All server options to launch

The launch mechanism has been modified but it still isn't perfect. Ideally:

• Ubiquity invokes a ruby command that returns the url of the local server
• Ubiquity invokes the local server
• Ubiquity sets itself up using the url by querying the server for the server configs.

The three step process is necessary because before the server exists, Ubiquity cannot know where to look for server configs. By splitting up the initial query and the invocation, you can re-use step 2 in manually launching a server. By having Ubiquity set itself up by querying an active server, step 3 can also be re-used when you want to specify a remote server for Ubiquity to use.

As it stands Tap-Ubiquity uses a file pass the url message from the process to the server. It's not idea because after the server launches there's a callback to delete the file. Better if step one simply prints out the url, which either gets read directly into Ubiquity, or gets dumped to a tempfile that Ubiquity manages.

Normally you don't want Ubiquity to be able to shutdown a server, because the url may have been set to a remote site that Ubiquity doesn't control. In step two, provide a Ubiquity-generated shutdown key that can trigger a kill.

At present step 3 doesn't seem necessary because Ubiquity only needs the uri (which is obviously provided or determined elsewhere). In the future step 3 is when Ubiquity gets information about what commands are available, for instance.

Here's a list of info to get:

|uri|base uri of the SERVER, ie http://localhost:8080/tap-ubiqui...| |dispatch path|dispatch| |query parameters|whatever Ubiquity may provide, args, selected text, etc.|