🚀 go-pugleaf

> Load the contents of the file, translate it to Unicode with
> MultiByteToWideChar(CP_UTF8, ...), and convert it to your 
> required
> code page with WideCharToMultiByte.
> </quote>
>
> In my bcb5 I cannot find these functions.

Strange. They are part of the Win32 API (library Kernel32.lib). 
Are you sure you can't use them?

Steve. 

Steve Aletto wrote:

> Strange. They are part of the Win32 API (library Kernel32.lib). 
> Are you sure you can't use them?

Mmm.. F1 did not find them and searching BCB's include directory
gave nothing. But indeed they are in win32.hlp. One has to
#include <winnls>

But then my filename is more than utf-8 encoded I think.
What has been done more ? I don't know where to start now.

Hans.

> But then my filename is more than utf-8 encoded I think.
> What has been done more ? I don't know where to start now.

It's base64. Go to 
http://makcoder.sourceforge.net/demo/base64.php, past 
c2FtZW4gbWV0IEFuZHLDqSBraWprZW4gbmFhciBQYXNjYWwuanBn and decode.

Probably your next question is how to decode it by code. If I'm 
not mistaken one way could be Indy IdDecoderMIME.

Steve. 

Steve Aletto wrote:

> It's base64. Go to 
> http://makcoder.sourceforge.net/demo/base64.php, past 
> c2FtZW4gbWV0IEFuZHLDqSBraWprZW4gbmFhciBQYXNjYWwuanBn and decode.

A ha. Indeed it gives:
samen met André kijken naar Pascal.jpg

It should be:
samen met Andr� kijken naar Pascal.jpg

Nearly ok. Now I only have to find a way to make an � of that é.

That looks like a multibyte ?

> Probably your next question is how to decode it by code. 

No. No problem. My base64 decoder gives also:
samen met André kijken naar Pascal.jpg

Thanks.

Hans.

> A ha. Indeed it gives:
> samen met André kijken naar Pascal.jpg
>
> It should be:
> samen met Andr� kijken naar Pascal.jpg

Because you need to decode UTF-8 after the base64.

Put Edit1, Edit2 and IdDecoderMIME1 on a form. Past the text to 
be decoded to Edit1. Then:

    Edit2->Text = IdDecoderMIME1->DecodeToString(Edit1->Text);
    int needed = MultiByteToWideChar(CP_UTF8, 0, 
Edit2->Text.c_str(), -1, NULL, 0);
    if (needed)
    {
        WCHAR *wc = new WCHAR[needed];
        MultiByteToWideChar(CP_UTF8, 0, Edit2->Text.c_str(), -1, 
wc, needed);
        Edit2->Text = String(wc);
        delete[] wc;
    }

Please notice I wrote this code only for test. It must be 
checked _very_ deeply.

Steve. 

"Hans Galema" <notused@notused.nl> wrote in message
news:43de245b$1@newsgroups.borland.com...

> The filename decodes to: samen met Andr� kijken naar Pascal.jpg
>
> Is c2FtZW4gbWV0IEFuZHLDqSBraWprZW4gbmFhciBQYXNjYWwuanBn the piece that
represents this?

It is part of it.  The fields used to describe the encoding are separated by
'?', where the encoding begins with '=?' and ends with '?='.  So, the
encoding breaks down into three fields as follows:

    UTF-8
    B
    c2FtZW4gbWV0IEFuZHLDqSBraWprZW4gbmFhciBQYXNjYWwuanBn

The 'UTF-8' is the encoding used for the original data.  The 'B' is the
transmission encoding for the encoded data.  In this case, 'B' means that
base64 is used.  So, to get the original filename, you have to first decode
the contents of the third field with base64, and then decode the resulting
data with UTF-8 to get the original value.

> In my bcb5 I cannot find these functions.

They are in the Win32 API.


Gambit

Steve Aletto wrote:

> Because you need to decode UTF-8 after the base64.

>     int needed = MultiByteToWideChar(CP_UTF8, 0, 
> Edit2->Text.c_str(), -1, NULL, 0);
>     if (needed)
>     {
>         WCHAR *wc = new WCHAR[needed];
>         MultiByteToWideChar(CP_UTF8, 0, Edit2->Text.c_str(), -1, 
> wc, needed);
>         Edit2->Text = String(wc);
>         delete[] wc;
>     }

This works very well.

> Please notice I wrote this code only for test. It must be 
> checked _very_ deeply.

I trust it blind.

Thanks so much.

Hans.

Remy Lebeau (TeamB) wrote:

>     UTF-8
>     B
>     c2FtZW4gbWV0IEFuZHLDqSBraWprZW4gbmFhciBQYXNjYWwuanBn
> 
> The 'UTF-8' is the encoding used for the original data.  The 'B' is the
> transmission encoding for the encoded data.  In this case, 'B' means that
> base64 is used.  So, to get the original filename, you have to first decode
> the contents of the third field with base64, and then decode the resulting
> data with UTF-8 to get the original value.

B == base64.

Very good info.

Thank you.

Hans.