Christophe Bliard

Thank you @pirj for the review. No worries for the time it took. I was not very available as well.

Anything particular that can be done to make the "CI / Prism" check pass and get that PR merged?

Thanks @ydah, that fix the Prism issue thanks to the example producing a syntax error you removed 2 weeks ago. Then there was [this error](https://github.com/rubocop/rubocop-rspec/actions/runs/12692150402/job/35376820002?pr=1948) poping on "Edge RuboCop: internal_investigation"...

Yes, I also searched for reference of this in the docs and in the Cucumber scenarios but without any success. The `args` are passed to [`RSpec::Core::ExampleGroup.set_it_up`](https://github.com/rspec/rspec-core/blob/2ba0f10fe68948e5ee9addd569b4694f0245aa71/lib/rspec/core/example_group.rb#L410) where it's modified by...

Nice findings! > I have some doubts that this weird out-of-order zoo of arguments is even accepted by RSpec if the variable holds a symbol: > > ```ruby > it...

Thanks for the clear examples. > will result in the lost docstring. I suggest not to sort variables. Nice catch. > Does this change to conclusion sound reasonable? Yes, perfect....

Actually, this is not correct because of this case: ```ruby metadata = { foo: :bar } it "works", :about, metadata do end ``` If the cop blindly changes the order...

Yep, I think I prefer args too actually (and actually, after writing a lengthy response, maybe not): Let's try it: ```ruby class Something def process_that_stuff(foo: nil) service = WorkPackages::SomeService.new(user:, work_package:)...

> This pattern could also be used for the contractable callbacks. Taking `after_perform` as an example it could be called and implemented like this: > > ```ruby > # calling...

Fix the flaky spec in another PR.