FAQ Search Today's Posts Mark Forums Read
» Video Reviews

» Linux Archive

Linux-archive is a website aiming to archive linux email lists and to make them easily accessible for linux users/developers.


» Sponsor

» Partners

» Sponsor

Go Back   Linux Archive > Debian > Debian Development

 
 
LinkBack Thread Tools
 
Old 02-05-2008, 07:57 PM
Lucas Nussbaum
 
Default rebuilding the archive in a dirty chroot: results

On 25/01/08 at 15:25 +0100, Lucas Nussbaum wrote:
> Hi,
>
> I've done two rebuilds of sid on i386.
> - one in a perfectly clean chroot, as I usually do
> - one in a chroot, where as many build-dependancies as impossible were
> installed (take the Sources file, extract the build-deps for all
> packages, and install as many packages as possible) (the chroot is
> named bdfh -- build daemon from hell)
>
> It is important that packages build in the same way (same binary
> packages generated) in both cases, because:
> - some maintainers still don't build in clean chroot
> - buildds don't necessarly cleanup all build-deps after build,
> so you don't know in which environment your packages are being
> built
>
> Then I compared the results, and I started crying.
>
> 236 packages built fine in the clean chroot, but failed to build in the
> dirty chroot. In some cases, it's caused by automake1.4 being
> installed, and not being removed by sbuild.
>
> 35 packages failed in both the clean and the dirty chroot, but with
> different reasons, apparently (I have a script that tries to "guess" the
> reason for a failed build, but it's far from being 100% reliable)
>
> 938 packages produced different binary packages according to debdiff. Of
> those, 477 produced different Depends line (caused by some features not
> being explicitely enabled, but not being explicitely disabled, usually).

Hi,

I did another rebuild. Thanks to the Debian GNU/kFreeBSD cabal, the
results are available on io.debian.net. If you want shell access, read
http://io.debian.net/ssh.html . If web access is enough, go to
http://io.debian.net/~lucas/bdfh-20080202/

Interesting files/dirs:
failed-packages.txt: list of packages that failed in one or the
other chroot, or which failed in both, but for a different reason.
failed-packages.dd-list.txt: same, dd-list sorted.
different-packages.txt: packages that generated different binary
packages, using debcmp.
different-packages.dd-list.txt: same, dd-list sorted.
logs.*/: build logs
pkg.*/: the generated packages
debcmp/: the output from debcmp

debcmp is an home-made script I use to compare packages, looking more
deeply than what debdiff does. It's available from
http://svn.debian.org/wsvn/collab-qa/tools/debcmp/ , if you want to
double-check the results.

My plan is to start filing bugs on the packages that failed to build,
and on those that differ in important ways (less/more files, less/more
Depends). It's already a lot of work. Is someone interested in helping
me with those tasks? Or maybe someone has an NM, and doesn't know what
he should do for T&S?

In the meantime, don't hesitate to fix your packages!
--
| Lucas Nussbaum
| lucas@lucas-nussbaum.net http://www.lucas-nussbaum.net/ |
| jabber: lucas@nussbaum.fr GPG: 1024D/023B3F4F |


--
To UNSUBSCRIBE, email to debian-devel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
 
Old 02-05-2008, 08:27 PM
Cyril Brulebois
 
Default rebuilding the archive in a dirty chroot: results

On 05/02/2008, Lucas Nussbaum wrote:
> I did another rebuild. Thanks to the Debian GNU/kFreeBSD cabal,

o/

> debcmp is an home-made script I use to compare packages, looking
> more deeply than what debdiff does. It's available from
> http://svn.debian.org/wsvn/collab-qa/tools/debcmp/ , if you want to
> double-check the results.

Having a look at some packages of mine, some suggestions:
- Please blacklist .pdf at least, timestamps stuff and so on are
likely to be at fault here.
- You could eventually no longer take the order of entries in shlibs
file into account. I'm not sure this order matters anyway.
- Harder, but you could try and ignore hunks where only a line
changed, which contains “Generated on $date”, that's likely to be
doxygen or a friend of its. Maybe ignoring .html files would do?
There's also docbook-to-man.
- I didn't check your script but I guess it might be enhanced WRT
file type detection, see for example synfigstudio log in debcmp/.

