffer correct instead of
     32     // using reinterpret_cast.
     33     UTF16 *targetStart = reinterpret_cast<UTF16*>(ResultPtr);
     34     ConversionFlags flags = strictConversion;
     35     result = ConvertUTF8toUTF16(
     36         &sourceStart, sourceStart + Source.size(),
     37         &targetStart, targetStart + 2*Source.size(), flags);
     38     if (result == conversionOK)
     39       ResultPtr = reinterpret_cast<char*>(targetStart);
     40     else
     41       ErrorPtr = sourceStart;
     42   } else if (WideCharWidth == 4) {
     43     const UTF8 *sourceStart = (const UTF8*)Source.data();
     44     // FIXME: Make the type of the result buffer correct instead of
     45     // using reinterpret_cast.
     46     UTF32 *targetStart = reinterpret_cast<UTF32*>(ResultPtr);
     47     ConversionFlags flags = strictConversion;
     48     result = ConvertUTF8toUTF32(
     49         &sourceStart, sourceStart + Source.size(),
     50         &targetStart, targetStart + 4*Source.size(), flags);
     51     if (result == conversionOK)
     52       ResultPtr = reinterpret_cast<char*>(targetStart);
     53     else
     54       ErrorPtr = sourceStart;
     55   }
     56   assert((result != targetExhausted)
     57          && "ConvertUTF8toUTFXX exhausted target buffer");
     58   return result == conversionOK;
     59 }
     60 
     61 bool ConvertCodePointToUTF8(unsigned Source, char *&ResultPtr) {
     62   const UTF32 *SourceStart = &Source;
     63   const UTF32 *SourceEnd = SourceStart + 1;
     64   UTF8 *TargetStart = reinterpret_cast<UTF8 *>(ResultPtr);
     65   UTF8 *TargetEnd = TargetStart + 4;
     66   ConversionResult CR = ConvertUTF32toUTF8(&SourceStart, SourceEnd,
     67                                            &TargetStart, TargetEnd,
     68                                            strictConversion);
     69   if (CR != conversionOK)
     70     return false;
     71 
     72   ResultPtr = reinterpret_cast<char*>(TargetStart);
     73   return true;
     74 }
     75 
     76 } // end namespace clang
     77