[libcxx-commits] [libcxx] [libc++] FreeBSD use fallback for atomic timed wait (PR #180400)

[Jessica Clarke via libcxx-commits]

jrtc27 wrote:

> > The original code was "fine", no? The only flaw was that it needed an intermediate cast via uintptr_t for the size-stuffed-in-a-pointer feature of the API?
> 
> @jrtc27 I think the it is probably fine adding `static_ast<uintptr_t>` here. But most importantly, I am removing all the Claude generated code (these few lines) to avoid some social media influencer trying to make use of it for any other purposes.

I hope that "some social media influencer" is not being used to refer to David or some other member of the community in the Mastodon thread that spawned this PR.

The current AI policy in LLVM is available from https://llvm.org/docs/AIToolPolicy.html, from which I'll make some quotes:

> LLVM’s policy is that contributors can use whatever tools they would like to craft their contributions, but there must be a human in the loop. Contributors must read and review all LLM-generated code or text before they ask other project members to review it. The contributor is always the author and is fully accountable for their contributions. Contributors should be sufficiently confident that the contribution is high enough quality that asking for a review is a good use of scarce maintainer time, and they should be able to answer questions about their work during review.

> Contributors are expected to be transparent and label contributions that contain substantial amounts of tool-generated content.

> Contributors should note tool usage in their pull request description, commit message, or wherever authorship is normally indicated for the work. For instance, use a commit message trailer like Assisted-by: . This transparency helps the community develop best practices and understand the role of these new tools.

The ways in which your PR fell short of this are as follows:

1. Neither you nor the reviewer were able to explain why the code was correct (or at least, none was given), just that it seemed to work. This could have been easily achieved by just searching for the word "timeout" in the _umtx_op(2) manpage.
2. There was no AI assistance declared by you in the PR description and thus eventual commit message, and this was not spotted by the reviewer.

In both cases there is a bigger problem here that a libc++ maintainer knowingly disregarded the project's AI policy.

https://github.com/llvm/llvm-project/pull/180400