Age | Commit message (Collapse) | Author |
|
|
|
describe output and use generic usage for runscript.
|
|
|
|
afterwards as the dependencies may need programs brought up by localmount. The dependencies will be regenerated correctly in the next runlevel.
|
|
|
|
|
|
rc_deptree_depend now returns the exact depends as listed. This is so we can support a new 'keywords' directive. We can add whatever here, like 'notimeout' which means that an init script doesn't timeout. This removes the hardcoded check on checkfs and checkroot.
|
|
|
|
Lowercase all configurable variables, non configurations remain uppercase.
Replace rc_env_bool with rc_yesno.
Split localmount info procfs (Linux) and dumpon, savecore (BSD)
|
|
|
|
|
|
|
|
|
|
Fixes Gentoo bug #176452 again.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
explictly with --netdev/--nonetdev
|
|
|
|
|
|
|
|
|
|
anymore
|
|
|
|
|
|
|
|
rc_service_schedule_start
|
|
|
|
|
|
|
|
rc_deptree_
|
|
|
|
|
|
things with just the one call making is more efficient.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
is so that programs can get information about the controlling terminal.
This change was triggered by bug #188506 where it's possible that
stdin, stdout and stderr didn't point to a terminal but ended up on one
via our pipes. Using a pty means that stdout and stderr always point to
a terminal, but we lose the ability to tell them apart.
If there is not a pty available then we use un-prefixed output as normal.
This change has also introduced the need for a signal pipe so that
SIGCHLD can exit the loop cleanly.
|