Debian Bug report logs - #452215
Add a check to make sure the .diff.gz is clean when debian/patches is used

Toggle useless messages

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

From: Raphael Geissert <atomo64@gmail.com>

To: Debian Bug Tracking System <submit@bugs.debian.org>

Subject: Add a check to make sure the .diff.gz is clean when debian/patches is used

Date: Tue, 20 Nov 2007 20:12:58 -0600

Package: lintian
Version: 1.23.36
Severity: wishlist

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

It would be nice to have some warnings in case the modification in .diff.gz go out of the scope of 
debian/ despite the existence of debian/patches (altough also reporting when the package build-depends 
on dpatch or quilt shouldn't harm).

This is based on the idea that if debian/patches is used it means the maintainer may have done something 
wrong (.patch left behind under some dir? what about a somedir.orig forgotten?).

Regards,
Raphael

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)

iD8DBQFHQ5QqYy49rUbZzloRAh1bAJ93uQQ5KLqWkHflBAkpXywf7VNvGgCePfop
rE+UYuuGEU0QiV3jdJHHgt4=
=ksKw
-----END PGP SIGNATURE-----

Message #14 received at 452215@bugs.debian.org (full text, mbox, reply):

From: Russ Allbery <rra@debian.org>

To: Raphael Geissert <atomo64@gmail.com>

Cc: 452215@bugs.debian.org

Subject: Re: Bug#452215: Add a check to make sure the .diff.gz is clean when debian/patches is used

Date: Tue, 20 Nov 2007 18:40:27 -0800

Raphael Geissert <atomo64@gmail.com> writes:

> Package: lintian
> Version: 1.23.36
> Severity: wishlist
>
> It would be nice to have some warnings in case the modification in
> .diff.gz go out of the scope of debian/ despite the existence of
> debian/patches (altough also reporting when the package build-depends on
> dpatch or quilt shouldn't harm).
>
> This is based on the idea that if debian/patches is used it means the
> maintainer may have done something wrong (.patch left behind under some
> dir? what about a somedir.orig forgotten?).

This is a very good idea.  I'll work on including something like that in
the next release.

-- 
Russ Allbery (rra@debian.org)               <http://www.eyrie.org/~eagle/>

Message #19 received at 452215-submitter@bugs.debian.org (full text, mbox, reply):

From: www-data <www-data@wolffelaar.nl>

To: control@bugs.debian.org, 452215-submitter@bugs.debian.org

Subject: Lintian bugs fixed in revision r1238

Date: Mon, 03 Mar 2008 11:49:02 +0100

package lintian
# Fixed in r1238 by he
tag 452215 + pending
thanks

These bugs are fixed in revision 1238 by he
Log message:
* check/patch-systems:
  + [HE] Warn if the .diff.gz contains changes while the package uses
    a patch system. (Closes: #452215)

Message #24 received at 452215@bugs.debian.org (full text, mbox, reply):

From: Russ Allbery <rra@debian.org>

To: 452215@bugs.debian.org

Subject: Re: lintian: r1238 - in trunk: checks debian testset testset/diffs/debian testset/diffs/debian/patches

Date: Mon, 03 Mar 2008 10:49:33 -0800

he@debian.org writes:

> * check/patch-systems:
>   + [HE] Warn if the .diff.gz contains changes while the package uses
>     a patch system. (Closes: #452215)

Some people have argued on debian-devel that the correct way to use a
patch system is to apply all the patches in the clean target rather than
in the build target so that the source package is shipped with all the
patches applied.  The rationale is that this allows someone preparing an
NMU to just ignore the patch system (at the cost of desynchronizing the
changes from the patch system for the maintainer).

I don't know if we want to allow for that or not.  I find the logic
dubious, but dpatch supports that mode of operation.

-- 
Russ Allbery (rra@debian.org)               <http://www.eyrie.org/~eagle/>

Message #29 received at 452215@bugs.debian.org (full text, mbox, reply):

From: Marc 'HE' Brockschmidt <he@ftwca.de>

To: 452215@bugs.debian.org

Subject: Re: Bug#452215: lintian: r1238 - in trunk: checks debian testset testset/diffs/debian testset/diffs/debian/patches

Date: Mon, 03 Mar 2008 22:08:28 +0100

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

Russ Allbery <rra@debian.org> writes:
> he@debian.org writes:
>> * check/patch-systems:
>>   + [HE] Warn if the .diff.gz contains changes while the package uses
>>     a patch system. (Closes: #452215)
> Some people have argued on debian-devel that the correct way to use a
> patch system is to apply all the patches in the clean target rather than
> in the build target so that the source package is shipped with all the
> patches applied.  The rationale is that this allows someone preparing an
> NMU to just ignore the patch system (at the cost of desynchronizing the
> changes from the patch system for the maintainer).

