/*
 * utf.h
 * 
 * This file is a part of NSIS.
 * 
 * Copyright (C) 2011-2020 Anders Kjersem
 * 
 * Licensed under the zlib/libpng license (the "License");
 * you may not use this file except in compliance with the License.
 * 
 * Licence details can be found in the file COPYING.
 * 
 * This software is provided 'as-is', without any express or implied
 * warranty.
 *
 */

#ifndef NSIS_UTF_H
#define NSIS_UTF_H

#include "Platform.h"

const WORD UNICODE_REPLACEMENT_CHARACTER = 0xfffd;

#define TSTR_INPUTCHARSET _T("ACP|OEM|CP#|UTF8|UTF16<LE|BE>")
#define TSTR_OUTPUTCHARSET _T("ACP|OEM|CP#|UTF8[SIG]|UTF16<LE|BE>[BOM]")

template<typename T> T S7ChLwr(T c) { return c>='A' && c<='Z' ? (T)(c|32) : c; }
template<typename T> T S7ChUpr(T c) { return c>='a' && c<='z' ? (T)(c-'a'+'A') : c; }
template<typename T> bool S7IsChEqualI(char ch,T cmp)
{
  return S7ChLwr((T)ch) == S7ChLwr(cmp);
}

inline bool IsValidUnicodeCodePoint(UINT32 c,bool StrictUTF32=false)
{
  // Unicode 6.1: 16.7 Noncharacters
  if ((c&0xfffe) == 0xfffe) return false; // ..FFFE & ..FFFF is reserved in each plane
  if (c >= 0xfdd0 && c <= 0xfdef) return false; // Reserved in BMP
  if (StrictUTF32 && c > 0x10ffff) return false;
  return true;
}
inline bool IsLeadSurrogateUTF16(unsigned short c) { return c >= 0xd800 && c <= 0xdbff; }
inline bool IsTrailSurrogateUTF16(unsigned short c) { return c >= 0xdc00 && c <= 0xdfff; }
inline UINT32 CodePointFromUTF16SurrogatePair(unsigned short lea,unsigned short tra)
{
  const UINT32 surrogate_offset = 0x10000 - (0xD800 << 10) - 0xDC00;
  return ((UINT32)lea << 10) + tra + surrogate_offset;
}

inline bool UTF8_GetTrailCount(unsigned char chFirst, unsigned char &cb)
{
  // This function should only be used to get a rough idea of how large the encoded
  // codepoint is, just because it returns true does not mean that it is valid UTF-8!
  cb = 0;
  if (0xC0 == (0xC0 & chFirst))
  {
    ++cb;
    if (0xE0 == (0xE0 & chFirst))
    {
      ++cb;
      if (0xF0 == (0xF0 & chFirst))
      {
        ++cb;
        if (0xF8 == (0xF8 & chFirst))
        {
          ++cb;
          if (0xFC == (0xFE & chFirst)) ++cb; else return 0xFC >= (0xFE & chFirst);
        }
      }
    }
    return true;
  }
  return (signed char)chFirst >= 0;
}

inline UINT InlineStrLenUTF16(const void*str)
{
  unsigned short *p = (unsigned short *) str;
  UINT cch = 0;
  for(;p[cch];) ++cch;
  return cch;
}

#ifdef MAKENSIS
#include <stdlib.h>
#include <stdio.h>
#include "tstring.h"
#include "util.h" // iconvdescriptor & my_fopen
#ifdef _WIN32
#include <io.h> // For _setmode
#include <fcntl.h> // For _O_BINARY
#endif

FILE* my_fopen(const TCHAR *path, const char *mode); // from util.h

void UTF16InplaceEndianSwap(void*Buffer, UINT cch);
UINT StrLenUTF16(const void*str);
bool StrSetUTF16LE(tstring&dest, const void*src);

UINT WCFromCodePoint(wchar_t*Dest,UINT cchDest,UINT32 CodPt);
#define DWCFBF_ALLOWOPTIMIZEDRETURN 0x80000000 // DupWCFromBytes can return input Buffer
wchar_t* DupWCFromBytes(void*Buffer,UINT cbBuffer,UINT32 SrcCP);
UINT DetectUTFBOM(void*Buffer,UINT cb);
UINT DetectUTFBOM(FILE*strm);
WORD GetEncodingFromString(const TCHAR*s, bool&BOM);
WORD GetEncodingFromString(const TCHAR*s);

class CharEncConv {
  char *m_Result;
  WORD m_TE, m_FE;
#ifdef _WIN32
  bool m_AllowOptimizedReturn; // Can Convert() return Src buffer?
  bool m_OptimizedReturn;
#else
  iconvdescriptor m_iconvd;
#endif
protected:
  size_t GuessOutputSize(size_t cbConverted);
  static bool IsWE(WORD Encoding) { return (WORD)-1 == Encoding; }
  static bool IsWE(UINT32 Encoding) { return (UINT32)-1 == Encoding; }
public:
  CharEncConv() : m_Result(0) { SetAllowOptimizedReturn(false); }
  ~CharEncConv() { Close(); }
  void Close()
  {
#ifdef _WIN32
    if (!m_OptimizedReturn)
#endif
      free(m_Result);
    m_Result = 0;
#ifndef _WIN32
    m_iconvd.Close();
#endif
  }
  void* Detach() { void *p = m_Result; m_Result = 0; return p; }
  bool Initialize(UINT32 ToEnc, UINT32 FromEnc);
  void* Convert(const void*Src, size_t cbSrc = -1, size_t*cbOut = 0);
#ifdef _WIN32
  void SetAllowOptimizedReturn(bool val = true) { m_AllowOptimizedReturn = val; }
#else
  void SetAllowOptimizedReturn(bool val = false) {}
#endif
  static BOOL IsValidCodePage(UINT cp);
};

