I do use commits as a metric as well when I evaluate stuff. But we talk about filesystems here, in particular about a filesystem that is battle-tested on very large installations and under development for more than 13 years. An answer purely on git commits would be completely useless. The only place where I rely for big parts on that metric is NPM modules :)

I agree with that for pretty much everything except filesystems. I moved to ZFS 11 years ago (on FreeBSD) where everything was pretty new and LVM on Linux started to become common.

I remember too well how careful I had to be with every single operation I did regarding partitions, RAID and filesystem. A wrong move and say goodbye to the filesystem, I've seen that with sysadmin colleagues on Linux on a regular base. With some exceptions ZFS prevents me of doing stupid things with my filesystem and that's exactly what I expect from something where I store all my (and others) data. And the fact it can do that is because it does most of these things under one roof.