Sean Doyle

@kamipo could you review? According to `git blame`, you reverted https://github.com/rails/rails/pull/33729 (authored by @kddnewton) and have opened https://github.com/rails/rails/pull/35116, so you might have context around inferring the association.

> Please perform checks on your side as well just to be sure. I'm happy to expand the test coverage on this branch. Could you share some (pseudo-)code to help...

Thank you for the review and merge @byroot! I hadn't had an opportunity to add coverage to ensure that this change closes https://github.com/rails/rails/issues/50782#issuecomment-1899009267. Are you able to verify that it...

@byroot also, https://github.com/rails/rails/pull/50280 is related to this issue, and ready for review.

@Edouard-chin thank you for sharing that. You're right, the callback only wraps `:setup`. I'm investigating further.

> It's not the case anymore now that we hooked into `Minitest#run`, the around callback wraps the whole test run (before and after any Minitest vanilla callbacks). > > This...

> > utilizing around blocks through [minitest-around](https://github.com/splattael/minitest-around/tree/master) the recommended 3rd party solution > > Thanks, I didn't know it was the recommended solution. Could be wrong but I find its...

> I dug a bit and Minitest doesn't make it easy, the only thing I could think of is the old `alias_method` chain trick, but it will need to be...

> This is a reluctant, but serious, suggestion. I've pushed up changes that amount to something like: ```ruby def send(name, ...) # :nodoc: if name.start_with?("test_") run_callbacks :test do capture_exceptions {...

Here's a rough benchmark: ```ruby # frozen_string_literal: true require "bundler/inline" require "benchmark/ips" gemfile(true) do source "https://rubygems.org" git_source(:github) { |repo| "https://github.com/#{repo}.git" } gem "activesupport" end require "active_support/all" class ControlTest < ActiveSupport::TestCase...