Simplify for anonymous functions?

[paulo-ferraz-oliveira]

I found myself doing a lot of these, today:

fun(Pattern) -> is_match_node(Pattern) end

becoming

fun is_match_node/1

I wonder how difficult it'd be to implement a "simplify it" rule for this...

[elbrujohalcon]

Yes!!! This is a great rule idea!
The algorithm should emit a warning if it finds an anonymous function where it's body is just a call to another function with the same parameters as the anonymous function in the exact same order.
I think that's achievable. We might miss some opportunities for simplification (e.g. if someone wrote fun(X, Y) -> erlang:apply(a, b, [X, Y]) end), but that's good because it allows people to work-around the rule if they really really want to avoid using -elvis(…). :P

[paulo-ferraz-oliveira]

#461