There are some enhanced abilities of IBO this may not be ideal with. I'll
need to consider this carefully.

Jason

----- Original Message -----
From: "Nando Dessena" <nandod@...>
To: <IBObjects@yahoogroups.com>
Sent: Friday, January 16, 2004 3:05 AM
Subject: [IBO] Issue with TIBODataset.CanModify


> Hello,
> I have discovered yet another difference in behaviour between
> TIBOQuery and good ol' TQuery (well, every query component that has a
> RequestLive property, actually). If you set TQuery.RequestLive to false
> then TQuery.CanModify always returns False. This has a number of
> implications, i.e. the data-aware controls set themselves to a
> read-only state and don't try modify operations on the dataset, which
> would throw exceptions.
>
> Unfortunately TIBODataset.CanModify doesn't take RequestLive into
> account. Here's the current (4.3Aa) code:
>
> function TIBODataset.GetCanModify: boolean;
> begin
> if Assigned( IB_Transaction ) and IB_Transaction.ReadOnly then
> Result := false
> else
> begin
> Result := InternalDataset.NeedToPost;
> if not Result then
> Result := not FReadOnly;
> end;
> {!!!
> begin
> CursorPosChanged;
> UpdateCursorPos;
> InternalDataset.RowNum := InternalDataset.BufferRowNum;
> Result := InternalDataset.CanModify;
> end;
> }
> end;
>
> And my proposed modification:
>
> function TIBODataset.GetCanModify: boolean;
> begin
> if not RequestLive then
> Result := False
> else if Assigned( IB_Transaction ) and IB_Transaction.ReadOnly then
> Result := false
> else
> begin
> Result := InternalDataset.NeedToPost;
> if not Result then
> Result := not FReadOnly;
> end;
> {!!!
> begin
> CursorPosChanged;
> UpdateCursorPos;
> InternalDataset.RowNum := InternalDataset.BufferRowNum;
> Result := InternalDataset.CanModify;
> end;
> }
> end;
>
> BTW I haven't got the foggiest idea why the call to
> InternalDataset.NeedToPost is necessary or even relevant here. Would
> very much like to find out.
>
> Does Jason or anyone else foresee any problems with this change? It is
> not a light change compatibility-wise, as it has the potential to
> expose logic bugs in applications, but IMHO it's the only right
> behaviour (incidentally it makes my apps work as they did before ;-)).
>
> If everybody thinks it is OK then I ask this change (or equivalent,
> interface-wise) be integrated in IBO's next subrelease.
> Thanks
> --
> Nando mailto:nandod@...