Hi,

> Another suggestion. You know buffer length for parameter in bytes.
> You know what charset set you are using and know how many bytes can
> characters occupy in this charset. You also know how to count
> characters in this charset. Given this information driver can easily
> trim returned string itself.
>
> This is a bit hacky way and can break some things (e.g. you declare
> CHAR(80) UNICODE column and store 200 ASCII characters in it, driver
> will trim some data), but it will work.

Yes, we can create a workaround for this issue in driver. However I'm
not inclined to provide any workaround inside driver if the problem
can be addressed in the engine. Each such workaround might have
consequences that are hard to debug.

Roman