The purpose of this change is to share this trait with other crates,
such as the forth-coming "semiring" crate that will be responsible for
handling some simple semiring operations as well as the querying, in
my plans.

Previously some incorrect forest nodes will be used for planting new
nodes.  I cannot fix the root cause of their presence in the
chain-rule machine.  But I can ignore them when they are encountered.

Of course I would like to really prevent them from existing, but still
cannot figure out how.

Not bug deals but adjustments of details.

Add an intentionally ambiguous grammar for testing purposes.

It seems to work fine.

This bumping of version is insignificant.  I just find it notable that
I seem to finally obtain a version without trivial bugs.  Hooray!

In the process of splitting, cloning, and planting the forest, I
forgot to check whether some cloned node of the node inquestion
satisfy the condition.  This used to cause forests that violate some
fundamental assumptions.  Now this is supposed to be fixed, but more
tests await us.

Now the binding part is finished.

What remains is a bug encountered when planting a fragment to the
forest which intersects a packed node, which would lead to invalid
forests.  This will also cause problem when planting a packed
fragment, but until now my testing grammars do not produce packed
fragments, so this problem is not encountered yet.

I am still figuring out efficient ways to solve this problem.

Add more directories under control of autotools.

Basic GNU standard files are added, and we now stop worrying about
monadic anamorphisms.

The current focus is on testing the correctness of the algorithm, so I
need convenient support for manipulating, interpreting, examining, and
per chance animating nondeterministic automata.