[PUSHED] work around "broken" uno implementation with ScRangeList

[Stephan Bergmann]

On 10/04/2012 08:52 PM, Markus Mohrhard wrote:
> [1] works around a problem where our uno implementation uses
> ScRangeList but can only deal with one range inside of the
> ScRangeList. The patch joins all the ranges that are generated when a
> column/ row is removed and generates one single range again.
>
> This patch fixes the subsequentcheck failure in 3-6.

That fix apparently made the difference between a failing and a working 
sc_unoapi test, so I pushed it now.

Stephan