On Wednesday 01 August 2007 19:48, Adriano dos Santos Fernandes wrote:

> Roman Rokytskyy escreveu:
> >> We greater field length in SQLDA to:
> >> - Replace broken UNICODE_FSS by UTF8 in system tables
> >> - Increase identifier length
> >
> > And I'd like to sqlschema in XSQLVARs, so that we won't change the thing
> > in the future.
>
> Ok.
>
> >> We also need a way to pass strings of unknown length.
> >
> > Isn't that limitation of the remote protocol handler?
>
> AFAIU, no, as data is transmited in BLR form.
>
> Another thing I'd like to see is a new way to represent CHAR columns.
>
> The current one, is totally dump for MBCS, as length in bytes of MBCS
> CHAR columns is not fixed.
>
> This can be done in one of this ways:
> 1) new type SQL_TEXT2 semantically identical to SQL_VARYING
> 2) using a subtype for SQL_VARYING
>
> I prefer (2), so it will just works, and one can differentiate between
> VARCHAR and CHAR columns using the subtype.

I believe we need some more changes in API to make at once. As soon as we plan
to have strings of unlimited length, we will do memory allocation ourself,
right? In that case no need to force user to allocate memory for SQLDA - we
may allocate it ourself. May be add a bit more user-friendly ways to deal
with parameter blocks. And I will prepare the proposal for the new form of
them - as a separate RFC.