Currently, the same goal instance implements consume and consume_cache. This has a few drawbacks, e.g., when we know that all we are going to do is to consume the cache to determine whether a goal has resolved already, we don't want to perform any setup in the __init__. However, if we continue to consume or if we only consume, then we also have more dependencies in __init__. Maybe these two things should be more separated.
Currently, the same goal instance implements consume and consume_cache. This has a few drawbacks, e.g., when we know that all we are going to do is to consume the cache to determine whether a goal has resolved already, we don't want to perform any setup in the
__init__
. However, if we continue toconsume
or if we onlyconsume
, then we also have more dependencies in__init__
. Maybe these two things should be more separated.