|
Home > Archive > Debian Developers > January 2006 > Debian derivatives and the Maintainer: field (again)
You are viewing an archived Text-only version of the thread.
To view this thread in it's original format and/or if you want to reply to
this thread please [click here]
| Author |
Debian derivatives and the Maintainer: field (again)
|
|
| Matt Zimmerman 2006-01-17, 6:06 pm |
| On Tue, Jan 17, 2006 at 12:46:52PM -0600, Adam Heath wrote:
> On Tue, 17 Jan 2006, Anthony Towns wrote:
>
>
> Debian developers set the Maintainer field to themselves(or a team), when they
> upload to Debian. The upstream author is only mentioned in the copyright
> file.
>
> Ubuntu should do something similiar. Set the Maintainer field to someone from
> their group, and mention debian in the copyright(or other appropriate place).
I would very much appreciate if folks would review
http://lists.debian.org/debian-deve...5/msg00260.html and consider the
points that I raise there. I put some effort into collating the issues
which came up the last time and presenting them.
It is important, in particular, to account for the fact that Ubuntu is not
the only Debian derivative, and that proposals like yours would amount to
Debian derivatives being obliged to fork *every source package in Debian*
for the sake of changing a few lines of text.
--
- mdz
--
To UNSUBSCRIBE, email to debian-project-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
| |
| Thomas Bushnell BSG 2006-01-17, 6:06 pm |
| Matt Zimmerman <mdz@debian.org> writes:
> It is important, in particular, to account for the fact that Ubuntu is not
> the only Debian derivative, and that proposals like yours would amount to
> Debian derivatives being obliged to fork *every source package in Debian*
> for the sake of changing a few lines of text.
Yes. Being a downstream modifier imposes costs. Debian meets those
costs, how about you?
--
To UNSUBSCRIBE, email to debian-project-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
| |
| Otavio Salvador 2006-01-17, 6:06 pm |
| Matt Zimmerman <mdz@debian.org> writes:
> I would very much appreciate if folks would review
> http://lists.debian.org/debian-deve...5/msg00260.html and consider the
> points that I raise there. I put some effort into collating the issues
> which came up the last time and presenting them.
In my point of view, maintainer field just need to be change when
Ubuntu does a non-trivial change on it. Otherwise, at least to me, is
OK to leave the maintainer field unchanged. Directly imported source
(that will be just recompiled by Ubuntu) doesn't need to be change
since it's the same source code that runs on Debian.
--
O T A V I O S A L V A D O R
---------------------------------------------
E-mail: otavio@debian.org UIN: 5906116
GNU/Linux User: 239058 GPG ID: 49A5F855
Home Page: http://www.freedom.ind.br/otavio
---------------------------------------------
"Microsoft gives you Windows ... Linux gives
you the whole house."
--
To UNSUBSCRIBE, email to debian-project-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
| |
| Adam Heath 2006-01-17, 6:06 pm |
| On Tue, 17 Jan 2006, Otavio Salvador wrote:
> In my point of view, maintainer field just need to be change when
> Ubuntu does a non-trivial change on it. Otherwise, at least to me, is
> OK to leave the maintainer field unchanged. Directly imported source
> (that will be just recompiled by Ubuntu) doesn't need to be change
> since it's the same source code that runs on Debian.
But linked against other libraries. The binary is downloaded from another
location(or installed from a different cd set). The program used to do the
download may be different.
While the above list may not be all inclusive, it's enough to warrant changing
the Maintainer field to something ubuntu specific.
Debian doesn't set the upstream author in the Maintainer field, when the
changes only amount to adding a debian directory to the upstream source.
--
To UNSUBSCRIBE, email to debian-project-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
| |
| Adam Heath 2006-01-17, 6:06 pm |
| On Tue, 17 Jan 2006, Matt Zimmerman wrote:
>
> I would very much appreciate if folks would review
> http://lists.debian.org/debian-deve...5/msg00260.html and consider the
> points that I raise there. I put some effort into collating the issues
> which came up the last time and presenting them.
>
> It is important, in particular, to account for the fact that Ubuntu is not
> the only Debian derivative, and that proposals like yours would amount to
> Debian derivatives being obliged to fork *every source package in Debian*
> for the sake of changing a few lines of text.
Modify the incoming processor, so that the Packages and Sources files get the
correct info.
--
To UNSUBSCRIBE, email to debian-devel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
| |
| Adam Heath 2006-01-17, 6:06 pm |
| On Tue, 17 Jan 2006, Matt Zimmerman wrote:
>
> I would very much appreciate if folks would review
> http://lists.debian.org/debian-deve...5/msg00260.html and consider the
> points that I raise there. I put some effort into collating the issues
> which came up the last time and presenting them.
>
> It is important, in particular, to account for the fact that Ubuntu is not
> the only Debian derivative, and that proposals like yours would amount to
> Debian derivatives being obliged to fork *every source package in Debian*
> for the sake of changing a few lines of text.
Actually, ignore my last mail.
I actually considered that you(ubuntu) would respond thusly. But, it doesn't
fly.
We don't allow J. Random Upstream to upload unchanged source into Debian. We
add meta-data, and set the Maintainer field appropriately. This is so
that Debian becomes the contact for the software, when it exists in
Debian. Debian derivaties need to do the same.
There really is no other way.
