[V4MD][SERVER] Cross platform text

[Robert Brenstein]

>  > Ha! We are back to the issue of needing both Language AND Character
>>  Encoding params for each text-containing field. Going to UTF16 will
>>  solve only the storage issue. You said earlier, Ruslan, that kernel
>>  will automatically convert from/to specified encoding on input and
>>  output. Or have you changed your mind and will require explicit
>>  conversions? With server things are getting more complicated since
>>  the same db may be serving Mac and Win clients at the same time. May
>>  be it is time to finally have someone on your team working on this
>>  issue :)
>
>Igor work on this issue.
>
>We already have include ICU library into 2.0 engine,
>And change a lots of code to support internally UTF16.
>
>ICU self can do conversions.
>
>So all we will have to do, (in the worse case),
>Just add this additional flags/encodings to
>     maybe server protocol,
>     fields
>     ...
>
>And I think right now we can fix mentioned problem in Vserver alpha,
>So it will work not worse of LOCAL SDKs.
>
>--
>Best regards,
>Ruslan Zasukhin      [ I feel the need...the need for speed ]

Sounds promising. Just please avoid the mistake you did in 1.x with 
assuming what encoding is good for the user. I mean assuming the 
encoding appropriate for client platform can be a default but there 
must be means to specify otherwise.

Robert