Working with X-Y-Diagrams and regressions I discovered some erroneous behaviour
concerning the diagram's axis scale and regression display:

- The x axis scale will automatically start at 0, even if there is no value of o
in the table (except the gap, which should be ignored!)
- A linear regression shown in the diagram is placed incorrectly if the axis
scale(s) are manually changed. It always assumes the axis to start at 0.

To make this much clearer I have an example document attached. You can also do
the following to reproduce the error:

Open a new document
Enter a XY table with at least one gap, e.g.
	X	Y
A	5	10
B	6	12
C	7	13
D		        <-gap
E	9	21
F	11	20
G	4	7
H	3	6

Calculate the regression: y=x*2.01-0.2
This works correctly. 
1.
Now insert an X-Y-diagram. The diagram will scale the x axis from 0 to 11,
although is should be 3 to 11. This is because of the data gap, but nonetheless
is an error of OOo.
2.
Now insert a linear regression. It will show up correctly.
Now manually rescale the x axis from 3 to 11. The data points will change
accordingly, but not the linear regression. It remains where it was and
consequently is WRONG.
3.
Now set the x axis back to automatic values.
Then enter some data into the gap (see how the diagram changes) and remove it
afterwards.
Now the x axis scales (automatically) correctly, but the regression is wrong
again. Funny, isn't it?

Result:
- A gap in the data set makes XY diagrams scales their axis' incorrectly
  (start at 0)
- A gap in the data set makes regressions to be drawn wrongly
  unless the x axis starts at 0.

I think this error is serious, because it simply produces wrong regression
diagrams - a big problem for serious scientific work!