--
To UNSUBSCRIBE, email to debian-project-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
| |
| Matt Zimmerman 2006-01-17, 6:06 pm |
| On Tue, Jan 17, 2006 at 03:07:25PM -0500, Stephen Frost wrote:
> You're already rebuilding the package, which I expect entails possible
> Depends: line changes and other things which would pretty clearly
> 'normally' entail different Debian package revision numbers; changing
> the Maintainer field at the same time is just not that hard,
> *especially* when you're rebuilding the package.
>
> You're implying that this is alot of work and it's just not. It's also
> not 'forking' in any real sense of the word. You don't even have to
> change the version number if you don't want to. When done in Debian,
> it's also not even a new source package (in general anyway) as the thing
> which has the Maintainer field is actually the patch.
You quite obviously haven't read
http://lists.debian.org/debian-deve...5/msg00260.html yet, where I
wrote (among other important things), "it would be fairly straightforward
for Ubuntu to override the Maintainer field in binary packages". I
explained exactly what is and isn't difficult and for whom.
If you're going to attack me, please do it on the basis of what I've
actually said. Honestly, I expected better from you, give that you've acted
like a human being toward me on IRC on several occasions in the past.
--
- mdz
--
To UNSUBSCRIBE, email to debian-project-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
| |
| Matt Zimmerman 2006-01-17, 6:06 pm |
| On Tue, Jan 17, 2006 at 03:50:09PM -0600, Adam Heath wrote:
> On Tue, 17 Jan 2006, Matt Zimmerman wrote:
>
>
> Modify the incoming processor, so that the Packages and Sources files get the
> correct info.
The .dsc and .diff.gz would still have the original values, and the
copyright file can't be modified this way.
--
- mdz
--
To UNSUBSCRIBE, email to debian-project-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
| |
| Matt Zimmerman 2006-01-17, 6:06 pm |
| On Tue, Jan 17, 2006 at 12:37:47PM -0800, Thomas Bushnell BSG wrote:
> Matt Zimmerman <mdz@debian.org> writes:
>
>
> Yes. Being a downstream modifier imposes costs. Debian meets those
> costs, how about you?
We really don't need a stand-in for Andrew Suffield while he's away.
--
- mdz
--
To UNSUBSCRIBE, email to debian-project-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
| |
| Florian Weimer 2006-01-17, 6:06 pm |
| * Matt Zimmerman:
> It is important, in particular, to account for the fact that Ubuntu is not
> the only Debian derivative, and that proposals like yours would amount to
> Debian derivatives being obliged to fork *every source package in Debian*
> for the sake of changing a few lines of text.
Such a change could be implemented in the toolchain. IIRC, you
rebuild everything anyway, so this wouldn't be such a terrible thing
to do.
FWIW, I think your implied assumption that all Debian derivatives
should be treated the same is flawed. Ubuntu is just not like any
other derivative, it's a significant operation on its own. Its
commercial backer is apparently able to pay quite a few Debian
developers, several of them among the core team. There is a
significant user base, and so on. Like it or not, Ubuntu is a bit
special.
--
To UNSUBSCRIBE, email to debian-project-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
| |
| Thomas Bushnell BSG 2006-01-17, 8:49 pm |
| Matt Zimmerman <mdz@ubuntu.com> writes:
> You quite obviously haven't read
> http://lists.debian.org/debian-deve...5/msg00260.html yet, where I
> wrote (among other important things), "it would be fairly straightforward
> for Ubuntu to override the Maintainer field in binary packages". I
> explained exactly what is and isn't difficult and for whom.
Notice that what you say, in response to what has been asked over and
over, is "my opinion is that changing the Maintainer field on
otherwise-unmodified source packages is too costly for derivatives in
general."
But you say nothing about why. You already have suitable automated
tools. Since you are rebuilding the package, you *must* change the
version number *anyway*. It is not correct to recompile, and leave
the version number alone. If you were not recompiling, then no
modification would be necessary.
Moreover, what about category (2), packages which are modified? Since
you are making a new source package *anyway*, why is it so expensive?
In response to your questions, as if they haven't been answered:
If a binary package is built by a third party from unmodified Debian
sources, should its Maintainer field be kept the same as the source
package, or set to the name and address of the third party?
If the third party has their own bug-tracking system, then the
Maintainer field should probably be changed. The original Debian
Maintainer should still be acknowledged.
Should Debian-derived distributions change the Maintainer field in source
packages which are modified relative to Debian? If so, should this be
done in all cases, or only if the modifications are non-trivial?
In absolutely every case, the Maintainer field should be changed if
you have altered the source in any respect.
Thomas
--
To UNSUBSCRIBE, email to debian-project-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
| |
| Matt Zimmerman 2006-01-17, 8:49 pm |
| On Wed, Jan 18, 2006 at 12:34:33AM +0100, Florian Weimer wrote:
> * Matt Zimmerman:
>
>
> Such a change could be implemented in the toolchain. IIRC, you
> rebuild everything anyway, so this wouldn't be such a terrible thing
> to do.
We don't rebuild every source package, which is what the proposal was about
(modifying source packages).
I outlined the options and their costs as I saw them here:
http://lists.debian.org/debian-deve...5/msg00260.html
> FWIW, I think your implied assumption that all Debian derivatives should
> be treated the same is flawed. Ubuntu is just not like any other
> derivative, it's a significant operation on its own. Its commercial
> backer is apparently able to pay quite a few Debian developers, several of
> them among the core team. There is a significant user base, and so on.
> Like it or not, Ubuntu is a bit special.
I can't accept this; if there is no principle here which should be applied
consistently, then it's entirely unfair to attack Ubuntu. Certainly, there
are things about Ubuntu which are unique, but none of them change the issues
at hand.
