Refactor initial `up` workflow

[tj]

More work on #482, but needs my weird react-ish term thingy to improve it further.

• make it look better
• make the CWD more obvious (if you try to deploy ~ etc)
• maybe help them set up AWS and ~/.aws/credentials

Please thumbs-up if you think helping create ~/.aws/credentials would have helped you. I'd like to streamline this as much as possible at some point, for non-AWS users.

[ksmithut]

Although I'm fairly comfortable setting up ~/.aws/credentials, I remember a time not too long ago when I had to dig through the AWS docs to figure out how to set it up. Putting a thumbs up for my past self.

[BTBurke]

I just started playing with Up today and I think the error handling leaves a lot to be desired from a user's POV. I still haven't been able to get a standard static site to deploy because of various errors returned by AWS API calls.

I started off with following the basic up.json setup and running up stack plan which causes a 409 error unless you run a naked up first.

So then I went ahead and started with just up, but every time I get a 409 response from the AWS API. These are returned to the user without much wrapping on the error to explain what happened.

Error: deploying: deploying: us-east-2: creating s3 bucket: BucketAlreadyOwnedByYou: Your previous request to create the named bucket succeeded and you already own it.
	status code: 409, request id: DECEB4F979325BA5, host id: b8OXJ+irpM3XHobLjV+m2CkMNeWJ7qTzEGs/Zpk/rRUlEEdWkQKjLcf60xe7+A2Xq8W0Y2rghgE=

So I can dive into my own AWS console and delete the bucket, but that doesn't improve matters. Changing the app name doesn't fix it.

At a minimum, I think from a first-time user's perspective:

• Most of the AWS errors need to be at least wrapped with a more helpful error message, and include a point in the right direction for how to fix it.

• It should be able to recover automatically, especially when working with resources that have already been created by up.

• There really needs to be an up clean command that will effectively nuke all up-created resources to start from scratch.

[tj]

@BTBurke ahh I've realized a few people run up stack plan before an initial up, I'll have to add some kind of guard against that.

Regarding nuking up stack delete at any point should work, I may have to ignore some errors from there though (since the stack wouldn't exist yet, only the s3 bucket in this case).

Agreed though, I'll have to collect the edge-cases and get some better errors going. Thanks for the feedback!

[BTBurke]

I got it to work after deleting a lot of the resources manually in the AWS console and then starting over with a more basic config. I think it was user error trying to deploy it in a multi-region setup. After switching to one region and starting over, it was smooth sailing from there.

Nice work with this. I'll take a look at some of the issues and see if I can lend a hand.

[tj]

Ahhh! I'll add an error for more than one region since that's not quite ready, and sounds good!