- Jul 29, 2022
-
-
Thomas Petazzoni authored
Fixes: https://gitlab.com/buildroot.org/buildroot/-/jobs/2781800458 Signed-off-by:
Thomas Petazzoni <thomas.petazzoni@bootlin.com>
-
Thomas Petazzoni authored
This bogus BR2_GLOBAL_PATCH_DIR value was added in commit 9dd5382d ("board/intel/galileo: fix build failure with host gcc 10") back in February. This should help fixing https://gitlab.com/buildroot.org/buildroot/-/jobs/2781800667 , as it complains with: WARN: defconfig ./configs/galileo_defconfig can't be used: Missing: BR2_GLOBAL_PATCH_DIR=board/intel/galileo/patches Signed-off-by:
-
Thomas Petazzoni authored
Fixes: https://gitlab.com/buildroot.org/buildroot/-/jobs/2781800747 Signed-off-by:
-
Thomas Petazzoni authored
Fixes: https://gitlab.com/buildroot.org/buildroot/-/jobs/2781800773 Signed-off-by:
-
- Jul 28, 2022
-
-
Julien Olivain authored
This is a simple test importing pyalsa, showing alsa library version and attempting to list cards. Signed-off-by:
Julien Olivain <ju.o@free.fr> Signed-off-by:
-
Thomas Petazzoni authored
Now that all hash files have been fixed, enable checking of hash spacing in check-package. Signed-off-by: -
Thomas Petazzoni authored
It's time to finally switch over globally to the new spacing format that we have agreed on for the hash file, with 2 spaces as a separator between fields. This commit was mechanically generated using: find . -type f -name '*.hash' | xargs sed -i 's%^md5[ \t]*\([^ \t]*\)[ \t]*\(.*\)$%md5 \1 \2%' find . -type f -name '*.hash' | xargs sed -i 's%^sha1[ \t]*\([^ \t]*\)[ \t]*\(.*\)$%sha1 \1 \2%' find . -type f -name '*.hash' | xargs sed -i 's%^sha256[ \t]*\([^ \t]*\)[ \t]*\(.*\)$%sha256 \1 \2%' find . -type f -name '*.hash' | xargs sed -i 's%^sha512[ \t]*\([^ \t]*\)[ \t]*\(.*\)$%sha512 \1 \2%' This commit can easily be backported on the LTS branch by re-running the same commands, if needed. Signed-off-by: -
In eudev and systemd, we have code that deletes the hwdb sources from the target - they are not useful since a binary hwdb is created from them. However, if eudev or systemd is not used, then those sources are not useful either. It's possible that other packages than eudev or systemd install hwdb files, which would be left on the system. Always remove the hwdb files. Note that we don't expect much space savings from this, but anything may help. It's certainly more consistent to do it always than just in eudev and systemd. We do this both from /usr/lib/udev (usual installation path for systemd) and in /etc/udev (usual installation path for eudev) because packages may install in either location. We keep the comment explaining why it's done in rootfs-pre-cmd instead of target-finalize - this was only present in eudev.mk. Signed-off-by:
Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be> Signed-off-by:
-
The HOST_EUDEV_SYSCONFDIR is a leftover from before eudev, when udev was still built from systemd sources. Now that we have a separate file for eudev and systemd, there's no need any more to have a condition there. In addition, its usage in systemd.mk has been removed. Remove the HOST_EUDEV_SYSCONFDIR variable and instead use /etc directly in eudev.mk. Signed-off-by:
-
Thomas Petazzoni authored
Use the more common ifeq ($(BR2_STATIC_LIBS),) form. Signed-off-by: -
It turns out that zsh has a number of additional TRY_RUN calls. We already set a few with cache variables, add a few more. They are found based on reading the configure.ac, aczsh.m4, and the configure output. Signed-off-by:
-
John Keeping authored
By default connman periodically saves network interface statistics into an mmap'd round robin file. These are not accessible over the D-Bus interface and the only method to interact with this is ResetCounters() which is marked as experimental. Continually writing statistics will increase drive wear and may not be desirable in some devices with limited flash erase cycles, so add an option to disable this. Note that connman does not provide any runtime control over this facility, it can only be disabled via the configure script. Signed-off-by:
