Improvements on routing url option

[JakeDawkins]

Related to changes introduced in #130

TODO (jake) -- fill in details here

[EverlastingBugstopper]

What I wrote to @JakeDawkins was that in my ideal world, this is what things would look like.

If --routing-url is not passed and there is already a routing url for a managed service, the API should leave it untouched
If --routing-url is not passed and this is the first time they're pushing a service, we should error and tell them they need to pass the --routing-url flag
If --routing-url is passed and it's the first time pushing a service, it should push the service and exit successfully.
If --routing-url is passed and it's the same routing url as what's currently in the registry, it should warn the user that the routing url will be unchanged, push the service, and exit successfully.
If --routing-url is passed and it's a different routing url than what's currently in the registry, rover should error and require the --force flag to be passed to update the routing URL.

---

@JakeDawkins had the following to say

If --routing-url is not passed and there is already a routing url for a managed service, the API should leave it untouched

I like this a lot ^^

If --routing-url is not passed and this is the first time they’re pushing a service, we should error and tell them they need to pass the flag

I kinda like this because if they intend on using managed federation, this would be helpful to them

If --routing-url is passed and it’s the first time pushing a service, it should push the service and exit successfully.
✅

If --routing-url is passed and it’s the same routing url as what’s currently in the registry, it should warn the user that the routing url will be unchanged, push the service, and exit successfully.

I don’t see that really being necessary. Especially since it’d probably require a query on our end first to see what the url is, and even then it technically could change between the first query and the mutation to update. Unless of course, we ask for this info in the response. I just don’t think it has a ton of tangible value

If --routing-url is passed and it’s a different routing url than what’s currently in the registry, rover should error and require the --force flag to be passed to update the routing URL.

This also seems unnecessary, since I imagine changing the url is a very common thing rather than a breaking dangerous thing. I think --force makes it feel like “You shouldn’t do this, but we’ll allow it if you’re sure”

[EverlastingBugstopper]

Re-reading this, I think I stand by my proposal despite @JakeDawkins' comments, with the exception of "emit a warning if --routing-url is passed and it's the same as what's in the registry".

By default, upsert, and require a --force flag to override an old routing url with a new one.

Perhaps the Gravity folks have thought about this a bit more with Launch Control though and we should solicit input from their end.

[JakeDawkins]

related to #121

[EverlastingBugstopper]

ssssoooo, i feel kinda dumb now since the API already handles this pretty much exactly! opening a pr shortly