I really suggest you postprocess your logs for the shlibs problem, it
would make it trivial to spot actual problem, also preventing from
missing some, hidden in an order modification.

> Or maybe someone has an NM, and doesn't know what he should do for
> T&S?

Not yet, already passed, too late.

--
Cyril Brulebois
 
Old 02-05-2008, 09:32 PM
Alexander Schmehl
 
Default rebuilding the archive in a dirty chroot: results

Hi!

* Lucas Nussbaum <lucas@lucas-nussbaum.net> [080205 21:57]:
> I did another rebuild. Thanks to the Debian GNU/kFreeBSD cabal, the
> results are available on io.debian.net. If you want shell access, read
> http://io.debian.net/ssh.html . If web access is enough, go to
> http://io.debian.net/~lucas/bdfh-20080202/
[..]
> In the meantime, don't hesitate to fix your packages!

I thought I did with fillets-ng 0.8.0-2, but now I'm slightly confused.

If I see correclty, the contents and dependencies are the same, but in different order,
so I think I fixed the libfridi problem successfully.

But I don't understand the following differences:

The normal buildlog has:
Files:
b96a42b062fd019fef7b37be44d6e784 818 games optional fillets-ng_0.8.0-2.dsc
4b56ebb7ed2ae32c2e409d0d3dc9e1e5 6752 games optional fillets-ng_0.8.0-2.diff.gz
52b045d74a42f364d69642da602f97f3 293038 games optional fillets-ng_0.8.0-2_i386.deb

While the bdfh buildlog lists:
Files:
f8b0f31fb8358cb43ab436ea1f524c10 818 games optional fillets-ng_0.8.0-2.dsc
254864c6a092d0f6cf07795e4b3f27ba 8780 games optional fillets-ng_0.8.0-2.diff.gz
3182e8fcaf977c11e4e81d4b7a191acb 293086 games optional fillets-ng_0.8.0-2_i386.deb

Okay, since the content of the .deb has a different order, the md5sum is
of course changed. The dsc changed it's md5sum, because the diff.gz
changed, looks like some autofoo stuff?

Why that? Both, the diff.gz and the .dsc, are part of the source
package, so why did they change at all?


Yours sincerely,
Alexander
 
Old 02-05-2008, 09:43 PM
Frank Lichtenheld
 
Default rebuilding the archive in a dirty chroot: results

On Tue, Feb 05, 2008 at 09:57:16PM +0100, Lucas Nussbaum wrote:
> On 25/01/08 at 15:25 +0100, Lucas Nussbaum wrote:
> I did another rebuild. Thanks to the Debian GNU/kFreeBSD cabal, the
> results are available on io.debian.net. If you want shell access, read
> http://io.debian.net/ssh.html . If web access is enough, go to
> http://io.debian.net/~lucas/bdfh-20080202/

http://io.debian.net/~lucas/bdfh-20080202/logs.bdfh/libnet-ssh-perl-perl_1.30-1_sid32-bdfh.buildlog
ends with:
Manifying blib/man3/NeE: Caught signal 'Terminated': terminating immediately

Any idea what that is about?

Gruesse,
--
Frank Lichtenheld <djpig@debian.org>
www: http://www.djpig.de/


--
To UNSUBSCRIBE, email to debian-devel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
 
Old 02-05-2008, 09:55 PM
Cyril Brulebois
 
Default rebuilding the archive in a dirty chroot: results

On 05/02/2008, Alexander Schmehl wrote:
> Why that? Both, the diff.gz and the .dsc, are part of the source
> package, so why did they change at all?

