summaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAgeFilesLines
* init: Fix path to external solversJulian Andres Klode2016-08-291-1/+1
| | | | | | | This accidentally had two apt in it. This fixes a regression from commit 8757a0f. Gbp-Dch: ignore
* don't loop on pinning pkgs from absolute debs by regexDavid Kalnischkies2016-08-292-6/+9
| | | | | | | | | | | | | | | | | | | | | An absolute filename for a *.deb file starts with a /. A package with the name of the file is inserted in the cache which is provided by the "real" package for internal reasons. The pinning code detects a regex based wildcard by having the regex start with /. That is no problem as a / can not be included in a package name… expect that our virtual filename package can and does. We fix this two ways actually: First, a regex is only being considered a regex if it also ends with / (we don't support flags). That stops our problem with the virtual filename packages already, but to be sure we also do not enter the loop if matcher and package name are equal. It has to be noted that the creation of pins for virtual packages like the here effected filename packages is pointless as only versions can be pinned, but checking that a package is really purely virtual is too costly compared to just creating an unused pin. Closes: 835818
* randomize acquire order for same type index filesDavid Kalnischkies2016-08-293-3/+41
| | | | | | | | | | | | | | | Without randomizing the order in which we download the index files we leak needlessly information to the mirrors of which architecture is native or foreign on this system. More importantly, we leak the order in which description translations will be used which in most cases will e.g. have the native tongue first. Note that the leak effect in practice is limited as apt detects if a file it wants to download is already available in the latest version from a previous download and does not query the server in such cases. Combined with the fact that Translation files are usually updated infrequently and not all at the same time, so a mirror can never be sure if it got asked about all files the user wants.
* Merge branch 'portability/freebsd'Julian Andres Klode2016-08-2744-119/+370
|\
| * CMake: Add missing iconv dependencyJulian Andres Klode2016-08-263-2/+28
| | | | | | | | | | | | | | | | | | | | | | | | FreeBSD has two iconv systems: It ships an iconv.h itself, and symbols for that in the libc. But there's also the port of GNU libiconv, which unfortunately for us, Doxygen depends on. This changes things to prefer a separate libiconv library over the system one; that is, the port on FreeBSD. Gbp-Dch: ignore
| * Make root group configurable via ROOT_GROUPJulian Andres Klode2016-08-266-8/+17
| | | | | | | | | | This is needed on BSD where root's default group is wheel, not root.
| * test: Use :$(id -gn) instead of :root (when run as root)Julian Andres Klode2016-08-262-4/+4
| | | | | | | | | | | | | | On BSD systems, the root group is wheel, not root, so let's just use the default group here. Gbp-Dch: ignore
| * test: Assert multi-arch in the chrootJulian Andres Klode2016-08-261-1/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The host system might not have a dpkg installed, which makes dpkg fail with: dpkg not recorded as installed, cannot check for multi-arch support! That's entirely useless of course. We want to know if dpkg could support multi-arch in our chroot, so we pseudo-install dpkg into the chroot and pretend it's version is one version higher than the minimum dpkg version, so dpkg --assert-multi-arch works on recent dpkgs. Gbp-Dch: ignore
| * changelog: Respect Dir setting for local changelog gettingJulian Andres Klode2016-08-262-4/+5
| | | | | | | | | | This fixes issues with chroots, but the goal here was to get the test suite working on systems without dpkg.
| * Lower-case uname -r output in kernel autoremove helperJulian Andres Klode2016-08-262-3/+3
| | | | | | | | | | This is needed on FreeBSD which has versions like 11.0-RC1, otherwise the tests would fail.
| * test: More portable check for dpkg versionsJulian Andres Klode2016-08-264-3/+6
| | | | | | | | | | | | | | This check should work regardless if dpkg was installed by dpkg or by a native package manager like RPM or pkg. Gbp-Dch: ignore
| * test: Use a file to determine TEST_DEFAULT_GROUPJulian Andres Klode2016-08-261-5/+2
| | | | | | | | | | | | | | This is more safe against sticky bits. For example, in FreeBSD all files created in /tmp have the group set to wheel. Gbp-Dch: ignore
| * apt-inst: debfile: Pass comp. Name to ExtractTar, not BinaryJulian Andres Klode2016-08-261-1/+1
| | | | | | | | | | | | In the old days, apt-inst used to use binaries, but now it uses the built-in support and matches using Name, and not a Binary.
