Debian Bug report logs - #477773
ucf: should respect dpkg-divert

Toggle useless messages

Message #5 received at submit@bugs.debian.org (full text, mbox, reply):

From: Timothy G Abbott <tabbott@MIT.EDU>

To: submit@bugs.debian.org

Subject: ucf: should respect dpkg-divert

Date: Fri, 25 Apr 2008 02:35:55 -0400 (EDT)

Package: ucf
Severity: normal

When a configuration file managed by ucf is diverted (using dpkg-divert), 
ucf attempts to modify the original path for the conffile, rather than the 
diverted path.

I realize that using dpkg-divert on configuration files is not a 
particularly common operation at the moment, but I think that may change 
in the future (http://wiki.debian.org/ConfigPackages suggests a workaround 
for this problem of diverting the ucf source file, but I think that 
solution reaches too far into the ucf-managed package's internals).

I think correct behavior here is for ucf to follow diversions so that it 
updates the undiverted version of the configuration file (this is what 
would happen with a conffile), rather than fighting with the package or 
local administrator that diverted the configuration file.

The tricky issue in implementing this feature is that the ucf interface 
doesn't specify what package is running ucf as, so there isn't a good way 
to achieve correct behaviour with diversions by packages.

I think the best way to do this would be to add an option -P<package name> 
that causes ucf to ignore diversions by <package name>, e.g. for the case 
that ucf is in the postinst script for <package name>.  Most packages 
would not need to specify this option (in particular, the ones that need 
to do so are precisely those that divert and then replace a configuration 
file with a ucf-managed version).

If a patch to follow diversions and add an option to specify the package 
name used for diversion exceptions would be acceptable to you, I'd be 
willing to write one.

Thanks,

	-Tim Abbott

Message #10 received at 477773@bugs.debian.org (full text, mbox, reply):

From: Manoj Srivastava <srivasta@golden-gryphon.com>

To: control@bugs.debian.org

Cc: 477773@bugs.debian.org

Subject: This is a wishlist ittem

Date: Sun, 27 Apr 2008 10:37:47 -0500

severity 477773 wishlist
thanks

Hi,

        Diverting configuration files is not supported. As a workaround,
 if a configuration file is diverted, the maintainer script can follow
 the diversion and let ucf operate on the desired destination file.

        manoj
-- 
Save the Whales -- Harpoon a Honda.
Manoj Srivastava <srivasta@acm.org> <http://www.golden-gryphon.com/>  
1024D/BF24424C print 4966 F272 D093 B493 410B  924B 21BA DABB BF24 424C

Message #19 received at 477773@bugs.debian.org (full text, mbox, reply):

From: Timothy G Abbott <tabbott@MIT.EDU>

To: 477773@bugs.debian.org

Cc: control@bugs.debian.org

Subject: Patch attached

Date: Sat, 7 Jun 2008 16:42:16 -0400 (EDT)

tags 477773 patch
thanks

Manoj,

Somehow I missed your reply back in April; you're right that this should 
have been reported as wishlist.

I see that this has been marked as wontfix.  In case you're willing to 
have this feature but aren't willing to write a patch to implement it, 
I've attached a patch to add this functionality to ucf.  The patch causes 
ucf to follow dpkg-divert, but supports configuring it to ignore 
diversions placed by a particular package by using the -P|--package option 
(this is in analogy with how dpkg itself follows diversions except those 
placed by the package currently being installed).

Only packages that themselves perform diversions on the files they manage 
via ucf would need to use the -P option.

By the way, the patch can be substantially cleaned up once #458012 is 
fixed.

	-Tim Abbott

Message #26 received at 477773@bugs.debian.org (full text, mbox, reply):

From: Adrien Cunin <adri2000@ubuntu.com>

To: Timothy G Abbott <tabbott@MIT.EDU>

Cc: 477773@bugs.debian.org

Subject: Patch missing

Date: Wed, 11 Jun 2008 18:38:07 +0200

[Message part 1 (text/plain, inline)]

Hello,

I'd be interested in your ucf patch, but it seems you forgot to attach it to
your mail. Could you please re-send it?

Thanks.

-- 
Adrien Cunin aka Adri2000

[signature.asc (application/pgp-signature, inline)]

Message #31 received at 477773@bugs.debian.org (full text, mbox, reply):

From: Timothy G Abbott <tabbott@MIT.EDU>

To: Adrien Cunin <adri2000@ubuntu.com>

Cc: 477773@bugs.debian.org

Subject: Re: Patch missing

Date: Wed, 11 Jun 2008 13:04:38 -0400 (EDT)

[Message part 1 (text/plain, inline)]

So it seems I did.  Sorry about that.  It is now attached.

	-Tim Abbott

On Wed, 11 Jun 2008, Adrien Cunin wrote:

> Hello,
>
> I'd be interested in your ucf patch, but it seems you forgot to attach it to
> your mail. Could you please re-send it?
>
> Thanks.
>
> -- 
> Adrien Cunin aka Adri2000
>

[ucf-dpkg-divert.patch (text/x-diff, attachment)]

Message #36 received at 477773@bugs.debian.org (full text, mbox, reply):

From: Dave Rawks <dave@pandora.com>

To: Ben Hutchings <ben@decadent.org.uk>

Cc: 596767@bugs.debian.org, 477773@bugs.debian.org

Subject: Re: Bug#596767: nfs-kernel-server does not check the ucf registry for conflicting package, registrations.

