I finally managed to reproduce the bug (or one of the bugs).

By scrolling to the very bottom and sorting according to any column,
it seems to sort, but the last items seen in the grid shouldn't be
last. Turns out the grid just hasn't updated the view properly or too
early somehow.

A delayed call to invalidateList after sorting seems at first to fix
this, but it's possible to get the grid into an invalid state despite
this.

Dispatching a reset event seemed better. Would be nice to figure out
the cause since this bandage solution hurts my eyes.

--- In flexcoders@yahoogroups.com, "kallebertell" <[EMAIL PROTECTED]>
wrote:
>
> 
> I'll speculate openly in case someone picks up on this.
> 
> Could there be some sort of race condition in the way AdvancedDataGrid
> works. Ie. redraws or updates the scroll bar too early in reaction to
> sorting? (ie. before the collection has had a chance to finish sorting
> itself)
> 
> Is this kind of bug even possible in the flash runtime, keeping in
> mind that it is single-threaded?
> 
> 
> --- In flexcoders@yahoogroups.com, "kallebertell" <kallebertell@>
> wrote:
> >
> > 
> > I'm facing a problem where an advanced data grid (adg) with a
> > hierarchical collection (~500-1000 rows) doesn't redraw properly after
> > sorting is applied to the underlying collection. 
> > 
> > The fun bit is that it isn't reproduceable. The occurance seems to be
> > completely random but somehow related to sorting, as it seems to rear
> > it's head only then (but not always).
> > 
> > Scrolling after the adg is messed up doesn't improve it either, but at
> > some point causes null reference exceptions onMouseRollOver. 
> > 
> > The advanced data grid is in a minimiziable container. If the
> > container is minimized and then restored, the adg looks normal again
> > (ie. redraws correctly).
> > 
> > Although this rarely happens on my dev machine it happens more often
> > for our customers, which might indicate that it's more frequent on
> > slower machines.
> > 
> > I'm looking for any clue or people who have also experienced this
> > weird behaviour in adg.
> > 
> > Thanks.
> >
>


Reply via email to