[Kst] new return codes...

[Andrew Walker]

To respond to the following:

> FYI there is no need to check for null return codes from operator new.  It
> will just throw an exception or crash soon after anyway.  That's the
reason
> for the big invisible change in kstlegend.

I'm not sure that I agree with this. If all new return codes are checked
and the code implemented carefully there is no reason that the user should
not be able to save and exit Kst gracefully under low memory conditions.
This should at least be the goal, whether or not it is achieved is another
thing. It seems counterproductive to deliberately not check the return
codes.

Andrew