Cop idea: Switch regex groups to non capturing if not needed

[choosen]

Is your feature request related to a problem? Please describe.

We don't need to use captured groups using ruby methods for matching content against regex or splitting. Using them will affect badly on performance of RE.
Exception is the back references in regex, so for sure we need regexp parser for that.

# bad
'str'.match? /(?<X>s)/
# => true

# bad
'str'.match? /(s)/
# => true

# good
'str'.match? /s/
# => true

Describe the solution you'd like

Add a cop to detect this problem, maybe also autofix ?

Describe alternatives you've considered

It can be optimized on ruby regex library without cop ? https://github.com/k-takata/Onigmo

Additional context

During the splitting we also don't check/fetch groups so it can be more beneficial to omit groups

It's not only regexp optimization because it is bound with ruby method using this regexps

require 'benchmark/ips'

regexp1 = %r{\s([\&/]\s)?}
regexp2 = %r{\s(?:[\&/]\s)?}
string = 'HR / People & People Ops'

Benchmark.ips do |x|
  x.report("with group") do
    string.split regexp1
  end

  x.report("without capture") do
    string.split regexp2
  end

  x.compare!
end

Calculating -------------------------------------
          with group    495.069k (± 2.7%) i/s -      2.496M in   5.045809s
     without capture    555.395k (± 1.4%) i/s -      2.797M in   5.037793s

Comparison:
     without capture:   555395.2 i/s
          with group:   495069.1 i/s - 1.12x  slower

[choosen]

somehow related with #7746

[stale]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contribution and understanding!

[stale]

This issues been automatically closed due to lack of activity. Feel free to re-open it if you ever come back to it.