[PATCH] D74185: Revert the revert of vectorization commits

Fri Feb 7 18:29:14 PST 2020

ABataev added a comment.

In D74185#1865361 <https://reviews.llvm.org/D74185#1865361>, @george.karpenkov wrote:

> > Even if it is not a real cause of the problem?
>
> Yes, of course. Many revisions are rolled back like this. It would be completely unsustainable to try to find the "real" culprit for a breaking change, and it's often not even clear where the "real" blame is. Taking this example to extreme: a commit in the compiler which breaks all executables due to triggering a CPU bug should still be rolled back, even if it perfectly fits the compiler contract.

I can investigate it thoroughly on Monday. Give me a Monday to investigate it and after that I will revert it myself, if no proper fix is found.

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D74185/new/

https://reviews.llvm.org/D74185