Re: upstart: please update to latest upstream version
2012/2/26 Uoti Urpala <email@example.com>:
> Cyril Brulebois wrote:
>> Uoti Urpala <firstname.lastname@example.org> (26/02/2012):
>> > Is there reason to believe this would be a particular problem with
>> > systemd? Most of the controversy I've seen surrounding Poettering has
>> > been due to people resisting the kind of change he has championed. I
>> > don't remember seeing much about him being "difficult to work with" in
>> > any sense that would make it hard to contribute to systemd (though not
>> > much to the contrary either).
>From talking to Lennart (I don't really work with him, so I can't
judge this), he is a understanding person if you can outline the
problem clearly. But he has a very strong view on doing things right,
e.g. writing portable code: He does not want portability patches in
systemd, because much invasive changes would be needed, making the
code more difficult to read (which might even lead to buggy code).
Therefore, discussion about areas where he already made clear that he
wouldn't support it is absolutely pointless.
Anyway, discussing other problems and code changes is constructive and
Personally, I'd consider Poettering a good upstream, one with a vision
and very strong opinions about some stuff, which is difficult
sometimes, but still someone to work with who is willing to make some
steps towards Debian too.
Please remember, I don't work with him, I just observe :P The systemd
package maintainers will know more.
In any case, as was mentioned before, these issues shouldn't have
priority now. Debian should pick the technically best solution, and
finding out which one is the best should be the priority. As soon as
this is decided somehow, we could think of how to support kFreeBSD and
all the other stuff.
Debian - as always - is extremely slow in making a decision on this,
but this time it might be an advantage: Fedora, openSUSE, Mageia, etc.
already switched to systemd (and Ubuntu, AFAIK the only distribution
using upstart at time) so we might get plenty of data on how the new
init systems behave and will face less problems when switching too.
(If this is decided)