]> git.saurik.com Git - apt.git/log
apt.git
12 years agomake these retry_write methods static so that they don't end up as symbols
David Kalnischkies [Thu, 22 Mar 2012 21:18:05 +0000 (22:18 +0100)] 
make these retry_write methods static so that they don't end up as symbols

12 years ago* debian/rules:
David Kalnischkies [Thu, 22 Mar 2012 18:29:53 +0000 (19:29 +0100)] 
* debian/rules:
  - do not sed in configure.in to set the version-number
* prepare-release:
  - add as a small script to lazy check and prepare releases

12 years agomerge with debian-sid (some version screw going on here)
David Kalnischkies [Thu, 22 Mar 2012 16:34:41 +0000 (17:34 +0100)] 
merge with debian-sid (some version screw going on here)

12 years agouse the variables as the rest of the tests to avoid escapes in the output
David Kalnischkies [Thu, 22 Mar 2012 12:42:59 +0000 (13:42 +0100)] 
use the variables as the rest of the tests to avoid escapes in the output

12 years ago* apt-inst/database.{cc,h}, apt-inst/deb/dpkgdb.{cc,h}:
David Kalnischkies [Thu, 22 Mar 2012 12:09:22 +0000 (13:09 +0100)] 
* apt-inst/database.{cc,h}, apt-inst/deb/dpkgdb.{cc,h}:
  - drop instead of fix as it is only needed if you want to reimplement dpkg
    and comes straight from the beginning of last decade (Closes: #663372)
* apt-inst/deb/debfile.cc:
  - {Extract,Merge}Control() is another instance of "lets reimplement dpkg"
    so shot of this code before someone ends up using this…

12 years ago* doc/apt-get.8.xml:
Mike Erickson [Thu, 22 Mar 2012 09:37:53 +0000 (10:37 +0100)] 
* doc/apt-get.8.xml:
  - typofix: respect → respecting, thanks Mike Erickson! (Closes: #664833)

12 years ago* debian/control
David Kalnischkies [Thu, 22 Mar 2012 00:02:22 +0000 (01:02 +0100)] 
* debian/control
  - bump Standards-Version to 3.9.3 (no changes needed)

12 years agouse the correct line to get the currently configure.in version
David Kalnischkies [Wed, 21 Mar 2012 23:59:51 +0000 (00:59 +0100)] 
use the correct line to get the currently configure.in version

12 years ago* debian/apt-utils.install:
David Kalnischkies [Wed, 21 Mar 2012 23:51:44 +0000 (00:51 +0100)] 
* debian/apt-utils.install:
  - ship the ftparchive, apt-extractemplates and apt-sortpkgs locales
    in the apt-utils package instead of the apt package

12 years agothe previously used VERSION didn't work everywhere so we are switching
David Kalnischkies [Wed, 21 Mar 2012 23:16:11 +0000 (00:16 +0100)] 
the previously used VERSION didn't work everywhere so we are switching
to the more standard PACKAGE_VERSION and make it work in every file

12 years agodo not update po and pot files in the process of the build as this
David Kalnischkies [Wed, 21 Mar 2012 22:47:54 +0000 (23:47 +0100)] 
do not update po and pot files in the process of the build as this
causes timestamp changes for the mo files which therefore can't
be refcounted by dpkg for your M-A: same packages
(Closes: #659333, LP: #924628)

The commit also enables a top-level 'make update-po' and does
all the needed changes to let this work now that update-po might
be called in a freshly checkout tree

12 years agofor these helpers we don't need too much checks,
David Kalnischkies [Tue, 20 Mar 2012 20:08:15 +0000 (21:08 +0100)] 
for these helpers we don't need too much checks,
but no warnings from gcc helps

12 years ago* methods/rred.cc:
David Kalnischkies [Tue, 20 Mar 2012 18:23:32 +0000 (19:23 +0100)] 
* methods/rred.cc:
  - check return of writev() as gcc recommends
* methods/mirror.cc:
  - check return of chdir() as gcc recommends
* apt-pkg/deb/dpkgpm.cc:
  - check return of write() a gcc recommends
* apt-inst/deb/debfile.cc:
  - check return of chdir() as gcc recommends
* apt-inst/deb/dpkgdb.cc:
  - check return of chdir() as gcc recommends

