http: prefer to return reference rather than shared ptr for clusterInfo()

[wbpcode]

Commit Message: http: prefer to return reference rather than shared ptr for clusterInfo()
Additional Description:

See #44025 for more detail.

Risk Level: low.
Testing: n/a.
Docs Changes: n/a.
Release Notes: n/a.
Platform Specific Features: n/a.

[ravenblackx]

Shouldn't this one be split into OptRef and SharedPtr variants too?

[wbpcode]

I will prefer to make this be another PR because we need evaluate the usage of the cluster() before changes.

[ravenblackx]

Shouldn't this one be using the optref variant since it's function-scoped?

[wbpcode]

/retest

[ravenblackx]

Suggested change

	.WillByDefault(Invoke([&callbacks]() -> OptRef<const Upstream::ClusterInfo> {
	return makeOptRefFromPtr<const Upstream::ClusterInfo>(callbacks.cluster_info_.get());
	}));
	ON_CALL(callbacks, clusterInfoSharedPtr())
	.WillByDefault(Invoke([&callbacks]() -> Upstream::ClusterInfoConstSharedPtr {
	return callbacks.cluster_info_;
	}));
	.WillByDefault([&callbacks]() -> OptRef<const Upstream::ClusterInfo> {
	return makeOptRefFromPtr<const Upstream::ClusterInfo>(callbacks.cluster_info_.get());
	});
	ON_CALL(callbacks, clusterInfoSharedPtr())
	.WillByDefault([&callbacks]() -> Upstream::ClusterInfoConstSharedPtr {
	return callbacks.cluster_info_;
	});

(Invoke is redundant in modern gmock.)

[wbpcode]

Although we can skip the Invoke, but because Invoke be used every where in the test, it may make sense to keep consistency?

[ravenblackx]

That's how you end up with keeping on replicating outdated code style forever. :)

Not too bothered in this PR though if you want to leave it. Maybe I should have an AI go through and remove all the mock Invokes repo-wide so we can stop cargo-cult-duplicating more of them in. :)

[ravenblackx]

This will still require a change if info gets updated to returning a ref (because this will try to make a copy), so it might be better to just put it all in one very long line for future compatibility, but at least this way it'll be a compile-time error rather than a thing that can be accidentally left "the wrong way". :)

[wbpcode]

/retest