Do you realize that Xandros, who maintains a Debian derivative which they
box and sell for US$50-$129 per copy, leaves the Maintainer field
unmodified, and as far as I'm aware, was doing so for a period of *years*
before Ubuntu even existed? This never seemed to bother anyone, and
personally, I don't think it's a big deal either.
Seriously, it's entirely unreasonable to single out Ubuntu on this issue.
--
- mdz
--
To UNSUBSCRIBE, email to debian-devel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
| |
| Bill Allombert 2006-01-17, 8:49 pm |
| On Tue, Jan 17, 2006 at 11:44:48AM -0800, Matt Zimmerman wrote:
> It is important, in particular, to account for the fact that Ubuntu is not
> the only Debian derivative, and that proposals like yours would amount to
> Debian derivatives being obliged to fork *every source package in Debian*
> for the sake of changing a few lines of text.
Ubuntu is different in that they rebuild all packages, not just the one
they changes.
Cheers,
--
Bill. <ballombe@debian.org>
Imagine a large red swirl here.
--
To UNSUBSCRIBE, email to debian-devel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
| |
| Matt Zimmerman 2006-01-17, 8:49 pm |
| On Tue, Jan 17, 2006 at 06:19:32PM -0600, Bill Allombert wrote:
> On Tue, Jan 17, 2006 at 11:44:48AM -0800, Matt Zimmerman wrote:
>
> Ubuntu is different in that they rebuild all packages, not just the one
> they changes.
The matter at hand above was source packages, not binary packages.
Besides which, do you honestly know which packages other Debian derivatives
rebuild? As a rule, they are far less communicative about their practices
than Ubuntu.
--
- mdz
--
To UNSUBSCRIBE, email to debian-devel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
| |
| Thomas Bushnell BSG 2006-01-17, 8:49 pm |
| Matt Zimmerman <mdz@debian.org> writes:
> Besides which, do you honestly know which packages other Debian derivatives
> rebuild? As a rule, they are far less communicative about their practices
> than Ubuntu.
How does the behavior of other Debian derivatives matter?
As a rule, those other derivatives do not cooperate with Debian. If
Ubuntu wants to be like them, fine, but don't say you cooperate with
Debian if that's what you want to do.
Thomas
--
To UNSUBSCRIBE, email to debian-devel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
| |
| Matt Zimmerman 2006-01-17, 8:49 pm |
| On Tue, Jan 17, 2006 at 04:09:50PM -0800, Thomas Bushnell BSG wrote:
> Notice that what you say, in response to what has been asked over and
> over, is "my opinion is that changing the Maintainer field on
> otherwise-unmodified source packages is too costly for derivatives in
> general."
>
> But you say nothing about why. You already have suitable automated
> tools.
I don't think you can speak to what tools we do or do not have. The fact
is, we import most Debian source packages unmodified, and do not have any
such tool for modifying them.
> Since you are rebuilding the package, you *must* change the version number
> *anyway*. It is not correct to recompile, and leave the version number
> alone.
I don't agree. This isn't even the case within Debian. Binary-only NMUs
don't modify the source package, even though the binaries are recompiled.
> Moreover, what about category (2), packages which are modified? Since you
> are making a new source package *anyway*, why is it so expensive?
If you re-read your own quote above, you'll see that I was talking about
otherwise-unmodified source packages, not source packages which were
modified anyway, and if you re-read
http://lists.debian.org/debian-deve.../msg00260.html, you'll see that
my second question simply asked whether this would be appropriate.
> In response to your questions, as if they haven't been answered:
So far I've received two clear responses in this thread. I do like jvw's
idea of setting up a poll, and that will be a much more effective way to
collect opinions on this. I've sent him my proposed options for the poll.
I do expect, however, for this decision to be taken with regard to all
Debian derivatives, and not to single out Ubuntu with a different set of
criteria.
--
- mdz
--
To UNSUBSCRIBE, email to debian-project-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
| |
| Thomas Bushnell BSG 2006-01-17, 8:50 pm |
| Matt Zimmerman <mdz@ubuntu.com> writes:
> I don't think you can speak to what tools we do or do not have. The fact
> is, we import most Debian source packages unmodified, and do not have any
> such tool for modifying them.
It's really a very short PERL script, or a simple modification in C to
the dpkg-building tools. Indeed, if you said, "hey, we would like to
do this, but need someone to write the tool for us" you might well
find volunteers.
>
> I don't agree. This isn't even the case within Debian. Binary-only NMUs
> don't modify the source package, even though the binaries are recompiled.
Actually, binary-only NMUs, after the first compilation, *do* get new
version numbers.
> I do expect, however, for this decision to be taken with regard to all
> Debian derivatives, and not to single out Ubuntu with a different set of
> criteria.
No other Debian derivative, as far as I'm aware, says that it
cooperates fully with Debian.
Thomas
--
To UNSUBSCRIBE, email to debian-project-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
| |
| Mike Bird 2006-01-17, 8:50 pm |
| On Tue, 2006-01-17 at 17:29, Thomas Bushnell BSG wrote:
> Matt Zimmerman <mdz@ubuntu.com> writes:
>
> Actually, binary-only NMUs, after the first compilation, *do* get new
> version numbers.
In Debian yes. Ubuntu recompiles the Debian source, in a
different environment and with different dependencies, then
uploads with exactly the same version as Debian.
Having two different package files with the exactly the same
name and different content and dependencies drove me crazy
for a while until we made our migration scripts smarter.