12 years ago* debian/libapt-pkg4.12:
David Kalnischkies [Tue, 20 Mar 2012 17:30:32 +0000 (18:30 +0100)] 
* debian/libapt-pkg4.12:
  - update symbols file

12 years ago* methods/makefile:
David Kalnischkies [Tue, 20 Mar 2012 16:08:08 +0000 (17:08 +0100)] 
* methods/makefile:
  - do not link rred against libz anymore as FileFd handles all
    this transparently now

12 years ago* apt-pkg/acquire-worker.cc:
David Kalnischkies [Tue, 20 Mar 2012 16:05:11 +0000 (17:05 +0100)] 
* apt-pkg/acquire-worker.cc:
  - check return of write() as gcc recommends
* apt-pkg/acquire.cc:
  - check return of write() as gcc recommends
* apt-pkg/cdrom.cc:
  - check return of chdir() and link() as gcc recommends
* apt-pkg/clean.cc:
  - check return of chdir() as gcc recommends
* apt-pkg/contrib/netrc.cc:
  - check return of asprintf() as gcc recommends

12 years agoupdate the versionnumber of the faked dpkg in the testcases for multiarch
David Kalnischkies [Tue, 20 Mar 2012 16:03:35 +0000 (17:03 +0100)] 
update the versionnumber of the faked dpkg in the testcases for multiarch
and add a comment about the need of this number

12 years agonow that stdout/stdlog are used consistently, the message order changed
David Kalnischkies [Tue, 13 Mar 2012 12:34:47 +0000 (13:34 +0100)] 
now that stdout/stdlog are used consistently, the message order changed

12 years agomerged from lp:~donkult/apt/experimental
Michael Vogt [Tue, 13 Mar 2012 12:29:52 +0000 (13:29 +0100)] 
merged from lp:~donkult/apt/experimental

12 years agorecheck dependencies in SmartUnpack after a change, too
David Kalnischkies [Tue, 13 Mar 2012 11:39:05 +0000 (12:39 +0100)] 
recheck dependencies in SmartUnpack after a change, too

12 years ago* apt-pkg/packagemanager.cc:
David Kalnischkies [Tue, 13 Mar 2012 10:37:15 +0000 (11:37 +0100)] 
* apt-pkg/packagemanager.cc:
  - recheck all dependencies if we changed a package in SmartConfigure
    as this could break an earlier dependency (LP: #940396)

12 years agoensure that the fullname of a package is displayed in the debug output
David Kalnischkies [Mon, 12 Mar 2012 18:30:48 +0000 (19:30 +0100)] 
ensure that the fullname of a package is displayed in the debug output

12 years agoimprove the testcode so that we do not depend on the order of unpacking
David Kalnischkies [Mon, 12 Mar 2012 18:29:59 +0000 (19:29 +0100)] 
improve the testcode so that we do not depend on the order of unpacking
of specific packages as long as the order is okay in general

12 years agonow that the package name 'apt' is a magic word enabling essential status
David Kalnischkies [Mon, 12 Mar 2012 18:02:57 +0000 (19:02 +0100)] 
now that the package name 'apt' is a magic word enabling essential status
for said package i can't use it anymore in the testcase, so use another name

12 years agoshow which dependency couldn't be satisfied in the debug output
David Kalnischkies [Wed, 7 Mar 2012 10:16:58 +0000 (11:16 +0100)] 
show which dependency couldn't be satisfied in the debug output

12 years ago* apt-pkg/packagemanager.cc:
Michael Vogt [Wed, 7 Mar 2012 08:20:31 +0000 (09:20 +0100)] 
* apt-pkg/packagemanager.cc:
  - fix inconsistent clog/cout usage in the debug output

12 years agoreleasing version 0.8.16~exp13
Michael Vogt [Tue, 6 Mar 2012 17:21:02 +0000 (18:21 +0100)] 
releasing version 0.8.16~exp13

12 years agomerged from lp:~donkult/apt/experimental
Michael Vogt [Tue, 6 Mar 2012 17:10:02 +0000 (18:10 +0100)] 
merged from lp:~donkult/apt/experimental

12 years agoadd Debug::pkgAcqArchive::NoQueue to disable package downloading
David Kalnischkies [Tue, 6 Mar 2012 16:58:16 +0000 (17:58 +0100)] 
add Debug::pkgAcqArchive::NoQueue to disable package downloading

12 years agomerge with apt/experimental
David Kalnischkies [Tue, 6 Mar 2012 16:36:59 +0000 (17:36 +0100)] 
merge with apt/experimental

* apt-pkg/packagemanager.cc:
  - fix bug in predepends handling - ensure that packages that needs
    unpackaging are unpacked before they are configured (LP: #927993)
[ Julian Andres Klode ]
* apt-pkg/deb/deblistparser.cc:
  - Set the Essential flag on APT instead of only Important
* apt-pkg/packagemanager.cc:
  - Do not use immediate configuration for packages with the Important flag
* Treat the Important flag like the Essential flag with those differences:
  - No Immediate configuration (see above)
  - Not automatically installed during dist-upgrade
  - No higher score for installation ordering

12 years ago* apt-pkg/acquire-item.cc:
Michael Vogt [Tue, 6 Mar 2012 16:22:44 +0000 (17:22 +0100)] 
* apt-pkg/acquire-item.cc:
  - remove 'old' InRelease file if we can't get a new one before
    proceeding with Release.gpg to avoid the false impression of a still
    trusted repository by a (still present) old InRelease file.
    Thanks to Simon Ruderich for reporting this issue! (CVE-2012-0214)

12 years ago* apt-pkg/packagemanager.cc:
Michael Vogt [Tue, 6 Mar 2012 13:06:38 +0000 (14:06 +0100)] 
* apt-pkg/packagemanager.cc:
  - fix bug in predepends handling - ensure that packages that needs
    unpackaging are unpacked before they are configured (LP: #927993)

12 years ago* apt-pkg/contrib/fileutl.cc:
David Kalnischkies [Tue, 6 Mar 2012 09:53:35 +0000 (10:53 +0100)] 
* apt-pkg/contrib/fileutl.cc:
  - do not warn about the ignoring of directories (Closes: #662762)

12 years agochangelog: Document scoring difference for Important flag
Julian Andres Klode [Mon, 5 Mar 2012 14:19:50 +0000 (15:19 +0100)] 
changelog: Document scoring difference for Important flag

* Treat the Important flag like the Essential flag with those differences:
  - No higher score for installation ordering

12 years ago* Treat the Important flag like the Essential flag with two differences:
Julian Andres Klode [Mon, 5 Mar 2012 14:12:31 +0000 (15:12 +0100)] 
* Treat the Important flag like the Essential flag with two differences:
  - No Immediate configuration (see above)
  - Not automatically installed during dist-upgrade

12 years ago* apt-pkg/packagemanager.cc:
Julian Andres Klode [Mon, 5 Mar 2012 14:10:54 +0000 (15:10 +0100)] 
* apt-pkg/packagemanager.cc:
  - Do not use immediate configuration for packages with the Important flag

12 years ago* apt-pkg/deb/deblistparser.cc:
Julian Andres Klode [Mon, 5 Mar 2012 13:57:11 +0000 (14:57 +0100)] 
* apt-pkg/deb/deblistparser.cc:
  - Set the Essential flag on APT instead of only Important

12 years agoas we parse datestrings from external sources a lot specify the length
David Kalnischkies [Sun, 4 Mar 2012 23:37:54 +0000 (00:37 +0100)] 
as we parse datestrings from external sources a lot specify the length
of the integer fields as well to avoid crashes in scanf as cppchecks warns:
"(warning) scanf without field width limits can crash with huge input data"

12 years agofix "(style) The scope of the variable 'count' can be reduced"
David Kalnischkies [Sun, 4 Mar 2012 23:31:52 +0000 (00:31 +0100)] 
fix "(style) The scope of the variable 'count' can be reduced"

12 years agomicro-optimize "(performance) Prefer prefix ++/-- operators for non-primitive types."
David Kalnischkies [Sun, 4 Mar 2012 23:28:39 +0000 (00:28 +0100)] 
micro-optimize "(performance) Prefer prefix ++/-- operators for non-primitive types."

12 years agoset char-limits for the scanf parsing previous crash-reports
David Kalnischkies [Sun, 4 Mar 2012 23:25:32 +0000 (00:25 +0100)] 
set char-limits for the scanf parsing previous crash-reports

12 years agofix "(performance) Possible inefficient checking for 'R' emptiness."
David Kalnischkies [Sun, 4 Mar 2012 23:21:25 +0000 (00:21 +0100)] 
fix "(performance) Possible inefficient checking for 'R' emptiness."

