Enforce ~consistent~ either hash syntax

[schmijos]

EnforcedShorthandSyntax: always is the default setting in Rubocop even though Bozhidar himself doesn't like the option.

I suggest to change the default option to consistent because I can see the advantage of the shorthand syntax when passing on options to deeper levels. But in any other context I found myself to be confused by it.

This change allows hashes to be in shorthand syntax only if the whole hash uses it consistently. See rubocop/rubocop#10776.

Update:
I changed my mind. Let's follow standardrb. Be consistent yourselves if you see fit.

[rnestler]

though Bozhidar himself doesn't like the option.

After reading this, I think I'd even vote to disallow it completely. But consistent style is probably enough.

[odthoma]

+1, this is also what I use in my project

[coorasse]

Can somebody explain to me the difference between consistent and either? Because I enabled the second one on vista.

[odthoma]

@coorasse

EnforcedShorthandSyntax: consistent

# bad
{foo: , bar: bar}

# good
{foo:, bar:}

# bad
{foo: , bar: baz}

# good
{foo: foo, bar: baz}

EnforcedShorthandSyntax: either

# good
{foo: foo, bar: bar}

# good
{foo:, bar:}

[coorasse]

I tried to run it with consistent on vista currently and I get the following for example:

C: [Correctable] Style/HashSyntax: Omit the hash value.
        let(:section_instance) { section.new(card_request: card_request) }

I don't want to be forced to write let(:section_instance) { section.new(card_request: card_request) }.

At the same time it does not complain about some_method(:hello, :b, card_request:, logger:).

As of now, I'd prefer either to consistent

[coorasse]

I found also an old discussion we had. Maybe irrelevant now: https://renuo.slack.com/archives/C013FEGN4D7/p1650550992375259

[coorasse]

either is also the default on standardrb: https://github.com/testdouble/standard/blob/main/config/base.yml#L1358

standardrb/standard#375
but consistent was not available back then.

[schmijos]

Switching to either, mainly because of this: rubocop/rubocop#10776 (comment)
and because of @coorasse mentioning standardrb.