New Rule: Consistent Variable Naming

[elbrujohalcon]

Consistent Variable Naming

Brief Description

This is a more sophisticated version of #266, but it's also more heuristic (thus the possibility of emitting false positives).
The idea is for Elvis to use some sound-alike algorithm to find variables with similar names, like MyVar and My_Var, or MyVar and MyVariable, or MyID and My_id, etc…

Should be on by default?

NO, unless we find a suitable default with a very low false positive rate.

Reasoning

The reasoning here is, again, similar to #266:

If two variables have the same name, it's very likely that they refer to the same concept and, therefore, it would make the lives of code readers easier if they share the exact same name, including casing and all.

Refactoring Proposal

Rename one or more of the variables.

Configuration

The rule must have some soundness_level parameter to adjust how much alike two variables need to be for them to be considered the same.

Examples

-module bad.

-export [my_fun/1, my_fun/2].

my_fun(MyVar) -> my_fun(MyVar, too).

my_fun(My_Var, too) -> {My_Var, too}.

-module good.

-export [my_fun/1, my_fun/2].

my_fun(MyVar) -> my_fun(MyVar, too).

my_fun(MyVar, too) -> {MyVar, too}.

-module also_good.

-export [my_fun/1, my_fun/2].

my_fun(MyVar) -> my_fun(MyVar, too).

my_fun(MyVar, MyOtherVar) -> {MyVar, MyOtherVar}.