[Mlir-commits] [mlir] [mlir][math] Fix `math.powf(a, b)` expansion for negative base (PR #193231)

[Zmicier Prybysh]

dimp-pl wrote:

Tagging reviewers:
@matthias-springer given your recent `math` dialect reviews, this might be in your area.
@bjacob you reviewed #170028, which was kind of similar  — that PR fixed UB in `math.ceil` expansion in a similar way to how this PR fixes IEEE semantics of `math.powf` expansion for negative bases.
@kuhar adding you for math-dialect continuity.

https://github.com/llvm/llvm-project/pull/193231