Date: Mon, 13 Sep 2010 15:41:13 -0700

On 09/13/2010 03:30 PM, Ben Hutchings wrote:
> On Mon, 2010-09-13 at 14:38 -0700, Dave Rawks wrote:
>> Package: nfs-kernel-server
>> Version: 1:1.1.2-6lenny2
>> Severity: important
>>
>>
>>         In the maintaner scripts included in nfs-kernel-server ucf is
>> used to detect whether the configfiles
>> managed by the scripts have been changed via a local edit. However the
>> postinst makes no inquiry against ucf
>> regarding whether another package has registered itself with any of the
>> files managed by nfs-kernel-server's
>> scripts. As such if another package ships for instance /etc/exports AND
>> also registers it's copy of the configfile
>> with ucf then subsequent calls to the "upgrade" section of
>> nfs-kernel-server's postinst script will result in
>> nfs-nerkel-server's default version of the configfile clobbering the
>> local copy even though ti is was
>> properly registered with ucf as belonging to the other package.
> [...]
> 
> It seems like we should register /etc/exports as belonging to
> nfs-kernel-server.  However, nfs-kernel-server will not be coinstallable
> with your other package since ucfr will fail.
> 
> Ben.
> 

Fair enough, however in the case that nfs-kernel-server detects that
another package has already registered /etc/exports it would seem
consistent to prompt for an interactive merge. Any presumption on the
part of the package that results in a locally managed config file being
overwritten without any prompt or warning is a pretty critical bug.
Especially in a server as core as nfs.

Additionally this type of situation should further support the cause
championed in bug #47773 against ucf
(http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=477773). And then
alternate packages which ship a version of /etc/exports could utilize
diversion to prevent having their changes clobbered. It seems that this
bug has been marked as "will not fix" but there is no note in the bug
report that correlates with that state.

-Dave

Message #41 received at 477773-close@bugs.debian.org (full text, mbox, reply):

From: Debian FTP Masters <ftpmaster@ftp-master.debian.org>

To: 477773-close@bugs.debian.org

Subject: Bug#477773: fixed in ucf 3.0040

Date: Wed, 20 May 2020 07:19:38 +0000

Source: ucf
Source-Version: 3.0040
Done: Manoj Srivastava <srivasta@debian.org>

We believe that the bug you reported is fixed in the latest version of
ucf, which is due to be installed in the Debian FTP archive.

A summary of the changes between this version and the previous one is
attached.

Thank you for reporting the bug, which will now be closed.  If you
have further comments please address them to 477773@bugs.debian.org,
and the maintainer will reopen the bug report if appropriate.

Debian distribution maintenance software
pp.
Manoj Srivastava <srivasta@debian.org> (supplier of updated ucf package)

(This message was generated automatically at their request; if you
believe that there is a problem with it please contact the archive
administrators by mailing ftpmaster@ftp-master.debian.org)


-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA512

Format: 1.8
Date: Tue, 19 May 2020 23:47:10 -0700
Source: ucf
Architecture: source
Version: 3.0040
Distribution: unstable
Urgency: medium
Maintainer: Manoj Srivastava <srivasta@debian.org>
Changed-By: Manoj Srivastava <srivasta@debian.org>
Closes: 477773 926661 930130 949314 952448
Changes:
 ucf (3.0040) unstable; urgency=medium
 .
   * This is a feature add release, for the most part
   * Bug fix: "Add support for creating files with default SELinux
     context", thanks to Christian Göttsche (Closes: #949314).
   * Bug fix: "please consider merge request #1 adding
     ucf_helper_functions", thanks to Marc Haber (Closes: #930130).
   * Bug fix: "ucf has /bin/bash shebangs but does not depend on bash",
     thanks to James Le Cuirot (Closes: #952448). It actually does not need
     bash, so this was easy to do.
   * Bug fix: "should respect dpkg-divert", thanks to Timothy G Abbott
     (Closes: #477773).
   * Bug fix: "[INTL:nl] Dutch translation of debconf messages", thanks to
     Frans Spiesschaert (Closes: #926661).
Checksums-Sha1:
 d37499ed097a60b583f65163e1aa3cc049047dca 1423 ucf_3.0040.dsc
 687101abef69af0cc16825227c70ac94dc2b79f1 70268 ucf_3.0040.tar.xz
Checksums-Sha256:
 81d2614c3ad81b3277d86446145198a1b774aeee8082e6fee69f5dff4b54d6ee 1423 ucf_3.0040.dsc
 6e1dcc178ea24ee05c79952bd26241191f5c2d9a29e255f66f4f11194b6a33d0 70268 ucf_3.0040.tar.xz
Files:
 a846166d27193980932bccbd5c6f3a90 1423 utils standard ucf_3.0040.dsc
 fddd0f742cb655f99718cb0bfbe833af 70268 utils standard ucf_3.0040.tar.xz

-----BEGIN PGP SIGNATURE-----
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=167i
-----END PGP SIGNATURE-----

Send a report that this bug log contains spam.