Guillem Jover [Tue, 3 May 2011 08:46:48 +0000 (3 10:46 +0200)]
libdpkg: Remove union in buffer_data and just use ptr
Guillem Jover [Tue, 3 May 2011 08:41:31 +0000 (3 10:41 +0200)]
libdpkg: Switch buffer API internally from descriptors to the file API
Guillem Jover [Thu, 2 Dec 2010 03:42:46 +0000 (2 04:42 +0100)]
libdpkg: New file based api
Guillem Jover [Sat, 10 Sep 2011 16:32:30 +0000 (10 18:32 +0200)]
Fix bug closure for “dpkg-deb: Do not warn on strange timestamps”
Spotted-by: Carl Fürstenberg <azatoth@gmail.com>
Sven Joachim [Fri, 9 Sep 2011 09:35:18 +0000 (9 11:35 +0200)]
German dpkg translation update
Update to 1022t.
Guillem Jover [Fri, 9 Sep 2011 07:02:27 +0000 (9 09:02 +0200)]
build: Ship debian/dpkg.prerm in the dist tarball
Regression introduced in commit
cf6b98d3738d1fe8b3119b71fd8321c12feb618c.
Guillem Jover [Fri, 9 Sep 2011 04:24:19 +0000 (9 06:24 +0200)]
dpkg-buildflags(1): Escape minus sign
Guillem Jover [Fri, 9 Sep 2011 03:55:37 +0000 (9 05:55 +0200)]
libdpkg: Do not generate the unused infodir in modstatdb_init()
Guillem Jover [Wed, 7 Sep 2011 04:38:41 +0000 (7 06:38 +0200)]
dpkg: Refactor control directory creation on unpack
Guillem Jover [Wed, 7 Sep 2011 04:20:43 +0000 (7 06:20 +0200)]
Do not include trailing slash for directory string literals
This makes it more natural to manage the directory names, and avoids
some doubled slashes.
Guillem Jover [Mon, 5 Sep 2011 14:08:51 +0000 (5 16:08 +0200)]
libdpkg: Move trigger note support to a new trignote module
This detangles the trigger note support from triglib and avoids code
using parsedb() to end up pulling the triglib and dbmodify modules.
This reduces the dpkg-deb binary size.
Guillem Jover [Tue, 6 Sep 2011 03:59:13 +0000 (6 05:59 +0200)]
libdpkg: Refactor trig_awaited_pend_head processing into new functions
This will allow accessing trig_awaited_pend_head from a different
module, w/o needing to pull in triglib.
Guillem Jover [Tue, 6 Sep 2011 03:54:36 +0000 (6 05:54 +0200)]
libdpkg: Rename trig_enqueue_awaited_pend to trig_awaited_pend_enqueue
Guillem Jover [Mon, 5 Sep 2011 14:03:21 +0000 (5 16:03 +0200)]
libdpkg: Move trig_name_is_illegal to a new trigname module
This detangles the function from the triglib module, avoiding
dpkg-trigger to pull it, and reducing its size.
Guillem Jover [Sun, 4 Sep 2011 14:07:29 +0000 (4 16:07 +0200)]
dpkg-deb: Do not warn on strange timestamps
The external extractor is only used on dpkg-deb -x (and friends), which
is only used by dpkg when extracting the control member. The internal
dpkg tar extractor never warns on timestamps, and there's no point for
dpkg to induce those warnings on the control extraction either.
Closes: #640298
Guillem Jover [Thu, 1 Sep 2011 13:17:35 +0000 (1 15:17 +0200)]
dpkg: Fix double “error:” string in missing PATH error output
Closes: #639997
Guillem Jover [Fri, 19 Aug 2011 16:07:13 +0000 (19 18:07 +0200)]
u-a: Update alternative links only if they change
There's no point in changing the links to the same target. This also
helps when systems might have a read-only file system mounted, but a
writable database.
Closes: #636700
Based-on-patch-by: Salvatore Bonaccorso <carnil@debian.org>
Signed-off-by: Guillem Jover <guillem@debian.org>
Guillem Jover [Thu, 8 Sep 2011 08:19:28 +0000 (8 10:19 +0200)]
u-a: Refactor alternative_path_classify from alternative_can_replace_path
Guillem Jover [Thu, 8 Sep 2011 17:12:20 +0000 (8 19:12 +0200)]
u-a: Do not try to readlink() a non-symlink
Return early in case the path is not a symlink, there's no point in
calling readlink() if we know it's going to fail.
