RE: [boost] Re: Boost.threads: interthread exceptionconveyance

From: Darryl Green (green_at_[hidden])
Date: 2002-08-27 20:11:25

> -----Original Message-----
> From: Alexander Terekhov [mailto:terekhov_at_[hidden]]
> 
> 
> Darryl Green wrote:
> > 
> > I still don't get it. What exactly is the point you are 
> trying to make? Are
> > you suggesting that one should or shouldn't, 
> 
> I'm suggesting that one shouldn't (no matter whether it can 
> be done or not,
> "conformance" aside for a moment).
> 

Ok - I understand that suggestion - and I agree to the extent that I don't
think it is always a good thing to do. I just happen to think it is not
always a bad thing to do.

> > can or can't write a terminate
> > handler that terminates the thread but not the process and 
> is such a program
> > conforming or not?
> 
> Well, let me put it this way: feel free to fool yourself into 
> believing that 
> it's somehow going to make your application safer/robuster 
> ("better"), using 
> the following {anti-}"pattern": 
> 
> < illustration only; NOT "real" code, but it doesn't screw up 
> terminate(),
>   "at least" >

Screw up terminate()? This is where I don't understand your posts - I can't
see that there is anything defined about what happens after the call (except
that it won't return) so how can it be "screwed up"?

> 
> extern "C" void* thread_routine_wrapper( void* arg )
> {
>   try {
>     return call_something( arg );
>   }
>   catch( const pthread_cancel_e& ) { throw; }
>   catch( const pthread_exit_e&   ) { throw; }
>   catch( ... ) {} // ignore 
>   return "HI_THERE_COMMA_TAKE_THIS_UNCAUGHT_EXCEPTION";
> }
> 

Thanks. I will (sometimes).

> regards,
> alexander.
>