I agree that's a funny coincidence. But, what about the change it wanted to commit? It is at least slightly interesting. It is doubly interesting that changing line 638 neither breaks nor fixes any test.

There's a tweet with a Claude screenshot with a bit more context (linked on the PR).

I don't know enough about the project to know if it makes any sense, but the Zig contributor seemed confused (at least about the title).

Perhaps the offset is always zero anyway in that scenario

But yeah hard to say