[Mlir-commits] [mlir] [mlir] Expose linalg vectorization without replacement (PR #144158)

[llvmlistbot at llvm.org]

Max191 wrote:

That SGTM! I was thinking the same thing about `FailureOr`. An additional benefit of using FailureOr, which I hadn't realized before, is that there will be less disruption in downstream projects, since it will be compatible with old `LogicalResult` implementations. I think everyone will be happy with this implementation.

Big thank you to Hanhan, for adding great additional context!

https://github.com/llvm/llvm-project/pull/144158