[PATCH] D47073: Document and Enforce new Host Compiler Policy

Wed Jan 16 09:52:28 PST 2019

jfb added a comment.

In D47073#1360048 <https://reviews.llvm.org/D47073#1360048>, @jyknight wrote:

> In D47073#1359998 <https://reviews.llvm.org/D47073#1359998>, @jfb wrote:
>
> > I suggest we abandon this thread. I'll start a separate llvm-dev thread and new patches to get this back on track.
>
>
> I concur that going back to the mailing list thread is appropriate. From my POV, the core problem here seems to be that the change from "date-based" to "requirements-based" was not truly agreed upon, despite the wording in this patch being updated to say it.
>
> That needs to be resolved before reviving a patch. Are we:
>
> 1. Choosing versions based on being 3 years old, or,
> 2. Choosing versions based on their feature-set (e.g. C++14 support right now).
>
>   From the last few comments, it would appear that both Erich and JF both continue to feel that the former should be the policy. And arguing about details when there seems to remain disagreement over that issue is obviously not working.

That's incorrect. I don't care what the policy is as long as it's sane. 3 years was proposed on the RFC, we went with it. Then nuance was asked for on this review, we went with it. Then it was asked that time was made a soft constraint, we went with it. Then more bikeshed happened. The version was absolutely chosen based on agreement that we're moving off C++11, and C++14 is the next version over. My updated email and patches will be clear on this, and I urge you Mehdi, and Chandler to try to drive this to a resolution ASAP. I think it would be unacceptable if LLVM 8 didn't have a soft-error, and I think the longer we want to cherry-pick such soft-error the more irresponsible we are.

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D47073/new/

https://reviews.llvm.org/D47073