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

From: Yitzhak Sapir (yitzhaks_at_[hidden])
Date: 2002-08-06 10:57:50

> -----Original Message-----
> From: Peter Dimov [mailto:pdimov_at_[hidden]]
> Sent: Tuesday, August 06, 2002 4:37 PM
> To: boost_at_[hidden]
> Subject: Re: [boost] Re: Re: Threads & Exceptions
> 
> Apart from that:
> 
> void join_and_delete(boost::thread * pt)
> {
>     pt->join();
>     delete pt;
> }
> 
> class oo_thread: noncopyable
> {
> private:
> 
>     shared_ptr<boost::thread> pt;
> 
> public:
> 
>     void start()
>     {
>         pt.reset(new boost::thread(bind(&oo_thread::run, this)),
> join_and_delete);
>     }
> 
> private:
> 
>     virtual void run() = 0;
> };
> 

One of the things I like about the Boost.Threads model which does not have a virtual run() function is that it makes you think about bugs like what I think is a bug in the above code, where run(), a virtual function could continue running and accessing the descendant of oo_thread after the destructor of the oo_thread's descendant has been called, but before pt has been destructed.

I do use "thread wrapper classes", but with a scoped_ptr, and I do it on a class by class basis, where operator() (my equivalent of run) is never virtual if it's in the same class as the owner of the scoped_ptr.