--Mike Bird
--
To UNSUBSCRIBE, email to debian-project-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
| |
| Thomas Bushnell BSG 2006-01-17, 8:50 pm |
| Mike Bird <mgb-debian@yosemite.net> writes:
> On Tue, 2006-01-17 at 17:29, Thomas Bushnell BSG wrote:
>
> In Debian yes. Ubuntu recompiles the Debian source, in a
> different environment and with different dependencies, then
> uploads with exactly the same version as Debian.
Right. I was contradicting Matt's statement that "This isn't even the
case within Debian."
--
To UNSUBSCRIBE, email to debian-project-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
| |
| Anthony Towns 2006-01-17, 8:50 pm |
| On Tue, Jan 17, 2006 at 04:38:29PM -0800, Matt Zimmerman wrote:
> On Tue, Jan 17, 2006 at 04:09:50PM -0800, Thomas Bushnell BSG wrote:
> I don't think you can speak to what tools we do or do not have. The fact
> is, we import most Debian source packages unmodified, and do not have any
> such tool for modifying them.
Huh? Of course you do -- it's called "make".
> I don't agree. This isn't even the case within Debian. Binary-only NMUs
> don't modify the source package, even though the binaries are recompiled.
However if a binNMU screws up a maintainer's package, the maintainer can
easily fix it, and doing so is just part of contributing to Debian. The
same thing applies when an autobuild on another architecture happens.
That's not the case if an Ubuntu rebuild screws things up.
Cheers,
aj
| |
| Matthew Garrett 2006-01-18, 2:50 am |
| Thomas Bushnell BSG <tb@becket.net> wrote:
> No other Debian derivative, as far as I'm aware, says that it
> cooperates fully with Debian.
Other than, say, the DCC Alliance?
--
Matthew Garrett | mjg59-chiark.mail.debian.devel@srcf.ucam.org
--
To UNSUBSCRIBE, email to debian-devel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
| |
| Thomas Bushnell BSG 2006-01-18, 2:50 am |
| Matthew Garrett <mgarrett@chiark.greenend.org.uk> writes:
> Thomas Bushnell BSG <tb@becket.net> wrote:
>
>
> Other than, say, the DCC Alliance?
I wasn't aware of them until just now. 
Interestingly, the DCC Alliance says that it wants to become part of
Debian.
Do you have information on their plans with respect to the issues
discussed in this thread?
Thomas
--
To UNSUBSCRIBE, email to debian-devel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
| |
| Matthew Garrett 2006-01-18, 2:50 am |
| Thomas Bushnell BSG <tb@becket.net> wrote:
> Matthew Garrett <mgarrett@chiark.greenend.org.uk> writes:
>
> I wasn't aware of them until just now.
Wow!
> Interestingly, the DCC Alliance says that it wants to become part of
> Debian.
>
> Do you have information on their plans with respect to the issues
> discussed in this thread?
The DCCA distribution is a mixture of packages from Sarge plus some
backports. In all cases, the Maintainer: field appears to be the same as
in Debian. Several derived distributions (gnuLinEx, Knoppix, Mepis,
Linspire, Progency, Sun-Wah, UserLinux (ha ha ha) and Xandros) are
supposed to be using these packages in an unmodified form, as I
understand it.
--
Matthew Garrett | mjg59-chiark.mail.debian.devel@srcf.ucam.org
--
To UNSUBSCRIBE, email to debian-devel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
| |
| Thomas Bushnell BSG 2006-01-18, 2:50 am |
| Matthew Garrett <mgarrett@chiark.greenend.org.uk> writes:
>
> The DCCA distribution is a mixture of packages from Sarge plus some
> backports. In all cases, the Maintainer: field appears to be the same as
> in Debian. Several derived distributions (gnuLinEx, Knoppix, Mepis,
> Linspire, Progency, Sun-Wah, UserLinux (ha ha ha) and Xandros) are
> supposed to be using these packages in an unmodified form, as I
> understand it.
Have they modified these packages?
What do they do about version numbering of recompilations? (Do they
recompile?)
Do they modify the packages at all?
It seems like they aren't doing the things which annoy me in the case
of Ubuntu, but it's possible I don't understand enough.
--
To UNSUBSCRIBE, email to debian-devel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
| |
| Matthew Garrett 2006-01-18, 2:50 am |
| On Tue, Jan 17, 2006 at 07:23:41PM -0800, Thomas Bushnell BSG wrote:
> Matthew Garrett <mgarrett@chiark.greenend.org.uk> writes:
>
> Have they modified these packages?
Some of them, yes. Mostly the backports.
> What do they do about version numbering of recompilations? (Do they
> recompile?)
They have to recompile the backports, at least. I haven't checked the
MD5s of the binaries, but that's easy enough to do.
> Do they modify the packages at all?
As stated, in some cases yes.
--
Matthew Garrett | mjg59@srcf.ucam.org
--
To UNSUBSCRIBE, email to debian-devel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
| |
| Matthew Garrett 2006-01-18, 2:50 am |
| On Tue, Jan 17, 2006 at 07:32:20PM -0800, Thomas Bushnell BSG wrote:
> Matthew Garrett <mjg59@srcf.ucam.org> writes:
>
>
> What happens to the maintainer field in these cases?
I haven't seen any that have been changed.
> Certainly, if they are modifying the packages, I would think the same
> there here applies as in the case of Ubuntu: they should reset the
> Maintainer field to point to themselves, and continue to give credit
> to the Debian developer in a suitable fashion.
The founder of Debian seems to disagree, but still. The DCCA situation
suggests that we need to define exactly what we want and make it clear
to all derived distributions that this is what we expect. This isn't
something that only affects Ubuntu - we're talking about a large number
of fairly major distributions.