| * test: Allow db_dump-5 instead of db_dumpJulian Andres Klode2016-08-261-3/+8
| | | | | | | | Gbp-Dch: ignore
| * Always pass a directory to find before an optionJulian Andres Klode2016-08-262-2/+2
| | | | | | | | | | | | | | On BSD systems, we cannot simply use find -name or stuff, we always have to pass a directory name first. Gbp-Dch: ignore
| * test: Fix invalid quoting in variable expansionJulian Andres Klode2016-08-261-6/+6
| | | | | | | | | | | | | | This breaks the tests with FreeBSD's shell, and is not needed - it works fine without it. Gbp-Dch: ignore
| * test: Use printf "%b\n" instead of echo for strings with '\'Julian Andres Klode2016-08-262-8/+8
| | | | | | | | | | | | | | | | | | | | | | | | | | Use of echo with special characters is not portable. On a normal POSIX system, the behavior with backslash escaped strings is implementation-defined. On an XSI-conformant system, they must be interpreted. A way out is the printf command - printf "%b" specifies that the following argument is to be printed with backslash escapes interpreted. Gbp-Dch: ignore
| * test: Make sure we always find a dpkg in status fileJulian Andres Klode2016-08-261-8/+8
| | | | | | | | | | | | | | Especially on non-Debian platforms, dpkg might not list itself on the host system, and thus dpkg --assert-multi-arch fails. Gbp-Dch: ignore
| * test: Avoid use of /proc/self/fdJulian Andres Klode2016-08-263-6/+11
| | | | | | | | | | | | | | | | | | Use /dev/fd in test-bug-712116-dpkg-pre-install-pkgs-hook-multiarch, skip test-no-fds-leaked-to-maintainer-scripts (it is not guaranteed that /dev/fd contains all file descriptors), and avoid the unneeded use of /proc/fd in another test case. Gbp-Dch: ignore
| * test: Explicitly pass --admindir=var/lib/dpkg to dpkgJulian Andres Klode2016-08-261-0/+2
| | | | | | | | | | | | | | | | Our test suite assumes that dpkg's admindir is var/lib/dpkg. This might not always be true; for example, on FreeBSD, it is located at /var/db/dpkg. Gbp-Dch: ignore
| * test: Fix building of noopchrootJulian Andres Klode2016-08-261-1/+5
| | | | | | | | Gbp-Dch: ignore
| * test: Get rid of debhelper rules.tiny example depJulian Andres Klode2016-08-261-1/+13
| | | | | | | | Gbp-Dch: ignore
| * test: Allow moreutils-parallel instead of parallelJulian Andres Klode2016-08-261-1/+5
| | | | | | | | | | | | That's what it's called on FreeBSD. Gbp-Dch: ignore
| * test: Allow stunnel to be used instead of stunnel4Julian Andres Klode2016-08-261-4/+9
| | | | | | | | | | | | This is needed for Fedora and FreeBSD. Gbp-Dch: ignore
| * test: Substitute GNU commands for other commands where availableJulian Andres Klode2016-08-261-2/+18
| | | | | | | | | | | | | | We are simply checking for gnuCMD and gCMD for each command we are interested in. Gbp-Dch: ignore
| * test: Use C locale instead of C.UTF-8Julian Andres Klode2016-08-261-1/+1
| | | | | | | | Gbp-Dch: ignore
| * Make directory paths configurableJulian Andres Klode2016-08-268-22/+45
| | | | | | | | | | | | | | This allows other vendors to use different paths, or to build your own APT in /opt for testing. Note that this uses + 1 in some places, as the paths we receive are absolute, but we need to strip of the initial /.
| * apt-key: Only use readlink -f for existing componentsJulian Andres Klode2016-08-261-1/+1
| | | | | | | | | | On FreeBSD, readlink -f requires the last component to exist.
| * Use C locale instead of C.UTF-8 for protocol stringsJulian Andres Klode2016-08-264-6/+6
| | | | | | | | | | | | The C.UTF-8 locale is not portable, so we need to use C, otherwise we crash on other systems. We can use std::locale::classic() for that, which might also be a bit cheaper than using locale("C").
| * CMake: Add support for libintlJulian Andres Klode2016-08-261-3/+9
| | | | | | | | | | | | | | | | | | | | | | | | | | This basically just links everything to libintl if USE_NLS is on. It would be better to just link those targets that are actually translated, but doing so is a huge PITA. Also move the include_directories() for the build-tree include/ directory to the top of the CMakeLists.txt, otherwise it only gets passed after Intl_INCLUDE_DIRS, which means we will built against installed apt-pkg headers (if any) instead of our own. Gbp-Dch: ignore
| * CMake: Handle Berkeley DB on FreeBSDJulian Andres Klode2016-08-262-3/+15
| | | | | | | | | | | | | | | | | | | | | | | | | | The BSD systems still ship their own db.h with a historical BSD implementation, which is preferred by CMake, as it searches its default path first. We thus have to disable the DEFAULT_PATH for the search, unfortunately. We also need to pass the correct include directory to the target. Furthermore, on FreeBSD the library is called db-<VERSION>, so let's add db-5 to the list of allowed names. Gbp-Dch: ignore
| * methods/connect.cc: Only use AI_IDN if definedJulian Andres Klode2016-08-261-0/+2
| | | | | | | | Gbp-Dch: ignore
| * CMake/private-download: Fix statfs includes on FreeBSDJulian Andres Klode2016-08-262-5/+13
| | | | | | | | | | | | | | On FreeBSD, we have to include sys/params.h and sys/mount.h, not sys/vfs.h. Gbp-Dch: ignore
| * CMake: Add FindLZ4 and FindLZMAJulian Andres Klode2016-08-263-3/+53
| | | | | | | | | | | | This makes things work with /usr/local on FreeBSD. Gbp-Dch: ignore
| * CMake: Do not use -lresolv if res_init exists in libcJulian Andres Klode2016-08-263-2/+10
| | | | | | | | Gbp-Dch: ignore
| * CMake: Do not hardcode -ldlJulian Andres Klode2016-08-261-1/+1
| | | | | | | | | | | | Does not exist on FreeBSD Gbp-Dch: ignore
| * CMake: Handle endian.h locations on other platformsJulian Andres Klode2016-08-263-0/+27
| | | | | | | | Gbp-Dch: ignore
| * CMake: Handle BSD platforms with sig_t instead of sighandler_tJulian Andres Klode2016-08-261-0/+13
| | | | | | | | | | | | | | Somewhat annoying, but OK. Might want to switch to something more clever to get rid of the typedef at all. Gbp-Dch: ignore
| * Add missing includes and external definitionsJulian Andres Klode2016-08-268-0/+10
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Several modules use std::array without including the array header. Bad modules. Some modules use STDOUT_FILENO and friends, or close() without including unistd.h, where they are defined. One module also uses WIFEXITED() without including sys/wait.h. Finally, environ is not specified to be defined in unistd.h. We are required to define it ourselves according to POSIX, so let's do that.
| * CMake: Define _WITH_GETLINE for FreeBSDJulian Andres Klode2016-08-231-0/+3
| | | | | | | | Gbp-Dch: ignore
* | Accept --autoremove as alias for --auto-removeJulian Andres Klode2016-08-262-1/+2
| | | | | | | | | | I probably missed that when I did the usability work. But better late than never.
