@tristanvb, I would have thought that the closing of this issue would have come along with a patch which both enables pylint's cyclic-import warning and deals with any circular imports?
We still have an instance of a cyclic import which I have linked to in the comment above.
To be clear, does the closing of this issue mean that we intend to ignore this one instance and keep the pylint warning disabled?
@tristanvb, I would have thought that the closing of this issue would have come along with a patch which both enables pylint's cyclic-import warning and deals with any circular imports?
That's one way to look at it, but I opened this issue with the understanding that python 3.4 just does not support circular imports; the test cases running python 3.4 are proving that to be incorrect - as such it's not really a problem anymore.
This does not mean that it is not messy to have circular imports: it's still desirable to change that - whether we want to track progress towards cleaning up this particular messyness in an issue, is another question (but rather tedious, it's now become a fairly minor stylistic point, nothing prevents it from being fixes without having an issue especially open for it).
@tristanvb does this mean we should open a different ticket or that this ticket should still be opened to consider @jennis case?