Guillem Jover [Thu, 8 Sep 2011 08:10:06 +0000 (8 10:10 +0200)]
u-a: Use xreadlink() instead of areadlink() in alternative_evolve()
The function does not expect the readlink call to fail.
Guillem Jover [Tue, 6 Sep 2011 02:19:03 +0000 (6 04:19 +0200)]
u-a: Split xreadlink() allocation code into areadlink()
Using a bool as an argument for a function already taking several
arguments is almost never a good interface, it's not clear from
the call sites what it refers to and as such prone to confusion.
Split the main xreadlink() code into a function that allocates but
can return NULL, and make xreadlink() use the common semantics of
never failing.
Guillem Jover [Thu, 25 Aug 2011 18:40:42 +0000 (25 20:40 +0200)]
dpkg: Print a message instead of asserting on readlink/stat size discrepancy
This will help the user diagnose which file was affected and in
correcting the situation, which in most cases is due to file system
breakage, or non POSIX compliance.
Closes: #639229
Guillem Jover [Thu, 25 Aug 2011 18:40:02 +0000 (25 20:40 +0200)]
dpkg: Change type of variable r from int to ssize_t in conffderef()
Guillem Jover [Mon, 2 May 2011 19:51:47 +0000 (2 21:51 +0200)]
libdpkg: Detangle buffer filter logic from write logic
Guillem Jover [Mon, 2 May 2011 19:51:47 +0000 (2 21:51 +0200)]
libdpkg: Add new BUFFER_FILTER_NULL filter action
Guillem Jover [Fri, 17 Jun 2011 12:11:43 +0000 (17 14:11 +0200)]
libdpkg: Rename BUFFER_WRITE_MD5 to BUFFER_FILTER_MD5
Guillem Jover [Mon, 2 May 2011 08:25:38 +0000 (2 10:25 +0200)]
libdpkg: Namespace buffer init and done functions with buffer_filter_
This makes it clear these functions are not to initialize normal
buffer operations, but just for when a filter is in place.
Guillem Jover [Tue, 3 May 2011 08:55:45 +0000 (3 10:55 +0200)]
libdpkg: Rename fd_null_copy() to fd_skip()
Guillem Jover [Fri, 9 Sep 2011 05:19:27 +0000 (9 07:19 +0200)]
libdpkg: Add a test case for fd_md5()
Guillem Jover [Wed, 7 Sep 2011 02:53:25 +0000 (7 04:53 +0200)]
Convert translation files to UTF-8
Consolidating the charset used for all files in the repository makes
diff actually useful, otherwise when showing a patch involving distinct
charsets it's not possible to display them correctly due to the mixture.
Kees Cook [Tue, 6 Sep 2011 06:34:49 +0000 (5 23:34 -0700)]
dpkg-buildflags(1): add initial hardening documentation
Document the various hardening options that can be enabled/disabled
via DEB_BUILD_MAINT_OPTIONS.
Improved-by: Raphaël Hertzog <hertzog@debian.org>
Signed-off-by: Kees Cook <kees@debian.org>
Signed-off-by: Raphaël Hertzog <hertzog@debian.org>
Raphaël Hertzog [Wed, 27 Jul 2011 20:10:49 +0000 (27 22:10 +0200)]
dpkg-buildflags: emit hardening build flags by default
All the hardening build flags supported by hardening-includes
are supported except that PIE is not enabled by default (just like
the corresponding gcc patch doesn't enable it by default).
Inspired by the work of Kees Cook <kees@debian.org>.
Raphaël Hertzog [Tue, 2 Aug 2011 12:15:17 +0000 (2 14:15 +0200)]
Dpkg::BuildOptions: enable usage of alternative variable names
Helge Kreutzmann [Thu, 8 Sep 2011 15:53:13 +0000 (8 17:53 +0200)]
Update German translation of manual pages
Update to 2038t.
Raphaël Hertzog [Sat, 3 Sep 2011 12:34:31 +0000 (3 14:34 +0200)]
dpkg-source(1): add example of extend-diff-ignore's usage
Suggested-by: Christoph Berg <myon@debian.org>
Closes: #640198
Peter Krefting [Sun, 4 Sep 2011 17:57:23 +0000 (4 18:57 +0100)]
Update Swedish translation.
man/po/sv.po: 2044t0f0u.
