I would LOVE to have such comments from competent and knowledgeable people about my work. I don't care about them being rude if they are right. work is not like disneyland.
If it takes a couple devs leaving in order to maintain security for the inarguably most shipped piece of software, be it.
I know I'm by no means a perfect coder, so I leave my fingers off kernel code. With kernel code, I carry responsibility, and part of acting responsibly is leaving my hands off stuff I don't understand!
Or is there a tolerance level for imperfectly meeting standards, below which it is not necessary to exclude?
I mean, you could, I suppose, require that submissions all include a coq proof of correctness (and of meeting standards), and then you wouldn't have to respond to submissions not meeting standards at all; any that didn't just wouldn't go through.
(Angle brackets meant to reduce ambiguity)
Do you feel the Linux kernel is perfect by any metric?
And what made you think that the code written by Sharp wasn't high quality?