Distribution quotes of the week

If i understand it right, and namespaces make my head hurt, is that under private crap all happens if something is mounted both inside and outside a namespace, and sysadmin is trying to unmount outside the namespace.

The problem though is that Linux have no one way streets. you either have a two way street between namespace and "base" (shared), or you have a roadblocks (private).

Still, this is something that should be fixed at the kernel level rather than have systemd make some kind of "executive" decision. But then they keep making decisions that indicate that they want to control the kernel, but don't want to fork the kernel (as likely they would not get anyone to take interest if they did).

Near-NULL (hell, sometimes literally NULL) commit messages and their essentially-equivalent tautological counter-parts are annoying. *Why* did you change that code? *What* led you to that code (motivation, problem report, use-case, etc.)? What options did you consider? *Why* did you select the option taken out of those? What is the *intended* result? How much testing was done? (Be honest - I much prefer it when people are honest that their testing has been limited; I can work with commits that are up-front about lack of testing; I'll have a much *worse* opinion of you, and be much more suspicious of your future contributions if you're silent and the patch turns out to have problems that clearly show no testing was done).