Comment opens up an attack vector (Score 0) 121
This new ai policy actually creates a fork attack vector they cant defend against. someone forks gcc. they pull bugs and feature requests from the gcc tracker and fix them using llm assistance. for 90% of bugs theres only one or two correct ways to fix it, thats just how code works. the fork publishes the fix. now gcc maintainers have three choices, all bad.
copy the fix and violate their own policy.
clean room rewrite it which is 10x the effort for the same result and now youve got a paper trail proving you didnt copy.
or just dont fix it and let the fork stay ahead on bug count. the community notices.
every bug report in the gcc tracker is now a potential trap. the only way to have avoided this was to never publish the policy in the first place, because you cant unring the bell. once its out there, any patch that looks like a fix the fork already published risks a taint accusation even if the gcc dev arrived at it independently. its the sco playbook in reverse and they walked right into it.