12 years agofix "(style) Checking if unsigned variable 'Minor' is less than zero."
David Kalnischkies [Sun, 4 Mar 2012 23:20:28 +0000 (00:20 +0100)] 
fix "(style) Checking if unsigned variable 'Minor' is less than zero."

12 years agoensure that (s)scanf doesn't parse a too long Code now that a previous
David Kalnischkies [Sun, 4 Mar 2012 23:19:50 +0000 (00:19 +0100)] 
ensure that (s)scanf doesn't parse a too long Code now that a previous
commit lifted the Line-length limit

12 years agofix two "(style) Variable 'Res' is assigned a value that is never used"
David Kalnischkies [Sun, 4 Mar 2012 22:58:23 +0000 (23:58 +0100)] 
fix two "(style) Variable 'Res' is assigned a value that is never used"

12 years agofix "(error) Possible null pointer dereference: BindAddr" by ensuring
David Kalnischkies [Sun, 4 Mar 2012 22:56:06 +0000 (23:56 +0100)] 
fix "(error) Possible null pointer dereference: BindAddr" by ensuring
that BindAddr isn't NULL after getaddrinfo()

12 years agofix "(style) Statements following return, break, continue, goto or throw
David Kalnischkies [Sun, 4 Mar 2012 22:50:30 +0000 (23:50 +0100)] 
fix "(style) Statements following return, break, continue, goto or throw
will never be executed." from cppcheck. The fd was closed only after a
return, so invert the order of lines and be happy

12 years agofix a bunch of cppcheck "(warning) Member variable '<#>' is not
David Kalnischkies [Sun, 4 Mar 2012 22:47:05 +0000 (23:47 +0100)] 
fix a bunch of cppcheck "(warning) Member variable '<#>' is not
initialized in the constructor." messages (no functional change)

12 years agohandle a SIGINT in all modes as a break after the currently running
David Kalnischkies [Sun, 4 Mar 2012 22:01:59 +0000 (23:01 +0100)] 
handle a SIGINT in all modes as a break after the currently running
dpkg transaction instead of ignoring it completely

12 years ago* apt-pkg/acquire-item.cc:
David Kalnischkies [Sun, 4 Mar 2012 21:50:21 +0000 (22:50 +0100)] 
* apt-pkg/acquire-item.cc:
  - remove 'old' InRelease file if we can't get a new one before
    proceeding with Release.gpg to avoid the false impression of a still
    trusted repository by a (still present) old InRelease file.
    Thanks to Simon Ruderich for reporting this issue! (CVE-2012-0214)

Effected are all versions >= 0.8.11

Possible attack summary:
- Attacker needs to find a user which has run at least one successful
 'apt-get update' against an archive providing InRelease files.
- Create a Packages file with his preferred content.
- Attacker then prevents the download of InRelease, Release and
  Release.gpg (alternatively he creates a valid Release file and sends
  this, the other two files need to be missing either way).
- User updates against this, getting the modified Packages file without
  any indication of being unsigned (beside the "Ign InRelease" and
  "Ign Release.gpg" in the output of 'apt-get update').
=> deb files from this source are considered 'trusted' (and therefore the
   user isn't asked for an additional confirmation before install)

12 years agoFrench: replace "étiquetage" by "épinglage" for "pinning"
bubulle@debian.org [Sat, 3 Mar 2012 15:19:16 +0000 (16:19 +0100)] 
French: replace "étiquetage" by "épinglage" for "pinning"

12 years agotestcase for CVE-2012-0214
Simon Ruderich [Sat, 3 Mar 2012 10:44:06 +0000 (11:44 +0100)] 
testcase for CVE-2012-0214

12 years ago* apt-pkg/packagemanager.cc:
David Kalnischkies [Sat, 3 Mar 2012 10:43:21 +0000 (11:43 +0100)] 
* apt-pkg/packagemanager.cc:
  - do not try to a void a breaks if the broken package pre-depends
    on the breaker, but let dpkg auto-deconfigure it

12 years agoallow msgtest to be used with only one parameter
David Kalnischkies [Sat, 3 Mar 2012 10:40:22 +0000 (11:40 +0100)] 
allow msgtest to be used with only one parameter

12 years agoshow in the debug output if we are looping in the avoid breaks
David Kalnischkies [Sat, 3 Mar 2012 09:08:19 +0000 (10:08 +0100)] 
show in the debug output if we are looping in the avoid breaks

