Ralph, the way you formulate it now is different from what I understood
originally. I thought it had to do with using cql.anywhere. If
truncation results in this error I would suggest to introduce a
diagnostic for it.
My personal predictable preference is that it should be possible to
accompany the diagnostic with the result of a scan starting at "com" in
this case (but I do not want to re-initiate that discussion).
Theo
>>> [log in to unmask] 30-8-04 14:53:01 >>>
> -----Original Message-----
> From: Marc Cromme [mailto:[log in to unmask]]
> Sent: Monday, August 30, 2004 3:45 AM
>
> Throwing in for my sixpence of opinion, I think this is not a good
idea:
> I think it's too specific to be useful for clients.
>
> - what reasonable behaviour should a client do when getting a
> 130: too many terms matched by masked query term ???
> how should it manipulate a query to take down the count of OR's (if
it
> indeed is
> implemented like that in your DB ??
> How should this be presented to the end user ??
The client can do little, other than let the user know what has
happened.
The user can do quite a bit. It they sent com*, maybe they could do
compu*.
> - It is much better to send back a more generic error message -
> something like
> a) sorry, I can not do this type of query at all, or
> b) sorry, this query is too complex - for whatever reasons.
Great, we're back to the old Fortran message, "syntax error", and it's
up to
the user to guess what they did wrong.
I think we can do better than that.
Ralph
|