Excerpts from Serge Hallyn's message of Wed Oct 19 18:38:35 UTC 2011:
> Hi Clint,
>
> is it really valid to require an upstart job for package X to check that
> other scripts installed by package X are in fact installed?
>
> If so this is trivial to add, but it seems over the top to worry about
> the upstart job failing (with no further ill effects on the host, IIUC)
> if the package is in a bad state. Put another way, you say the upstart
> job should 'exit gracefully', but really the upstart job should fail
> under this condition!
>
Its required by policy. Upstart jobs are conffiles, and so are not
removed until a package is *purged*.. even though the binaries are.
Excerpts from Serge Hallyn's message of Wed Oct 19 18:38:35 UTC 2011:
> Hi Clint,
>
> is it really valid to require an upstart job for package X to check that
> other scripts installed by package X are in fact installed?
>
> If so this is trivial to add, but it seems over the top to worry about
> the upstart job failing (with no further ill effects on the host, IIUC)
> if the package is in a bad state. Put another way, you say the upstart
> job should 'exit gracefully', but really the upstart job should fail
> under this condition!
>
Its required by policy. Upstart jobs are conffiles, and so are not
removed until a package is *purged*.. even though the binaries are.