12 years agoeanup the ordering-code avoiding a break (no function change)
David Kalnischkies [Sat, 3 Mar 2012 09:02:06 +0000 (10:02 +0100)] 
eanup the ordering-code avoiding a break (no function change)

12 years ago* apt-pkg/packagemanager.cc:
Michael Vogt [Fri, 2 Mar 2012 21:01:51 +0000 (22:01 +0100)] 
* apt-pkg/packagemanager.cc:
  - when calculating pre-dependencies ensure that both unpack and
    configure are considered (instead of only configure) LP: #927993

12 years agomerged patch from lp:~uusijani/apt/uusi-branch:
Michael Vogt [Thu, 1 Mar 2012 11:51:15 +0000 (12:51 +0100)] 
merged patch from lp:~uusijani/apt/uusi-branch:
Correct fi translation for hash sum mismatches (lp:420403)
Thanks to Jani Uusitalo

12 years agofix the string in the testcases
David Kalnischkies [Mon, 27 Feb 2012 11:51:31 +0000 (12:51 +0100)] 
fix the string in the testcases

12 years ago* cmdline/apt-cache.cc:
David Kalnischkies [Sat, 25 Feb 2012 18:44:55 +0000 (19:44 +0100)] 
* cmdline/apt-cache.cc:
  - correct --pre-depends option by using dash consistently (LP: #940837)

12 years agoimprove 'error' message for packages which are only referenced
David Kalnischkies [Sat, 18 Feb 2012 20:20:57 +0000 (21:20 +0100)] 
improve 'error' message for packages which are only referenced
e.g. in a Depends line and are now requested for removal

12 years ago* cmdline/apt-get.cc:
David Kalnischkies [Sat, 18 Feb 2012 19:44:31 +0000 (20:44 +0100)] 
* cmdline/apt-get.cc:
  - if a package can't be removed as it is not installed, suggest to
    the user an (installed) multiarch silbing with 'Did you mean?'

12 years agouse pdiff for Translation-* files if available (Closes: #657902)
David Kalnischkies [Sat, 18 Feb 2012 18:48:04 +0000 (19:48 +0100)] 
use pdiff for Translation-* files if available (Closes: #657902)

Beware: pdiffs for Translation-* are only acquired if their
availability is advertised in the Release file.

12 years ago* apt-pkg/acquire-item.cc:
David Kalnischkies [Sat, 18 Feb 2012 18:40:13 +0000 (19:40 +0100)] 
* apt-pkg/acquire-item.cc:
  - drop support for i18n/Index file (introduced in 0.8.11) and use
    the Release file instead to get the Translations (Closes: #649314)
* ftparchive/writer.cc:
  - add 'Translation-*' to the default patterns

i18n/Index was never used outside debian - and even here it isn't used
consistently as only 'main' has such a file. As the Release file now
includes the Translation-* files we therefore drop support for i18n/Index.

A version supporting it was never part of a debian release and still
supporting it would mean that we get 99% of the time a 404 as response
to the request anyway and confuse archive maintainers who want to
provide all files APT tries to acquire.

12 years agoensure that the cache-directories are really directories before
David Kalnischkies [Sat, 18 Feb 2012 18:17:57 +0000 (19:17 +0100)] 
ensure that the cache-directories are really directories before
trying to get a list of included files from them

12 years ago* apt-pkg/indexrecords.cc:
David Kalnischkies [Sat, 18 Feb 2012 17:23:09 +0000 (18:23 +0100)] 
* apt-pkg/indexrecords.cc:
  - do not create empty Entries as a sideeffect of Lookup()

12 years ago* apt-pkg/cachefile.cc:
David Kalnischkies [Sat, 11 Feb 2012 21:36:03 +0000 (22:36 +0100)] 
* apt-pkg/cachefile.cc:
  - clean up lost atomic cachefiles with 'clean' (Closes: #650513)

12 years agojust in case: correct the header-guard for https to not match http
David Kalnischkies [Sat, 11 Feb 2012 20:26:42 +0000 (21:26 +0100)] 
just in case: correct the header-guard for https to not match http

12 years agoremove the arbitrary MAXLEN limit for response lines (Closes: #658346)
David Kalnischkies [Sat, 11 Feb 2012 20:25:57 +0000 (21:25 +0100)] 
remove the arbitrary MAXLEN limit for response lines (Closes: #658346)

