[Reportbug-maint] Bug#741930: reportbug: add current init system information

Cameron Norman camerontnorman at gmail.com
Sat Nov 8 17:46:28 UTC 2014

On Wed, Nov 5, 2014 at 3:10 AM, Sandro Tosi <morph at debian.org> wrote:
> Hello,
> On Wed, Nov 5, 2014 at 1:09 AM, Cameron Norman <camerontnorman at gmail.com> wrote:
>> A few notes I have:
>> 1. With Jessie and on, with sysvinit /sbin/init //will// be a link,
>> not the true init file. This would lead to unknowns when the init was
>> actually sysv.
> care to explain a bit better? I just upgraded a Wheezy VM to testing
> and (except some issues) once I replaced systemd with sysvinit-core
> /sbin/init *is* a regular file:

You are correct. I thought that the sysvinit-core package just
installed a link to the sysvinit package's /lib/sysvinit/init, but
that was not correct.

>> 2. With Upstart, /sbin/init is not a link, so that third test would
>> give a false positive for sysvinit when it was actually Upstart
>> (assuming the Upstart check gave a false negative).
> it should not be a false negative, do you have a situation in mind
> where it might happen?

No, but if the Upstart check is a false negative you should give an
unknown reading, not sysvinit. This is a minor issue, probably will
never occur.

>> 3. Maybe you should embed the check for Upstart, so that you do not
>> have to source all of the init functions, and if that file is ever not
>> available you still get the correct check.
> lsb init functions are part of lsb-base, a required package

Yeah, you are right.

Best wishes,
Cameron Norman

More information about the Reportbug-maint mailing list