See the build log:
,----
| gpg: Signature made Tue Jan 29 21:33:22 2008 CET using DSA key ID 00D8CD16
| gpg: Can't check signature: public key not found
| dpkg-source: extracting fillets-ng in fillets-ng-0.8.0
| dpkg-source: unpacking fillets-ng_0.8.0.orig.tar.gz
| dpkg-source: applying ./fillets-ng_0.8.0-2.diff.gz
| dpkg-buildpackage: source package fillets-ng
| dpkg-buildpackage: source version 0.8.0-2
| dpkg-buildpackage: source changed by Alexander Schmehl <tolimar@debian.org>
| dpkg-buildpackage: host architecture i386
| /usr/bin/fakeroot debian/rules clean
| QUILT_PATCHES=debian/patches quilt --quiltrc /dev/null pop -a -R || test $? = 2
| No patch removed
| rm -rf .pc debian/stamp-patched
| dh_testdir
| dh_testroot
| rm -f build-stamp
| # Add here commands to clean up after the build process.
| [ ! -f Makefile ] || /usr/bin/make distclean
| dh_clean debian/fillets-ng.png
| dpkg-source -b fillets-ng-0.8.0
| dpkg-source: building fillets-ng using existing fillets-ng_0.8.0.orig.tar.gz
| dpkg-source: building fillets-ng in fillets-ng_0.8.0-2.diff.gz
| dpkg-source: building fillets-ng in fillets-ng_0.8.0-2.dsc
| debian/rules build
`----

There you are. -b/-B aren't used, the source package gets (re)built,
and due to gzip magic, the diff.

Cheers,

--
Cyril Brulebois
 
Old 02-05-2008, 10:29 PM
Lucas Nussbaum
 
Default rebuilding the archive in a dirty chroot: results

On 05/02/08 at 23:43 +0100, Frank Lichtenheld wrote:
> On Tue, Feb 05, 2008 at 09:57:16PM +0100, Lucas Nussbaum wrote:
> > On 25/01/08 at 15:25 +0100, Lucas Nussbaum wrote:
> > I did another rebuild. Thanks to the Debian GNU/kFreeBSD cabal, the
> > results are available on io.debian.net. If you want shell access, read
> > http://io.debian.net/ssh.html . If web access is enough, go to
> > http://io.debian.net/~lucas/bdfh-20080202/
>
> http://io.debian.net/~lucas/bdfh-20080202/logs.bdfh/libnet-ssh-perl-perl_1.30-1_sid32-bdfh.buildlog
> ends with:
> Manifying blib/man3/NeE: Caught signal 'Terminated': terminating immediately
>
> Any idea what that is about?

Yes, for some reason, that build just blocked for hours. I don't know
what happened. I'll try to reproduce it when filing bugs.
--
| Lucas Nussbaum
| lucas@lucas-nussbaum.net http://www.lucas-nussbaum.net/ |
| jabber: lucas@nussbaum.fr GPG: 1024D/023B3F4F |


--
To UNSUBSCRIBE, email to debian-devel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
 
Old 02-05-2008, 10:33 PM
Lucas Nussbaum
 
Default rebuilding the archive in a dirty chroot: results

On 05/02/08 at 23:32 +0100, Alexander Schmehl wrote:
> Hi!
>
> * Lucas Nussbaum <lucas@lucas-nussbaum.net> [080205 21:57]:
> > I did another rebuild. Thanks to the Debian GNU/kFreeBSD cabal, the
> > results are available on io.debian.net. If you want shell access, read
> > http://io.debian.net/ssh.html . If web access is enough, go to
> > http://io.debian.net/~lucas/bdfh-20080202/
> [..]
> > In the meantime, don't hesitate to fix your packages!
>
> I thought I did with fillets-ng 0.8.0-2, but now I'm slightly confused.
>
> If I see correclty, the contents and dependencies are the same, but in different order,
> so I think I fixed the libfridi problem successfully.
>
> But I don't understand the following differences:
>
> The normal buildlog has:
> Files:
> b96a42b062fd019fef7b37be44d6e784 818 games optional fillets-ng_0.8.0-2.dsc
> 4b56ebb7ed2ae32c2e409d0d3dc9e1e5 6752 games optional fillets-ng_0.8.0-2.diff.gz
> 52b045d74a42f364d69642da602f97f3 293038 games optional fillets-ng_0.8.0-2_i386.deb
>
> While the bdfh buildlog lists:
> Files:
> f8b0f31fb8358cb43ab436ea1f524c10 818 games optional fillets-ng_0.8.0-2.dsc
> 254864c6a092d0f6cf07795e4b3f27ba 8780 games optional fillets-ng_0.8.0-2.diff.gz
> 3182e8fcaf977c11e4e81d4b7a191acb 293086 games optional fillets-ng_0.8.0-2_i386.deb
>
> Okay, since the content of the .deb has a different order, the md5sum is
> of course changed. The dsc changed it's md5sum, because the diff.gz
> changed, looks like some autofoo stuff?