12 years agosave the universe by not printing messages about apport if a package
David Kalnischkies [Sat, 11 Feb 2012 20:01:35 +0000 (21:01 +0100)] 
save the universe by not printing messages about apport if a package
with this name is not installed (Closes: #619646)

12 years agofix the hold-testcase as it has problems with 'foreign' operations
David Kalnischkies [Sat, 11 Feb 2012 18:46:52 +0000 (19:46 +0100)] 
fix the hold-testcase as it has problems with 'foreign' operations

12 years agocorrectly ignore already (un)hold packages
David Kalnischkies [Sat, 11 Feb 2012 17:54:48 +0000 (18:54 +0100)] 
correctly ignore already (un)hold packages

12 years ago* cmdline/apt-mark.cc:
David Kalnischkies [Fri, 10 Feb 2012 18:34:35 +0000 (19:34 +0100)] 
* cmdline/apt-mark.cc:
  - detect if dpkg has multiarch support before calling --set-selections

12 years agoFix the testcases to work with and configure dpkg correctly in a
David Kalnischkies [Fri, 10 Feb 2012 14:56:50 +0000 (15:56 +0100)] 
Fix the testcases to work with and configure dpkg correctly in a
multi-arch environment

It's not a complete and the "fixed" test is fixed more like a hack
as we have communication problems with dpkg if dpkg and APT disagree
on the interpretation of the native architecture, see also:
http://lists.debian.org/debian-dpkg/2012/02/msg00051.html

12 years agoensure that architectures are not added multiple times
David Kalnischkies [Fri, 10 Feb 2012 14:01:31 +0000 (15:01 +0100)] 
ensure that architectures are not added multiple times

12 years agocall dpkg --assert-multi-arch with execvp instead of execv
David Kalnischkies [Thu, 9 Feb 2012 17:06:29 +0000 (18:06 +0100)] 
call dpkg --assert-multi-arch with execvp instead of execv

12 years agoDanish translation update
bubulle@debian.org [Sun, 5 Feb 2012 07:08:59 +0000 (08:08 +0100)] 
Danish translation update

12 years agorework previous patch to avoid changing the inline code
Michael Vogt [Fri, 3 Feb 2012 12:10:34 +0000 (13:10 +0100)] 
rework previous patch to avoid changing the inline code

12 years ago* apt-pkg/contrib/fileutl.h:
Michael Vogt [Fri, 3 Feb 2012 11:14:39 +0000 (12:14 +0100)] 
* apt-pkg/contrib/fileutl.h:
  - fix compat with FileFd::OpenDescriptor() in ReadOnlyGzip mode

12 years ago* apt-pkg/contrib/fileutl.h:
Michael Vogt [Fri, 3 Feb 2012 10:56:29 +0000 (11:56 +0100)] 
* apt-pkg/contrib/fileutl.h:
  - fix compat with FileFd::OpenDescriptor() in ReadOnlyGzip mode

12 years agoFix IndexCopy::CopyPackages and TranslationsCopy::CopyTranslations to
Michael Vogt [Tue, 31 Jan 2012 16:50:58 +0000 (17:50 +0100)] 
Fix IndexCopy::CopyPackages and TranslationsCopy::CopyTranslations to
handle compressed files again (LP: #924182, closes: #658096)

12 years agomerged from lp:~mvo/apt/mvo
Michael Vogt [Tue, 31 Jan 2012 10:22:24 +0000 (11:22 +0100)] 
merged from lp:~mvo/apt/mvo

12 years agomerged from lp:~donkult/apt/experimental/
Michael Vogt [Tue, 31 Jan 2012 10:21:02 +0000 (11:21 +0100)] 
merged from lp:~donkult/apt/experimental/

12 years ago* apt-pkg/deb/dpkgpm.cc:
Michael Vogt [Mon, 30 Jan 2012 19:58:13 +0000 (20:58 +0100)] 
* apt-pkg/deb/dpkgpm.cc:
  - fix crash when a package is in removed but residual config state
    (LP: #923807)

12 years agomerged from donkult
Michael Vogt [Mon, 30 Jan 2012 19:22:18 +0000 (20:22 +0100)] 
merged from donkult

12 years ago* apt-pkg/aptconfiguration.cc:
David Kalnischkies [Mon, 30 Jan 2012 18:17:58 +0000 (19:17 +0100)] 
* apt-pkg/aptconfiguration.cc:
  - chroot if needed before calling dpkg --print-foreign-architectures

