Convert wide_string to string (as the same byte array)

Convert wide_string to string (as the same byte array)

[Erich]

A newbie question: I need to convert a wide_string to a (platform/
endian independent) string that represents all the bytes of the
wide_string. How do you do that?

Just to make this clear, I'm not looking for a lossy conversion based
on mapping wide_characters to characters, but for a lossless
conversion.

The reason is that I have an older package that takes a key as string
and need to use it with wide_strings as keys.

Re: Convert wide_string to string (as the same byte array)

[Adam Beneschan]

On Feb 24, 2:01 pm, Erich <j...@peppermind.com> wrote:
> A newbie question: I need to convert a wide_string to a (platform/
> endian independent) string that represents all the bytes of the
> wide_string. How do you do that?

Each Wide_Character of the Wide_String will have to map to two
Characters of the String.  Also, I assume you want things to work so
that when your older package does a lexicographic comparison on
Strings, it will return the correct result based on how the original
Wide_Strings compared.

You'll probably want something like this, for each Wide_Character,
where S is the String and WS is the Wide_String:

   S (J)     := Character'Val (Wide_Character'Pos (WS (I)) / 256);
   S (J + 1) := Character'Val (Wide_Character'Pos (WS (I)) mod 256);

I'll let you fill in the rest of the details.

If you knew you were working with a big-endian machine, I think you
could just do an Unchecked_Conversion on the whole string, but you did
say endian-independent.

                       -- Adam


> Just to make this clear, I'm not looking for a lossy conversion based
> on mapping wide_characters to characters, but for a lossless
> conversion.
>
> The reason is that I have an older package that takes a key as string
> and need to use it with wide_strings as keys.

Re: Convert wide_string to string (as the same byte array)

[Erich]

> You'll probably want something like this, for each Wide_Character,
> where S is the String and WS is the Wide_String:
<snip>

Thanks a lot! Seems to work fine.

 function To_String (WS : Wide_String) return String is
      S : String (1 .. WS'Length*2);
      J : Positive;
   begin
      J := 1;
      for I in WS'Range loop
         S (J)     := Character'Val (Wide_Character'Pos (WS (I)) /
256);
         S (J + 1) := Character'Val (Wide_Character'Pos (WS (I)) mod
256);
         J := J + 2;
      end loop;
      return S;
   end To_String;


Best,

Erich

Re: Convert wide_string to string (as the same byte array)

[Randy Brukardt]

"Erich" <john@peppermind.com> wrote in message 
news:f88cc8ca-183a-40c7-a01c-2adc1137d845@b18g2000vbz.googlegroups.com...
>A newbie question: I need to convert a wide_string to a (platform/
> endian independent) string that represents all the bytes of the
> wide_string. How do you do that?

An alternative to Adam's solution would be to use the Ada2012 encoding 
functions (A.4.11), specifically Ada.Strings.UTF_Encoding.Wide_Strings, and 
use a UTF-8 encoding. That would be shorter, but not fixed length, so 
whether that would work for you depends on the API you are feeding these 
into.

                                           Randy.

Re: Convert wide_string to string (as the same byte array)

[Adam Beneschan]

On Monday, March 5, 2012 5:58:48 PM UTC-8, Randy Brukardt wrote:
> 
> An alternative to Adam's solution would be to use the Ada2012 encoding 
> functions (A.4.11), specifically Ada.Strings.UTF_Encoding.Wide_Strings, and 
> use a UTF-8 encoding. That would be shorter, but not fixed length, so 
> whether that would work for you depends on the API you are feeding these 
> into.

This may seem like a dumb question, but does that preserve order?

                        -- Adam

Re: Convert wide_string to string (as the same byte array)

[Randy Brukardt]

"Adam Beneschan" <adam@irvine.com> wrote in message 
news:5368448.8.1331049289886.JavaMail.geo-discussion-forums@pbbpr1...
> On Monday, March 5, 2012 5:58:48 PM UTC-8, Randy Brukardt wrote:
>>
>> An alternative to Adam's solution would be to use the Ada2012 encoding
>> functions (A.4.11), specifically Ada.Strings.UTF_Encoding.Wide_Strings, 
>> and
>> use a UTF-8 encoding. That would be shorter, but not fixed length, so
>> whether that would work for you depends on the API you are feeding these
>> into.
>
> This may seem like a dumb question, but does that preserve order?

My understanding was that UTF-8 was designed so that ordinary byte 
comparison operations would work "properly" on UTF-8 strings (presuming no 
"overlong encodings" are used; there is no point in such things, it's like 
including NOPs in your generated instructions). That's surely true if only 
equality is involved; I believe it is also true for ordering, but as I've 
never tried it I don't want to say for absolutely certain.

                                            Randy.