I have to admit that I did not follow that thread on -devel, somewhere
after the 200th mail, I lost interest.

> I don't know if we want to allow for that or not.  I find the logic
> dubious, but dpatch supports that mode of operation.

I have to admit that I consider using a patch system in this way to be a
bug. I think we might want to upload with this check and see if there
are too many "false" positives.

OTOH, what might be more important would be to introduce checks for
other patch systems, but I fear I don't know enough about them to add
some checks (especially after seeing how much I didn't know about
dpatch...)

Marc
-- 
BOFH #179:
multicasts on broken packets

[Message part 2 (application/pgp-signature, inline)]

Message #34 received at 452215@bugs.debian.org (full text, mbox, reply):

From: Russ Allbery <rra@debian.org>

To: 452215@bugs.debian.org

Subject: Re: Bug#452215: lintian: r1238 - in trunk: checks debian testset testset/diffs/debian testset/diffs/debian/patches

Date: Mon, 03 Mar 2008 13:42:53 -0800

Marc 'HE' Brockschmidt <he@ftwca.de> writes:
> Russ Allbery <rra@debian.org> writes:

>> I don't know if we want to allow for that or not.  I find the logic
>> dubious, but dpatch supports that mode of operation.

> I have to admit that I consider using a patch system in this way to be a
> bug. I think we might want to upload with this check and see if there
> are too many "false" positives.

I'm okay with that.  I just wanted to warn in advance, since I expect
there will be at least some controversy over it (even though the check has
also been requested by others).

> OTOH, what might be more important would be to introduce checks for
> other patch systems, but I fear I don't know enough about them to add
> some checks (especially after seeing how much I didn't know about
> dpatch...)

Adding more tests for quilt is on my list, but behind various other things
(such as Policy work).

-- 
Russ Allbery (rra@debian.org)               <http://www.eyrie.org/~eagle/>

Message #39 received at 452215-close@bugs.debian.org (full text, mbox, reply):

From: Russ Allbery <rra@debian.org>

To: 452215-close@bugs.debian.org

Subject: Bug#452215: fixed in lintian 1.23.46

Date: Tue, 04 Mar 2008 02:47:03 +0000

Source: lintian
Source-Version: 1.23.46

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

lintian_1.23.46.dsc
  to pool/main/l/lintian/lintian_1.23.46.dsc
lintian_1.23.46.tar.gz
  to pool/main/l/lintian/lintian_1.23.46.tar.gz
lintian_1.23.46_all.deb
  to pool/main/l/lintian/lintian_1.23.46_all.deb



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 452215@bugs.debian.org,
and the maintainer will reopen the bug report if appropriate.

Debian distribution maintenance software
pp.
Russ Allbery <rra@debian.org> (supplier of updated lintian 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@debian.org)


