The parameter table was primarily sorted after the first parameter which was varied.
This resulted in a wrong assignment of the pixels in the *parameter* images;
the *chi2* image was, however, unaffected.
Squared parameter images can be flipped at the diagnoal (switching x- and y-axis),
non-squared, but rectangular parameter images were completely mixed up.

Example:

Assume that there are three parameters p1, p2, p3.
After calling

   get_confmap(p1, 1, 3, 3,  p2, 1, 2, 2)

the parameter table was (prior to this bug fix):

   p1 p2 p3
   1  1  p3(1,1)
   1  2  p3(1,2)
   2  1  p3(2,1)
   2  2  p3(2,2)
   3  1  p3(3,1)
   3  2  p3(3,2)

and the parameter image for p3, which is just created from reshaping the p3 column, was:

   p2=2   p3(2,2) p3(3,1) p3(3,2)
   p2=1   p3(1,1) p3(1,2) p3(2,1)

          p1=1    p1=2    p1=3

After this bug fix, the parameter table will be:

   p1 p2 p3
   1  1  p3(1,1)
   2  1  p3(2,1)
   3  1  p3(3,1)
   1  2  p3(1,2)
   2  2  p3(2,2)
   3  2  p3(3,2)

giving the correct parameter image after reshaping:

   p2=2   p3(1,2) p3(2,2) p3(3,2)
   p2=1   p3(1,1) p3(2,1) p3(3,1)

          p1=1    p1=2    p1=3

I'm sorry for any inconvenience this bug may have created.

Manfred