It is. See
http://io.debian.net/~lucas/bdfh-20080202/debcmp/fillets-ng_0.8.0-2.log
--
| Lucas Nussbaum
| lucas@lucas-nussbaum.net http://www.lucas-nussbaum.net/ |
| jabber: lucas@nussbaum.fr GPG: 1024D/023B3F4F |
 
Old 02-06-2008, 05:43 AM
Reinhard Tartler
 
Default rebuilding the archive in a dirty chroot: results

Lucas Nussbaum <lucas@lucas-nussbaum.net> writes:

> I did another rebuild. Thanks to the Debian GNU/kFreeBSD cabal, the
> results are available on io.debian.net. If you want shell access, read
> http://io.debian.net/ssh.html . If web access is enough, go to
> http://io.debian.net/~lucas/bdfh-20080202/

thanks for the rerun of the bdfh rebuilds. One of my packages ftbfs in
that chroot, because it failed to remove build-conflicts:
http://io.debian.net/~lucas/bdfh-20080202/logs.bdfh/pong2_0.1.2-2_sid32-bdfh.buildlog


I wonder why sbuild fails here, can someone explain that? I thought that
adding a build conflict on libssl-dev was a good idea, but it seems to
cause an FTBFS now?


--
Gruesse/greetings,
Reinhard Tartler, KeyID 945348A4


--
To UNSUBSCRIBE, email to debian-devel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
 
Old 02-06-2008, 06:45 AM
Raphael Hertzog
 
Default rebuilding the archive in a dirty chroot: results

On Wed, 06 Feb 2008, Reinhard Tartler wrote:
> Lucas Nussbaum <lucas@lucas-nussbaum.net> writes:
>
> > I did another rebuild. Thanks to the Debian GNU/kFreeBSD cabal, the
> > results are available on io.debian.net. If you want shell access, read
> > http://io.debian.net/ssh.html . If web access is enough, go to
> > http://io.debian.net/~lucas/bdfh-20080202/
>
> thanks for the rerun of the bdfh rebuilds. One of my packages ftbfs in
> that chroot, because it failed to remove build-conflicts:
> http://io.debian.net/~lucas/bdfh-20080202/logs.bdfh/pong2_0.1.2-2_sid32-bdfh.buildlog
>
> I wonder why sbuild fails here, can someone explain that? I thought that
> adding a build conflict on libssl-dev was a good idea, but it seems to
> cause an FTBFS now?

It looks like sbuild uses dpkg to remove build conflicting packages
instead of apt-get and obviously fails given that dozen of -dev packages
depend on it... (this is the build daemon from hell, remember ).

Cheers,
--
Raphal Hertzog

Le best-seller franais mis jour pour Debian Etch :
http://www.ouaza.com/livre/admin-debian/


--
To UNSUBSCRIBE, email to debian-devel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
 

Thread Tools




All times are GMT. The time now is 06:58 PM.

VBulletin, Copyright ©2000 - 2014, Jelsoft Enterprises Ltd.
Content Relevant URLs by vBSEO ©2007, Crawlability, Inc.
Copyright 2007 - 2008, www.linux-archive.org