I'm trying the qflll function Pari for lattice reduction. However, I'm not getting correct answers.

First, the wikipedia example:

https://en.wikipedia.org/wiki/Lenstr...tion_algorithm
Code:

(20:32) gp > qflll([1,-1,3;1,0,5;1,2,6])
%50 =
[-4 5 0]
[-1 1 1]
[ 1 -1 0]

Second, the matlab example:

http://www.mathworks.com/help/symboloddslotic/mupad_ref/lllint.html
Code:

(20:32) gp > qflll([1, 2, 3;4, 5, 6])
%51 =
[-1 4]
[ 1 -3]
[ 0 0]

As you can see, both results are incorrect. What am I doing wrong?