EinsteinInitialData/Exact: the param.ccl says that exact_eps is in terms of the grid spacing, the code uses it as an absolute value.

[Gabriele Bozzola]

Here we can see that the docstring of the parameter exact_eps in EinsteinInitialData/Exact says that the size of the stencil for the finite difference depends on the grid spacing. However, when this is used, e.g., here, the grid is not taken into account.

‌

[Roland Haas]

Thank you for reporting. The obvious fix will be to make the parameter documentation fit the code. Do you want to propose a pull request for this?

[Roland Haas]

Pull request is here:

https://bitbucket.org/einsteintoolkit/einsteininitialdata/pull-requests/12/exact-fix-exact_eps-docstring

[Roland Haas]

Applied as git hash 1d82f008 "Exact: Fix exact_eps docstring" of einsteininitialdata