Helge Kreutzmann [Wed, 24 Aug 2011 17:35:38 +0000 (24 19:35 +0200)]
Update German translation of manual pages
Update to 2038t.
Guillem Jover [Tue, 23 Aug 2011 03:24:39 +0000 (23 05:24 +0200)]
dpkg-source: Make single-debian-patch not abort on upstream changes
The single-debian-patch option usage in Dpkg::Source::Package::V2
was typoed as single_debian_patch, and thus not avoiding aborting
on upstream changes. Instead of fixing the typo, let's just not
expose single-debian-patch in Dpkg::Source::Package::V2, as it's
Dpkg::Source::Package::V3::quilt specific and as such a layer
violation, the code will just set auto_commit when single-debian-patch
is specified, which is the actual intended behaviour. Reflect this
explicitly in the man page.
Regression introduced in commit
46b688e4824f00842568b517b119eec34a5f7286.
Reported-by: Bernhard R. Link <brlink@debian.org>
Guillem Jover [Tue, 23 Aug 2011 01:55:27 +0000 (23 03:55 +0200)]
Dpkg::Source::Package: Use “-z” instead of “not -s”
Helge Kreutzmann [Mon, 22 Aug 2011 18:57:14 +0000 (22 20:57 +0200)]
Update German scripts translation
Update to 523t.
Helge Kreutzmann [Mon, 22 Aug 2011 18:33:18 +0000 (22 20:33 +0200)]
Update German translation of manual pages
Update to 2038t.
Raphaël Hertzog [Sun, 21 Aug 2011 09:11:02 +0000 (21 11:11 +0200)]
dpkg: correctly handle a hardlink pointing to a conffile
Normal files tagged as a conffile are not tagged with
fnnf_deferred_rename and due to this the hardlinking code
was assuming that the target is in its final location
already. This is obviously not the case for a conffile.
Fix the code so that the hardlinking logic takes into account
the fact that conffiles are unpacked in .dpkg-new and stay
there until the configuration phase.
Reported-by: Igor Pashev <pashev.igor@gmail.com>
Sven Joachim [Sun, 21 Aug 2011 16:37:50 +0000 (21 18:37 +0200)]
German dpkg translation update
Update to 1022t.
Peter Krefting [Sat, 20 Aug 2011 21:08:09 +0000 (20 22:08 +0100)]
Update Swedish translation.
man/po/sv.po: 1978t50f16u (WIP).
Guillem Jover [Sat, 20 Aug 2011 13:36:00 +0000 (20 15:36 +0200)]
u-a: Do not build strings from parts in alternative_check_install_args
This makes the work of translators harder.
Guillem Jover [Sat, 20 Aug 2011 10:55:33 +0000 (20 12:55 +0200)]
u-a: Refactor alternative argument checks into new function
Guillem Jover [Fri, 19 Aug 2011 20:59:53 +0000 (19 22:59 +0200)]
u-a: Refactor --install argument checking into a new function
Guillem Jover [Fri, 19 Aug 2011 20:32:19 +0000 (19 22:32 +0200)]
u-a: Move alternative map loading to the actions that make use of them
This will reduce start up time, by not doing unneeded operations.
Guillem Jover [Fri, 19 Aug 2011 16:38:10 +0000 (19 18:38 +0200)]
u-a: Split alternative map loading into two functions
This will allow only loading them when needed, instead of doing it
for any action.
Guillem Jover [Sat, 20 Aug 2011 10:21:22 +0000 (20 12:21 +0200)]
u-a: Refactor scandir() namelist free into new altdb_free_namelist()
Guillem Jover [Sat, 20 Aug 2011 10:30:35 +0000 (20 12:30 +0200)]
u-a: Rename get_all_alternatives() to altdb_get_namelist()
This makes it clear we are referring to the filenames in the database
directory.
Guillem Jover [Fri, 19 Aug 2011 15:50:25 +0000 (19 17:50 +0200)]
u-a: Do not check if linknames can be updated if --force is in use
The action is going to proceed in any case, no point wasting resources
doing the checks. Also this removes a possible point of non-overridable
failure.
Guillem Jover [Fri, 19 Aug 2011 10:59:35 +0000 (19 12:59 +0200)]
Dpkg::Compression: Change default values depending on the compressor used
This makes dpkg-source use more appropriate default compression values
depending on the compressor used.
