I finally found the time to test the 4.8 release (vs. 4.6, which I
used until now).

IMHO there is something wrong with the Client Character Set handling.
(You pass the Client Character Set using the CharSet property of
TIB_Connection.)

I have a Firebird 2.0.1 database with (default) character set UTF8
(this is the Server Charset).

In my application I want to process all strings from/to the database
as UTF-8 strings. So I specify 'UTF8' for the CharSet property of my
TIB_Connection. I now expect to get a UTF-8 string from the database
when I do a SELECT and I will pass UTF-8 strings in my INSERT/UPDATE
commands. As the database also is UTF-8, no transliteration or
transformation has to be done, neither by IBO, nor by the fbclient.dll
nor by the database itself. This used to work correctly with IBO4.6.

However, when I use IBO 4.8.7, there seems to be a transformation of
strings somewhere (probably in IBO) that leads to malformed UTF-8
strings in the database. Only when I use an empty string for the
CharSet property, everything seems to work fine. But now, there's no
Client Character Set defined which may lead to unexpected results for
fields that have another character set than the default character set
...

Is that a bug or is it my misconception?


Regards

Stefan

--
Stefan Heymann, Tübingen, Germany