[ofiwg] ofiwg item: supporting other OS's

Jeff Hammond jeff.science at gmail.com
Fri Jan 8 21:03:30 PST 2016


On Fri, Jan 8, 2016 at 8:36 PM, Jason Gunthorpe <
jgunthorpe at obsidianresearch.com> wrote:

> On Fri, Jan 08, 2016 at 07:47:38PM -0800, Jeff Hammond wrote:
> >     C11 support requires gcc 4.9 for full compiler features (yes, I know
> >     the library side is missing) and that is now nearly a two
> >     year old compiler.
>
> > Does GCC 4.9 support atomics? I don't have a binary handy...
>
> Claims to, it is very similar to the C++11 atomics which are there.
>

The preprocessor macro claimed to support them at least a year before they
did.  The bug report was rejected because the website noted the lack of
feature support.  Apparently, documenting bugs makes them no longer bugs in
the GCC world :-\  I no longer take GCC seriously...


> 4.9 is a significant compiler because it is the first g++ release that
> really supports C++11 and all the polishing in C++14 very well.
>
> I think we will see it or a slightly newer one made widely available
> because of that. I already am working on C++ code bases that are
> incompatible with 4.8 because of that, and it is unquestionably worth
> it for C++.


C++ programmer love showing off how smart they are by using the latest and
greatest-according-to-Herb-Sutter features at the expense of platform
portability.  "Hey look, my new C++ code is so awesome that it can only be
compiled with the Clang trunk!" :-)

On the other hand, C programmers love showing off how smart they are by
writing code that runs everywhere and getting stuff done, without having to
wait for 25 compiler bugs to be fixed.


> > I don't know that the group will go for anything past C99, but we should
> try to
> > enumerate the list of C11 features we would use if it were permitted.
>
> Realistically public header files probably have to remain at a C89/C99
> compatible level, which puts a crimp on a lot of new features :(
>
> What features?  I'm still not clear on what C11 features you find
essential.

We should not need atomics or threads in the public API.

Generic selection?  Anonymous unions?

Jeff



-- 
Jeff Hammond
jeff.science at gmail.com
http://jeffhammond.github.io/
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openfabrics.org/pipermail/ofiwg/attachments/20160108/7e4faa76/attachment.html>


More information about the ofiwg mailing list