It’s interesting how both of these guys who seems to both be firmly in the “don’t break shit for the users” camp, and both are very adamant about quality of code in the kernel can’t get along, saddens me a bit.
I’m excited about bcachefs, I’m even planning a migration from btrfs, but I need to battle test my offsite backup a bit more before having the guts to walk into bcachefs land.
Might wait for 6.18,6.19 or something and follow any future drama
From what I’ve read, Kent expects others to just take his word for it, when he says his code wont break anything.
The kernel has long had practices around merging and releasing, specifically so that it no longer has to rely on contributors simply promising that their contributions have been tested and confirmed safe.
But Kent has repeatedly skirted or straight up ignored those practices.
This isn’t about not agreeing on code needing to be reliable. It’s about one person refusing to work with an established way of achieving that when contributing to an upstream effort.
He’s been told how to contribute again, and again, and again. And every time he takes it like it’s a personall affront to his credibility.
bachefs is probably a couple years from stability. I certainly wouldn’t use it for something I wasn’t willing to rebuild from scratch. And Linus said he’s dropping it out of 6.17, the patches were allowed for 6.16 but that was the last. You’ll be patching manually going forward unless he changes his mind.
It’s interesting how both of these guys who seems to both be firmly in the “don’t break shit for the users” camp, and both are very adamant about quality of code in the kernel can’t get along, saddens me a bit.
I’m excited about bcachefs, I’m even planning a migration from btrfs, but I need to battle test my offsite backup a bit more before having the guts to walk into bcachefs land.
Might wait for 6.18,6.19 or something and follow any future drama
From what I’ve read, Kent expects others to just take his word for it, when he says his code wont break anything.
The kernel has long had practices around merging and releasing, specifically so that it no longer has to rely on contributors simply promising that their contributions have been tested and confirmed safe.
But Kent has repeatedly skirted or straight up ignored those practices.
This isn’t about not agreeing on code needing to be reliable. It’s about one person refusing to work with an established way of achieving that when contributing to an upstream effort.
He’s been told how to contribute again, and again, and again. And every time he takes it like it’s a personall affront to his credibility.
bachefs is probably a couple years from stability. I certainly wouldn’t use it for something I wasn’t willing to rebuild from scratch. And Linus said he’s dropping it out of 6.17, the patches were allowed for 6.16 but that was the last. You’ll be patching manually going forward unless he changes his mind.