🚀 go-pugleaf

Asger Jorgensen wrote:
> Hi there
> 
> Is there an API function that can translate a UTF-8 string
> to an asci string in the curent local codepage ?
> 
> This situation is that I recieve some files from 
> companies here in my country, and because of an international
> standard in these files, they are in the UTF-8 format.
> 

There is no need for API. VCL offers function Utf8ToAnsi in system unit 
(that is in BDS 2006). If you have older version of BCB than you can use 
two api's to accomplish this conversion:

MultiByteToWideChar - this one converts multibyte string to unicode
WideCharToMultiByte - this one can convert unicode to ansi

So with MultiByteToWideChar convert utf-8 to unicode and than with 
WideCharToMultiByte convert that unicode to ansi string.

Hi Darko

Thanks for Your answer

In article <46e54543@newsgroups.borland.com>, kiklop@fibertel.com.ar says...
> There is no need for API. VCL offers function Utf8ToAnsi in system unit 
> (that is in BDS 2006). If you have older version of BCB than you can use 
> two api's to accomplish this conversion:

I only have BCB5 pro
> 
> MultiByteToWideChar - this one converts multibyte string to unicode
> WideCharToMultiByte - this one can convert unicode to ansi
> 
> So with MultiByteToWideChar convert utf-8 to unicode and than with 
> WideCharToMultiByte convert that unicode to ansi string.

Is there realy no function that translate UTF-8 to good old ascii 
in one pass ?? I would very much like it to be fast.

Thanks in advance
Asger

Asger Jorgensen wrote:

> Is there realy no function that translate UTF-8 to good old ascii 
> in one pass ?? I would very much like it to be fast.

No. Windows is like that. You have to write your own function using 
these two API's.

Asger Jorgensen <Junk@Asger-P.dk> writes:

> In article <46e54543@newsgroups.borland.com>, kiklop@fibertel.com.ar says...
>>
>> There is no need for API. VCL offers function Utf8ToAnsi in system unit 
>> (that is in BDS 2006). If you have older version of BCB than you can use 
>> two api's to accomplish this conversion:
>
> I only have BCB5 pro

Have you checked whether you have Utf8ToAnsi()?

 
>> MultiByteToWideChar - this one converts multibyte string to unicode
>> WideCharToMultiByte - this one can convert unicode to ansi
>> 
>> So with MultiByteToWideChar convert utf-8 to unicode and than with 
>> WideCharToMultiByte convert that unicode to ansi string.
>
> Is there realy no function that translate UTF-8 to good old ascii 
> in one pass ?? I would very much like it to be fast.

You don't need a conversion from UTF-8 to ASCII; all ASCII characters
have the same code in UTF-8 and ASCII.


In your original post, you wrote "to an asci string in the curent
local codepage", which is a contradiction, since code pages have
characters with codes >= 128, which ASCII doesn't.

If you need support for those non-ASCII characters, Utf8ToAnsi() comes
in.

Thomas Maeder [TeamB] wrote:

> You don't need a conversion from UTF-8 to ASCII; all ASCII characters
> have the same code in UTF-8 and ASCII.

You don't need a conversion from UTF-8 to ASCII if all charactes in 
utf-8 string are ascii characters, otherwise this is not true. He said 
this in his first message:

 > This situation is that I recieve some files from
 > companies here in my country, and because of an international
 > standard in these files, they are in the UTF-8 format.

So these files may contain non ascii characters and thus conversion is 
needed.

Hi Thomas 
Thanks for Your reply

In article <m23axmtstn.fsf@glue.ch>, maeder@glue.ch says...
> 
> Have you checked whether you have Utf8ToAnsi()?

Yes I have checked and no I dont have it

> You don't need a conversion from UTF-8 to ASCII; all ASCII characters
> have the same code in UTF-8 and ASCII.

Sorry this is my mistake, I am just so use to call it ascii,
but we use codepage 850 and we have 3/6 spcial character here 
in Denmark, but when it comes to names we have a some foreign
characters also.
So I need to convert UTF-8 to a string of chars in our codepage.

Is there a way to get the Utf8ToAnsi() function, for use in my BCB5 ??

Thanks in advance 
Asger

Asger Jorgensen wrote:
> 
> Is there a way to get the Utf8ToAnsi() function, for use in my BCB5 ??

Not unless you write it yourself.

But here is a STL implementation I wrote

#include <windows.h>
#include <string>
#include <vector>
#include <iostream>

std::string utf82ansi (const std::string &str, UINT codepage = CP_ACP) {
   std::string result;
   if (!str.empty()) {
     int buffLen = ::MultiByteToWideChar( CP_UTF8,
                                          0U,
                                          str.c_str(),
                                          str.size(),
                                          NULL,
                                          0);
     if (buffLen > 0) {
         std::vector<wchar_t> buffer(buffLen);
         std::size_t res = ::MultiByteToWideChar( CP_UTF8,
                                                  0U,
                                                  str.c_str(),
                                                  str.size(),
                                                  &buffer[0],
                                                  buffer.size());
         if (res > 0) {
             std::vector<char> sbuffer (buffLen);
             res = ::WideCharToMultiByte( codepage,
                                          0U,
                                          &buffer[0],
                                          buffer.size(),
                                          &sbuffer[0],
                                          sbuffer.size(),
                                          NULL,
                                          NULL);
             if (res > 0) {
                 result.assign(sbuffer.begin(),sbuffer.end());
             }
         }
     }
   }
   return result;
}

int main() {
     std::cout << utf82ansi("<some utf-8 text>") << std::endl;
     return 0;
}

Hi Darko

Thank You for the code!

That will sure come in handy.

Kind regards
Asger

Hi Darko

Just one question


>      int buffLen = ::MultiByteToWideChar( CP_UTF8,
>                                           0U,
>                                           str.c_str(),
>                                           str.size(),
>                                           NULL,
>                                           0);

What does 0U stand for ??

Everything is working very well though, I just like to know
what is happening.

Thanks again for the code.

Kind regards
Asger

0 unsigned
similarly 0L = 0 long
HTH Pete

"Asger Jorgensen" <Junk@Asger-P.dk> wrote in message 
news:MPG.2150ddcffe0fa6339897e9@newsgroups.borland.com...
> Hi Darko
>
> Just one question
> What does 0U stand for ??

Asger Jorgensen <Junk@Asger-P.dk> writes:

> Just one question
>
> What does 0U stand for ??

It's a literal with type unsigned int and value 0. 0 without the
suffix would have type int.

Thanks for explaining

Kind regards
Asger