class WCToUTF16LEHlpr {
  unsigned short* m_s;
public:
  WCToUTF16LEHlpr() : m_s(0) {}

  bool Create(const TCHAR*in, unsigned int codepage = CP_ACP)
#if !defined(_WIN32) || !defined(_UNICODE)
  ;
#else
  {
    m_s = (unsigned short*) in;
    return true;
  }
#endif
  void Destroy()
  {
#if !defined(_WIN32) || !defined(_UNICODE)
    free(m_s);
#endif
  }
  const unsigned short* Get() const { return m_s; }
  UINT GetLen() const { return StrLenUTF16(m_s); }
  UINT GetSize() const { return (GetLen()+1) * 2; }
  unsigned short* Detach() { unsigned short *r = m_s; m_s = 0; return r; }
  void CopyTo(unsigned short*Dest) const { memcpy(Dest, Get(), GetSize()); }
};

class NStreamEncoding {
protected:
  WORD m_cp;
public:
  enum {
    ACP = CP_ACP,
    OEMCP = 1,
    UTF16LE = 1200,
    UTF16BE = 1201,
    UTF32LE = 12000,
    UTF32BE = 12001,
    UTF8 = CP_UTF8,
    UNKNOWN = (0xffff-0),
    AUTO = (0xffff-1),
    CPCOUNT = (0xffff-2) // Must be less than our other magic numbers
  };

  NStreamEncoding() { Reset(); }
  NStreamEncoding(WORD cp) { Reset(), SetCodepage(cp); }
  WORD GetCodepage() const { return m_cp; }
  void SetCodepage(WORD cp) { m_cp = cp; }
  void SafeSetCodepage(WORD cp)
  {
    if (NStreamEncoding::AUTO==cp) cp = GetPlatformDefaultCodepage();
    SetCodepage(cp);
  }
  void Reset() { SetCodepage(GetPlatformDefaultCodepage()); }
  WORD GetPlatformDefaultCodepage() const
  {
#ifdef _WIN32
    return ACP;
#else
    return UTF8;
#endif
  }
  bool IsUTF8() const { return UTF8==GetCodepage(); }
  bool IsUTF16() const { return (UTF16LE|1)==(GetCodepage()|1); }
  bool IsUTF16LE() const { return UTF16LE==GetCodepage(); }
  bool IsUTF16BE() const { return UTF16BE==GetCodepage(); }
  bool IsUnicode() const { return IsUnicodeCodepage(GetCodepage()); }
  void GetCPDisplayName(TCHAR*Buf) { GetCPDisplayName(m_cp, Buf); }

  static UINT GetCodeUnitSize(WORD cp)
  {
    if ((UTF16LE|1)==(cp|1)) return 2;
    if ((UTF32LE|1)==(cp|1)) return 4;
    return 1;
  }
  static bool IsUnicodeCodepage(WORD cp)
  {
    return UTF8==cp || (UTF16LE|1)==(cp|1) || (UTF32LE|1)==(cp|1); 
  }
  static void GetCPDisplayName(WORD CP, TCHAR*Buf);
};

class NStream {
public:
  enum {
    OK = 0,
    ERR_BUFFEROVERFLOW,
    ERR_IOERROR,
    ERR_INVALIDENCODING,
    ERR_UNSUPPORTEDENCODING,
  };
  static bool IsNewline(wchar_t chW, bool HandleUnicodeNL)
  {
    if (L'\n'==chW || L'\r'==chW) return true;
    if (HandleUnicodeNL)
    {
      // www.unicode.org/standard/reports/tr13/tr13-5.html#UNICODE NEWLINE GUIDELINES
      if (L'\f'==chW) return true; // FF/Form Feed
      if (L'\v'==chW) return true; // VT/Vertical Tab
      // NOTIMPLEMENTED: NEL/Next Line/U+0085
      // NOTIMPLEMENTED: LS/Line Separator/U+2028
      // NOTIMPLEMENTED: PS/Paragraph Separator/U+2029
    }
    return false;
  }
#ifdef _WIN32
  static bool SetBinaryMode(int fd) { return -1 != _setmode(fd, _O_BINARY); }
  static bool SetBinaryMode(FILE*f) { return SetBinaryMode(_fileno(f)); }
#else
  static bool SetBinaryMode(int fd) { return true; }
  static bool SetBinaryMode(FILE*f) { return true; }
#endif
};

class NBaseStream {
protected:
  FILE* m_hFile;
  NStreamEncoding m_Enc;

public:
  NBaseStream(FILE *hFile = 0) : m_hFile(hFile) {}
  ~NBaseStream() { Close(); }

