[Help-glpk] Are my integer constraints being ignored?

[Alan Larkin]

When solving a MILP with GLPK via Matlab (glpkmex) I am getting non-integer
solutions. I could accept that (I wouldnt be happy but I could accept it) if
there was some indication that no integer solution could be found. Instead,
the integer constraint is seemingly just ignored. Any explanation? The first
few lines of output are below. Thanks.

lpx_simplex: original LP has 512 rows, 496 columns, 11334 non-zeros
lpx_simplex: presolved LP has 454 rows, 467 columns, 11276 non-zeros
lpx_adv_basis: size of triangular part = 454
      0:   objval =   7.218422676e+00   infeas =   1.000000000e+00 (0)
    200:   objval =   1.495206743e+04   infeas =   4.956196015e-02 (0)
    400:   objval =   3.576502704e+04   infeas =   4.079243578e-03 (0)
    432:   objval =   4.615788519e+04   infeas =   1.482525588e-17 (0)
*   432:   objval =   4.615788519e+04   infeas =   4.796163466e-14 (0)
*   600:   objval =   1.821209863e+03   infeas =   0.000000000e+00 (0)
*   800:   objval =   3.589917305e+02   infeas =   5.329070518e-15 (0)
*   934:   objval =   4.076064312e+01   infeas =   7.371880884e-13 (0)
OPTIMAL SOLUTION FOUND

P.S. Could someone please explain to me what infeas, the (sum of) primal
infeasibilities, means?

Alan