Confused if the behavior of when_top_level_zoneless_env_zoned_single_route_empty is actually intended. It essentially has workers_dev = true in the top level of the toml but then uses an environment where workers_dev is not set. Should the environment inherit the top level workers_dev and then create a zoneless (workers.dev) route or create a zoned route since the environment doesn't have wrangler_dev set?

@jspspike - looking at this now. An aside, but relevant to your PR... would you add a commit that updates this error message:

⚠️ No depolyment routes specified, worker will not be triggered. Please specify your deployment routes or set wrangler_dev = true inside of your configuration file in order to trigger your worker. For more information, see: https://developers.cloudflare.com/workers/cli-wrangler/configuration#keys

So that instead of wrangler_dev = true it is corrected to print workers_dev = true?

https://github.com/cloudflare/wrangler/blob/4e18717d7c632335db8198252d311c40d95d67d8/src/settings/toml/tests/deployments.rs#L808-L828

The test you mention (embedded above) generates a wrangler.toml manifest with the following configuration:

name = "top_level_zoneless_env_zoned_single_route_empty"
type = "webpack"
account_id = "fakeaccountid"
workers_dev = true

[env.test]
route = ""
zone_id = "samplezoneid"

Looking at this, I would assume the intention of these settings is to have no Zoned route, since there is no route defined despite the zone_id being present, but to inherit the workers_dev key into the environment. I'm not sure how the script naming works there (maybe append -test to the value in name?).

I assume this should be inherited based on the way we document the configuration, seen here:

Maybe @Electroid could comment on this assumption - and if we're in agreement, I think the test might need to be changed to expect 1 zoneless route (the workers.dev test route)? This would also mean some code is changed somewhere to inherit this top-level workers_dev key.

Going off of the docs linked I made workers_dev get passed down to environments. This made some tests need to be changed where both zoned and zoneless routes would be created if the environment got workers_dev = true from the top level and also had a zone/routes in the environment.

I went through the changed tests and tried to verify that the behavior is intended, but any changed tests should be looked at closely by the reviewer to make sure some intended behavior isn't changed since this will affect a lot of things in terms of what routes are deployed on publish

I tested this with some variations and it all looked fine!