delphi XE7 HttpEncode 编码问题

近期在做网址编码相关的工作，发现在用 XE5 编译的时候，一切正常，拿到 XE7下就结果错误了。百度了下，谷歌了下，有人提出，但是，我没有找到答案，也许都没有碰到这个问题，也许都己经自己默默的解决了，在此小记一下，方便后人，也方便自己查寻。

例子：原字符 "过年"

httpencode('过年') 结果：

XE5为 %B9%FD%C4%EA

XE7 调用相当函数结果 %E8%BF%87%E5%B9%B4

百思不得其解啊，折腾了很长时间，后来终于想到是不是此函数官方更新修改了，（没办法，人比较笨）

于是查看源码：

XE5的 web.httpapp 中（分string 和 ansistring，我用ansistring 版本得到期望的结果）：

 function HTTPEncode(const AStr: string): string;

 // The NoConversion set contains characters as specificed in RFC 1738 and

 // should not be modified unless the standard changes.

 const

   NoConversion = [Ord('A')..Ord('Z'), Ord('a')..Ord('z'), Ord('*'), Ord('@'),

                   Ord('.'), Ord('_'), Ord('-'), Ord('')..Ord(''), Ord('$'),

                   Ord('!'), Ord(''''), Ord('('), Ord(')')];

 var

   Sp, Rp: PChar;

 begin

   SetLength(Result, Length(AStr) * );

   Sp := PChar(AStr);

   Rp := PChar(Result);

   while Sp^ <> # do

   begin

     if Ord(Sp^) in NoConversion then

       Rp^ := Sp^

     else

       if Sp^ = ' ' then

         Rp^ := '+'

       else

       begin

         FormatBuf(Rp, , string('%%%.2x'), , [Ord(Sp^)]);

         Inc(Rp,);

       end;

     Inc(Rp);

     Inc(Sp);

   end;

   SetLength(Result, Rp - PChar(Result));

 end;

 function HTTPDecode(const AStr: string): string;

 var

   Sp, Rp, Cp: PChar;

   S: string;

 begin

   SetLength(Result, Length(AStr));

   Sp := PChar(AStr);

   Rp := PChar(Result);

   Cp := Sp;

   try

     while Sp^ <> # do

     begin

       case Sp^ of

         '+': Rp^ := ' ';

         '%': begin

                // Look for an escaped % (%%) or %<hex> encoded character

                Inc(Sp);

                if Sp^ = '%' then

                  Rp^ := '%'

                else

                begin

                  Cp := Sp;

                  Inc(Sp);

                  if (Cp^ <> #) and (Sp^ <> #) then

                  begin

                    S := Char('$') + Cp^ + Sp^;

                    Rp^ := Char(StrToInt(string(S)));

                  end

                  else

                    raise EWebBrokerException.CreateFmt(sErrorDecodingURLText, [Cp - PChar(AStr)]);

                end;

              end;

       else

         Rp^ := Sp^;

       end;

       Inc(Rp);

       Inc(Sp);

     end;

   except

     on E:EConvertError do

       raise EConvertError.CreateFmt(sInvalidURLEncodedChar,

         [Char('%') + Cp^ + Sp^, Cp - PChar(AStr)])

   end;

   SetLength(Result, Rp - PChar(Result));

 end;

 function HTTPEncode(const AStr: AnsiString): AnsiString;

 // The NoConversion set contains characters as specificed in RFC 1738 and

 // should not be modified unless the standard changes.

 const

   NoConversion = ['A'..'Z','a'..'z','*','@','.','_','-',

                   ''..'','$','!','''','(',')'];

 var

   Sp, Rp: PAnsiChar;

 begin

   SetLength(Result, Length(AStr) * );

   Sp := PAnsiChar(AStr);

   Rp := PAnsiChar(Result);

   while Sp^ <> # do

   begin

     if Sp^ in NoConversion then

       Rp^ := Sp^

     else

       if Sp^ = ' ' then

         Rp^ := '+'

       else

       begin

         System.AnsiStrings.FormatBuf(Rp^, , AnsiString('%%%.2x'), , [Ord(Sp^)]);

         Inc(Rp,);

       end;

