Bug#811377: closed by Dmitry Bogatov <KAction@debian.org> (Bug#811377: fixed in sysvinit 2.88dsf-60)
Hello Dmitry Bogatov,
On Fri, Oct 26, 2018 at 08:55:12PM +0000, Dmitry Bogatov wrote:
> [2018-10-26 13:58] Ben Hutchings <ben@decadent.org.uk>
[...]
> > Benda created a git repository at <https://salsa.debian.org/debian/sysvinit>.
>
> My repository is fork of 'debian/sysvinit'. As far as I can tell, the
> following work was done on debian/sysvinit:
>
> * import history from alioth up to 2.89-59.10
> * create (very) incomplete upgrade to 2.90 in branch dgit/experimental
>
> What I done:
>
> * incorporated NMU 2.89-59.11
> * modernized package a bit
>
> > You seem to have missed that Ian and Benda already started to adopt
> > sysvinit. (While they missed closing this bug.)
>
> Sysvinit have undecided maintainance status for years, which prevented
> me from taking actions previously. But now danger is grave and immediate
> (see recent discussions about removal of sysvinit on debian-devel@), so
> I took libery to take action right now.
I'd like to echo what Ben already said that there's no immediate
danger of removal. However I appreciate seeing someone take action
sooner rather than later. Thanks for stepping up!
>
> I am sorry, if I stepped on someone's toes. I would greatly appericate
> any help in maintaining sysvinit.
Recruitment will be as big of a challenge in your maintainer role
as technical work. I hope you manage to find people you can work
together with that actually do work rather than just talk.
I'd like to suggest you put the alioth mailinglist back as maintainer
and put yourself as uploader. That will allow casual bystanders to
have a chance to follow progress from the sidelines (and hopefully
that way build up an interest in stepping in). If you want to
aquire admin access to the mailing list, that can probably be
arranged by contacting either Ian Jackson or pere.
Also working towards using the debian/sysvinit repo would likely
be a good move. It seems like you're using a -guest account on
salsa, but at the same time you're a DD?! You should have an
account on salsa matching your debian username (kaction?) which will
allow you full access to debian/sysvinit. Try using that.
>
> Oh, and surely, while I did my best to not introduce any functional
> changes in 2.89-60, brave souls to test upload are more then welcome.
Unfortunately you seem to have been a little to eager with
"(lintian?) cleanups" where you broke a few things.
Please remember to always decide for yourself rather than listen
to what lintian has to say. Lintian is usually good at "regular"
packages but very often wrong about "special" packages (like sysvinit,
et.al.). Also always be very careful and understand all possible
consequenses when changing LSB headers in central init scripts.
Here are some examples that I overheard a discussion about:
You seem to have introduced loops with your changes in:
https://salsa.debian.org/iu-guest/sysvinit/commit/7f1224311e24078be99864bd14f146d56d99804c
(insserv will tell you about it. Just try to install any package which
has an init script to see the warnings.)
bootlogd will not catch entire boot process since your changes in:
https://salsa.debian.org/iu-guest/sysvinit/commit/daa1e40ec7c3fb136a3f93e6f1a82dd699ec245e
Your change to rmnologin init script likely also cause subtle breakage.
You might want to consider reverting all the LSB header changes and
adding lintian overrides where needed instead.
Regards,
Andreas Henriksson
PS. I'm personally looking forward to a fix for #799329 finally being
incorporated. That'll make helping out with testing easier.
Reply to: