Add grmtools section to .y files

[ratmice]

During development I'd added the %grmtools section to a few of the examples, to test it.
But forgot to add them to all the rest, this adds those, and removes the .yacckind calls in build.rs.

Tested using the following shell script, which never specifies a -y flag and also performing cargo run in the examples,

echo -n "/* /* */ */" | ./target/debug/nimbleparse lrpar/examples/start_states/src/comment.{l,y} -
echo -n 5+++ | ./target/debug/nimbleparse ./lrpar/examples/clone_param/src/param.{l,y} -
echo -n 1+1 | ./target/debug/nimbleparse ./lrpar/examples/calc_storaget/src/calc.{l,y} -
echo -n 1+1 | ./target/debug/nimbleparse ./lrpar/examples/calc_parsetree/src/calc.{l,y} -
echo -n 1+1 | ./target/debug/nimbleparse ./lrpar/examples/calc_ast/src/calc.{l,y} -
echo -n 1+1 | ./target/debug/nimbleparse ./lrpar/examples/calc_actions/src/calc.{l,y} -

All looked good, the only question I have is about calc_parsetree which when I cargo run the compile time version,
I don't seem to see a parse tree emitted (just the calc computation), which is not what I had expected, but I'm now uncertain of whether that behavior has changed with these recent patches.

I will need to build a prior release and check the behavior against that.

[ratmice]

So, I did test every 0.x.0 release from 0.13.0 to 0.5.0 and all had the same output for cargo run in calc_parsetree.
So I believe that this is behaving correctly, and perhaps just me misunderstanding the expected behavior.

[ratmice]

I think that should cover all of the examples/documentation, assuming we want to go so far as removing the places where it was set but could be inferred.

[ltratt]

the only question I have is about calc_parsetree which when I cargo run the compile time version, I don't seem to see a parse tree emitted

The idea of calc_parsetree is to use grmtools' generic parse tree internally: in essence the calc_* examples are mostly intended to be "let's show you different ways of doing the same thing".