Bug#928919: patch: Do not hide errors from initscripts
Dmitry Bogatov
KAction at debian.org
Wed May 22 00:45:07 BST 2019
[2019-05-18 23:26] Thorsten Glaser <t.glaser at tarent.de>
> > If initscript is called during postinst and fail -- yes, it interrupts
> > installation. And it is bug in package.
>
> No, I can install a package and not have it start perfectly fine.
You have exit status 6 for case of package not being configured.
Otherwise, I strongly disagree with "failure to start service fine".
> Worse is breakage on package *removal*, especially in situations
> where the system is wrecked (e.g. package install was aborted)
> anyway.
>
> Breaking apt is more evil than some dæmon not starting during
> the system run (on large upgrades, a reboot will be done anyway,
> so the problem is very transient).
Yeah, having non-unstallable package is evil. But I think hiding
the fact, that initscript is broken is even more evil.
But with these matters of good and evil we got off the main discussion:
what is wrong with having initscripts from bin:initscripts more sensible
exit values, given they are not invoked in maintainer scripts?
--
Note, that I send and fetch email in batch, once every 24 hours.
If matter is urgent, try https://t.me/kaction
--
More information about the Debian-init-diversity
mailing list