Subject | Re: [IBO] Hc Test release seems to work good, but some question and opinions required |
---|---|
Author | Geoff Worboys |
Post date | 2002-05-13T02:36:54Z |
> Just curios to know IB_UpdateBar OnDataChange, that I've removed butIn one of my applications I built the ability to dynamically alter
> Jason has kept. Is there any situation where the IB_UpdateBar should
> react to such an event? I did not figured out any, but I've not done
> exhaustive tests or deep thinking. And like as TIB_Date
> demonstrates, you can't think enough;). Any thoughts?
whether a particular record is readonly or not according to the value
of a field. I built my own derivation to TIB_Query to handle this
automatically and I dont use updatebar in this particular situation.
However the same concept could be built just using provided event
handlers and the ability of updatebar to respond to that situation may
be important. (I've not studied that situation in detail, but it
seems likely.)
> So I've found very annoying having to specify a whole query syntax,What is wrong with just putting the following into LockSQL ??
> while it would be very good the one generated automatically by IBO
> except that it should work only with the desired field. So if you
> specify a LockSQL without the "UPDATE" keyword, I assume it's in the
> form of LOCK_FIELD=LOCK_VALUE, ie LOCK_FLG='Y', and let IBO build
> the other parts of the query, if possible. Do you think that is
> convenient, are there any drawbacks? Any other interested in this
> feature?
UPDATE MYTABLE SET
LOCKFIELD='Y'
WHERE MYKEY = :OLD_MYKEY
Set up your standard SQL so that you dont select LOCKFIELD at all,
which will prevent IBO from ever assigning to a normal insert or
update.
> but would not be better "publish" FPickDate as a whole? (or aNo opinion other than to say that compatibility may be serious to
> TPickDate descendant with only relevant properties?) That way you
> could reduce IB_Date properties number, have them better organized
> (you immediately know that that property is about PickDate), and
> have more control upon PickDate too. Maybe compatibility could
> suffer, but should be nothing serious,
someone with a hundred forms in their application. (There are ways
around this but it starts to make the code messy.)
--
Geoff Worboys
Telesis Computing