We have multiple testnet runtimes configured for the node. A new pallet I'm building is only available for testing in one testnet. There is also a new custom RPC I have written to call into the new pallet. How do I prevent errors caused by some of the runtimes not implementing the RPC interface because they don't have the new pallet included/configured when building the node?
I wonder if cargo features might help here, but not sure how to use it