V4RB2 // storing encoding

[erne]

On 22 ott 2004, at 21:01, Ruslan Zasukhin wrote:

> On 10/22/04 8:16 PM, "erne" <ernestogiannotta at tiscalinet.it> wrote:
>
>> Well, I can live with this behaviour
>> but I still don't get sense of it
>>
>> what's the point in defining a storing encoding
>> if I don't I/O data in that format by default?
>
> Hmm.
>
> Storage Enoding -- you say HOW you want keep strings on disk
>     Japan guys will prefer UTF16
>     roman language group will prefer UTF8
>     Russian will prefer Cyrillic Mac or Win.
>

sure, that is why I expect to deal with that data in my preferred 
encoding
I really can appreciate that Vale will convert any string fed to her 
into preferred encoding, this is smart
But I'd expect to get back my data in that encoding and not have to 
convert them from UTF8

>
> But this is not related in any case to IO encoding,
> You can give to V4RB string in UTF8 or UTF16 and Valentina do next
>
>     INSIDE of KERNEL it will convert it to UTF16.
>
>     ALWAYS.
>     ALWAYS
>     ALWAYS.
>
> Because inside of kernel all Valentina algorithms use UTF16.
>
> When Valentina ed tore something to disk it convert from UTF16 to
> StorageEncoding. When it read from disk then back to UTF16.
>
>
> You see?
>
>
>     REALBasic  <------->   engine  <--------> disk
>                    IO               Storage
>

yep! I can understand this
btw have you figures of what the overhead of this conversion?
i.e. will we gain significant advantage in using UTF16 for storage?


Cool Runnings,
Erne.
|er| musical box
|ne| a media store