Re: [boost] Re: Boost.Threads draft library submission

From: Beman Dawes (bdawes_at_[hidden])
Date: 2001-03-15 16:20:03

At 03:18 PM 3/15/2001, williamkempf_at_[hidden] wrote:

 >I haven't missed the problem.  The C++ "volatile" with platform
 >specific memory barriers (a concept I've been considering for a later
 >phase of Boost.Threads) will make "atomic_t" compliant on platforms
 >that must have the poor "mutex implementation".  In practice I expect
 >we'll never really use the poor implementation.

Here is a question to ask yourself.  Does the platform's native API provide 
a call which correctly performs the function desired?  Does that function 
work correctly when called by various C++ programs?

If the answer is yes (as it usually is on modern operating systems), then 
presumably compilers have used the licence granted by the C++ standard's 
1.9 paragraph 6 to treat that function as "observable behavior".  If they 
didn't, you would not be able to safely call the various system API's from 
C++.

So as long as a C++ program relies only on that "observable behavior" I'm 
assuming we are OK.

For a system which doesn't guarantee even its own API's to work from C++, 
then all bets are off.

Am I missing something?

--Beman