Prompted-by: Timo Juhani Lindfors <timo.lindfors@iki.fi>
Guillem Jover [Sat, 20 Aug 2011 07:51:13 +0000 (20 09:51 +0200)]
Do not use possessive «'s» after ‘s’ or ‘x’
Guillem Jover [Wed, 17 Aug 2011 14:44:18 +0000 (17 16:44 +0200)]
Fix spelling errors
Found by codespell.
Guillem Jover [Sun, 14 Aug 2011 22:03:00 +0000 (15 00:03 +0200)]
dpkg: Clarify --update-avail usage error message
Unify the reference to the Packages-file to match the one used on the
man pages and on the --help output.
Closes: #628055
Guillem Jover [Sun, 14 Aug 2011 22:01:53 +0000 (15 00:01 +0200)]
dpkg-parsechangelog: Remove duplicate program name from usage error output
Guillem Jover [Sun, 14 Aug 2011 21:45:36 +0000 (14 23:45 +0200)]
Expand numeric literals and word abbreviation in usage error output
This avoids unneeded abbreviations, uses correct english, and should
help translators.
Guillem Jover [Sun, 14 Aug 2011 21:28:07 +0000 (14 23:28 +0200)]
dpkg-deb(1): Clarify the default compression-levels
Guillem Jover [Sun, 14 Aug 2011 20:45:20 +0000 (14 22:45 +0200)]
deb-src-control(5): Use linux-any in example control stanza
Instead of using the negated list of not supported architectures, just
use the positive form, which should be considered current best practice.
Guillem Jover [Sun, 14 Aug 2011 20:43:05 +0000 (14 22:43 +0200)]
dselect(1): Correct configuration directory reference
Use /etc/dpkg/dselect.cfg.d/ instead of /etc/dpkg/dpkg.cfg.d/.
Guillem Jover [Sun, 14 Aug 2011 20:41:16 +0000 (14 22:41 +0200)]
man: Use minus signs and hyphens consistently
Any string that can be copy & pasted into control files or into a
terminal, for example, needs to correctly use minus signs (escaped
dashes), instead of the default hyphen.
Helge Kreutzmann [Sat, 20 Aug 2011 06:31:37 +0000 (20 08:31 +0200)]
Regenerate .pot files and merge .po files with them
Peter Krefting [Wed, 17 Aug 2011 19:59:37 +0000 (17 20:59 +0100)]
Update Swedish translation.
po/sv.po: 1023t0f0u.
man/po/sv.po: 1824t180f40u (WIP).
scripts/po/sv.po: 524t0f0u
Guillem Jover [Sat, 13 Aug 2011 01:19:24 +0000 (13 03:19 +0200)]
dpkg: Defer hardlink renames to avoid breakage on unpack
This fixes two issues related to making the new file contents available
through an alternative but final path before
1) the one "carrying" the content has been fsync()ed, to avoid getting
garbage on system failure.
2) the deferred fsync() on the main path has been performed, avoiding
a possible error when reopening the file for writing if it's currently
under execution.
Deferring the hardlink is safe as they are always after the tar entry
containing the actual file data.
Regression introduced in commit
9cd41fdda1c27169c52d73b3b3ce71991d724994.
Closes: #635683
Reported-by: Niko Tyni <ntyni@debian.org>
Based-on-patch-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Guillem Jover <guillem@debian.org>
Raphaël Hertzog [Fri, 12 Aug 2011 20:35:46 +0000 (12 22:35 +0200)]
Dpkg::Deps: fix simplification logic of union dependencies
The code merging dependencies was wrongly updating the dependency that's
not integrated in the resulting field.
Reported-by: Andreas Metzler <ametzler@downhill.at.eu.org>
Guillem Jover [Tue, 9 Aug 2011 05:19:09 +0000 (9 07:19 +0200)]
libdpkg: Use a generic pointer to pass the parsed object data around
This will allow code not parsing binary package control files to be
able to easily reuse the parser, by supplying their own object
structure.
Unfortunately this sacrifices the type-safety of the parse_stanza()
and parse_field() functions, but that only affects one variable,
which should be an acceptable tradeoff given the gain in code reuse.
Christian PERRIER [Mon, 8 Aug 2011 18:29:13 +0000 (8 20:29 +0200)]
Changelog for French translation update
Christian PERRIER [Mon, 8 Aug 2011 18:27:40 +0000 (8 20:27 +0200)]
Fix Sylvestre's translation
Sylvestre Ledru [Mon, 8 Aug 2011 13:44:01 +0000 (8 15:44 +0200)]
Update of the french localization
Sylvestre Ledru [Mon, 8 Aug 2011 13:42:56 +0000 (8 15:42 +0200)]
Fix a french typo
Guillem Jover [Fri, 5 Aug 2011 04:23:59 +0000 (5 06:23 +0200)]
libdpkg: Initialize ps->pkg and ps->pkgbin to NULL in parse_open()
This guarantees the parsedb_state variable is in a sane state after
the parse_open() call.
Guillem Jover [Fri, 5 Aug 2011 03:21:01 +0000 (5 05:21 +0200)]
libdpkg: Refactor parse_open() and parse_close() from parsedb()
This will allow code using parse_stanza() to avoid having to reimplement
the opening and closing logic.
Guillem Jover [Fri, 5 Aug 2011 04:03:02 +0000 (5 06:03 +0200)]
libdpkg: Close the file descriptor just after slurping it in parsedb()
Raphaël Hertzog [Thu, 4 Aug 2011 19:33:23 +0000 (4 21:33 +0200)]
dpkg-source: simplify message pointing to dpkg-source --commit
The supplementary arguments suggested are awkward and are generally not
needed unless your package is huge and takes a long time to uncompress
and diff.
The best advice is thus “dpkg-source --commit” without any parameter.
Maintainers of big packages will read the manual package and learn
by themselves how to reuse the generated patch.
Raphaël Hertzog [Wed, 3 Aug 2011 20:02:33 +0000 (3 22:02 +0200)]
dpkg-source: don't always print "local changes have been recorded in a new patch"
The message should only be printed if such a patch has really
been created.
Reported-by: Sven Joachim
Helge Kreutzmann [Wed, 3 Aug 2011 19:19:38 +0000 (3 21:19 +0200)]
Fix typo
Helge Kreutzmann [Wed, 3 Aug 2011 19:14:41 +0000 (3 21:14 +0200)]
Update German translation of manual pages
Update to 2047t.
Helge Kreutzmann [Tue, 2 Aug 2011 19:55:50 +0000 (2 21:55 +0200)]
Update German translation of manual pages
Update to 2037t7f3u.
Helge Kreutzmann [Tue, 2 Aug 2011 17:32:57 +0000 (2 19:32 +0200)]
Update German scripts translation
Update to 524t.
Sven Joachim [Tue, 2 Aug 2011 09:57:17 +0000 (2 11:57 +0200)]
German dpkg translation update
Update to 1023t.
Raphaël Hertzog [Tue, 2 Aug 2011 09:20:54 +0000 (2 11:20 +0200)]
dpkg-source: clean up temporary file in case of interruption
Navite packages (both "1.0" and "3.0 (native)") failed to remove
the temporary tarball if the process is interrupted during its
generation.
Reported-by: Cyril Brulebois <kibi@debian.org>
Raphaël Hertzog [Tue, 2 Aug 2011 09:06:11 +0000 (2 11:06 +0200)]
Move changelog entry to attribute it correctly to me
Guillem Jover [Tue, 2 Aug 2011 03:05:30 +0000 (2 05:05 +0200)]
libdpkg: Remove obsolete variable references in parsedb() comment
The warnto and warncount variables were removed in commits
ba908783e306ba38457a84fd6d975ca69e351e56 and
8451065cf85cb8a17dcd17e0cd2ca1fdef7cb749 respectively.
Guillem Jover [Tue, 2 Aug 2011 02:54:42 +0000 (2 04:54 +0200)]
libdpkg: Make parse_stanza() a public function
Make public also the required parse helper macros and structures. At
least parse_EOF() will be needed for other callers to get out of the
loop when parsing multi-stanza files.
Guillem Jover [Tue, 2 Aug 2011 02:50:09 +0000 (2 04:50 +0200)]
libdpkg: Detangle fieldinfos from struct field_state
Define the fieldencountered array in the stack, and switch the array
in struct field_state to be a pointer to that array. This way we do not
pull the fieldinfos symbol (specific to the parsedb() function) to the
now generic struct field_state.
