Revision history [back]

I need to evaluate $L$-functions of the form $$\sum_{n=0}^{\infty} \Big( \frac{D}{n} \Big) n^{-s}$$ at integers, where $\Big( \frac{\cdot}{\cdot} \Big)$ is the Kronecker symbol. This is apparently what quadratic_L_function__exact(k,D) does. However I find this unreliable when D is not squarefree.

For example, the value quadratic_L_function__exact(2,4) = pi^2 / 6 is incorrect; the sum should actually be pi^2 / 8. However, the value quadratic_L_function__exact(2,12) = 1/18 * sqrt(3) * pi^2 seems correct. I can't tell what's going on.

I need to evaluate $L$-functions of the form $$\sum_{n=0}^{\infty} \Big( \frac{D}{n} \Big) n^{-s}$$ at integers, where $\Big( \frac{\cdot}{\cdot} \Big)$ is the Kronecker symbol. This is apparently what quadratic_L_function__exact(k,D) does. However I find this unreliable when D is not squarefree.

For example, the value quadratic_L_function__exact(2,4) = pi^2 / 6 is incorrect; not what I need; the sum should series above is actually be pi^2 / 8. However, the value quadratic_L_function__exact(2,12) = 1/18 * sqrt(3) * pi^2 seems correct. I can't tell what's going on.