[ruby-core:65051] Re: [ruby-trunk - Bug #10231] Process.detach(pid) defines new singleton classes every call

[e-mail] wrote:
> FWIW, JVM always does a waiter thread to avoid multiple calls to
> waitpid (which as you know returns results exactly once). The waiter
> thread is not unusual in that regard.

Right.  The problem I want to solve is when independently-maintained
pieces of code have conflicting calls to waitpid:

	th = Thread.new { Process.waitpid(-1) }

	p Process.waitpid(fork {})
	p th.value

The above behaves inconsistently depending on scheduling.  Presumably
the thread waiting on a single pid should beat the thread waiting on all
threads.  And it may not be solvable for waiting on process groups...

So the above combination of waitpid calls is likely broken no matter
what.