Adriano dos Santos Fernandes wrote:

>
> For *all* current FB users (who that don't depent on broken UNICODE_FSS)
> will cause these problems:
>
> 1) Slow (because conversions will always be made).

Yes, there will be an extra processing step, but not a very expensive
one, and one that will be performed at the client side before or after a
network transfer.

> 2) Break all UDFs that use strings.

We can do conversions in and out for legacy UDF's.

> 3) Break all external tables.

External tables already have specific meta data. The can be maintained
in other character sets.

> 4) Don't simplify the engine because conversions between charsets is
> done like conversions between numbers and strings.

I think it does simplify the engine in that it allows a number of string
functions (substring, strlen, etc.) to be handled generically where they
are not character set dependent.

> 5) Waste of disk space and memory.

Probably not, and certainly not if we use a better compression algorithm.

>
> Continue thinking. You'll found more problems.
>

No doubt.

Regards,


Ann