Re: fix for BUG #3720: wrong results at using ltree

I wrote:
> I dunno, that doesn't really seem clearer to me (although some of it
> might be that you expended no effort on making the comments match
> the new code logic).

... although looking closer, this formulation does have one very nice
advantage: for the typical non-star case with high = low = 1, the
only recursive call is a tail recursion, so it ought to consume less
stack space than what I wrote.

Let me see what I can do with the comments.

            regards, tom lane