Hello Helen,

> I was watching your messages in fb-devel and came to the

conclusion that

> you are getting this warning from ib_sql because it has no way to

know that

> your view is a join of four tables, and not a table.

No way ?
It could check RDB$VIEW_RELATIONS.

By the way, it would have to do it recursively, because
the resulting *relation*-names are not necessarily tables but might
be views themselves.

> I can't see why this should be a problem - a warning isn't an

error - and
The problem is that this warning is false and therefore very
misleading.

At least, this warning does not read "I have no real idea about how
big your dbkey should be and I have no way to caclulate it but I
warn you that I would prefer 8 over your 152"
Indstead it reads "DBKey_Length should be 8 but is 152 instead".
I understand this warning as "There is something wrong in your
database. Your db-key length is 19 times too big. That might work
for you at the moment but you should look after it soon."

Of course, NOW that I know how to judge it, it is no problem any
longer.
I posted that message just as a proposal for improvements:
To save other people the mess I had.
Additionally there is a slight chance that Jason has overlooked this
(rare?) constellation not only in the metadata-extract part of
ib_sql but in the ibobjects-access components as well (I don't know
much about the inner workings of ib-objects but I guess that he is
making heavy use of dbkeys; so for example allocating a 8-byte
buffer because he expects the dbkey to be of such size could be
fatal if the key is 32 bytes size instead.)

Kind regards
Frank