Re: using upstart in Debian
Steve Langasek <vorlon@debian.org> writes:
> On Mon, May 27, 2013 at 11:58:13AM -0700, Nikolaus Rath wrote:
>> As I said, there isn't a bug anywhere here. Once you understand what's
>> going on, this all makes sense. But I don't consider this a very good
>> design.
>
> Well, I don't think it's a bad design that third-party jobs can see and use
> upstart events that were created primarily for internal consumption; if
> anything, the problem is in documenting these in a way that doesn't make it
> clear that they are not a general-purpose interface for services and should
> not be used except in special cases.
But exactly the same can happen between other jobs as well, it's not
specific to mountall or the mounted event. In my opinion the problem is
not that third-party jobs can see upstart internal events, but the way
that upstart has defined events in the first place. To me, this leads to
confusing semantics all over the place.
To give a different example, if I look at something like "start on
event1 and event2" then the first thing that comes to my mind is "so
this job is only going to start if these two events happen to occur
simultanously by some coincidence?"[1]. As I understand, Upstart is
handling this by giving lengths a duration. But the duration is defined
by how other jobs combine this event with others, yet affects the
program that tries to emit the event.
For this reason, the systemd way of declaring jobs seems much more
natural to me. I don't want my X server to start when dbus is available
the the home directory is mounted (as upstart forces me to think), I
want dbus to be started and the home directories mounted when I request
an X session. If I am not mistaken, only systemd allows me to express
the later .
Best,
-Nikolaus
[1] Don't get me wrong, the upstart documentation does a good job of
explaining this. My point is not that it's not well documented, but
that it is counterintuitive.
--
»Time flies like an arrow, fruit flies like a Banana.«
PGP fingerprint: 5B93 61F8 4EA2 E279 ABF6 02CF A9AD B7F8 AE4E 425C
Reply to: