Re: [Gomp-discuss] Release candidate

[Lars Segerlund]

 I agree with scott, also if we do the library right we don't have to change 
it, some of the language semantics should be handled in the frontends, ( 
actually all ... ).

 If not I do think we would have done a sincerely stupid implementation. 
 / Lars Segerlund.


On Tue, 02 Nov 2004 10:21:42 -0500
Scott Robert Ladd <address@hidden> wrote:

> Biagio Lucini wrote:
> > Agreed. Moreover, a merge of the C/C++ and Fortran Specifications is in 
> > progress. It is quite likely that OpenMP 3.0 will look pretty much the same 
> > for Fortran and C. From our point of view, this means that  - in the worst 
> > hypothesis - we won't have to write more than one runtime lib :-)
> 
> The OpenMP mailing list archive is private to subscribers, but here's 
> the pertinent part for those who aren't on that list. From the October 
> 2004 newsletter:
> 
> [quote]
> * The merge of OpenMP Fortran, C, and C++ specifications is progressing
> 
> A "final draft" is now being reviewed by the committee members. There
> are still discussions on how much is flushed when a lock routine is
> called, and on private variables in Fortran statement functions.
> The ARB considers a public review period desirable after the final
> internal review, as it promotes inclusiveness and openness.
> [/quote]
> 
> For now, I'm proceeding with the 2.0 specification, since I doubt the 
> proposed changes will invalidate too much existing code.
> 
> -- 
> Scott Robert Ladd
> site: http://www.coyotegulch.com
> blog: http://chaoticcoyote.blogspot.com
> 
> 
> _______________________________________________
> Gomp-discuss mailing list
> address@hidden
> http://lists.nongnu.org/mailman/listinfo/gomp-discuss