Re: Plug treesit.el into other emacs constructs

[Stefan Monnier]

>> In my view ";" is not a substree.  It's the node of a substree.
>> We can't actually move over a proper subtree in that case because there
>> is no substree whose left boundary starts right before the ";", so the
>> closest is to move over the ";" *plus* its right child.
>
> Ah, so by “smallest subtree” you basically mean “smallest non-leaf node”?

Hmm... I don't think so, tho I guess it depends on what is your notion
of "node" and "leaf".  In my book, "a = x + 1; b = y" is a tree of the
form:
                      ;
                     / \
                    /   \
                   =     =
                  /|     |\
                 / |     | \
                a  +     b  y
                  / \
                 /   \
                x     1

So there is no ";" subtree at all (there is a subtree with ";" in its
root but it contains all of "a = x + 1; b = y").  OTOH there is an "x"
subtree and it would be the "smallest subtree" if we're immediately to
the left of this "x" and we're doing a `forward-sexp`.  So my "smallest
subtree" can definitely be a "leaf node" since I consider "x" to be
a leaf node.

[ And the ";" node at the top can have N children, by the way.  ]

If point is to the left of the semi-colon, it can be considered to be:
A) right after "1"
B) right after "x + 1"
C) right after "a = x + 1"
But in order to be able to "move forward" you need to move up to the ";"
node, so only C makes sense, and the smallest meaningful subtree
over which we can move is the "b = y" subtree, since in order to move
only over "b" we'd have to go down inside the "=" node, so we'd be
moving over a subtree that's not connected to our starting node.


        Stefan