  FILE* GetHandle() const { return m_hFile; }
  NStreamEncoding& StreamEncoding() { return m_Enc; }
  bool IsEOF() const { return feof(m_hFile) != 0; }
  bool IsError() const { return ferror(m_hFile) != 0; }
  bool IsUnicode() const { return m_Enc.IsUnicode(); }

  void Close()
  {
    FILE*hF = Detach();
    if (hF) fclose(hF);
  }
  
  bool OpenFileForReading(const TCHAR* Path, WORD enc = NStreamEncoding::AUTO)
  {
    return Attach(my_fopen(Path, "rb"), enc);
  }
  bool OpenFileForReading(const TCHAR* Path, NStreamEncoding&Enc)
  {
    return OpenFileForReading(Path, Enc.GetCodepage());
  }

  FILE* Detach() 
  {
    FILE *hFile = m_hFile;
    m_hFile = 0;
    return hFile;
  }
  bool Attach(FILE*hFile, WORD enc, bool Seek = true);

  UINT ReadOctets(void*Buffer, UINT cbBuf)
  {
    size_t cb = fread(Buffer, 1, cbBuf, m_hFile);
    return (UINT) cb;
  }
  bool ReadOctets(void*Buffer, UINT*pcbBuf)
  {
    UINT cbReq = *pcbBuf, cb = ReadOctets(Buffer, cbReq);
    *pcbBuf = cb;
    return cbReq == cb;
  }
  bool ReadOctet(void*Buffer) { return 1 == ReadOctets(Buffer, 1); }
  bool ReadInt16(void*Buffer) { return 2 == ReadOctets(Buffer, 2); }
};

class NIStream : public NBaseStream {
public:
  bool OpenStdIn(WORD enc = NStreamEncoding::AUTO)
  {
    return Attach(stdin, enc, false);
  }
  bool OpenStdIn(NStreamEncoding&Enc)
  {
    return OpenStdIn(Enc.GetCodepage());
  }
};

class NOStream : public NBaseStream {
public:
  NOStream(FILE *hFile = 0) : NBaseStream(hFile) {}

  bool CreateFileForWriting(const TCHAR* Path, WORD enc = NStreamEncoding::AUTO)
  {
    return Attach(my_fopen(Path, "w+b"), enc);
  }
  bool CreateFileForWriting(const TCHAR* Path, NStreamEncoding&Enc)
  {
    return CreateFileForWriting(Path, Enc.GetCodepage());
  }
  bool CreateFileForAppending(const TCHAR* Path, WORD enc = NStreamEncoding::AUTO)
  {
    return Attach(my_fopen(Path, "a+b"), enc);
  }
  bool CreateFileForAppending(const TCHAR* Path, NStreamEncoding&Enc)
  {
    return CreateFileForAppending(Path, Enc.GetCodepage());
  }

  bool WriteOctets(void*Buffer, size_t cbBuf)
  {
    return cbBuf == fwrite(Buffer, 1, cbBuf, m_hFile);
  }
  bool WriteBOM(NStreamEncoding&Enc)
  {
    static const unsigned char u8b[] = {0xEF,0xBB,0xBF};
    static const unsigned char u16lb[] = {0xFF,0xFE}, u16bb[] = {0xFE,0xFF};
    switch(Enc.GetCodepage())
    {
    case NStreamEncoding::UTF8: return WriteOctets((void*) u8b, sizeof(u8b));
    case NStreamEncoding::UTF16LE: return WriteOctets((void*) u16lb, sizeof(u16lb));
    case NStreamEncoding::UTF16BE: return WriteOctets((void*) u16bb, sizeof(u16bb));
    }
    return false;
  }
  bool WriteString(const wchar_t*Str, size_t cch = -1);
  bool WritePlatformNLString(const wchar_t*Str, size_t cch = -1);
};

class NStreamLineReader {
protected:
  NIStream &m_Strm;
  wchar_t m_PrevNL;

public:
  NStreamLineReader(NIStream &Strm) : m_Strm(Strm), m_PrevNL(0) {}

  NIStream& GetStream() { return m_Strm; }
  NStreamEncoding& StreamEncoding() { return m_Strm.StreamEncoding(); }
  bool IsEOF() const { return m_Strm.IsEOF(); }
  bool IsUnicode() const { return m_Strm.IsUnicode(); }

  UINT ReadLine(wchar_t*Buffer, UINT cchBuf);
  tstring GetErrorMessage(UINT Error, const TCHAR*Filename=0, UINT Line=0);

protected:
  bool CompleteLine(wchar_t*&BufWC, UINT cchWC, UINT&cchRemain, bool HandleUnicodeNL)
  {
    const wchar_t chW = *BufWC;
    BufWC += cchWC, cchRemain -= cchWC;
    if (0 == --cchWC) // We only care about code points that fit in a single wchar_t
      return NStream::IsNewline(chW, HandleUnicodeNL);
    return false;
  }
};

#endif // MAKENSIS
#endif // NSIS_UTF_H