The #fetch utility acts in (roughly) the same way as Ruby's Hash#fetch function.
This addition should hopefully make constructors more rigorous and help to
prevent a few classes of errors.
These events are being removed, as they're essentially useless in most cases, as
the events will already have been emitted before the work block is run to set up
handlers for them.
Also removes the find{Robot,RobotConnection,RobotDevice} methods. Additionally,
stops passing the Master through to the Robot constructor, we can just
require('./cylon') if we need to access anything in there.
This commit reframes the Utils module to be more self-contained, and now only
the Number prototype methods and the every/after/constantly methods are exposed
globally.
I'm unsure about whether these methods can be more scoped such that people can
turn them off if they want, but that's potentially something to look into in the
future.
This commit also updates most of the other source files in Cylon, requiring the
Utils module rather than assuming the methods are in the global namespace.
A similar batch of commits will need to be made for all other Cylon modules to
make sure they don't explode when they can't find #subclass or something like
that.
To make the previously-global methods available, both in submodules and for
people looking for them, all the utility functions are exported in Cylon.Utils.
Robots can now take a break and have #play functions instead of #work.
Internally they're treated the same, but it's a fun easter egg!
Example:
Cylon.robot({
name: 'Jack',
connections: [],
devices: [],
play: function play(my) {
every((5).seconds(), function() {
console.log("All work and no play makes %s a dull boy.", my.name);
});
}
});
I sincerely apologize for the repeated reference to The Shining, blame
@deadprogram.
Closes#189.
Andrew Stewart