The Code of Conduct committee has published the July 2023 transparency report:
https://llvm.org/coc-reports/2023-07-15-report.html
1 Like
I don’t know if it is appropriate to comment here, but I don’t know where else?
I’m surprised that the CoC committee considers that
Committing patches without public review using LLVM’s review tools (Phabricator) is against the LLVM Developers’ Policy.
Here is the policy: LLVM Code-Review Policy and Practices — LLVM 18.0.0git documentation
It only states that significant changes must be reviewed before being committed. And it even clearly states that “Code can be reviewed either before it is committed or after.”
Note that this likely does not change the analysis and the conclusion of the CoC committee in this instance, since the post-commit review would require interaction with the original author of the patch, but still better to stick to what the policy is actually.
11 Likes
The more surprising part to me here is not so much whether the code was reviewed (I would expect if an intermediary contributed the code - they would be responsible for its pre or post commit review, as they would with any other code they contribute) but a licensing issue - I was pretty sure we had a policy (maybe it’s not written down) that you can’t contribute someone else’s code, because it’s not clear you have the right to do so/the project has the right to include that code under the project’s license, etc…
Thank you for questions/feedback. The July CoC report has been modified to try to reflect the intentions of the CoC committee statement regarding following developer policy.
The statement now reads:
Committing patches with public review not using LLVM’s review tools (Phabricator) is against the LLVM Developers’ Policy. All patches must be reviewed according to the LLVM Developers’ Policy - Public review on Phabricator
While we hope this helps clear up some confusion, we are unable to comment further on this particular statement due to confidentiality reasons.
1 Like