On Tue, 13 Mar 2007, Jason Wharton wrote:

> Randal,
>
>> Turning on request live as I saw for other peoples
>> issues in the archives made this work sort of...
>> but on my original bde query does not use request
>> live so I was grasping straws and this one worked.
>
> One thing to beware of is when using cached updates it is often desirable to
> include the old values comparison in the WHERE clause as a means to prevent
> overwriting values that others have updated. You may not care either, but
> this is a common strategy to avoid conflict resolution. Setting RequestLive
> to true leaves the generation of the SQL for handling updates to IBO and it
> may not do things exactly the way you want.

I moved the queries all directly under the IBOQuery, so now I dont have to
use live updates. Seems this is some weirdness with the IBOUpdate object,
but I now know how to avoid it as I stated in an email I sent after this
one.

>
>> The other part to the picture (the sort of part)
>> is that when I close the query and change the SQL
>> in the IBOQuery it don't do an implied cancelupdates
>> when I close the query. I must explicitly call
>> cancelupdates, then do the close, then change the
>> sql and prepare it for use, the bde held my hand
>> on that one. No biggie though, those are easy
>> enough to find in my code since I did do this in
>> functions called setupqueryXX, clearqueryXX under
>> the datamodule for each query.
>
> I consider this a bug in IBO and I will add this to my list of issues to
> resolve. Mind you, this isn't a list in publishable for. It is a folder
> where emails go with potential issues that I work on when time allows.

Thanks, I was wondering about that one...I thought that it should notice
the change in the sql.

thanks,
Randal Carpenter