[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: threadedness



I agree with Stuart that errno should follow a threaded model, because the LSB
does specify threads.  Also, there are community efforts to reduce the amount of
exceptions on Linux to the beloved :-) pthreads standard; therefore,
strengthening the need of a threaded model for errno.

http://www.linuxbase.org/spec/gLSB/gLSB/libpthread.html

Any objections?

George (gk4)

anderson@freestandards.org wrote:
> 
> Another issue that has come up as a result the 1.1 work, is wether we should
> assume a threaded model or not. In particular, this affects errno, and h_errno,
> both of which are simply macros to hide a function that obtains the thread
> specific values. The functions behind these macros have been added. If we
> assume a threaded model, then errno and h_errno are no longer interfaces, but
> become preprocessor symbols instead.
> 
>         #define errno (*_errno_location())
> 
> Since we are specifying pthreads, I think that we should assume a threaded
> model, provide these symbols as #defines instead of externs.
> 
> Opinions?
> 
>                                 Stuart
> 
> anderson@metrolink.com                         anderson@freestandards.org
> Metro Link, Inc.                               Free Standards Group
>                                                Linux Standards Base
> 
> --
> To UNSUBSCRIBE, email to lsb-discuss-request@lists.linuxbase.org
> with subject of "unsubscribe". Trouble? Email listmaster@lists.linuxbase.org

-- 
George Kraft IV
gk4@austin.ibm.com



Reply to: