Użytkownik Paul McNett napisał:
> 2)
> Setting ScanRequeryChildren doesn't actually requery any children that I can 
> see[1],
> all it does is set the current parent, which should be a pretty fast 
> operation, and
> definitely desirable (can't think of why you would want to purposefully not 
> set the
> current parent).
>
> I think that at one time, children were being requeried needlessly resulting 
> in bad
> performance, and that the operation of setting the current parent would 
> result in an
> implicit requery(), but I don't think that's happening anymore.
>
> Should we:
>    1) remove ScanRequeryChildren completely
> -or-
>    2) Keep it, default to False like we claim, and add it as a separate 
> concept from
> setCurrentParent()?
>
>
> [1] The only case where a child would be requeried is if there wasn't already 
> a
> cursor for that child, and child.RequeryOnLoad is True.
>    

Hi.

I'm using my patch from ticket #1351 with the following rules:
1) default value of ScanRequeryChildren is True
2) to avoid retrieving unneeded data, I turn it of in isAnyChanged() mehod
3) I have set updateChildren to false for all row pointer restoring code.
This works for me.

BTW, I think you shouldn't use RowNumber = val internally
for row pointer restoration, but in some places you do it,
like in restorePosition(), but in other you don't - see getFieldVal() 
method.

-- 
Regards
Jacek Kałucki


_______________________________________________
Post Messages to: [email protected]
Subscription Maintenance: http://leafe.com/mailman/listinfo/dabo-dev
Searchable Archives: http://leafe.com/archives/search/dabo-dev
This message: http://leafe.com/archives/byMID/[email protected]

Reply via email to