John Keeping <john@metanate.com> Signed-off-by:
-
Fabrice Fontaine authored
aplay needs alsa-lib mixer since bump to version 4.0.0 in commit e843433e and https://github.com/Arkq/bluez-alsa/commit/7050c9a81895500fe2d59ec6e4859633d4353e36: In file included from alsa-mixer.c:11: alsa-mixer.h:16:21: error: unknown type name 'snd_mixer_t'; did you mean 'snd_timer_t'? 16 | int alsa_mixer_open(snd_mixer_t **mixer, snd_mixer_elem_t **elem, | ^~~~~~~~~~~ | snd_timer_t So select BR2_PACKAGE_ALSA_LIB_MIXER instead of building aplay conditionally to keep backward compatibility Fixes: - http://autobuild.buildroot.org/results/d3faac6e49fe9cc21091a5dfa44780b8c63189be Signed-off-by:
Fabrice Fontaine <fontaine.fabrice@gmail.com> Signed-off-by:
-
Fabrice Fontaine authored
Drop fourth patch (already in version) https://github.com/FreeRDP/FreeRDP/releases/tag/2.8.0 Signed-off-by:
-
Fabrice Fontaine authored
Instead of having the user need to figure out that 2 alsa-lib features are needed for mpv ALSA support, have the top-level mpv option select the relevant alsa-lib features, when alsa-lib is enabled as already done for pulseaudio in commit 97cd3680 Signed-off-by:
-
Fabrice Fontaine authored
Instead of having the user need to figure out that 2 alsa-lib features are needed for wine ALSA support, have the top-level wine option select the relevant alsa-lib features, when alsa-lib is enabled as already done for pulseaudio in commit 97cd3680 Signed-off-by:
-
Fabrice Fontaine authored
Instead of having the user need to figure out that 2 alsa-lib features are needed for mediastreamer ALSA support, have the top-level pulseaudio option select the relevant alsa-lib features, when alsa-lib is enabled as already done for pulseaudio in commit 97cd3680 Signed-off-by:
-
Fabrice Fontaine authored
Commit ea746f31 forgot to select pcre2 resulting in the following build failure: Makefile:575: *** pcre2 is in the dependency chain of libmodsecurity that has added it to its _DEPENDENCIES variable without selecting it or depending on it from Config.in. Stop. Fixes: - http://autobuild.buildroot.org/results/6528d2611bd1a45c1e94bc6b866de9c33dd90a7b Signed-off-by:
-
Fabrice Fontaine authored
- Drop patch (already in version) - Update hash of LICENSE (update in year: https://github.com/Reference-LAPACK/lapack/commit/e7ad0370d4c3bb81d663abb773e49094f68bd5d6) http://www.netlib.org/lapack/lapack-3.10.1.html Signed-off-by:
-
Fabrice Fontaine authored
pulseaudio needs speexdsp, not speex resulting in the following build failure since commit 7752afbe which explicitly enable or disable speex: ../output-1/build/pulseaudio-16.1/meson.build:713:2: ERROR: Dependency "speexdsp" not found, tried pkgconfig and cmake Fixes: - http://autobuild.buildroot.org/results/8d4c5d5bfd65744753ca75180dd45a01c90f0b91 Signed-off-by:
-
Fabrice Fontaine authored
Enable MD4 in openssl to avoid the following build failure raised since the addition of the package in commit 736c4c16: In file included from src/lib/md4.c:16: src/lib/md4.c: In function 'fr_md4_calc': src/freeradius-devel/md4.h:76:33: error: unknown type name 'MD4_CTX'; did you mean 'FR_MD4_CTX'? 76 | # define FR_MD4_CTX MD4_CTX | ^~~~~~~ Fixes: - http://autobuild.buildroot.org/results/a8aa12a8129056391f975820ea3d1d62241ac051 Signed-off-by:
-
Fabrice Fontaine authored
PR_CAP_AMBIENT is only available since kernel 4.3 and https://github.com/torvalds/linux/commit/58319057b7847667f0c9585b9de0e8932b0fdb08 resulting in the following build failure since bump to version 1.7 in commit 07b7f886 and https://github.com/rfjakob/earlyoom/commit/9edc5cfc9ad794ed59baec5e6ee8d1060456a315: main.c: In function 'main': main.c:119:11: error: 'PR_CAP_AMBIENT' undeclared (first use in this function) prctl(PR_CAP_AMBIENT, PR_CAP_AMBIENT_CLEAR_ALL, 0, 0, 0); ^ Fixes: - http://autobuild.buildroot.org/results/5a0ad7e1f6de3c37e40c37d8f583fc699a457753 Signed-off-by:
-
Norbert Lange authored
the builtin string =~ regexp operator needs to load the regex module. This operator is used very often in zsh scripts, so link it statically to avoid the overhead of dynamic loading. Signed-off-by:
Norbert Lange <nolange79@gmail.com> Signed-off-by:
-
Norbert Lange authored
The autoconf scripts seem to assume that no dynamic loader is available if crosscompiling. This results in only a small set of modules being enabled and linked statically. For comparison, debian patches the fallback to enable dynamic modules (unconditionally). This commit enables modules if shared libraries are enabled. Signed-off-by:
-
Norbert Lange authored
this directory is used by the rpm package manager, and packages like systemd will install "macros" for this system. It should be deleted just like the similar /usr/share/aclocal directory from Autoconf. Signed-off-by:
-
Norbert Lange authored
Since the source HWDB files are already removed from the target, it makes little sense to keep the service around. If accidently triggered it could create an empty database. The service is installed in both staging and host directory as well and not removed from there, so a user could craft his own solutions if it's necessary to update the database. Signed-off-by:
-
Norbert Lange authored
Up to now, the source files of the hwdb was only cleaned if systemd-hwdb was used. However, those files should only be used by the hwdb anyway -- otherwise we shouldn't clean them even when the hwdb is enabled. Therefore, remove them even if hwdb is disabled. hwdb files can be put in two places: /usr/lib/udev, or /etc/udev. The former is where systemd-compliant packages would install them, but the latter is where they were originally placed. In systemd-land, /etc should only contain local information, so ideally it should be empty in the rootfs. The reality, however, is that many packages still install files in /etc. Thus, there are probably also packages that install hwdb files in /etc/udev. Therefore, clean the hwdb files both in /etc/udev and in /usr/lib/udev. Probably, we should do this cleaning even when systemd is not selected, since non-systemd packages may install files there which are not relevant without systemd-hwdb (or eudev). However, that's a bigger change which needs a separate patch (and a bit more consideration). While we're at it, get rid of the HOST_EUDEV_SYSCONFDIR variable. This is a leftover from when udev was still extracted from the systemd package (i.e. before eudev became a separate package). Signed-off-by:
-
Norbert Lange authored
udevadm has been deprecated for a while for building the HWDB, systemd-hwdb should be used in its place. This tool further has a switch to generate the database in /usr which is a better place for a one-time created file than the default /etc/udev. The systemd-hwdb wasn't enabled yet in the host build, so do that. Signed-off-by:
-
Norbert Lange authored
Since many years now, /run is the standard runstate directory for Linux systems, and Buildroot follows this standard. Encode this as the default for autotools packages. Note that the configure script bundled with a package may be generated with an older version of Autotools, that may not yet support the --runstatedir option. This will give an "Unsupported arguments" warning and the option will be ignored. We may have better luck with passing runstatedir as a variable instead of an option, but chances are that the configure script doesn't know that variant either (it wouldn't print a warning in that case). Since the warning is actually preferrable, pass it as an option after all. Signed-off-by:
-
Norbert Lange authored
Since many years now, /run is the standard runstate directory for Linux systems, and Buildroot follows this standard. Encode this as the default for CMake packages. Note that we don't really care for host packages. Signed-off-by:
-
