On Thu, Apr 11, 2019 at 10:42:38AM +0000, Dmitry Bogatov wrote: > [2019-04-08 12:54] Jesse Smith <jsm...@resonatingmedia.com> > > On 4/8/19 12:38 PM, Dmitry Bogatov wrote: > > > Maybe, but many of us are accustomed that /sbin/halt turns off the > > > computer, > > > so here comes confusion. > > > > That is certainly true, but I'd like to point out that /sbin/halt only > > turns off the computer because Debian modifies halt's behaviour. If you > > run /sbin/halt without Debian's modifications, the traditional action > > (stop without powering off) occurs. I'd almost consider this a bug since > > /sbin/halt should be used to stop the system while /sbin/poweroff should > > be used to, well, turn off the power to the system.
A computer being able to control its own power is a relatively new thing. Systems of old had only "halt", and without a separate command, it was changed to power off once that functionality appeared in the hardware. For PCs, it was at the time of AT -> ATX. > I believe at least in some of RPM-based distributions /sbin/halt also > turns off computer. But it is just outdated memories from my studentship > times :) Your memory agrees with mine. Making "halt" not turn off capable hardware is a systemd change. On one hand, it follows the name of these commands. On the other, I can't quite fathom the purpose of that infinite loop. If someone for some weird reason wanted the final hang, you can kexec that or something -- it's a really odd case. Meow! -- ⢀⣴⠾⠻⢶⣦⠀ ⣾⠁⢠⠒⠀⣿⡁ Did ya know that typing "test -j8" instead of "ctest -j8" ⢿⡄⠘⠷⠚⠋⠀ will make your testsuite pass much faster, and fix bugs? ⠈⠳⣄⠀⠀⠀⠀