Subj : Re: pthread_cancel
To   : comp.programming.threads
From : ptjm
Date : Tue May 10 2005 04:14 am

In article <d5nbq1$mqp$1@balena.cs.interbusiness.it>,
Giancarlo Niccolai  <noone@nothing.it> wrote:

% that it calls or calls not write() to perform an output to a stream. I
% don't have news about printf being explicitly declared NOT a cancellation
% point by posix, so, what happens inside printf is to be considered
% "random".

POSIX contains a list of functions in which cancellation points shall
occur and a larger list of functions in which cancellation points may
occur. It says explicitly that `An implementation shall not introduce
cancelation points into any other functions specified in this volume of
IEEE Std 1003.1-2001.', so don't assume a function can be a cancellation
point based on not seeing language to the contrary (although in this
case, printf is in the second list, so you're correct that it may or
may not be a cancellation point).

-- 

Patrick TJ McPhee
North York  Canada
ptjm@interlog.com

.