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

Tue Jan 15 16:02:45 PST 2019

mehdi_amini added a comment.

In D47073#1358663 <https://reviews.llvm.org/D47073#1358663>, @jfb wrote:

> `s/two/one/` : I've talked to Chandler already, and he was leaning on others (in large parts, me) to help Erich with the review

I'm not gonna speak for him, I just still see a red mark on this patch, admittedly the patch has been updated since.

> Further, Erich addressed your comments repeatedly. It seems like you're asking for more?

Yes, not everything has been addressed AFAICT: like how the specific version of the compilers have been picked.

>> The fact that the process of discussing which version of the compiler to bump to should happen *before* adding a check, and it hasn't happened for this patch which *actually* change the minimum versions.
> 
> It happened months ago in the LLVM dev RFC, and was re-discussed in-person at the dev meeting.

No: the RFC does not include this discussion. If you want to claim this, then please source it precisely.

>>> There's agreement that we'll move away from C++11, soon. We'd be irresponsible if we didn't tell developers who build LLVM. That's the only thing the patch does: tell people what we've agreed to do?
>> 
>> 
>> 
>> 1. I asked *multiple* times in this patch *why* and *how* these particular version of the compiler where picked. I still don't know why clang-3.5 would trigger a cmake error.
> 
> Read the email thread?

It does not motivate the reason to pick clang 3.6, and basing it *only* on the arbitrary time-based choice: this is not acceptable IMO.

This was mentioned in the RFC as well but unanswered AFAICT: http://lists.llvm.org/pipermail/llvm-dev/2018-May/123261.html

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

https://reviews.llvm.org/D47073