Guillem Jover [Tue, 2 Aug 2011 02:24:10 +0000 (2 04:24 +0200)]
libdpkg: Rename parse helper *_mmap functions to parse_*
This abstracts the method used to load the file contents, and it also
namespaces it for the parser.
Guillem Jover [Thu, 14 Jul 2011 15:21:55 +0000 (14 17:21 +0200)]
libdpkg: Refactor RFC-822 stanza parsing into new parse_stanza()
This generalizes the stanza parsing so that it can be used by other
parts of the code.
Guillem Jover [Thu, 14 Jul 2011 02:37:31 +0000 (14 04:37 +0200)]
libdpkg: Move dataptr and endptr from parsedb() to struct parsedb_state
Guillem Jover [Thu, 14 Jul 2011 01:58:45 +0000 (14 03:58 +0200)]
libdpkg: Add pkg and pkgbin to struct parsedb_state
We are passing these variables as arguments most of the time for
parse_error() and parse_warn() calls which are not run-time frequent,
this also unties somewhat the code from the package structures.
Guillem Jover [Tue, 2 Aug 2011 02:56:00 +0000 (2 04:56 +0200)]
libdpkg: Remove trailing semicolons from getc_mmap and ungetc_mmap
Those are unnecessary as these are function macros which need to be
followed by semicolons on the call sites.
Guillem Jover [Tue, 26 Apr 2011 05:35:16 +0000 (26 07:35 +0200)]
dpkg-deb: Add --raw-extract option combining --control and --extract
This allows to extract the control and file system members of a .deb
into a single directory using the same layout dpkg-deb -b expects for
building a package.
Closes: #552123
Guillem Jover [Mon, 20 Jun 2011 17:11:20 +0000 (20 19:11 +0200)]
dpkg-deb: Add new -v, --verbose option
Change --extract to honour --verbose and behave as --vextract.
Closes: #293280
Raphaël Hertzog [Mon, 1 Aug 2011 20:27:08 +0000 (1 22:27 +0200)]
Regenerate .pot files and merge .po files with them
Raphaël Hertzog [Mon, 1 Aug 2011 20:01:20 +0000 (1 22:01 +0200)]
Revert "dpkg-buildpackage: support for Build-Features: build-arch"
This reverts commit
14d48ef9abc2ce2d394e9ae4d69d4ba68b551620. Given
that Guillem does not want this field and that the technical committee
has not yet decided on the proper approach to enable
build-arch/build-indep, it's safer to revert this change for now and allow
us to proceed with a release.
Conflicts:
man/deb-src-control.5
Raphaël Hertzog [Wed, 27 Jul 2011 21:20:00 +0000 (27 23:20 +0200)]
debian/rules: use the in-tree dpkg-buildflags to inject build flags
Note that $(shell ...) does not receive variables defined
within make with "export VAR=foo" so that we have to include
DEB_CFLAGS_MAINT_APPEND in the command-line.
Raphaël Hertzog [Wed, 27 Jul 2011 17:18:04 +0000 (27 19:18 +0200)]
dpkg-buildflags: support a "strip" operation
The strip operation is useful to drop some options from the
returned build flags.
Raphaël Hertzog [Wed, 27 Jul 2011 16:49:46 +0000 (27 18:49 +0200)]
dpkg-buildflags: implement support of DEB_<flag>_MAINT_<op> env variables
Those variables are meant to be used by the package maintainer within
debian/rules to alter the resulting build flags if needed.
Raphaël Hertzog [Wed, 27 Jul 2011 15:45:40 +0000 (27 17:45 +0200)]
dpkg-buildflags: implement --export=configure
The goal is to make it easy to set compilation flags on a ./configure
command line for example like this:
./configure $(shell dpkg-buildflags --export=configure)
The output thus looks like this:
CFLAGS="-g -O2" CPPFLAGS="" CXXFLAGS="-g -O2" FFLAGS="-g -O2" LDFLAGS=""
Raphaël Hertzog [Wed, 27 Jul 2011 11:21:53 +0000 (27 13:21 +0200)]
Revert "dpkg-buildflags: support debian/buildflags"
This reverts commit
316df0a76f2ef91badbccb6fef0bfa7ae23fac59.
A recent discussion concluded that it was not a proper interface
for maintainers to override/extend build flags.
See http://lists.debian.org/debian-dpkg/2011/07/msg00033.html
Conflicts:
scripts/Dpkg/BuildFlags.pm