Rubocop suggests using same casecmp code even with the logic flipped

[Souravgoswami]

Expected behavior

When comparing two case-insensitive strings with != or ==, rubocop suggests using less readable casecmp('...').zero? for performance reason.

There are two problems.

• First of all, if there's so tight performance concern, I'd not use zero?, I'd instead use == 0, which is much faster (and more readable to all programmers).

• The second problem is that it suggests the same code no matter if the condition is flipped. For example:

Use input.strip.casecmp('something').zero? instead of input.strip.downcase == 'something'.

2nd case where I just flipped the logic with !=:

Use input.strip.casecmp('something').zero? instead of input.strip.downcase != 'something'.

In inverted case (2nd), it suggest the exact same code as the 1st case.

Steps to reproduce the problem

The error will be raised by a code like this, even if it's inverted:

some_string.strip.downcase == 'something'

RuboCop version

$ rubocop -V
1.32.0 (using Parser 3.1.2.0, rubocop-ast 1.19.1, running on ruby 2.7.2 x86_64-linux)
  - rubocop-performance 1.14.3
  - rubocop-rails 2.15.2
  - rubocop-rspec 2.12.1
  - rubocop-thread_safety 0.4.4