--
Matthew Garrett | mjg59@srcf.ucam.org
--
To UNSUBSCRIBE, email to debian-devel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
| |
| Thomas Bushnell BSG 2006-01-18, 2:50 am |
| Matthew Garrett <mjg59@srcf.ucam.org> writes:
>
> Some of them, yes. Mostly the backports.
What happens to the maintainer field in these cases?
Certainly, if they are modifying the packages, I would think the same
there here applies as in the case of Ubuntu: they should reset the
Maintainer field to point to themselves, and continue to give credit
to the Debian developer in a suitable fashion.
Thomas
--
To UNSUBSCRIBE, email to debian-devel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
| |
| Hamish Moffatt 2006-01-18, 2:50 am |
| On Tue, Jan 17, 2006 at 04:54:36PM -0800, Thomas Bushnell BSG wrote:
> Matt Zimmerman <mdz@debian.org> writes:
>
>
> How does the behavior of other Debian derivatives matter?
>
> As a rule, those other derivatives do not cooperate with Debian. If
> Ubuntu wants to be like them, fine, but don't say you cooperate with
> Debian if that's what you want to do.
To be fair, co-operation and attribution are really separate issues.
We do need to be consistent about each. Any complaint we have about
co-operation with Ubuntu should not mean we have special requirements
with regard to attribution.
Hamish
--
Hamish Moffatt VK3SB <hamish@debian.org> <hamish@cloud.net.au>
--
To UNSUBSCRIBE, email to debian-devel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
| |
| Wouter Verhelst 2006-01-18, 2:50 am |
| On Tue, Jan 17, 2006 at 04:54:36PM -0800, Thomas Bushnell BSG wrote:
> Matt Zimmerman <mdz@debian.org> writes:
>
>
> How does the behavior of other Debian derivatives matter?
How does it not matter?
--
..../ -/ ---/ .--./ / .--/ .-/ .../ -/ ../ -./ --./ / -.--/ ---/ ..-/ .-./ / -/
.../ --/ ./ / .--/ ../ -/ ..../ / -../ ./ -.-./ ---/ -../ ../ -./ --./ / --/
-.--/ / .../ ../ --./ -./ .-/ -/ ..-/ .-./ ./ .-.-.-/ / --/ ---/ .-./ .../ ./ /
.../ .../ / ---/ ..-/ -/ -../ .-/ -/ ./ -../ / -/ ./ -.-./ ..../ -./ ---/ .-../
---/ --./ -.--/ / .-/ -./ -.--/ .--/ .-/ -.--/ .-.-.-/ / ...-.-/
--
To UNSUBSCRIBE, email to debian-devel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
| |
| Otavio Salvador 2006-01-18, 5:56 pm |
| Adam Heath <doogie@debian.org> writes:
> * 1 FETCH (BODY[TEXT] {1008}
> On Tue, 17 Jan 2006, Otavio Salvador wrote:
>
>
> But linked against other libraries. The binary is downloaded from another
> location(or installed from a different cd set). The program used to do the
> download may be different.
Using this as rule, then all Debian CDD distributions would need to
recompile all sources to change the maintainer field. This include
Debian-EDU, Debian-BR-CDD and others. That's what you think is
correct?
In case of CDDs, the only exception is it isn't build against other
libraries but it is installed by different cd set and downloaded from
another location in many cases.
--
O T A V I O S A L V A D O R
---------------------------------------------
E-mail: otavio@debian.org UIN: 5906116
GNU/Linux User: 239058 GPG ID: 49A5F855
Home Page: http://www.freedom.ind.br/otavio
---------------------------------------------
"Microsoft gives you Windows ... Linux gives
you the whole house."
--
To UNSUBSCRIBE, email to debian-devel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
| |
| Andreas Tille 2006-01-18, 5:56 pm |
| On Wed, 18 Jan 2006, Otavio Salvador wrote:
>
> Using this as rule, then all Debian CDD distributions would need to
> recompile all sources to change the maintainer field.
I'm sorry if we agree that a CDD is what we defined under
http://people.debian.org/~tille/cdd...t.en.html#s-CDD
then no change is necessary.
"To clarify a common misunderstanding: Custom Debian Distributions are
not forks from Debian. They are completely included, and if you obtain
the complete Debian GNU/Linux distribution, you have all available
Custom Debian Distributions included."
(I know that the name CDD is very confusing and many people think that
it is something else.)
> In case of CDDs, the only exception is it isn't build against other
> libraries but it is installed by different cd set and downloaded from
> another location in many cases.
If it is a CDD than it is installed from a Debian mirror and nothing else.
Kind regards
Andreas.
--
http://fam-tille.de
--
To UNSUBSCRIBE, email to debian-devel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
| |
| Otavio Salvador 2006-01-18, 5:56 pm |
| Andreas Tille <tillea@rki.de> writes:
>
> If it is a CDD than it is installed from a Debian mirror and nothing else.
Debian-EDU is available in Debian but also outside of it since they
need to do more updated that aren't allowed in our stable versions. In
that case, they would need to recompile all source again to change
the maintainer field.
--
O T A V I O S A L V A D O R
---------------------------------------------
E-mail: otavio@debian.org UIN: 5906116
GNU/Linux User: 239058 GPG ID: 49A5F855
Home Page: http://www.freedom.ind.br/otavio
---------------------------------------------
"Microsoft gives you Windows ... Linux gives
you the whole house."
