Layout/DotPosition with EnforcedStyle: trailing false positive

[LelandAM]

This commit that updated support of heredocs does not support our usage of them.

We call a method with a heredoc as an argument and call methods on the result.

For example, prior to this commit this code did not register an offense but after it does.

method_call(
  <<~HEREDOC
    something
  HEREDOC
).method_name

This also has changed behavior for multi-line method calls with non-heredoc arguments

method_call(
  'foo'
).method_name

---

Expected behavior

When the below spec is added to the "Trailing dots style" context in dot_position_spec.rb we would expect it to pass.

it 'does not register an offense with method calls after the heredoc argument' do
      expect_no_offenses(<<~RUBY)
        something(
          <<~HERE
            something
          HERE
        ).method_name
      RUBY
    end

Actual behavior

The spec fails

Steps to reproduce the problem

Run rubocop on a file containing either of these snippets

method_call(
  <<~HEREDOC
    something
  HEREDOC
).method_name

or

method_call(
  'foo'
).method_name

RuboCop version

$ [bundle exec] rubocop -V
1.22.1 (using Parser 3.0.2.0, rubocop-ast 1.12.0, running on ruby 2.7.2 x86_64-darwin20)
  - rubocop-performance 1.11.5
  - rubocop-rake 0.6.0
  - rubocop-rspec 2.5.0

[dvandersluis]

This might be fixed by #10168, can you try getting the HEAD of rubocop and running it?

[LelandAM]

Thanks for getting back! I saw that, tried testing on HEAD and still had the problem.

[pfeffer]

We're seeing the same issue - throwing Layout/DotPosition warning on things like:

var = {
   key: 'some', 
   key1: 'other'
}.compact

[dvandersluis]

#10207 was merged today, all the examples in this issue are fixed now!

[hajee]

Is a release containing this fix planned for shortly? I'd love to use the update instead of changing all Gemfiles.

[koic]

Yeah, the next bugfix release is planned. Please wait a moment. Thank you.

[bbatsov]

I just released 1.22.3.