Overview
Currently, the MVVM Toolkit source generators are propagating the cancellation token in each callback to all calls that need one, but they're not checking for cancellation themselves that much. This can increase overhead when concurrent invocations are triggered, as previous runs of the generators will take longer before they get to a point where they can detect they've been cancelled. We should update them to add more cancellation checks in places that make sense.
Overview
Currently, the MVVM Toolkit source generators are propagating the cancellation token in each callback to all calls that need one, but they're not checking for cancellation themselves that much. This can increase overhead when concurrent invocations are triggered, as previous runs of the generators will take longer before they get to a point where they can detect they've been cancelled. We should update them to add more cancellation checks in places that make sense.