FYI - I found the problem - thought I would pass the info along for
posterity.

My issue was caused by custom header renderers - if every column used a
custom renderer, then I would get this problem.  Turns out my custom
renderers
were not setting explicitHeight or measuredHeight, so the
mx:DataGrid.calculateHeaderHeight() would end up with a value of 0.  Setting
a value for measuredHeight
on my custom renderer fixed the problem.  In the cases where I had some
grids that worked correctly, it turns out that they had at least on header
renderer that
extended mx:Checkbox or mx:DataGridItemRenderer. Apparently those classes
return something for measuredHeight.

Not sure what changed between Flex 2 and Flex 3.2 to cause this - some bug
fix I guess.

I've read the excellent blog on item renderers (
http://blogs.adobe.com/aharui/item_renderers/) and other sources about the
component lifecycle, but I haven't found a
discussion about requirements for custom renderers to set measuredHeight or
measuredwidth - can anyonne elaborate on that topic?

Thanks

On Wed, Jan 7, 2009 at 10:18 AM, Glenn Jones <[email protected]> wrote:

> I recently upgraded from Flex 2.0.1 to 3.2. I have a class that extends
> mx:DataGrid. In Flex 3.2, some instances of this grid class are rendered
> just fine, but some instances (about half maybe) are drawn incorrectly - the
> first row of data values is drawn on top of the column header row. At least,
> that's the visual result.
>
> So far I haven't been able to isolate a root-cause or even a good clue to
> separate instances that work right from those that don't.
>
> Has anyone else seen something like this?
>
> I've reviewed the issues noted here:
>
> http://learn.adobe.com/wiki/display/Flex/Backwards+Compatibility+Issues
> http://butterfliesandbugs.wordpress.com/page/2/
>
> But I don't **think** any of them are related to this issue:
>
> - my flex2 code did not use/override DataGrid.drawHeaderBackground
> - I've tried to locate and correct all of the changes for decreasing
> rowIndex and rowCount/lockedRowCount by 1
>
> Any suggestions are much appreciated.
>
> Thanks,
> - Glenn
>

Reply via email to