* | treat .ddeb files like .deb, especially for dpkgDavid Kalnischkies2016-08-253-2/+31
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Ubuntu uses *.ddeb files for their debug packages, but the interface we are using since f495992428a396e0f98886c9a761a804aa161c68 to talk to dpkg isn't supporting *.ddeb files. This used to work previously as apt itself isn't caring about the filenames at all and if they are explicitly mentioned dpkg will accept all, too. It might or might not be a good idea to patch dpkg, too, but regardless of it happening, we don't want to couple us to closely to dpkg for this minor feature but testing for this at runtime as it would delay shipping the fix for the too long commandlines further. It is also questionable if it is really a good idea to allow any file extension to be used here (like .foobar in the testcase), but we used to and we tend to avoid breaking existing usecases if we can help it. As a bonus, this also allows the installation of ddeb files directly from the commandline as you can with deb files already. We continue to ignore udeb through as the user-mistake to useful ratio is too high. LP: #1616909
* | show apt-key warnings in apt updateDavid Kalnischkies2016-08-253-0/+31
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | In 105503b4b470c124bc0c271bd8a50e25ecbe9133 we got a warning implemented for unreadable files which greatly improves the behavior of apt update already as everything will work as long as we don't need the keys included in these files. The behavior if they are needed is still strange through as update will fail claiming missing keys and a manual test (which the user will likely perform as root) will be successful. Passing the new warning generated by apt-key through to apt is a bit strange from an interface point of view, but basically duplicating the warning code in multiple places doesn't feel right either. That means we have no translation for the message through as apt-key has no i18n yet. It also means that if the user has a bunch of sources each of them will generate a warning for each unreadable file which could result in quite a few duplicated warnings, but "too many" is better than none. Closes: 834973
* | apt-key: warn instead of fail on unreadable keyringsDavid Kalnischkies2016-08-252-26/+55
| | | | | | | | | | | | | | | | | | | | | | apt-key has inconsistent behaviour if it can't read a keyring file: Commands like 'list' skipped silently over such keyrings while 'verify' failed hard resulting in apt to report cconfusing gpg errors (#834973). As a first step we teach apt-key to be more consistent here skipping in all commands over unreadable keyrings, but issuing a warning in the process, which is as usual for apt commands displayed at the end of the run.
* | do not restore selections for already purged packagesDavid Kalnischkies2016-08-243-7/+19
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | In most cases apt was already skipping the (re)setting of packages as to be removed/purged if dpkg had told us that it already did, but we haven't dealt with it in the most obvious of the cases: Selections set for packages we touched in this operation which either restores selections even dpkg would have overridden or e.g. tries to restore a purge selection for a package which was just purged – does not happen with apt itself as it isn't using selections in this way, but higher frontends like aptitude do. The result in the later case is a warning printed by dpkg that we try to set selections for an unknown package, which is harmless per se, but can be confusing for users and we really shouldn't cause warnings in dpkg if we can help it. Reported-By: Guillem Jover on IRC
* | improve code & doc for aquire weak/loop failingDavid Kalnischkies2016-08-241-6/+6
| | | | | | | | | | | | Improve-Upon: 2e2865ae53a65c00dd55a892d5b48458f3110366 Reported-By: Julian Andres Klode Gbp-Dch: Ignore
* | do fail on weakhash/loop earlier in acquireDavid Kalnischkies2016-08-248-50/+70
|/ | | | | | | | | | | | | | | | | The bugreport shows a segfault caused by the code not doing the correct magical dance to remove an item from inside a queue in all cases. We could try hard to fix this, but it is actually better and also easier to perform these checks (which cause instant failure) earlier so that they haven't entered queue(s) yet, which in return makes cleanup trivial. The result is that we actually end up failing "too early" as if we wouldn't be careful download errors would be logged before that process was even started. Not a problem for the acquire system, but likely to confuse users and programs alike if they see the download process producing errors before apt was technically allowed to do an acquire (it didn't, so no violation, but it looks like it to the untrained eye). Closes: 835195
* CMake: Do not add po/ if USE_NLS is OFFJulian Andres Klode2016-08-232-2/+6
| | | | | | Previously, we would have generated all the translations, but not turn them on in the code. Instead, move the Translation crap into po/ and disable po/ alltogether if USE_NLS if OFF.
* CMake: Discover docbook stylesheet in other locationsJulian Andres Klode2016-08-234-6/+34
| | | | | Distributions seem to install this stuff all over the place, so let's add a common list of paths we know about.
* install-progress: Call the real ::fork() in our fork() methodJulian Andres Klode2016-08-231-1/+1
| | | | | | We basically called ourselves before, creating an endless loop. Reported-By: clang
generated by cgit v1.2.3-70-g09d2 (git 2.46.0) at 2024-11-01 06:31:22 +0000