The upstream glibc 5 year average of recorded reviews is hovering at ~50% and it wobbles around that mark. I wonder if there is a human or statistical reason for this, the commits without Reviewed-by: are from some of the most experienced developers. It seems we've hit a limit on the human compliance of asking for, gathering, and recording consensus for a change. To go further than 50% certainly looks like you'd need a technically enforced process interlock (with waivers).
=> More informations about this toot | More toots from codonell@fosstodon.org
The RBs pickup after 2019 as developers continued to use the process to record consensus on a change. While it goes up quickly from 2019 it plateaus quickly into what I expect is a traditional "cost of compliance" problem.
2019: 256 RB, 678 no RB, 27% recorded review
2020: 550 RB, 783 no RB, 41% recorded review
2021: 834 RB, 744 no RB, 52% recorded review
2022: 737 RB, 575 no RB, 56% recorded review
2023: 519 RB, 584 no RB, 47% recorded review
2024: 678 RB, 563 no RB, 54% recorded review
=> More informations about this toot | More toots from codonell@fosstodon.org
text/geminiThis content has been proxied by September (3851b).