Dimitry Sibiryakov wrote:

>> Why it is a bug?
> Because it allows an invalid object to exist in database after commit. Such behavior is
> a bug, bringing database into not usable state.

I understand what you are saying, and it would be useful if we had a warning.
However just as there is no requirement for the UDF to exist before creating the
connection, there is no way of ensuring it has not later been deleted or
corrupted? My usual problem here is having moved to new hardware the correct UDF
gets missed. That is often these days because the new hardware is Linux and the
old Windows, so I am having to change the UDF entry anyway, and the database is
not in an unusable state, just some queries fail to work which is often why I
find the problem :) The data is perfectly safe?

I would like to run the update script before doing the backup and restore, but
the Linux UDF will not work on Windows, and I would want to avoid having Linux
files in a windows install anyway. The bug to me is having to wait until I have
moved the database to complete the installation of the update simply because the
UDF is not yet present ... as Bruce reported.

--
Lester Caine - G8HFL
-----------------------------
Contact - http://lsces.co.uk/wiki/?page=contact
L.S.Caine Electronic Services - http://lsces.co.uk
EnquirySolve - http://enquirysolve.com/
Model Engineers Digital Workshop - http://medw.co.uk
Rainbow Digital Media - http://rainbowdigitalmedia.co.uk