* Sami Haahtinen <ressu@ressukka.net> [2006-01-15 11:27]:
> Gerfried Fuchs wrote:
>> It's also about false statements like "We sync our packages to Debian
>> regularly," because that simply doesn't happen for quite a lot of us,
>> otherwise all these heated discussions wouldn't happen.
>
> They have their own timetable. They do their stabilization differently
> than debian does. Ubuntu freezes the packages at a certain point in time
> and only does manual syncs after that. "Regularly" could be once a year
> and still be regular.
>
> What do you want? decide which packages get to certain ubuntu release?
> Didn't they just offer you that chance?
I want that they inform their upstream about changes they do to their
releases. This is what this whole fuzz is about. And I'm not the only
one. Why should some upstream developers who don't even know that they
are forked look around the net to find things that might be useful to
them? Shouldn't rather the people that fork try to get the changes back
into their upstream, to _ease their own work_? I mean... if they want to
keep their patches forever and want to stumble into problems every now
and then for taking a look why the patch doesn't apply anymore it's fine
with me. But, they are saying that "they" sync it back, not that I would
have to run around to most of the time find nothing.... That's not the
way it's supposed to work.
What I want: If someone changes something, they are on the call anyway.
They know already that they changed something, what's so difficult to
inform the involved parties? Especially as long as they keep the
Maintainer field intact and let it look like the changes were done by
me? Yes, it's in the changelog, but keeping the Maintainer means that I
am still maintaining their fork, which I simply don't do since I have no
access to upload there.
> As i see it utnubu is the middle ground for debian and ubuntu people.
> It's something that debian people want to do to keep up with Ubuntu.
Yes, because Ubuntu isn't able or willing or whatever to keep up with
Debian. It's still working the wrong way round, it's sort of
reverse-engineering for hardware, because they know what they changed
already, we have to work it out at first, just to find most of the
time... nothing.
> I see utnubu as a good thing, it solves problems that the people behind
> utnubu want to get solved. They decided to do the work instead of
> throwing it back to Ubuntu and saying "It's your problem to make me
> agree with you". As utnubu page says:
> "We are about cooperation, not confrontation, with Ubuntu."
It's not about agreeing. It's about working on stuff that is unique in
the Free Software community: The people who are confronted with changed
things have to actively pull the changes back, not the way around it
works like everywhere else: That people who change something push it
back.
> co-operation needs co-operation from both parties!
Again, like I mentioned, I never was addressed about cooperation, so I
never had the chance to turn it down. And I am very sure I'm not the
only one in that state.
> You are not forced to pull anything from Ubuntu.
Uh? But this is what it is all about. I _am_ sort of forced because
they don't push their changes, like it would rather be expected.
> But you should remember that the packages that are being worked on
> outside of the ubuntu main are maintained by a small group (when
> compared to the people in debian) of people. They have limited time to
> push all changes to upstream and usually the changes are just for the
> packaging anyway.
If they have limited time it should be in _their own_ interest to push
the changes back. Hell, have you never stumbled upon a patch that simply
doesn't apply anymore? It's a lot of work to take a look what's going
wrong now again, whereas it is next to no work sending a small mail with
"I changed this or that, maybe you'd like to take a look at it." It's
about investing into the future, but some people only seem to work only
for today. That's also the reason why we have so many duplicated
security advisories because people don't think about the future but only
copy stuff because it's the easier approach for now....
> Also, you should remember that there are people that have said that they
> don't want to be in contact with ubuntu.
So this counts for everyone now? I don't think that the people that
have said that they don't want to be in contact with ubuntu are the ones
complaining about not hearing from them. This would be very strange,
don't you think so?
> So it's not an easy thing to notify debian people about the changes in
> their packages when some people get offended by the notification
> itself.
Why not? Either maintainer a list of bad-DDs or don't take it personal.
I also don't take it personal if my upstreams sort-of ignore me most of
the time. That doesn't mean that I don't contact them from time to time,
because I care about the users of those packages. If ubuntu rather likes
to shy away and forget about users.... well. Maybe they haven't adopted
to our social contact, they propably came up with their own
interpetation...
> If you have a solution for this, let me know. Or better yet, let the
> Ubuntu people know.
And where? I was never contacted in any way, so I wouldn't know where
to contact them. And why should I care for them more than they care for
me?
> You do realize that your work is out there for anyone to take and to
> modify.
Yes, and I also realize that my work usually gets feedback from people
who see the need of some changes. Do you realize that people that modify
should send back the changes, or do you think that people whose code get
modified should go around and pull it?
> I agree that for the modified packages it should be more clear that
> the package has been modified by ubuntu and the maintainer or some
> other field should reflect that.
I'm not even really asking for that. I'm just asking for a notification
about what changes had been done and what for. Nothing more, nothing
less. If the patches are ubuntu-only, fine. If they would make sense
upstream, what's the problem sending them upstream?
> But again, some people are offended if the maintainer field is changed
> to something ubuntu specific for the modified packages.
Why? I don't maintain any package within ubuntu, especially not if it
had been changed. We don't run around saying that the Linux kernel
packages are maintained by Linus, or that the GNU stuff is maintained by
RMS, so why would we want them to do that? (And yes, elaborate on that
RMS and Linus aren't really maintaining them neither... it'd just show
how serious I can take you.)
> And about pulling the changes, did you notice these:
> http://packages.qa.debian.org/libm/libmetakit2.4.9.3.html
No, didn't notice this. I even didn't notice why you put the PTS in, I
overlooked the link there, it wasn't really announced. But, the package
is a good example: Have you taken a look at the diff? I can only *hope*
that this is the reason why I haven't heard anything from them, because
the patch doesn't make much sense: Added trailing whitespace, running
autoconf on build-time; the only "real" change is there changing from
python 2.3 to python 2.4, which upstream decided to do with the 2.4.9.4
release....
If the patches are always that meaningful I start to understand why so
many people are objecting to accepting them. Though I'm not saying that
I'd object them, if I would be addressed, with reasoning (which their
changelog don't really contain) -- which never happened.
> Ubuntu side:
> https://launchpad.net/people/alfie/+packages
So it's about pulling again....
> I had a hard time finding your packages that were modified in ubuntu,
And you would expect me to do this? Maybe even on a regular basis?
Sheesh... Even if the "all hail ubuntu" people have a hard time to find
something, how is the rest of the world expected to do it?
> so maybe that's something ubuntu people should work on. Other htan
> that, you should easily be able to pull changes to your packages from
> there
... if they would make sense. If they would have a reasoning applied.
And if I wouldn't have to go and look every now and then to find...
nothing, most of the time.
> A good indicator that your package has been modifies in ubuntu is the
> string ubuntu in the package version.
So you are saying, appart from that, there is no real indicator? So
basically I'd have to pull both their .dsc and .orig.tar.gz to check?
Hey, sounds like fun, for finding nothing most of the time...
So long,
Alfie
--
No, really, I surveyed everyone on the planet just last night. Everyone thinks
it rocks. I'm pretty tired now actually, and I'm _damn_ sick of people who let
their dogs in their front yards.
-- Anthony Towns, debian-devel, <20020516105403.GB15515@azure.humbug.org.au>
Attachment:
signature.asc
Description: Digital signature