delphi - Unicode 字符串中的奇数字符

Question

在实现 MP3 ID3 v2 时我遇到了一些问题。除了这个问题之外，我的大部分工作都可以正常进行，这可能与此完全无关。无论如何，我使用下面的代码来处理检索涉及文本的标题标签的数据。

我遇到的是(我猜？)我在一些不同的字符串中遇到了 Unicode 字符。我尝试在下面对其进行转换，并且它有效。但我在字符串前面得到 $3F 作为字符，在后面得到 $3F$3F 。我可以对下面的代码做些什么来解析它们，还是我必须自己做？如果有帮助的话，这些文件是由 iTunes 编码的。

function Id3v2_string(currp: pointer; datasize: integer): string;
{ handles string processing for ID3v2 data }
  const
    IS_TEXT_UNICODE_UNICODE_MASK = $0F;
  var
    outstr: string;
    uscan: integer;
  begin
    outstr := '';
    SetLength(outstr, datasize);
    uscan := IS_TEXT_UNICODE_UNICODE_MASK;
    if IsTextUnicode(currp, datasize, @uscan) then
      outstr := WideCharToString(currp)
    else
      move(currp^, outstr[1], datasize);
    Result := outstr;
  end;

请注意，我真的对此媒体库不感兴趣，因为我想做的只是编辑 ID3 标签而不是播放文件 - 除了像这样的一些小问题外，实现已经完成。

Answer 1

取决于 ID3 的版本使用 v2 时，文本字符串前面可能会也可能不会有一个字节来告诉您字符串的实际编码。不要使用 IsTextUnicode() 来猜测编码是什么(特别是因为它可以报告 false results )。

在 ID3 v2 到 v2.3 中，没有编码字节，文本为 ISO-8859-1 或 UCS-2，以及 UCS -2 字符串始终以 BOM 开头，以便您了解字节顺序。例如:

// prior to Delphi 2009 - String is Ansi
function Id3v2_string(currp: Pointer; datasize: Integer): String; 
var
  W: WideString;
  I: Integer;
  Ch: WideChar;
begin 
  Result := '';
  if (datasize >= SizeOf(Word)) and ((PWord(currp)^ = $FEFF) or (PWord(currp)^= $FFFE)) then begin
    // UCS-2 with BOM
    W := WideCharLenToString(PWideChar(Integer(currp) + SizeOf(Word)), (datasize - SizeOf(Word)) div SizeOf(WideChar)); 
    if PWord(currp)^ = $FFFE then begin
      // BE, convert to LE
      for I := 1 to Length(W) do begin
        Ch := W[I];
        W[I] := WideChar(((Word(Ch) and $FF) shl 8) or (Word(Ch) shr 8));
      end;
    end;
  end else begin
    // ISO-8859-1
    I := MultiByteToWideChar(28591, 0, PAnsiChar(currp), datasize, nil, 0);
    if I > 0 then begin
      SetLength(W, I);
      MultiByteToWideChar(28591, 0, PAnsiChar(currp), datasize, PWideChar(W), I);
    end;
  end;
  Result := TrimRight(W);
end; 

.

// Delphi 2009+ - String is Unicode
function Id3v2_string(currp: Pointer; datasize: Integer): String; 
var
  Enc: TEncoding;

  function Convert(P: Pointer; Size: Integer): String;
  var
    Buf: TBytes;
  begin
    SetLength(Buf, Size);
    if Size > 0 then Move(P^, Buf[0], Size);
    Result := Enc.GetString(Buf);
  end;

begin 
  Result := '';
  if (datasize >= SizeOf(Word)) and ((PWord(currp)^ = $FEFF) or (PWord(currp)^ = $FFFE)) then begin
    // UCS-2 with BOM
    if PWord(currp)^ = $FFFE then begin
      // BE
      Enc := TEncoding.BigEndianUnicode;
    end else begin
      // LE
      Enc := TEncoding.Unicode;
    end;
    Result := Convert(PWord(currp)+1, datasize - SizeOf(Word));
  end else begin
    // ISO-8859-1
    Enc := TEncoding.GetEncoding(28591);
    try
      Result := Convert(currp, datasize);
    finally
      Enc.Free;
    end;
  end;
end; 

ID3 v2.4 将 UCS-2 切换为 UTF-16，并添加了对 UTF-8 和 UTF- 的支持16BE 无 BOM，例如:

// prior to Delphi 2009 - String is Ansi
function Id3v2_string(currp: Pointer; datasize: Integer; Encoding: Byte): String; 
var
  W: WideString;
  I: Integer;
  Ch: WideChar;
begin 
  Result := '';

  case Encoding of
    $00: begin
      // ISO-8859-1
      I := MultiByteToWideChar(28591, 0, PAnsiChar(currp), datasize, nil, 0);
      if I > 0 then begin
        SetLength(W, I);
        MultiByteToWideChar(28591, 0, PAnsiChar(currp), datasize, PWideChar(W), I);
      end;
    end;
    $01: begin
      // UTF-16 with BOM
      SetString(W, PWideChar(Integer(currp) + SizeOf(Word)), (datasize - SizeOf(Word)) div SizeOf(WideChar));
      if PWord(currp)^ = $FFFE then begin
        // BE, convert to LE
        for I := 1 to Length(W) do begin
          Ch := W[I];
          W[I] := WideChar(((Word(Ch) and $FF) shl 8) or (Word(Ch) shr 8));
        end;
      end;
    end;
    $02: begin
      // UTF-16BE without BOM, convert to LE
      SetString(W, PWideChar(currp), datasize div SizeOf(WideChar));
      for I := 1 to Length(W) do begin
        Ch := W[I];
        W[I] := WideChar(((Word(Ch) and $FF) shl 8) or (Word(Ch) shr 8));
      end;
    end;
    $03: begin
      // UTF-8
      I := MultiByteToWideChar(65001, 0, PAnsiChar(currp), datasize, nil, 0);
      if I > 0 then begin
        SetLength(W, I);
        MultiByteToWideChar(65001, 0, PAnsiChar(currp), datasize, PWideChar(W), I);
      end;
    end;
  end;
  Result := TrimRight(W);
end;

.

// Delphi 2009+ - String is Unicode
function Id3v2_string(currp: Pointer; datasize: Integer; Encoding: Byte): String; 
var
  Enc: TEncoding;

  function Convert(P: Pointer; Size: Integer): String;
  var
    Buf: TBytes;
  begin
    SetLength(Buf, Size);
    if Size > 0 then Move(P^, Buf[0], Size);
    Result := Enc.GetString(Buf);
  end;

begin 
  Result := '';

  case Encoding of
    $00: begin
      // ISO-8859-1
      Enc := TEncoding.GetEncoding(28591);
      try
        Result := Convert(currp, datasize);
      finally
        Enc.Free;
      end;
    end;
    $01: begin
      // UTF-16 with BOM
      if PWord(currp)^ = $FFFE then begin
        // BE
        Enc := TEncoding.BigEndianUnicode;
      end else begin
        // LE
        Enc := TEncoding.Unicode;
      end;
      Result := Convert(PWord(currp)+1, datasize - SizeOf(Word));
    end;
    $02: begin
      // UTF-16BE without BOM
      Enc := TEncoding.BigEndianUnicode;
      Result := Convert(currp, datasize);
    end;
    $03: begin
      // UTF-8
      Enc := TEncoding.UTF8;
      Result := Convert(currp, datasize);
    end;
  end;
  Result := TrimRight(Result);
end;