Jerome Kelleher

I've updated this so that it's a clean build against upstream now @petrelharp. I think it needs a bit of documentation though - I don't really follow what the algorithm...

I think the docs build is because of a stale cache. The simplest thing is probably to create a new PR. I've squashed my original commits down to one.

It's fairly esoteric knowledge @bhaller which I'm not sure is worth keeping a PR open in perpetuity for. We've learned the key lessons from it, I think, and moved on?

If it's useful we should merge it I guess - I didn't think too hard about it tbh. Keeping it here in limbo isn't an answer. Does anyone want to...

I can finish it up too @petrelharp, if you'd prefer to have a look through and post some review comments.

I'm not sure "breakpoint" here is a great bit of terminology as @hyanwong points out. It can be confused with ``ts.breakpoints()`` which is quite a different thing. How about ``terminal_positions``...

I think it's straightforward enough logically, we just treat the reference sequence like imaginary entries in the Site table which have no mutations. So, we should get the same result...

What we get is: ``` Traceback (most recent call last): File "/home/jk/tmp/bug.py", line 17, in next(seqs) File "/home/jk/.local/lib/python3.10/site-packages/tskit/trees.py", line 5572, in alignments raise ValueError( ValueError: Missing data not currently supported...

> I'm all for implementing some more Fst methods, by adding a method= argument, though - people do like their different measures of Fst! Agreed!

Lots of ink has been spilled over what Fst is/isn't/should be interpreted/etc, and our documentation isn't the place to settle these questions. I think the documentation should - a) State...