Autocorrect of HashSytnax does not work in some case

[iblue]

Given the following file

FactoryGirl.build :user, :email => "[email protected]"
FactoryGirl.build :user, :email=> "[email protected]"

When I run rubocop --only HashSyntax --auto-correct then I get

Inspecting 1 file
C

Offenses:

example.rb:1:26: C: [Corrected] Use the new Ruby 1.9 hash syntax.
FactoryGirl.build :user, :email => "[email protected]"
                         ^^^^^^^^^
example.rb:2:26: C: [Corrected] Use the new Ruby 1.9 hash syntax.
FactoryGirl.build :user, :email=> "[email protected]"
                         ^^^^^^^^

1 file inspected, 2 offenses detected, 2 offenses corrected

But it does not correct :email=> "[email protected]". I found the comment at https://github.com/bbatsov/rubocop/blob/master/lib/rubocop/cop/style/hash_syntax.rb#L41 which says that it does not autocorrect in this cases, but I don't understand why.

[jonas054]

If HashSyntax and SpaceAroundOperators were to change the code simultaneously, it would lead to destroyed code. The special logic was added in this commit.

But you've exposed a bug. HashSyntax avoids making its correction if it thinks that SpaceAroundOperators is about to make a change. It does this by looking at configuration. It's just that --only doesn't disable all other cops by changing their configuration, so this is not the right way to find out if SpaceAroundOperators is active.

Actually I think there's another way to solve the problem with interference between cops now. There's a loop around auto-correction that picks up exceptions caused by conflicting corrections and tries to do them one by one instead. This will work if we make the correction ranges larger so that they cause "clobbering".

I will try to fix this.

[iblue]

👍 Great! Thank you very much!