resolve MATLAB version compatibility #1

README.md

@@ -10,6 +10,11 @@ Author: Bowei Wu, 2020/6/28
 
 Also contain supporting functions modified from Alex Barnett's [BIE3D](https://github.com/ahbarnett/BIE3D) package
 
+### Note on MATLAB version
+
+- If you are using a MATLAB version before R2017b, please rename the `Vecnorm.m` function to lower case `vecnorm.m` to use this code.
+- In case you would like to use the Epstein zeta function `epstein_zeta(s,...)` with a complex `s` (which is not needed for the purpose of quadrature correction), the `igamma` function from the Symbolic Math Toolbox is required.
+
 ### Example
 
 ![](convergence_on_patch.png)

Vecnorm.m

@@ -0,0 +1,4 @@
+function vn = vecnorm(A)
+% Alternative vecnorm function (in case MATLAB version earlier than R2017b)
+% return column-wise 2-norm for the matrix A.
+vn = sqrt(sum(A.^2));

epstein_zeta.m

@@ -50,7 +50,7 @@
 QB      = @(i,j) L*i^2+2*M*i*j+N*j^2;
 
 % Determine summation cutoffs based on decay of incomplete gamma func
-lambda = min((E+G)/2 - sqrt((E-G).^2+4*F.^2)/2,[],'all'); % min eigenvalue of Q
+lambda = min(min((E+G)/2 - sqrt((E-G).^2+4*F.^2)/2)); % min eigenvalue of Q
 n = floor(sqrt(33/pi./lambda))+3;
 
 % summation, exclude origin