Peter Korsgaard authored
Signed-off-by:Peter Korsgaard <peter@korsgaard.com>
-
Peter Korsgaard authored
Signed-off-by:
-
Fabrice Fontaine authored
gdal needs json-c, not libjson, resulting in the following build failure since the addition of the package in commit 1e64fa29: configure: error: could not find json-c/json.h While at it, also drop unrecognized options: configure: WARNING: unrecognized options: --disable-gtk-doc, --disable-gtk-doc-html, --disable-doc, --disable-docs, --disable-documentation, --with-xmlto, --with-fop, --disable-dependency-tracking, --enable-ipv6, --disable-nls, --without-dods-root, --without-fme, --without-grass, --without-libgrass, --without-ingres, --without-jasper, --without-charls, --without-mdb, --without-mongocxx, --without-mrsid_lidar, --without-perl Fixes: - http://autobuild.buildroot.org/results/f977675125167806e844a176a4c8c8873557b709 Signed-off-by:
-
Fabrice Fontaine authored
license hash is wrong since the addition of the package in commit 1e64fa29: ERROR: LICENSE.TXT has wrong sha256 hash: ERROR: expected: b82e6cca0b13f5db2f22ab667f22254fb1f4b135ea73d5bd6238ef89aff31f6c ERROR: got : fac0116e799ca4de79b076ee04863b5e2b69541ea3664b9bb108c81a5355f16a ERROR: Incomplete download, or man-in-the-middle (MITM) attack Fixes: - http://autobuild.buildroot.org/results/e443db6d7ff3ae867035d169853b880b5e13d181 Signed-off-by:
-
Marcin Niestroj authored
Second newline at the end of licence file was removed upstream, so update license file hash. [1] https://github.com/pytest-dev/pytest-asyncio/commit/4c7da65d6fcf9d725eccba28ad1ed2083524ee16 Signed-off-by:
Marcin Niestroj <m.niestroj@grinn-global.com> Signed-off-by:
-
Marcin Niestroj authored
There is just a starting year in licence file now (see [1]), but licence itself did not change. Also fixup the spacing in the hash file to match the "new" coding style. [1] https://github.com/pytest-dev/pytest/commit/ed24f29e6017183ea22f3136aa180a8e3280cf8a Signed-off-by:
-
Marcin Niestroj authored
Use the same TF-A as orangepi_zero_plus2 board. Fixes: https://gitlab.com/buildroot.org/buildroot/-/jobs/2720001176 Signed-off-by:
-
Sen Hastings authored
sortGrid() has been rewritten to dynamically generate stylesheets with explicit grid-row properties to re-order the rows, instead of removing and reinserting the cells. Performance *should* now be comperable to sorttable.js. Signed-off-by:
Sen Hastings <sen@phobosdpl.com> Signed-off-by:
-
Sen Hastings authored
Like all good problems, disparate pieces work together to create a "synergistically" hairy mess. The sortGrid() overhaul highlighted a flaw in pkg-stats allowing for duplicate package class names across rows. As an example, boot/barebox/barebox.mk and boot/barebox/barebox/barebox.mk both get the classname ._barebox and so sortGrid() sticks them on the same line giving a table with a vestigal row sticking out of the right side like some kind of appendage. Also I neglected to add a "_" to the current version column's cells pkgname class so instead of "._pkgname" we had ".pkgname" and so the cells were not collected properly as part of the row. These issues explain the formatting weirdness. package classnames are now ".path_to_package_makefile" without suffix (.mk) (so ._boot_barebox_barebox and ._boot_barebox_barebox_barebox instead of ._barebox) in order to guarantee uniqueness. and what was *accidentally* class="centered current_version data .barebox" is now class="centered current_version data ._boot_barebox_barebox" just like *all* the other cells in the row. :p Signed-off-by:
-
Christian Stewart authored
nerdctl is a CLI for containerd (package docker-containerd) which is drop-in compatible with the Docker Daemon CLI. This allows using the lighter weight containerd daemon directly, instead of via the additional docker daemon. It also implements rootless mode. https://github.com/containerd/nerdctl Signed-off-by:
Christian Stewart <christian@paral.in> Signed-off-by:
-
