Synthetic call-forwarding API comparison
This is a controlled comparison using a disposable test account. Every account identifier, user identifier, extension identifier, telephone number, timestamp, and request identifier below is a placeholder. The purpose is to document the expected API state transitions without exposing customer or production data.
The comparison covers the older extension call-handling route and the current user call-forwarding route. The useful distinction is between adding a destination to the user’s available call-handling choices and activating forwarding for the entire number. Those outcomes look similar in a request body but are different in the desktop client and in a subsequent read request.
For the setup, assume one licensed test user with an assigned direct number, business hours enabled, no holiday override, and no pre-existing external contacts. The OAuth application belongs to the same disposable account and has the administrative Zoom Phone read and write scopes. No delegated or cross-account access is involved.
Before making a change, record the current business-hours configuration, the direct-number assignment, the user’s extension identifier, and the result returned by the call-forwarding read route. This establishes whether a later difference comes from the write operation, an inherited site policy, or a client display delay.
For the legacy comparison, the request identifies the extension and business-hours setting, then supplies one external E.164 destination with a short label. A successful test should produce a visible forwarding banner and a control that lets the user disable the active forwarding state. The destination value used here is synthetic and cannot receive calls.
For the current comparison, the request identifies the user and the business-hours period, enables one call-forwarding entry, and supplies the same synthetic E.164 destination. Empty holiday and break identifiers are omitted rather than sent as account data. The external-contact object is included only when a disposable contact was created for the test.
A transport success is recorded separately from a state change. After each write, wait for the documented propagation interval and issue a fresh read instead of relying on a cached client view. Capture the response status, request identifier, returned enable value, destination value, and ordering of the call-handling entries.
The desktop-client check should be performed only with the disposable user. Sign out and back in once, then compare the banner, the call-handling list, and the forwarding toggle. This helps distinguish an API persistence issue from a stale local view. No real user should be asked to place or receive a call for this comparison.
The web-portal check should use a new browser session and the same disposable account. Verify the business-hours selector, confirm that no holiday or closed-hours rule is active, and note whether the destination is shown as an available choice or as the currently active route. Screenshots should redact all identifiers before sharing.
The readback sequence should use the same user identifier that was supplied to the write. If the endpoint returns a destination but the client does not show active forwarding, compare the enable flag and the applicable hour type. If the endpoint omits the destination, compare the write response and request identifier before retrying.
The test should also distinguish direct-number forwarding from extension call handling. A route that merely adds an option may be functioning as designed even though it does not reproduce the legacy banner. The open question is which supported operation activates forwarding for the complete user and number while preserving the user’s ability to disable it.
Potential configuration differences include an account-level policy that locks call handling, a number assigned through a site rather than directly to the user, a destination that requires an external-contact record, or a business-hours schedule inherited from the site. Each condition should be changed separately so the result remains attributable.
For error handling, record validation responses without repeatedly submitting the same write. A permission error should be checked against the token’s account and scopes. A not-found response should be checked against the user and extension mapping. A success response followed by unchanged state should be investigated through the request identifier.
A minimal sanitized request body should contain only the enable flag, placeholder destination, placeholder label, and any required external-contact identifier. Optional empty fields should be removed for a second comparison. This makes it easier to determine whether an empty override field changes how the route interprets business hours.
The expected outcome for this synthetic test is a precise explanation of whether the current route activates forwarding or only configures an available destination. If a different route is required, the useful answer would include its hour-type semantics and the supported way to restore normal routing afterward.
The following checklist keeps the comparison reproducible: confirm the disposable license, confirm the assigned number, capture the initial read, submit one legacy-form comparison where still available, restore state, submit one current-form request, wait once, perform a fresh read, refresh the client, and restore the original test configuration.
No credentials, access tokens, customer identifiers, real telephone numbers, recordings, call logs, or personal information belong in this diagnostic. Placeholder values should remain visibly marked, and the final state of the disposable user should be restored after the comparison is complete.