[stdlib] Add UInt __invert__

[martinvuyk]

Add UInt __invert__

[soraros]

I believe this won't compile if you add a test case as we can't convert a negative IntLiteral to UInt.

[helehex]

This looks good, although I wonder why __invert__() doesn't directly use the mlir ops. I noticed that with Int as well.
Maybe we should do something similar to Int.__neg__(), to be consistent, but it's not super important.

[martinvuyk]

Hi @helehex , well I think it's just an xor however you express it 🤷‍♂️ . I just want the functionality to do stuff with hashes

[helehex]

Yeah it's just a question about consistency. Hopefully this can get merged soon since it's been missing for a while.

[helehex]

Maybe you should add a test (even though int doesn't have one).

Also, if you use mlir ops, the implementation could be the same between Int and UInt, ie.

return __mlir_op.`index.xor`(
    value.value,
    __mlir_op.`index.constant`[value = __mlir_attr.`-1:index`](),
)

works for Uint and Int, the difference is just the implicit conversion on return.

[martinvuyk]

Maybe you should add a test (even though int doesn't have one).

Honestly this makes no difference, it's trivial. Inverting bits doesn't need a test IMO, that would be testing the CPU basically.

works for Uint and Int, the difference is just the implicit conversion on return.

When #3753 gets merged it will fail. And the code is already what gets added by the xor operation

@jackos can we merge this directly?

[helehex]

Both UInt and Int are really just wrappers around the mlir.index type, so i dont see why it would fail when that gets merged, theres no conversion between Int and UInt happening.

It may be trivial, but having a test is nice regardless to make sure it compiles, and to ensure nothing else breaks it down the line.

[helehex]

That being said, I'd be fine with it getting merged as is, it's just some suggestions.

[martinvuyk]

i dont see why it would fail when that gets merged, theres no conversion between Int and UInt happening.

yeah you're right I got confused and thought I changed the mlir index constructor in that PR but it was just the Int one.

It may be trivial, but having a test is nice regardless to make sure it compiles, and to ensure nothing else breaks it down the line.

Yeah it would be nice, I'm just a bit frustrated and I don't want to extend this PR further

[jackos]

When #3753 gets merged it will fail. And the code is already what gets added by the xor operation

Hi @martinvuyk that one is already failing a bunch of internal tests, I'll leave comments in that PR.

But this is separate I'll add a test on the internal PR and see if all the other internal tests pass.

[jackos]

!sync

[modularbot]

✅🟣 This contribution has been merged 🟣✅

Your pull request has been merged to the internal upstream Mojo sources. It will be reflected here in the Mojo repository on the main branch during the next Mojo nightly release, typically within the next 24-48 hours.

We use Copybara to merge external contributions, click here to learn more.

[modularbot]

Landed in f789084! Thank you for your contribution 🎉