Re: [boost] Boost.Threads, N2178, N2184, et al

From: Roland Schwarz (roland.schwarz_at_[hidden])
Date: 2007-03-26 16:43:57

Peter Dimov wrote:
> (Internally there is a flag being set, of course.) 

Isn't this simply an implementation detail?
What I think cancel is: instruct the thread to take a
different route. (In our case one that ends it.)

Possibly seeing a thread as an object is not the
very best model. A thread when viewed in this manner
never can be const. It's instruction pointer usually
continuously is changing.

Roland