Subj : Re: I write a multithread program with pthread, and meet a program
To   : comp.programming.threads
From : Giancarlo Niccolai
Date : Wed Apr 20 2005 11:13 am

David Butenhof wrote:

> That is, they may intend that you disable
> cancelability around any area of code that might do this; and while you
> can argue which makes most sense there's no objective authority to say
> they're wrong.

Well, I never stated about correctness of one or the other possible
solution. I just stated that as long as cout uses write(), and unless c++
makes a future statement about the contrary, cout will be a cancellation
point. Assuming that cout absolutely uses write() is arbitrary, and
probably should not be taken for granted; what I meant is just that IF cout
uses write(), and if C++ standard doesn't specify for the contrary, then
cout will be a cancellation point.

Since, at least atm, on posix compliant systems using write() seems the most
reasonable way to implement cout, then I would suggest to prepare c++ code
to prevent the case in which, if cout is a cancellation point, the program
misbehaves. That is, cout must not be taken for a sure cancellation point
(i.e. you should add your own cancellation point anyhow, or avoid using
cout if you wish to be cancelled while writing), yet is higly advisable to
prevent a program to fail in case it is via cancellation disabling or
cleanup setting. Hope my point makes sense...

Best regards,
Giancarlo Niccolai

.