--
To UNSUBSCRIBE, email to debian-devel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
| |
| Andreas Tille 2006-01-18, 5:56 pm |
| On Wed, 18 Jan 2006, Otavio Salvador wrote:
> Debian-EDU is available in Debian but also outside of it since they
Well, that's a "temporary" hack until we have implemented solutions which
makes this superfluous.
Kind regards
Andreas.
--
http://fam-tille.de
--
To UNSUBSCRIBE, email to debian-devel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
| |
| Stephen Frost 2006-01-18, 5:56 pm |
| * Matt Zimmerman (mdz@ubuntu.com) wrote:
> On Tue, Jan 17, 2006 at 03:07:25PM -0500, Stephen Frost wrote:
>
> You quite obviously haven't read
> http://lists.debian.org/debian-deve...5/msg00260.html yet, where I
> wrote (among other important things), "it would be fairly straightforward
> for Ubuntu to override the Maintainer field in binary packages". I
> explained exactly what is and isn't difficult and for whom.
Wow, is this ever silly. Of course I read it and I appreciate your
position that it's more work than not doing anything different from what
you're doing now but I simply disagree about it and it seems like a
pretty straight-forward solution to implement. I also understand that
not all Debian derivatives are changing the Maintainer field and that
Debian's not specifically chastising them for it. There are reasons for
each though. Other Debian derivatives aren't (or at least, don't seem)
as popular so it's less of an issue; other derivatives don't come across
as pulling resources away from Debian (which Ubuntu seems to be doing,
reality aside, that's the perception); other derivatives didn't ask and
sometimes that's just the burden you have to bear when you're actively
trying to do the right thing; other derivatives (some portion of them
anyway, I expect) don't recompile packages (which makes leaving the
Maintainer field alone somewhat less of an offense to some).
> If you're going to attack me, please do it on the basis of what I've
> actually said. Honestly, I expected better from you, give that you've acted
> like a human being toward me on IRC on several occasions in the past.
Funny, I didn't think I was attacking you at all. Rereading what you
quoted above I really don't see how that's an attack and I'm afraid
perhaps you've gotten a little sensitive on this. I'm happy enough to
excuse that as I'm sure you've gotten a fair number of poor reactions
from others. Looking through my other emails on the subject it seems
perhaps unkind of me to say you're ignoring the answer but, well, that's
how it's coming across. :/
Thanks,
Stephen
| |
| Stephen Frost 2006-01-18, 5:56 pm |
| * Matt Zimmerman (mdz@ubuntu.com) wrote:
> On Wed, Jan 18, 2006 at 12:34:33AM +0100, Florian Weimer wrote:
>
> I can't accept this; if there is no principle here which should be applied
> consistently, then it's entirely unfair to attack Ubuntu. Certainly, there
> are things about Ubuntu which are unique, but none of them change the issues
> at hand.
Personally I think the principle *should* be applied consistently but as
a volunteer and with generally not much time I'm not going to hunt down
every Debian derivative out there, see what they do and complain at them
if they're not doing it the right way. I doubt it'd have any effect in
the majority of the cases anyway. Ubuntu, by trying to do the right
thing (which many of us appreciate) and by asking the question of what
*should* be done has put themselves in a position where if they don't do
what 'should' be done, regardless of what others do, they're going to
seem like bad guys.
Also, I'm afraid, given Ubuntu's popularity and the impression
(unfounded or not) that Ubuntu is taking resources away from Debian is
going to mean Ubuntu will be held to a higher standard than other
derivatives. I think many of us would like to see Ubuntu be the
best derivative and always do the right thing and that's why there's
more pressure on Ubuntu than other derivatives.
> Seriously, it's entirely unreasonable to single out Ubuntu on this issue.
Perhaps so, but then Ubuntu's just another derivative and not the
derivative many of us would like to see it be, and I expect the
derivative that Ubuntu itself would like to be from a PR standpoint.
Thanks,
Stephen
| |
| Kevin B. McCarty 2006-01-18, 5:56 pm |
| -----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Anthony Towns wrote:
> On Tue, Jan 17, 2006 at 04:38:29PM -0800, Matt Zimmerman wrote:
>
> However if a binNMU screws up a maintainer's package, the maintainer can
> easily fix it, and doing so is just part of contributing to Debian. The
> same thing applies when an autobuild on another architecture happens.
> That's not the case if an Ubuntu rebuild screws things up.
Let me make it clear that this is not only a hypothetical worry, it's
actually happened. (Sorry to people who already saw my similar email
sent to -project, but I thought the point was worth repeating.) One of
my packages (binary package paw, from source package cernlib) has
seriously broken functionality in Breezy because it was compiled with a
buggy version of gcc. The breakage did not appear in Debian until later
[1], since Ubuntu switched to gcc-3.4 before Debian switched from 3.3 to
4.0. Once the breakage occurred in Debian I promptly uploaded a
workaround and filed a bug on gcc [2].
Since paw is not very widely used, no one was bitten by the bug in
Ubuntu until recently. An Ubuntu user emailed me about it [3] upon
finding my name in the package maintainer field (and also asked upstream
about it). If the Maintainer field included something like
ubuntu-motu@lists.ubuntu.com, instead of keeping my name and email, I
imagine the question would have worked its way to me eventually, but
without first making it look like I must be clueless not to have fixed
such an obvious bug.
