New rule: Avoid duplicate keys in Maps

[elbrujohalcon]

Avoid Duplicate Keys in Maps

Brief Description

The idea is to emit a warning in situations like the following ones (notice the duplication of keys everywhere):

Map = #{key1 => value1, key2 => value2, key1 => value3}.
OtherMap = Map#{key2 := new_value2, key3 => new_value3, key2 := new_value4, key3 => new_value5},
#{key1 := V1, key2 := V2, key1 := V3} = OtherMap.

Should be on by default?

YES

Reasoning

These are perfectly valid Erlang expressions but are often the result of a typo or some confusion from the developers and they can easily be simplified to produce exactly the same output.

Examples

-module bad.

-export [my_fun/0].

my_fun() ->
    Map = #{key1 => value1, key2 => value2, key1 => value3}.
    OtherMap = Map#{key2 := new_value2, key3 => new_value3, key2 := new_value4, key3 => new_value5},
    #{key1 := V1, key2 := V2, key1 := V3} = OtherMap,
    {V1, V2, V3}.

-module good.

-export [my_fun/0].

my_fun() ->
    Map = #{key2 => value2, key1 => value3}.
    OtherMap = Map#{key2 := new_value4, key3 => new_value5},
    #{key1 := V1 = V3, key2 := V2} = OtherMap,
    {V1, V2, V3}.

Origin (#281)

Inspired by the duplicate_key rule from https://coffeelint.github.io/

[elbrujohalcon]

You are probably thinking about the _successive maps_ rule. It’s a similar one, but different to this one.

On Tue, 28 Feb 2023 at 22:03, Paulo F. Oliveira ***@***.***> wrote: I was sure to have seen this somewhere. Was it in the formatter's issues? — Reply to this email directly, view it on GitHub <#282 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAAW3WLHEZC3TDN772DVNVDWZZR2RANCNFSM6AAAAAAVKYDDNI> . You are receiving this because you authored the thread.Message ID: ***@***.***>

-- Sent from Gmail Mobile by Brujo Benavides

[paulo-ferraz-oliveira]

Ah, you got it by email and replied 😄 My comment was "I'm sure there was something along the lines of this rule. Maybe I saw it in the formatter".

I was talking about (found it!) this: #156

[paulo-ferraz-oliveira]

(which is not the same, that one deals with assignment to different variables, from the same key)