Manodeep Sinha

If I disable the duplicate cell-checking in [here](https://github.com/manodeep/Corrfunc/blob/master/utils/gridlink_impl.c.src#L460-L462), then the bug disappears. However, we need to work through the logic and make sure we are not double-counting under any circumstances...

@lgarrison That is definitely what is happening here. Consider zmax = Lbox/2, and we have 2 bins. In that case, the top half of `iz=0` can be within `zmax` of...

@lgarrison If we revert the change, then Corrfunc will report an error for `(nx, ny, nzbins < 3)` and @beckermr will not be able to run this example

Thanks for the offer! What kind of functionality did you need - could you provide a bit more details about the custom binning strategy?

@lgarrison looks good. I don't think any of the combine pair counts sort of utilities will have to change. I would simply change the internal API to be in arbitrary-but-contiguous...

@Christopher-Bradshaw Thanks! @lgarrison Since the pair-counts are designed to be the same for an auto-corr of `D1` and a cross-corr of `D1` with `D1`, shouldn't the conversion from pair-counts to...

@lgarrison Would your gist to demonstrate how to correctly account for pair-counts in a `Rmin=0.0` bin work?

> Also, just checking that by natural estimator you mean DD/RR? @Christopher-Bradshaw Yup!

@Christopher-Bradshaw @lgarrison Does anyone recollect what needs to happen with this PR?

Another thing to do would be to find the final sorted order (i.e., account for `sort_on_z`) while performing this initial partitioning of particles into cells.