References:
[1] the Debian bug report on paw: http://bugs.debian.org/324902
[2] the Debian bug report I filed on gcc: http://bugs.debian.org/325050
[3] the Ubuntu bug report on paw:
https://launchpad.net/distros/ubunt...rnlib/+bug/6588
(the user who filed the bug was nice enough to add my emailed response
to him as the second reply in the Launchpad entry)
regards,
- --
Kevin B. McCarty <kmccarty@princeton.edu> Physics Department
WWW: http://www.princeton.edu/~kmccarty/ Princeton University
GPG: public key ID 4F83C751 Princeton, NJ 08544
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org
iD8DBQFDznTYfYxAIk+Dx1ERAn3UAKDI/W2yvJOcQyaH7UXeaps+cVCW1gCbBqjo
nfcPVa0Yk+bz2hG/oXd8MM8=
=xHNq
-----END PGP SIGNATURE-----
--
To UNSUBSCRIBE, email to debian-devel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
| |
| Thomas Bushnell BSG 2006-01-18, 5:56 pm |
| Wouter Verhelst <wouter@debian.org> writes:
> On Tue, Jan 17, 2006 at 04:54:36PM -0800, Thomas Bushnell BSG wrote:
>
> How does it not matter?
Because "tu quoque" is a fallacy, and "is quoque" is even more one.
The fact that many people do a bad thing does not make the badness of
John Doe's doing it any better.
--
To UNSUBSCRIBE, email to debian-devel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
| |
| Otavio Salvador 2006-01-18, 5:56 pm |
| Andreas Tille <tillea@rki.de> writes:
> On Wed, 18 Jan 2006, Otavio Salvador wrote:
>
>
> Well, that's a "temporary" hack until we have implemented solutions which
> makes this superfluous.
But exist!
--
O T A V I O S A L V A D O R
---------------------------------------------
E-mail: otavio@debian.org UIN: 5906116
GNU/Linux User: 239058 GPG ID: 49A5F855
Home Page: http://www.freedom.ind.br/otavio
---------------------------------------------
"Microsoft gives you Windows ... Linux gives
you the whole house."
--
To UNSUBSCRIBE, email to debian-devel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
| |
| Goswin von Brederlow 2006-01-18, 5:56 pm |
| Matt Zimmerman <mdz@ubuntu.com> writes:
> On Tue, Jan 17, 2006 at 04:09:50PM -0800, Thomas Bushnell BSG wrote:
>
> I don't think you can speak to what tools we do or do not have. The fact
> is, we import most Debian source packages unmodified, and do not have any
> such tool for modifying them.
Don't you run wanna-build, buildd and sbuild? It is easy enough to
change the maintainer field with that.
>
> I don't agree. This isn't even the case within Debian. Binary-only NMUs
> don't modify the source package, even though the binaries are recompiled.
They obviously do. The version is bumped and a new changelog entry is
added.
MfG
Goswin
--
To UNSUBSCRIBE, email to debian-project-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
| |
| Goswin von Brederlow 2006-01-18, 5:56 pm |
| Mike Bird <mgb-debian@yosemite.net> writes:
> On Tue, 2006-01-17 at 17:29, Thomas Bushnell BSG wrote:
>
> In Debian yes. Ubuntu recompiles the Debian source, in a
> different environment and with different dependencies, then
> uploads with exactly the same version as Debian.
>
> Having two different package files with the exactly the same
> name and different content and dependencies drove me crazy
> for a while until we made our migration scripts smarter.
>
> --Mike Bird
Andreas Barth has some patches for the debian policy and packaging
manual from me under review that also include this situation.
In short it adds (explains the existing) an optional fourth (sub)part
to the debian version:
[epoch:]upstream_version[-debian_revision[+branch]]
The debian_revision may contain an additional branch suffix denoting a
"fork" in the debian version number. I suggest 4 types of branches [abcs]:
- 1.2-3+a0.ubuntu.1 - recompile of a package without changes
- 1.2-3+b1 - debian binary only recompile
- 1.2-3+c0.ubuntu.1 - patched source based on the debian version
- 1.2-3+s1.sarge.1 - debian security upload
If that gets accepted into policy I suggest asking all debian based
distributions to use the 'a' or 'c' branch to correctly flag
recompiles and patching. Using the distribution name in the branch
should give an unique enough version to avoid any confusion about
the origin. An unbranched version should always mean the binary is
unmodified (the md5sum matches debians).
MfG
Goswin
--
To UNSUBSCRIBE, email to debian-project-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
| |
| Andreas Tille 2006-01-18, 5:56 pm |
| On Wed, 18 Jan 2006, Otavio Salvador wrote:
>
> But exist!
Sure they exist, but the statement you made about the maintainer field
was simply wrong, because it makes no sense to change the maintainer
field of Debian internal packages.
Kind regards
Andreas.
--
http://fam-tille.de
--
To UNSUBSCRIBE, email to debian-devel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
| |
| Ian Murdock 2006-01-18, 8:49 pm |
| Matthew Garrett wrote:
>
> The founder of Debian seems to disagree, but still. The DCCA situation
> suggests that we need to define exactly what we want and make it clear
> to all derived distributions that this is what we expect. This isn't
> something that only affects Ubuntu - we're talking about a large number
> of fairly major distributions.
Come on, Matthew, that's a slight misrepresentation, now isn't it?
Matthew is right that this has been common practice for as long as
derivatives have been around (6-7 years now)--it's just that Ubuntu
takes derivation much further than any of the rest
of us have done, so the problem is a bit more pronounced for Ubuntu.
Fact is, the potential for confusion here never even occurred to
me when we started doing this at Progeny. Quite the contrary to what
Matthew suggests, it seems to me that changing the Maintainer
field is a perfectly reasonable thing to do now that I'm aware of it.
