Add comments to explain verification logic

lib/minitest/verify.rb

@@ -12,25 +12,33 @@ class << self
 
     def verify_fails_without(&block)
       if @current_caller
+        # If @current_caller is set, we are in the verification phase.
+        # Evaluate the block unless it is the one currently being verified.
         block.call unless caller(1..1).first == @current_caller[0]
       else
+        # If @current_caller is not set, we're in the normal test phase.
+        # Collect the caller (there might be multiple per test) and evaluate the block.
         callers << caller
         block.call
       end
     end
 
     def run
+      # If verification is disabled, run the test normally.
       return super unless Verify.enabled
 
       super
 
+      # If there are normal failures, don't run verification.
       return Result.from(self) if failures.any?
 
       begin
+        # For each caller, run the test again and verify that it fails.
         while (@current_caller = callers.shift)
           with_verification { super }
         end
       rescue VerificationFailedError
+        # If verification fails, break out of the loop.
       end
 
       Result.from(self)
@@ -45,14 +53,19 @@ def callers
     def with_verification
       yield
 
+      # Fail verification if there is an unexpected error.
+      # Encountering an unexpected error doesn't imply the test is not a false negative, it might just be broken.
       if failures.any? { |f| f.is_a?(Minitest::UnexpectedError) }
         raise VerificationFailedError
       end
 
+      # Remove all assertion failures so the failing test passes.
+      # If at least one was removed, the test is not a false negative.
       if failures.reject! { |f| f.is_a?(Minitest::Assertion) }
         return
       end
 
+      # If there were no assertion failures, we probably have a false negative.
       exception = Minitest::Assertion.new("Expected at least one assertion to fail.")
       exception.set_backtrace(@current_caller)
       failures << exception