[mmporong]

Process::Status#success? returns nil when the process has not exited, for example when it was killed by a signal. The RDoc comment above the signature already lists that case (success? -> true, false, or nil), but the return type was bool:

pid = spawn("sleep 5")
Process.kill(:KILL, pid)
_, status = Process.wait2(pid)
status          # => #<Process::Status: pid 2166906 SIGKILL (signal 9)>
status.success? # => nil

This changes the return type to bool? and adds Process_StatusInstanceTest#test_success?, which checks the true (exit), false (exit 1) and nil (child killed with SIGKILL) cases with assert_send_type.

Tested with Ruby 4.0.9 (ruby -Ilib bin/test_runner.rb test/stdlib/Process_test.rb, with the extension compiled from this checkout):

	Result
new test, current signature	1 failure: Call trace does not match one of method definitions at the () -> nil assertion
new test, () -> bool?	4 tests, 65 assertions, 0 failures

rbs validate passes, and rubocop (with rubocop-on-rbs and the stdlib test cop) reports no offenses on the two changed files.

AI disclosure: this change was prepared with Claude Code (Claude Opus 5.5), which compared the signature with the RDoc and the runtime behavior, made the change, wrote the test and ran the checks above.

[ksss]

Thank you!