12 years ago* methods/http{s,}.cc:
David Kalnischkies [Mon, 30 Jan 2012 15:26:10 +0000 (16:26 +0100)] 
* methods/http{s,}.cc:
  - if a file without an extension is requested send an 'Accept: text/*'
    header to avoid that the server chooses unsupported compressed files
    in a content-negotation attempt (Closes: #657560)

12 years ago* apt-pkg/contrib/configuration.cc:
David Kalnischkies [Mon, 30 Jan 2012 12:13:29 +0000 (13:13 +0100)] 
* apt-pkg/contrib/configuration.cc:
  - do not stop parent transversal in FindDir if the value is empty

See http://lists.debian.org/deity/2012/01/msg00053.html , too.

12 years ago* apt-pkg/depcache.cc:
David Kalnischkies [Sun, 29 Jan 2012 14:25:02 +0000 (15:25 +0100)] 
* apt-pkg/depcache.cc:
  - if a M-A:same package is marked for reinstall, mark all it's installed
    silbings for reinstallation as well (LP: #859188)

12 years ago* apt-pkg/algorithms.cc:
Colin Watson [Sun, 29 Jan 2012 13:47:34 +0000 (14:47 +0100)] 
* apt-pkg/algorithms.cc:
  - use a signed int instead of short for score calculation as upgrades
    become so big now that it can overflow (Closes: #657732, LP: #917173)

12 years agoadd the testcase i wrote for revno 2209.1.2, to ensure that breaks on
David Kalnischkies [Sun, 29 Jan 2012 12:57:16 +0000 (13:57 +0100)] 
add the testcase i wrote for revno 2209.1.2, to ensure that breaks on
not-installed versions are ignored while searching for breakage loops

12 years agoadd a very simple testcases for the previous commited fix for
David Kalnischkies [Sun, 29 Jan 2012 12:54:40 +0000 (13:54 +0100)] 
add a very simple testcases for the previous commited fix for
Breaks handling in algorithms.cc by Colin Watson

12 years ago* apt-pkg/algorithms.cc:
Colin Watson [Sun, 29 Jan 2012 12:53:25 +0000 (13:53 +0100)] 
* apt-pkg/algorithms.cc:
  - don't break out of the main-resolver loop for Breaks to deal with all
    of them in a single iteration (Closes: #657695, LP: #922485)

12 years ago* cmdline/apt-get.cc:
Steve Langasek [Sun, 29 Jan 2012 12:24:17 +0000 (13:24 +0100)] 
* cmdline/apt-get.cc:
  - for cross-build-dependencies M-A: none should be DEB_HOST_ARCH,
    not DEB_BUILD_ARCH (Closes: #646288)

12 years agoensure that dpkg binary doesn't have the chroot-directory prefixed
David Kalnischkies [Sun, 29 Jan 2012 12:10:38 +0000 (13:10 +0100)] 
ensure that dpkg binary doesn't have the chroot-directory prefixed

12 years ago* apt-pkg/deb/dpkgpm.cc:
David Kalnischkies [Sun, 29 Jan 2012 11:59:42 +0000 (12:59 +0100)] 
* apt-pkg/deb/dpkgpm.cc:
  - chroot if needed before dpkg --assert-multi-arch

12 years agoreleasing version 0.8.16~exp12
Michael Vogt [Tue, 24 Jan 2012 13:02:04 +0000 (14:02 +0100)] 
releasing version 0.8.16~exp12

12 years agoPolish (Michał Kułach). Closes: #656908
bubulle@debian.org [Mon, 23 Jan 2012 20:42:03 +0000 (21:42 +0100)] 
Polish (Michał Kułach). Closes: #656908

12 years agomerged from bzr+ssh://bazaar.launchpad.net/~donkult/apt/experimental/
Michael Vogt [Fri, 20 Jan 2012 09:27:03 +0000 (10:27 +0100)] 
merged from bzr+ssh://bazaar.launchpad.net/~donkult/apt/experimental/

12 years ago* test/libapt/run-tests:
David Kalnischkies [Fri, 20 Jan 2012 01:00:48 +0000 (02:00 +0100)] 
* test/libapt/run-tests:
  - hurd doesn't have dmesg yet and we don't really need it either,
    so use with $0 a more stable data source for hashsumming