Jonathan W. Zaleski

If someone would like to submit a PR to buff up the `retry_criteria_valid?` method (e.g. change the ordering to support this case) I'd be happy to review / merge if...

We copy a number of class instance-variables to the receiver in the `inherited` method. Should we maybe just copy these as well? I think @lantins is probably the best resource...

It seems like right now we support both. Perhaps we should just buff up the `inherited` method to copy over all of the instance vars we use in `resque-retry` however...

My vote would be for all or none. If `resque-retry` can function [properly] w/o the `inherited` method I would love to remove it.

I think that the second approach might actually be favorable. Unless of course there is a reliable/easy way to do the arity check up front, during the registration process. If...

My concern here was allowing callbacks to blow up within the internals of `resque-retry`. I am all for blowin up early if we can do it with 100% confidence. Perhaps...

Excellent. Glad we're in agreement.

Got `rbx-3.27` to work on El Capitan with the following: ``` brew install homebrew/versions/llvm35 rvm install rbx-3.27 -- --llvm-config /usr/local/opt/llvm35/bin/llvm-config-3.5 ``` I would assume this should work for other versions...

It looks like you can access this through the `retry_attempt` accessor which is initialized in code, but backed in `redis`, you could also reconstruct the key if necessary and query...

@alessiocancian added a decent quality of life (I feel) change with this PR, also addressed a bug w/ the `disabledIndexes` behavior on _Android_. What do you think?