[PATCH] D79578: [mlir] NFC: Rename LoopOps dialect to SCF (Structured Control Flow)

Mehdi AMINI via Phabricator via llvm-commits llvm-commits at lists.llvm.org
Tue May 12 11:49:28 PDT 2020


mehdi_amini added a comment.

In D79578#2031204 <https://reviews.llvm.org/D79578#2031204>, @ftynse wrote:

> In D79578#2029832 <https://reviews.llvm.org/D79578#2029832>, @PeteSteinfeld wrote:
>
> > > I don't think this is the request. The request is just "don't break the build". 
> > >  The pre-merge testing will test flang as well, but you would have had to rebase the patch after you landed the dependent ones first so that the pre-merge testing could work (I think marking the dependencies in Phabricator would have worked also).
> > > 
> > > In general trying to get the pre-merge testing green before pushing changes is safer.
> >
> > Yes, the request is to not break the build.
>
>
> This is standard policy. So is reverting the breaking changes. Sometimes we do break stuff and there are technical solutions to decrease the probability of this happening.
>
> I am however opposed to having to contact someone other than people who would normally review the patch. If something is not covered by `check-flang` (which I forgot to run, I admit, that's why we need automation!) or any other simple command, I think it is unreasonable to block us from working and wait for somebody to think about how the changes might affect dependent projects.


This is not what was asked: it was mentioned that some people can *help*, I think if you can't figure out yourself how to not break flang or if you don't understand a flang failure in the pre-merge testing for example.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D79578/new/

https://reviews.llvm.org/D79578





More information about the llvm-commits mailing list