Re: [Chicken-users] trying to gracefully stop threads

[Michael Greenly]

Fortunately I can't imagine I would ever actual use thread-sleep! like this, rather the thread would be waiting on I/O, so it appears I have no obstacles.  

Thank you for taking the time to look at it.

On Sat, Dec 7, 2013 at 2:16 PM, Christian Kellermann <address@hidden> wrote:

* Christian Kellermann <address@hidden> [131207 21:11]:

> Hi Michael,
>
> * Michael Greenly <address@hidden> [131207 20:23]:
> > My assumption is that the thread would be started and recurs until done
> > becomes false at which time that thread would exit cleanly and allow the
> > join in the primary thread to continue.
> >
> > The 'done' flag only becomes false when INT is signaled and handled
> >
> > Instead I get a "Error: uncaught exception: #<condition:
> > (join-timeout-exception)>".
>
> This may be a bug you have found. If you run the example in csi
> withtout the signal handler and set done to #t it works as expected.
>
> However as I see the thread is found in a "blocked" state when the
> signal handler gets called. This in turn causes the timeout condition
> to be raised. This shouldn't happen as I understand it, as the user
> clearly didn't set any timeout.

Ah well the blocked state probably originates from being in
thread-sleep!...  Removing that makes the example work as expected
modulo the busy loop...

So the question is how do we determine whether we actually ran into
a timeout...

(Am I thinking out loud again?)

--
In the world, there is nothing more submissive and weak than
water. Yet for attacking that which is hard and strong, nothing can
surpass it. --- Lao Tzu

--
Michael Greenly
http://logic-refinery.com