Re: XSH/XRAT conflict over cancellation points

[Dave Butenhof <yyyyyyyyyyyyyy@xxxxxx>]

Alexander Terekhov wrote:

> > Should that be filed as a defect?
> >    
> The real defect here is that POSIX still doesn't defer to 
> ISO/IEC 14882. Thread cleanup handlers shall be defined 
> semantically equivalent to dismissible C++ "scope guards" 
> of the local scopes formed by pthread_cleanup_push/pop() 
> and thread exit/cancel shall be defined semantically 
> equivalent to unspecified C++ exception (POSIX.C++ profile 
> shall specify them, of course). No brain-damaged volatiles 
> needed.
>  
Alexander, you make it progressively more difficult to compose civil 
answers to your foolish tirades. One of these days I'll just give up on 
you entirely.

You are welcome to undertake the effort of creating and driving a POSIX 
workgroup to develop a C++ binding to POSIX. Really. But you should know 
by now that POSIX 1003.1 is not that binding. It does not reference C++ 
and cannot, and your comment is completely irrelevant and meaningless. 
Even if there is at some time in the future a C++ binding, 1003.1 still 
will not reference ISO/IEC 14882. So just drop it.

We are chartered and limited here with defining the behavior of the 
POSIX interfaces as defined for the ISO/IEC C language. That's why we 
don't have exceptions. When this silly limitation of the C language is 
lifted (ha), we can fix up the definition of POSIX cancellation.

Finally, "volatile" is NOT in any way brain damaged. The volatile 
keyword has a clear definition, and there are many cases in which it's 
both necessary and sufficient. It is simply neither necessary nor 
sufficient for sharing a variable between concurrent execution streams. 
That some persist in attempting to use it that way is not the fault of 
'volatile' nor of the committee that defined it, but of the individuals 
who have either willfully or ignorantly misinterpreted the standards.

--
/--------------------[ yyyyyyyyyyyyyy@xxxxxx ]--------------------\
| Hewlett-Packard Company       Tru64 UNIX & VMS Thread Architect |
|     My book: http://www.awl.com/cseng/titles/0-201-63392-2/     |
\----[ http://homepage.mac.com/dbutenhof/Threads/Threads.html ]---/