[cfe-commits] [PATCH] First OpenMP patch

[Douglas Gregor]

On Dec 17, 2012, at 6:42 AM, Chandler Carruth <chandlerc at google.com> wrote:

> On Mon, Dec 17, 2012 at 5:45 AM, Hal Finkel <hfinkel at anl.gov> wrote:
> ----- Original Message -----
> > From: "Dmitri Gribenko" <gribozavr at gmail.com>
> > To: "Alexey Bataev" <a.bataev at gmx.com>
> > Cc: "Hal Finkel" <hfinkel at anl.gov>, cfe-commits at cs.uiuc.edu, "mahesha llvm" <mahesha.llvm at gmail.com>, "benny kra"
> > <benny.kra at gmail.com>
> > Sent: Monday, December 17, 2012 6:45:07 AM
> > Subject: Re: [cfe-commits] [PATCH] First OpenMP patch
> >
> > On Mon, Dec 17, 2012 at 11:48 AM, Alexey Bataev <a.bataev at gmx.com>
> > wrote:
> > > Dmitry, Hal,
> > > Thank you for your comments and sorry for the delay, I was on a
> > > vacation.
> > > I've made some fixes according to your comments.
> > > Hal, I've change the sentence in the doc. Now there is only one
> > > warning, if
> > > any OpenMP pragma is found.
> > > Dmitry, I've changed the processing of the -fno-openmp flag. If
> > > -fno-openmp
> > > is specified, the option -Wno-source-uses-openmp is passed to the
> > > frontend.
> > > Option -Wsource-uses-openmp is on by default.
> >
> > Hello Alexey,
> >
> > This patch looks good to me.
> 
> This also looks good to me. In nobody objects in the next day or so, please commit.
> 
> Sorry that this got lost Hal, but I have said on another thread about this patch (but with a different author) that I don't really think we should add documentation and the beginnings of support for -fopen-mp without first having a clear discussion and document describing the expected design of OpenMP support in Clang.
> 
> Essentially, I think this patch is starting ast step 2, 3, or 4 rather than step 1.
> 
> I think we're actually imagining OpenMP working generally the same way these days (based on the discussion we had on the aforementioned thread -- another reason I really dislike forking threads), but I'd like to get that documented, and a general roadmap of how the support is going to be added to Clang and by whom laid out first. And I think based on *that* discussion, Doug needs to sign off on OpenMP as a viable language extension for Clang to support. I think it is viable, and would support it, but it's not obvious under the current rules. Does this make sense to folks? I think it's the next step.

OpenMP is something that comes up time and time again, and I think it's fairly clear that it's a reasonable language extension for Clang to support. Given our previous discussion threads, I think we do have a basic shared understanding of how the implementation should proceed. However, that does need to be captured in a design document to be discussed on cfe-dev.

> Finally, I would point out that I still have some concerns over the contributors of OpenMP support not being significant contributors to Clang in general, and not contributing to the support and maintenance burden of the system as a whole. I have not yet seen significant changes there, although I'm hopeful they'll be forthcoming. I think that is an essential component to getting these patches in and adding OpenMP to the Clang.

I don't agree with this. As with any contributor looking to do non-trivial work, we'll need to discuss more of the design up front, so we're certain that we're heading in the right direction, and we will be reviewing patches very carefully pre-commit to make sure that what goes into the tree is solid and fits in with that design.

	- Doug

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-commits/attachments/20121219/c64cd4d5/attachment.html>