A function deployment workflow includes a runtime, an entry point, a trigger and source content. Those settings describe how the eventual handler should be hosted. They don't prove that the handler processes normal events, boundary cases or retries correctly.
These tasks save a gen2 configuration using a fictional archive URI. The browser doesn't fetch that archive or execute Google Cloud function code. Think separately about workload identity, dependencies and repeated events. A retry may arrive after an earlier side effect completed, so the implementation needs a stable business identifier and an idempotency strategy. Production completion requires a valid package and independent behavior tests.
Separate hosting configuration from handler and retry behavior.