uniaxial strain isn't correctly handled in POSCAR generation

[meldontaragon]

When generating a series of POSCAR files with uni-axial strain the second line of the POSCAR unit cell matrix is always slight wrong. As an example, he is the hexagonal unit cell for base lattice parameter of a = 5.000 where the second generated structure is -0.02 Ang compression along only the a-axis.

Initial cell (a=5.000 Ang)

   5.000000000                0.000000000                0.000000000
   -2.500000000                4.330127019                0.000000000
   0.000000000                0.000000000                23.600000000

Compressed cell (a=4.980 Ang)

   4.980000000                0.000000000                0.000000000
   -2.490000000                4.330127019                0.000000000
   0.000000000                0.000000000                23.600000000

As can be checked quickly, while the a-axis is compressed correctly, the b-axis is not the proper 5.000 Ang and instead there is a small compression along the x direction of that axis.