     Inc(Rp);

     Inc(Sp);

   end;

   SetLength(Result, Rp - PAnsiChar(Result));

 end;

在XE7中

web.httpapp:

function HTTPEncode(const AStr: string): string;

begin

  Result := TNetEncoding.URL.Encode(AStr);

end;

查看 system.netencoding

找到

function TNetEncoding.DoEncode(const Input: array of Byte): TBytes;

begin

  Result := TEncoding.UTF8.GetBytes(DoEncode(TEncoding.UTF8.GetString(@Input[])));

end;

查看类定义：  
TURLEncoding = class(TNetEncoding)

  protected

    function DoDecode(const Input: string): string; overload; override;

    function DoEncode(const Input: string): string; overload; override;

  end;

查看函数代码：

function TURLEncoding.DoEncode(const Input: string): string;

// The NoConversion set contains characters as specificed in RFC 1738 and

// should not be modified unless the standard changes.

const

  NoConversion = [Ord('A')..Ord('Z'), Ord('a')..Ord('z'), Ord('*'), Ord('@'),

                  Ord('.'), Ord('_'), Ord('-'), Ord('')..Ord(''), Ord('$'),

                  Ord('!'), Ord(''''), Ord('('), Ord(')')];

  procedure AppendByte(B: Byte; var Buffer: PChar);

  const

    Hex = '0123456789ABCDEF';

  begin

    Buffer[] := '%';

    Buffer[] := Hex[B shr  + Low(string)];

    Buffer[] := Hex[B and $F + Low(string)];

    Inc(Buffer, );

  end;

var

  Sp, Rp: PChar;

  MultibyteChar: TBytes;

  I, ByteCount: Integer;

begin

  // Characters that require more than 1 byte are translated as "percent-encoded byte"

  // which will be encoded with 3 chars per byte -> %XX

  // Example: ?character

  //   Multibyte representation: C391 (2 bytes)

  //   URL encode representation: %C3%91

  //

  // So the worst case is 4 bytes(max) per Char, and 3 characters to represent each byte

  SetLength(Result, Length(Input) *  * );

  Sp := PChar(Input);

  Rp := PChar(Result);

  SetLength(MultibyteChar, );

  while Sp^ <> # do

  begin

    if Ord(Sp^) in NoConversion then

    begin

      Rp^ := Sp^;

      Inc(Rp)

    end

    else if Sp^ = ' ' then

    begin

      Rp^ := '+';

      Inc(Rp)

    end

    else

    begin

      if (Ord(Sp^) < ) then

        // Single byte char

        AppendByte(Ord(Sp^), Rp)

      else

      begin

        // Multi byte char

        ByteCount := TEncoding.UTF8.GetBytes([Sp^], , , MultibyteChar, );

        for I :=  to ByteCount -  do

          AppendByte(MultibyteChar[I], Rp);

      end

    end;

    Inc(Sp);

  end;

  SetLength(Result, Rp - PChar(Result));

end;

似乎有点不同。

目前是我自己建立一个函数，复制XE5版本的代码放在XE7里面调用，得到希望的结果的。

代码：

function MyHTTPEncode(const AStr: AnsiString): AnsiString;

// The NoConversion set contains characters as specificed in RFC 1738 and

// should not be modified unless the standard changes.

const

  NoConversion = ['A'..'Z','a'..'z','*','@','.','_','-',

                  ''..'','$','!','''','(',')'];

var

  Sp, Rp: PAnsiChar;

begin

  SetLength(Result, Length(AStr) * );

  Sp := PAnsiChar(AStr);

  Rp := PAnsiChar(Result);

  while Sp^ <> # do

  begin

    if Sp^ in NoConversion then

      Rp^ := Sp^

    else

      if Sp^ = ' ' then

        Rp^ := '+'

      else

      begin

        System.AnsiStrings.FormatBuf(Rp^, , AnsiString('%%%.2x'), , [Ord(Sp^)]);

        Inc(Rp,);

      end;

    Inc(Rp);

    Inc(Sp);

  end;

  SetLength(Result, Rp - PAnsiChar(Result));

end;

巴特西

delphi XE7 HttpEncode 编码问题

最新文章

热门文章