Re: New experimental sysvinit
Anyone see any problems with this scheme? Any better ideas?
This might be a silly question, but what would happen if DELAYLOGIN=yes
and /var/lib/initscripts/nologin does not exist?
I surmise that it's created somewhere in 0000runmefirst. After all, the
usual state of a running system is that the file doesn't exist...
And that is probably not what I would expect. What about doing something
NOLOGIN=boot nologin exists during boot
NOLOGIN=always nologin always exists
NOLOGIN=never nologin never exists
...and leave handling of actual file/symlink to the script? This way you
would be able to change the implementation and avoid breakage and
confusion of manual messing with files.