[Mlir-commits] [llvm] [mlir] [MLIR] Generalize expand_shape to take shape as explicit input (PR #90040)

[Oleksandr Alex Zinenko]

https://github.com/ftynse requested changes to this pull request.

I have concerns with layering here. This adds a dependency on the arithmetic dialect from dialect utils. It's not exactly clear where such a code should go since it seems to be shared between memref and tensor, maybe somewhere next to type inference.

https://github.com/llvm/llvm-project/pull/90040