Redis [QUESTION] Are replica reads supposed to work with client side caching

I've done a dive into the code, and don't see anywhere invalidations would be triggered due to a replica update, is this true, that they don't get triggered when a replica updates from its master? Or did I just miss where this is happening? (Tried to trace every path that leads to trackingInvalidateKey* functions).

Comment From: zuiderkwast

Have you tried it? It shouldn't be too hard to check it by modifying a test case.

Comment From: sjpotter

you're right, tested and it works as expected. Now to not only believe that it works, to understand the code why it works.