Guideline proposal: Use proc.call() over proc[] or proc.()

[dirkbolte]

There are three different ways to call a proc or lambda:

l = ->(v)  { puts v }
l[1]
l.(1)
l.call(1)

According to http://www.ruby-doc.org/core-2.0/Proc.html , they all seem to be equivalent.

l[1] looks like addressing an array index instead of calling a method, therefore it does not do what I would expect from reading the code.
´l.(1) looks like a method call where the method name is missing. When I saw it for the first time, it was hard to understand it. This way of calling a proc is just mentioned as syntactic sugar in two method documentations, so if you are not familiar with this construct, it is most likely just confusing and thereby impacts readability.
l.call(1) is pretty easy to understand. Although it uses more characters than the other two examples, I prefer to have easily readable code.

[fuadsaud]

Hum, I like #(), but I agree that #call is more readable.

[equivalent]

l.call(1) looks more object oriented, where l.(1) looks more like something taken from Brainf**k language

so +1

[lee-dohm]

👍 I agree that #call is the most readable and is probably the best choice in the common case.

[bbatsov]

👍 I'd suggest creating a PR for that rule.