-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Format: 1.7
Date: Mon, 03 Mar 2008 18:21:32 -0800
Source: lintian
Binary: lintian
Architecture: source all
Version: 1.23.46
Distribution: unstable
Urgency: low
Maintainer: Debian Lintian Maintainers <lintian-maint@debian.org>
Changed-By: Russ Allbery <rra@debian.org>
Description: 
 lintian    - Debian package checker
Closes: 410042 421011 452215 463474 464992 465132 465192 465258 466501 466617 466665 466701 466714 466979 467019 467425 467642 468362 468370 468759 468927 469024 469106 469222
Changes: 
 lintian (1.23.46) unstable; urgency=low
 .
   The "five Lintian committers!" release.
 .
   * checks/binaries:
     + [RA] Anchor file matches for Perl libraries and debugging symbols so
       as to not match partial paths.  Based on a patch from Niko Tyni.
       (Closes: #466501)
     + [RA] Also check for make -i when checking if errors from make clean or
       distclean are ignored.  Thanks, Jörg Sommer.  (Closes: #469024)
   * checks/changelog-file:
     + [RA] Adjust for the extra space added by Parse::DebianChangelog when
       checking line length.  Thanks, Robert Millan.  (Closes: #467019)
   * checks/common_data.pm:
     + [RA] Accept (and ignore for now) the new Checksums-Sha1,
       Checksums-Sha256, and Checksums-Md5 fields in source packages.
       Thanks, Raphael Hertzog.  (Closes: #466979)
   * checks/control-file:
     + [RA] Skip empty dependencies for stronger-dependency-implies-weaker.
   * checks/control-files{.desc,}:
     + [FL] Warn about empty control files. I can't see any use for
       them. Exclude udebs though, since they might differ there.
   * checks/copyright-file{.desc,}:
     + [RA] Tests for copyright notices must be case-insensitive.  Patch
       from Chris Lamb.  (Closes: #464992)
     + [RA] When looking for cases where Copyright is separated from a year
       by a newline, also allow (C) before or after the newline.  Thanks,
       Michael Meskes.  (Closes: #465258)
     + [RA] Further explain why lintian looks for a valid copyright notice.
       (Closes: #466714)
     + [RA] Suggest an override if only the Debian packaging or some other
       part of the package not linked with OpenSSL is covered by the GPL.
       Thanks, Andrew Pollock.  (Closes: #469106)
   * checks/cruft{.desc,}:
     + [FL] Add hg to the list of <vcs>-control-dir tags.
     + [RA] In the long description of outdated-autotools-helper-file,
       point to autotool-dev's documentation.  (Closes: #465132)
     + [RA] Remove now-unnecessary workarounds for calling File::Find with
       symlink starting directory.
   * checks/debian-readme{.desc,}:
     + [RA] Warn of references to /usr/doc in README.Debian.  Thanks,
       Guillem Jover.  (Closes: #465192)
   * checks/files{.desc,}:
     + [FL,RA] Merge all the tags for vcs ignore files, inventory files,
       and vcs control dirs. Since the explanations don't actually differ,
       different tags make no sense. While at it, add the mercurial
       equivalents as well and update some of the tag descriptions. Partly
       based on a patch by Chris Lamb.
   * checks/md5sums{.desc,}:
     + [FL] Remove tag md5sums-control-file-is-empty in favour of
       a more general tag in control-files.
   * checks/manpages.desc:
     + [RA] Mention in binary-without-manpage that if the man pages are
       provided by another package on which this package depends, a lintian
       override is appropriate.
   * checks/menu-format:
     + [RA] When the menu or desktop file invokes a quoted command via an
       su program, don't try to check whether the command is in the
       package.  Thanks, Paul Wise.  (Closes: #466665)
   * checks/menus{,desc}:
     + [RA] Check that doc-base files are encoded in UTF-8.  Thanks, Robert
       Luberda.  (Closes: #468759)
     + [RA] Check doc-base sections against the new canonical list.
       Thanks, Robert Luberda.  (Closes: #463474)
   * check/patch-systems:
     + [HE] Warn if the .diff.gz contains changes while the package uses
       a patch system. (Closes: #452215)
   * checks/scripts{.desc,}:
     + [FL] Warn if a maintainer script is a valid shell script but
       doesn't seem to have any actual code.  Based on an idea
       by Justin Pryzby.  (Closes: #410042)
     + [RA] Update regexes based on checkbashisms to remove additional
       false positives and add additional checks.  Thanks, Adam D. Barratt.
     + [RA] Add kaptain.  Thanks, Tobias Toedter.  (Closes: #466701)
     + [RA] Allow zsh-beta as an alternative to zsh.  Thanks, Joost van
       Baal.  (Closes: #468362)
     + [RA] Add yorick.  Thanks, Thibaut Paumard.  (Closes: #468370)
     + [RA] Add warnings for dpkg --assert* flags for features that have
       been guaranteed for nearly a decade.  Thanks, Bernhard R. Link.
       (Closes: #467642)
   * checks/shared-libs{,desc}:
     + [RA] Rename sharedobject-in-library-directory-not-actually-a-shlib
       to sharedobject-in-library-directory-missing-soname for clarity.
       Thanks, Peter Eisentraut.  (Closes: #467425)
 .
   * data/doc-base/sections:
     + [RA] New file listing known doc-base sections.
 .
   * debian/dirs:
     + [RA] Install the base directory for the Lintian root here.
   * debian/rules:
     + [RA] Redo the installation rules to support arbitrary nesting under
       the copied directories.  As a side effect, install
       lib/Tags/ColonSeparated.pm, unbreaking the experimental
       colon-separated output format.
 .
   * frontend/lintian:
     + [CW] Make the presence of an Ubuntu release name in the version number
       trigger the Ubuntu distribution field checks.
     + [RA] Fix option parsing bug leading lintian to incorrectly reject -a
       or -p without a package.
 .
   * lib/Dep.pm:
     + [RA] Skip empty dependencies, avoiding false positives for
       stronger-dependency-implies-weaker.  Patch from Adam D. Barratt.
       (Closes: #469222)
   * lib/Lintian/Data.pm:
     + [RA] New module for loading and querying lists of keywords.
   * lib/Read_pkglists.pm:
     + [RA] Extract Uploaders from the source package list.
   * lib/Spelling.pm:
     + [RA] Add another spelling correction.
     + [RA] Merriam-Webster and the OED have both given up on publically
       vs. publicly, so we should as well.  (Closes: #466617)
 .
   * reporting/html_reports:
     + [RA] Load the correct configuration file.
     + [RA] Include packages for which one is an uploader in one's
       maintainer page, marked accordingly.  (Closes: #421011)
   * reporting/templates/*:
     + [TK] Correct Lintian maintainers address in page footers.
   * reporting/templates/maintainer.html:
     + [RA] Include packages for which the maintainer is an uploader.
 .
   * unpack/list-srcpkg:
     + [RA] Include Uploaders in the package list.
   * unpack/unpack-srcpkg-l2:
     + [RA] Don't rely on the output format of dpkg-source to learn the
       source directory and create a symlink.  Instead, tell dpkg-source to
       unpack the source directly into the unpacked directory.  Discard
       output of dpkg-source to allow for the current version, which prints
       things even with -q (fixed in later versions).  Thanks, Raphael
       Hertzog.  (Closes: #468927)
Files: 
 403e0452a59c1ff32b2f9140c625ec19 904 devel optional lintian_1.23.46.dsc
 5452c83a18e27584ad026b6dc4e8787a 395985 devel optional lintian_1.23.46.tar.gz
 1b6cd600f83ebf4fd97e07845efa43c2 328062 devel optional lintian_1.23.46_all.deb

