[Openmcl-devel] Unix signal handling

[Gary Byers]

On Wed, 7 Jul 2010, Scott L. Burson wrote:

> On Tue, Jul 6, 2010 at 11:39 PM, Ron Garret <ron at flownet.com> wrote:
>
>       So my model of how thread scheduling works must be badly broken.
> 
> 
> There's some relevant information in section 16.1 of the manual, though it
> doesn't quiiiite seem to answer the question for me either -- since the GC
> has access to the global TCR list, and has suspended all the other threads,
> why can't it track down the exception context of any thread that was in an
> exception handler and find the relevant roots?  But I can easily imagine
> that there is a complicated, uninteresting answer :-)
> 
> -- Scott
>

The 3rd paragraph of section 16.1.2 is as lucid as it is fascinating, so it
couldn't be what you're referring to.  (It's actually describing what happens
on a synchronous signal - as caused by an exception - but the same things
have to happen on an asynchronous signal as well.)