rescue memcached errors in a consistent way

[grosser]

@rafaelfranca

a small dry up PR without lots of strings attached ... I hope :D

[rails-bot]

r? @kaspth

(@rails-bot has picked a reviewer for you, use r? to override)

[arthurnn]

I think this is a good step. I would like to see, maybe in a near future, the memcached returning some kinda object, so who is calling can decide what is the success and error path. That would make easier to implement resilient memcached calls.
cc @jnunemaker because I know you ❤️ those conversations.

[kaspth]

When passing in false or nil, it feels like you're turning something off. Mitigate by making this a keyword argument with _to suffix (since ´fallback: false` still sounds like turning something off):

def rescue_error(fallback_to: nil)
  # ...
end

(Note, I think we can safely default it to nil)

[grosser]

I renamed it to rescue_errors_with nil to not use an extra keyword arg / allocate a hash / be simple ... looks good ?

[ArnonHongklay]

Cool!

[arthurnn]

thanks