-ian
--
Ian Murdock
317-863-2590
http://ianmurdock.com/
"Don't look back--something might be gaining on you." --Satchel Paige
--
To UNSUBSCRIBE, email to debian-devel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
| |
| Matthew Garrett 2006-01-18, 8:49 pm |
| On Wed, Jan 18, 2006 at 08:43:48PM -0500, Ian Murdock wrote:
> Fact is, the potential for confusion here never even occurred to
> me when we started doing this at Progeny. Quite the contrary to what
> Matthew suggests, it seems to me that changing the Maintainer
> field is a perfectly reasonable thing to do now that I'm aware of it.
Ah, my apologies. I'd assumed it was something that you'd probably
have thought about, so I'll happily withdraw that.
--
Matthew Garrett | mjg59@srcf.ucam.org
--
To UNSUBSCRIBE, email to debian-devel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
| |
| Jonas Smedegaard 2006-01-18, 8:49 pm |
| -----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
On Wed, 18 Jan 2006 16:54:22 -0200
Otavio Salvador <otavio@debian.org> wrote:
> Andreas Tille <tillea@rki.de> writes:
>=20
>=20
> But exist!
I consider Debian-edu as in the development proces of becoming a CDD,
exactly because they still need such hacks.
Until all of the packages needed for a Debian-edu release is contained
completely within Debian (possibly a mix of stable, testing and
unstable), they need to host some packages outside of Debian and is
thus not a CDD by the definitions set by debian-custom.
- Jonas
- --=20
* Jonas Smedegaard - idealist og Internet-arkitekt
* Tlf.: +45 40843136 Website: http://dr.jones.dk/
- Enden er n=E6r: http://www.shibumi.org/eoti.htm
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2 (GNU/Linux)
iD8DBQFDzvZWn7DbMsAkQLgRAlL/AJ9mv14M5TB+sgUZ+THN46DZaPjfLgCdF4MB
BPEOJWK6tbFRaKcVEFsChnY=3D
=3DbL72
-----END PGP SIGNATURE-----
| |
| Thomas Bushnell BSG 2006-01-19, 2:57 am |
| Ian Murdock <imurdock@imurdock.com> writes:
> Fact is, the potential for confusion here never even occurred to
> me when we started doing this at Progeny. Quite the contrary to what
> Matthew suggests, it seems to me that changing the Maintainer
> field is a perfectly reasonable thing to do now that I'm aware of it.
Glad to hear.
Is there anyone from Debian who thinks that changing the Maintainer
field is a bad idea in these cases (remember that this isn't about
credit, because we would certainly request that the Debian maintainer
still be mentioned as such in a suitable fashion)?
Thomas
--
To UNSUBSCRIBE, email to debian-devel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
| |
| Goswin von Brederlow 2006-01-19, 8:12 am |
| Wouter Verhelst <wouter@debian.org> writes:
> On Wed, Jan 18, 2006 at 08:57:51PM +0100, Goswin von Brederlow wrote:
>
> Yes. And then the source used to build that binNMU is thrown away. It's
> a *binary* NMU, you don't see a sourceful upload with that.
Which means the Maintainer field in the binary package could easily be
changed.
MfG
Goswin
--
To UNSUBSCRIBE, email to debian-project-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
| |
| Marc Haber 2006-01-19, 8:12 am |
| On Tue, 17 Jan 2006 16:03:05 -0800, Matt Zimmerman <mdz@ubuntu.com>
wrote:
>Do you realize that Xandros, who maintains a Debian derivative which =
they
>box and sell for US$50-$129 per copy, leaves the Maintainer field
>unmodified, and as far as I'm aware, was doing so for a period of =
*years*
>before Ubuntu even existed? This never seemed to bother anyone, and
>personally, I don't think it's a big deal either.
Xandros does not employ a significant number of people in important
single-point-of-failure-positions in Debian, most notably not the
people who are notoriously known for not doing the job they have
volunteered for.
Additionally, Xandros doesn't have nearly the user base that Ubuntu
has, and they are not nearly as loud PR-wise as Ubuntu is.
Greetings
Marc
--=20
-------------------------------------- !! No courtesy copies, please !! =
-----
Marc Haber | " Questions are the | Mailadresse im =
Header
Mannheim, Germany | Beginning of Wisdom " | =
http://www.zugschlus.de/
Nordisch by Nature | Lt. Worf, TNG "Rightful Heir" | Fon: *49 621 =
72739834
| |
| Matt Zimmerman 2006-01-19, 6:25 pm |
| On Thu, Jan 19, 2006 at 02:15:15PM +0100, Marc Haber wrote:
> On Tue, 17 Jan 2006 16:03:05 -0800, Matt Zimmerman <mdz@ubuntu.com>
> wrote:
>
> Xandros does not employ a significant number of people in important
> single-point-of-failure-positions in Debian, most notably not the
> people who are notoriously known for not doing the job they have
> volunteered for.
Apart from its questionable accuracy, this is a red herring and has nothing
to do with how derivatives should treat the Maintainer field.
> Additionally, Xandros doesn't have nearly the user base that Ubuntu
> has, and they are not nearly as loud PR-wise as Ubuntu is.
Likewise, I don't think that the popularity of a derivative is an important
consideration on this point. What exactly do you consider "loud" PR?
Ubuntu doesn't exactly mount campaigns; what messaging there is is by word
of mouth. Other Debian derivatives buy ad space on Google keywords like
"Debian".
--
- mdz
--
To UNSUBSCRIBE, email to debian-devel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
|
|
|
|
|