Error in first sample in GibbsRectangleMNormal

[César Pérez Álvarez]

If we compile this code several times:

Matrix mu = Col(0.0);
Matrix cov = Col(1);
Matrix c1 = Col(-1/0);
Matrix c2 = Col(1/0);
Matrix MCsample = GibbsRectangleMNormal(mu, cov, c1, c2, 1, 1);

Always sample value is -8.29131843138. If we try with this code

Matrix mu = Col(0.0);
Matrix cov = Col(1);
Matrix c1 = Col(-1/0);
Matrix c2 = Col(1/0);
Matrix MCsample = GibbsRectangleMNormal(mu, cov, c1, c2, 1, 100);

And the first sample, that we get, is the upper value. I think that problem is the limits (-inf, +inf). If we try with others value not +/-inf the sampling works well.
If we calculate the probabability density without this value, this density is the expected one N(0,1).