Allow resize -s to take a string size rather than ID

[kura]

While writing a bash completion script for Tugboat I discovered that using tugboat resize droplet -s 1GB fails as it expects the ID of the size from tugboat sizes

I'm not sure if the --size flag is any different but I think it would make more sense for both to support the string representation of the size.

In terms of the bash completio it makes a lot more sense since you cannot display the Human readable string representation to the user but actually submit the ID to the command.

[pearkes]

I think this is a good idea. Sizes aren't every going to return a huge number of results, so we can definitely just parse the output searching for a string match and grab the corresponding ID. It's not perfect but I think the only real choice we have, based on the API as it is now.

[kura]

Cool. Glad you agree. Sadly I am awful at Ruby otherwise I'd do the work myself.

Just FYI, my bash completion are at https://github.com/kura/tugboat-bash-completion under MIT license, feel free to link, submodules or ignore at will.

Once the changes are in place for resize I can finish that last completion.

The only downside right now is that nothing is cached within tubrug so each command that requires a droplet or image name requires an API call which takes time. Ruby itself may be a factor in this too, not sure since tubrug help is fast enough.

[petems]

Closed by #161 (New API uses string slugs)