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)

iD8DBQFHzLQs+YXjQAr8dHYRAnPkAKCopY4X5qkpIBolLn9KYZxJc5VfUgCbBkTv
dJeSxNp7SYHQDJ/YVYhix0M=
=PMlY
-----END PGP SIGNATURE-----

Message #44 received at 452215@bugs.debian.org (full text, mbox, reply):

From: Joerg Jaspert <joerg@ganneff.de>

To: Russ Allbery <rra@debian.org>

Cc: 452215@bugs.debian.org

Subject: Re: Bug#452215: lintian: r1238 - in trunk: checks debian testset testset/diffs/debian testset/diffs/debian/patches

Date: Tue, 04 Mar 2008 08:56:37 +0100

>>> I don't know if we want to allow for that or not.  I find the logic
>>> dubious, but dpatch supports that mode of operation.
>> I have to admit that I consider using a patch system in this way to be a
>> bug. I think we might want to upload with this check and see if there
>> are too many "false" positives.

FWIW I give a +1 for this warning. IMO one shouldn't ship the changes from
the patch in the .diff.gz too. Thats just wrong. Better would be to have
a common patch/unpatch target defined in policy.
And then the normal maintainer should follow the W (maybe upgrade to E
unless NMU detected) and the NMUer can ignore it, ie run "foo patch" and
then edit the stuff, so the NMU ends up in diff.gz

(And jftr - I lost interest in that thread way before HE did)

-- 
bye, Joerg
Mal verlierst Du und mal gewinnen die anderen: Immer im Wechsel!

Message #49 received at 452215@bugs.debian.org (full text, mbox, reply):

From: "Bernhard R. Link" <brlink@debian.org>

To: 452215@bugs.debian.org

Subject: Re: Bug#452215: lintian: r1238 - in trunk: checks debian testset testset/diffs/debian testset/diffs/debian/patches

Date: Tue, 4 Mar 2008 20:11:15 +0100

* Joerg Jaspert <joerg@ganneff.de> [080304 08:31]:
> 
> >>> I don't know if we want to allow for that or not.  I find the logic
> >>> dubious, but dpatch supports that mode of operation.
> >> I have to admit that I consider using a patch system in this way to be a
> >> bug. I think we might want to upload with this check and see if there
> >> are too many "false" positives.
> 
> FWIW I give a +1 for this warning. IMO one shouldn't ship the changes from
> the patch in the .diff.gz too. Thats just wrong. Better would be to have
> a common patch/unpatch target defined in policy.

A future better solution might be good, but currently I just hate any
package making me jump through hops to just see the actual source, so
my packages don't do such uglyness like requiring people to call "patch"
targets.

But I agree a general warning is nice, though it should be easily
circumventable. (Though I guess it always is via renaming the patches
dir to something else).

Hochachtungsvoll,
	Bernhard R. Link

Message #54 received at 452215@bugs.debian.org (full text, mbox, reply):

From: Russ Allbery <rra@debian.org>

To: "Bernhard R. Link" <brlink@debian.org>

Cc: 452215@bugs.debian.org

Subject: Re: Bug#452215: lintian: r1238 - in trunk: checks debian testset testset/diffs/debian testset/diffs/debian/patches

Date: Tue, 04 Mar 2008 11:45:42 -0800

"Bernhard R. Link" <brlink@debian.org> writes:

> But I agree a general warning is nice, though it should be easily
> circumventable. (Though I guess it always is via renaming the patches
> dir to something else).

Source lintian overrides also work.

-- 
Russ Allbery (rra@debian.org)               <http://www.eyrie.org/~eagle/>

Send a report that this bug log contains spam.