I took at quick look at the code (and, BTW, this part of the code was not changed with my recent polar-alignment changes).
From what I can see, there is no issue with tracking during polar alignment. In fact, the code explicitly turns on tracking
if it isn't already tracking (though after Start is clicked).
However, the Start button is explicitly disabled during slewing and re-enabled when slewing is completed:
I'll bet that, as I mentioned earlier, for some reason the end-of-slew status-change message was missed,
so the Start button was never re-enabled. Then, I think that when you turn off tracking, that causes a change the mount's status,
and that causes the setMountStatus method to be called again
and this causes the Start button to be re-enabled!
To show that this theory might be correct, I'll bet that if you did the following, it would also work.
After your slew is complete, if the Start button is not enabled, then disable tracking, and then a few seconds later re-enable tracking
(so that you've really done nothing but "shake things up").
My guess is that after all that the Start button will be enabled, with tracking on, and you'd be able to polar align.
If this is the case, I can forward this issue to someone else who knows more about mount/Ekos communications than I do.