[Mlir-commits] [mlir] [MLIR] Update Maintainers.md: add Python bindings under "Core" (PR #153410)

[Jacques Pienaar]

jpienaar wrote:

> For lack of a better place, we put python bindings under "core" because it's not egress/tensor related, but we can very well create another "unrelated" or "standalone" categories (possibly without "top" maintainers), so that we can bundle the loose parts of MLIR

I don't think we should, this would mean more things end up being secondary responsible for smaller group (e.g., if we have many small stand-alone ones, they all become lead maintainer secondary).

As folks keep emphasizing (and the terminology is bad), this is not a hierarchy nor control. This is fallback nestings of responsibility to keep working. Python bindings may not be as core to some folks as cmake, but it is a large entry and usage path for MLIR proper (e.g., I have at least one projects that use only it and core, nothing in tensor or egress). And so fallback responsibility path makes sense to me in isolation. But again, it's fallback, it doesn't apply here unless Maks and Rolf are non-responsive, but general procedural element for future too. Could discuss later and in separate thread, but the fallback responsibility nesting makes sense to me.

Yes this PR crossed with the other, but one can only hold so much state in head at one point, so that's understandable and just propagating reported state.

https://github.com/llvm/llvm-project/pull/153410