public inbox for linux-kbuild@vger.kernel.org
 help / color / mirror / Atom feed
From: maximilian attems <max@stro.at>
To: Frans Pop <elendil@planet.nl>
Cc: linux-kbuild@vger.kernel.org,
	Andres Salomon <dilinger@debian.org>,
	tytso@mit.edu, sam@ravnborg.org
Subject: Re: Comments on deb-pkg patch series
Date: Wed, 1 Apr 2009 18:23:20 +0200	[thread overview]
Message-ID: <20090401162320.GY3901@baikonur.stro.at> (raw)
In-Reply-To: <200904011807.57574.elendil@planet.nl>

[ adding Theodore Ts'o on CC ]

Sam can you please merge 1-6 of the series as those are not contested.
thanks.

On Wed, Apr 01, 2009 at 06:07:56PM +0200, Frans Pop wrote:
> Below some comments on the patch series submitted yesterday by Maximilian 
> Attems. I was not subscribed to the kbuild list, so apologies for 
> breaking the thread. It would have been nice if the patches had been CCed 
> to lkml for general review.
> 
> I have some patches of my own that I'll submit later today.

great.

the patches were submitted to the relevant subsystem,
no need to flood lkml with such.
 
> FYI: Like Max I am a DD, but unlike him I'm not a member of the kernel 
> teamm. I have however been using the deb-pkg target intensively over the 
> past year and a half for all my kernel testing on 4 different arches.
> 
> General comment:
> It looks to me as if this patch series is trying to make the deb-pkg 
> target into something it is not. It is not a target that produces perfect 
> and Debian policy-compliant packages. Instead it is a very basic method 
> to create an installable kernel image package direct from upstream 
> source.

i strongly disagree, why shouldn't it build policy complian packages.
 
> [PATCH 1/7] deb-pkg: Beautify changelog
>             http://marc.info/?l=linux-kbuild&m=123851278623264&w=2
> 
> > -  * A standard release
> > +  * New upstream release
> 
> In my own patch series I have an alternative, which IMO better matches the 
> purpose of deb-pkg:
> -  * A standard release
> +  * Custom built Linux kernel.
> 
> The name and email changes seem somewhat overengineered to me, but 
> otherwise no objection.

well this nitpicking, anyway no strong objection on a follow up
to my patch anything but "A standard release" sounds better.
usually you will need make deb-pkg due to newer upstream
and is the standard opening of our changelogs ;)
 
> [PATCH 2/7] deb-pkg: Fix Provides field
>             http://marc.info/?l=linux-kbuild&m=123851274923192&w=2
> 
> No objection.
> 
> [PATCH 3/7] deb-pkg: bump standards version
>             http://marc.info/?l=linux-kbuild&m=123851275023204&w=2
> 
> As deb-pkg only creates binary packages and does not have a source 
> package, the created package is not actually source compliant. Instead of 
> updating the Standards-Version field we could also simply drop it (as it 
> is not strictly required). IMO it's fairly bogus anyway and would make 
> for one less thing to maintain.
> 
> No strong objection though.

as we want it policy compliant better say to which one.
 
> [PATCH 4/7] deb-pkg: Fix Section and Source field
>             http://marc.info/?l=linux-kbuild&m=123851275123210&w=2
> 
> I strongly disagree with this patch.
> 
> linux-2.6 is the source package for official Debian kernels and packages 
> built using deb-pkg are NOT built from that source package.
> IMO there's no need to change it (the field is required and thus cannot 
> simply be dropped). If it does want changing for some reason I'd suggest
> "linux-upstream" or similar.

no,
just checkout linux-2.6 git and you'll get per default a matching
linux-2.6 dir, so your arg does not stand.
 
> [PATCH 5/7] deb-pkg: Generate a debian/copyright
>             http://marc.info/?l=linux-kbuild&m=123851274923195&w=2
> 
> As the generated package is not policy compliant anyway, I see see no real 
> reason to burden it with a copyright file. No strong objection though.
> 
> > +Copyright: 1991 - 2008 Linus Torvalds and others.
> s/2008/2009/
> > +git://git.eu.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6.git
> s/eu.//
> 
> Also, the git reference is somewhat random as deb-pkg can just as well be 
> used to build kernels from any other source tree (stable, mm, tip, ...).

well a follow up can s/2008/2009/
better have the most important git tree mentioned then none.
 
> [PATCH 6/7] deb-pkg: Fix generated packagename
>             http://marc.info/?l=linux-kbuild&m=123851275023201&w=2
> 
> This is not actually a "fix". There's nothing really wrong with the 
> current package name, and I actually like the fact that packages built 
> using deb-pkg are in a somewhat different namespace than the official 
> Debian kernel image packages.
> 
> I'd prefer to leave this unchanged, but have no hard objection.

no the name is wrong, but as you don't have  an objection...
 
> [PATCH 7/7] deb-pkg: generate changelog, copyright and control on demand
>             http://marc.info/?l=linux-kbuild&m=123851275123207&w=2
> 
> NAK!
> 
> This completely breaks the most common use case of deb-pkg. This patch 
> would mean that every package would get identical (and incorrect) version 
> info in the Debian maintainer files unless you manually clean the debian 
> directory before each build.
> One of the really great things of deb-pkg is that you can simply 
> repeatedly call it after checking out different branches (and cross-build 
> for different arches) or during bisections without having to worry about 
> such things.

big non non to your arguments.
this is explicitly been asked for make deb-pkg,
private follow ups to
https://lists.linux-foundation.org/pipermail/ksummit-2008-discuss/2008-June/000191.html

an rm -rf debian would regenerate them anyway, but i see your args
and be interested how one can support both, the throwaway generation
and keeping specific debian/ files.

-- 
maks


  reply	other threads:[~2009-04-01 16:26 UTC|newest]

Thread overview: 11+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2009-04-01 16:07 Comments on deb-pkg patch series Frans Pop
2009-04-01 16:23 ` maximilian attems [this message]
2009-04-01 17:07   ` Frans Pop
2009-04-01 17:32     ` maximilian attems
2009-04-01 17:53       ` Frans Pop
2009-04-01 17:57         ` maximilian attems
2009-04-01 18:35           ` Frans Pop
2009-04-01 18:47             ` maximilian attems
2009-04-01 19:11               ` Frans Pop
2009-04-01 19:21                 ` maximilian attems
2009-04-05 19:38   ` Sam Ravnborg

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20090401162320.GY3901@baikonur.stro.at \
    --to=max@stro.at \
    --cc=dilinger@debian.org \
    --cc=elendil@planet.nl \
    --cc=linux-kbuild@vger.kernel.org \
    --cc=sam@ravnborg.org \
    --cc=tytso@mit.edu \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox