While debugging general performance/reliability issues in our test suite, I thought I would see what effect disabling Billy's cache had on our test suite. (We don't intentionally use it.)
So I configured with cache = false, but was still surprised to see log entries about calculated cache keys for requests.
It appears ProxyHandler was unconditionally calculating the scope and key, the latter of which seems expensive to perform.
Admittedly, my understanding of the overall system is minimal, but it seems like this change wouldn't have any unwanted side effects, and skip unused method calls for users who have cache = false.
While debugging general performance/reliability issues in our test suite, I thought I would see what effect disabling Billy's cache had on our test suite. (We don't intentionally use it.)
So I configured with
cache = false
, but was still surprised to see log entries about calculated cache keys for requests.It appears ProxyHandler was unconditionally calculating the scope and key, the latter of which seems expensive to perform.
Admittedly, my understanding of the overall system is minimal, but it seems like this change wouldn't have any unwanted side effects, and skip unused method calls for users who have
cache = false
.