Plot Summary

Options

All options are arguments, but not all arguments are options.

$ nurl -I --insecure https://topsyturvey.onion/mikado.mkv

In the example above, the program nurl has three arguments, two of which are options. Options are arguments that are one letter long and start with a dash (-), these are "short options," or many letters long and start with a double dash (--), these are long options. Arguments that are not options are called "positional" arguments, because they have no name, so can only be referred to by their position following the command.

Operetta would parse the above example as follows:

For the program to receive these values, it calls the start function with a callback.

Simple, right? And quite enough for many programs. But that is not all, oh no that is not all!

Parameters

All parameters are options but not all options are parameters.

$ nysql --database secret_database --host=control.onion -usmart -p Iheart99

Sometimes options take a value. We call these Parameters.

The above shows the four valid forms to set values. Without any further instruction, Operetta would parse the above as follows:

Uhgg. That's probably not what we want. It got --host right, because that is the most unambiguous form for a parameter to take, a long option connected to a value by an equal sign. However the rest, what a mess! Since it doesn't know that --database and -p are parameters, it treats "secret_database" and "Iheart99" as positional arguments, and since short options can be chained together, Operetta thinks "usmart" is a chain of 6 options. We're going to have to give operetta more information to handle these correctly.

We use the parameters function to tell Operetta some things about our parameters, first we pass a list of options, i.e. ['-D','--database'], this gives the long and short form of the option, then we give a description.

Now, we get the follow values:

Much better! Note that the key for the value is always the first item in the options list passed, so -D is present, even though --database was used.

Help

What's more is now that we have descriptions, operetta will automatically bind the options -h and --help to show these descriptions as help.

Nifty, huh? But what about plain old options? We may want to give these descriptions too. For example, in our earlier nurl example, we may want to provide descriptions for -I and --insecure. We can use the options function for this.

If you really insist, you can can override -h and --help using either the options or parameters function, you can then then get the help output by calling the usage function, either with or without a callback.

We can add a banner above line that says "Usage."

Now we get the following Help:

There you go! Now you can add options and parameters to your program and have it display nice help with the descriptions. That's all you need right? But that is not all operetta can do! Oh no, that is not all!

Events

Sometimes you don't just want all the options parsed and dumped to a single callback as a values object, but you wold rather have an event triggered for each option. Here is where Operetta Sings!

The operetta object is an EventEmitter, so you can bind events with the on option.

Since -k is just an option, value will always be true when this event is called, in the case of a parameter, value will be the value passed or null if none was passed.

While using the on function works, the preferred way to set a callback is to pass it as the third argument to either the options or parameters function.

So there you have it, Options, Parameters, Help and Events. Surely that's the end of this interminable readme file? No! That's not all. And stop calling me Shirley.

Subcommands

Sometimes programs have different commands, each with their own options, i.e.

If the program nit has many subcommands, i.e. clone, commit, push then each of these could have their own options and help. Operetta has a command function that allows you to define these and get a new instance of operetta for each command.

Now, if you called help without a subcommand:

$ nit -h

You get a list of the subcommands.

However, if you call help on commit:

$ nit commit --help

You get the descriptions of the options defined for commit.

And yes, if you really want, subcommand can have subcommands:

Now you could do:

$ nit submodule add http://nitorious.onion/nitorious.nit

Coda

So, options, parameters, help, events and subcommands. Shirley, you're thinking operetta must be some big, baroque, bloated, blob of blubbery JavaScript! Well, here's what SLOCcount has to say:

That's right, small and cheap. So far it's only got One Hundred and Seven Lines of Code. So get it while it's small, before I add thousands of lines to support such must-have features as sending and receiving email and impersonating a teenager in IRC channels.

And yes, I called you Shirley.