[Kst] You cannot load this much data

[Nicolas Brisset]

> Yes, 1GB should be enough. But maybe the algorithm for calculating
> the needed memory
> gets somehow confused.
Yes, I think it gets *massively confused*.
What I don't understand though is that I've already loaded vectors with close to *3 million* points each from ASCII files without problem.
I'll test it again with a recent version to see if we have a regression there, which would be a pity because I'm actually going to have to use kst for large data samples soon...

And by the way, what is preventing us from releasing 2.0.5, apart from this newly reported bug? 

Nicolas