Re: [boost] Re: Re: Threads & Exceptions

From: William E. Kempf (williamkempf_at_[hidden])
Date: 2002-08-13 17:02:55

From: "Victor A. Wagner, Jr." <vawjr_at_[hidden]>
> >Again, I don't understand your thinking.  I don't see how the issue at
hand
> >(unhandled exceptions in the main thread resulting in terminate() being
> >called, possibly with out thread unwinding) can be solved by a set/test,
> >atomic or not.  I also don't see how synchronization plays a part here
> >either.
>
> The ONLY complaint I've heard you voice is "race condition"... later you
> mentioned "without language support"
> This may not directly bear on something _I_ never considered a problem in
> the first place...attempting to propagate an exception out of the "main
> program".  I thought it bore on your worries about polling and "race
> condition".

Victor, I'm afraid you've been mixing arguments here.  I never claimed that
the "race condition" problem could be solved through language support, nor
was the "race condition" involved in any of the discussion surrounding the
invocation of undefined behavior.  At this point, I think we'd be much
better to drop the argument, since I'm not sure we can recover from these
misunderstandings.

Bill Kempf