On 11/7/18 7:04 PM, Jonathan de Boyne Pollard wrote: > > Jesse Smith is saying that invoking "halt -p" runs /sbin/halt, and > that works as designed and as documented in the halt(8) manual page. > Whereas it does not invoke /etc/init.d/halt. If "halt -p" runs > /etc/init.d/halt for you, then your PATH is somewhat odd and you are > running the wrong thing.
That is not what I'm saying. And the person reporting the bug is not running the wrong command. I think you're misunderstanding how /etc/init.d/halt gets called. > > This is genuinely an /sbin/halt bug, not an /etc/init.d/halt one. Its > do_shutdown() function needs another parameter, derived from > do_poweroff. This bug report should be re-opened. Both of these assertions are mistaken. /etc/init.d/halt gets called as part of the shutdown process. That's where it is getting run and the script runs halt, passing the necessary parameters to match the bahavour defined in /etc/default/halt. It does not offer a way to accept parameters from other, outside sources such as halt or shutdown. If it did then we could set a flag to indicate an override parameter.
