Potential ways to optimize width_dimensions() algorithm

[ArniDagur]

See the following commit in my (rather poor) Python implementation of this algorithm: ArniDagur/guide-key.nvim@fbd9b5e

As you can see, I tried both binary search and looping downwards from the theoretical maximum number of lines, to 1. Both seemed to improve performance, but there are almost certainly better ways to go about it.

I thought about basing binary search hi/lows off of the theoretical maximum number of lines, but the problem there is that I'm not sure how to efficiently calculate the theoretical minimum. This occurs when the items are sorted from highest to lowest (or the other way around).

@ogham I would love to hear your ideas and thoughts.