BATS v1.6.0 has a regression which can result in test case failures being
ignored when teardown code executes successfully. This patch applies the
accepted bug-fix until the next BATS release.
Also, fix a comment that wasn't updated during the BATS upgrade, to match the
correct release (v1.6.0) associated with the SRCREV.
Signed-off-by: Richard Neill <richard.neill@arm.com>
Signed-off-by: Armin Kuster <akuster808@gmail.com>
The upstream repo location has changed, it's now
https://git.enlightenment.org/old/legacy-imlib2
It's not clear when or why it happened, but the the commit hash we use
in SRCREV exists in the 'new' location, so let's at least update the
SRC_URI for now, and fix this warning:
WARNING: imlib2-1.7.1-r0 do_fetch: Failed to fetch URL git://git.enlightenment.org/legacy/imlib2.git;protocol=https;branch=master, attempting MIRRORS if available
Signed-off-by: Nicolas Dechesne <nicolas.dechesne@linaro.org>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
(cherry picked from commit 62becef109)
Signed-off-by: Nicolas Dechesne <nicolas.dechesne@linaro.org>
Signed-off-by: Armin Kuster <akuster808@gmail.com>
Instead of hardcoding 4K page size, query the system and use the value for
memory mapping.
Signed-off-by: Denys Dmytriyenko <denis@denix.org>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
(cherry picked from commit 3af1d119de)
Signed-off-by: Armin Kuster <akuster808@gmail.com>
Some build server may provide sendmail like /usr/lib/sendmail which
may result the VAL_MTA to be /usr/lib/sendmail during do_configure,
but the sendmail is actually /usr/sbin/sendmail on the target.
Fixes:
# echo "invalide address email" | s-nail -s "test mail" -r test@test.comtester@test.com
s-nail: Cannot start /usr/lib/sendmail: executable not found (adjust *mta* variable)
/home/root/dead.letter 10/234
s-nail: ... message not sent
# which sendmail
/usr/sbin/sendmail
Signed-off-by: Mingli Yu <mingli.yu@windriver.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
(cherry picked from commit f334a00947)
Signed-off-by: Armin Kuster <akuster808@gmail.com>
rateconv.c looks like BSD-2 but it has second clause modified
significantly, therefore be specific about it.
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Cc: Alistair Francis <alistair.francis@wdc.com>
(cherry picked from commit d901de434d)
Signed-off-by: Armin Kuster <akuster808@gmail.com>
The install target is checking for install paths on build host assuming
thats where it will run too, which breaks cross builds. Add a patch to
aide the install from recipe.
Signed-off-by: Khem Raj <raj.khem@gmail.com>
(cherry picked from commit bbc6fa72c3)
Signed-off-by: Armin Kuster <akuster808@gmail.com>
These files are not needed when init system is not systemd
Signed-off-by: Khem Raj <raj.khem@gmail.com>
(cherry picked from commit f0d0034e26)
Signed-off-by: Armin Kuster <akuster808@gmail.com>
Changelog for released lsscsi-0.32 [20210505] [svn: r167]
- improve NVMe device parsing (e.g. /dev/nvme0c1n2)
- print nr_hw_queues when available for SCSI hosts
- Remove blank line after NVMe device name with -HL
- collect_disk_wwn_nodes: Fix WWN string copy
- make WWN printing for NVMe more consistent with
output from SCSI devices (e.g. with -u and -t)
- logic to select best SCSI id (--scsi_id) to output
- clean up warnings for gcc-10
- build with autoconf 2.70
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
(cherry picked from commit 4812d74acc)
Signed-off-by: Armin Kuster <akuster808@gmail.com>
april 2022
- 1.6
- configure script must error-out when (n)curses isn't found
- Move SIGWINCH handling from handler to NCURSES
- Make configure.ac compliant with autoconf >= 2.70
- Check that malloc did not return NULL
- Fix a possible memory leak in findFile()
- Prevent division by zero on empty files.
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
(cherry picked from commit c4d4e192f4)
Signed-off-by: Armin Kuster <akuster808@gmail.com>
It does not fail anymore as the qemu options has been sorted
Signed-off-by: Khem Raj <raj.khem@gmail.com>
(cherry picked from commit 1fa927eba9)
Signed-off-by: Armin Kuster <akuster808@gmail.com>
in case usrmerge is enabled, unit file is already in expected location
Signed-off-by: Khem Raj <raj.khem@gmail.com>
(cherry picked from commit b8e97f5b2a)
Signed-off-by: Armin Kuster <akuster808@gmail.com>
This helps compile gjs, external icu is needed with gcc-12 but not with
clang
Signed-off-by: Khem Raj <raj.khem@gmail.com>
(cherry picked from commit 78180b6c21)
Signed-off-by: Armin Kuster <akuster808@gmail.com>
This takes care of rootdir being different when using usrmerge
Signed-off-by: Khem Raj <raj.khem@gmail.com>
(cherry picked from commit 8cec1f1fd9)
Signed-off-by: Armin Kuster <akuster808@gmail.com>
default SBINDIR is hardcoded to /sbin which does not work when using
usrmerge
Signed-off-by: Khem Raj <raj.khem@gmail.com>
(cherry picked from commit 8c7ffffa07)
Signed-off-by: Armin Kuster <akuster808@gmail.com>
The gps-utils-python package was introduced in commit 66721940 (gpsd:
split python utils from gps-utils), except it was never added to
PACKAGES, which resulted in the following error if anything depends on
gpsd:
ERROR: Nothing RPROVIDES 'gps-utils-python' (but .../gpsd_3.23.1.bb
RDEPENDS on or otherwise requires it)
Also remove a no longer relevant comment.
Signed-off-by: Peter Kjellerstedt <peter.kjellerstedt@axis.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
since sdbus-c++-libsystemd outputs only static libraries, it will
not be installed inside the SDK.
add RDEPENDS:${PN}-dev = "" to remove the dependency to the empty package
when installing the sdbus-c++-libsystemd-dev into the SDK
without this change the SDK is missing sdbus-c++-dev and sdbus-c++-libsystemd-dev
Signed-off-by: Ben Fekih, Hichem <hichem.benfekih@ifm.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
ARM needs to use system icu, otherwise build fails with gcc-12
firefox-91.8.0/intl/icu/source/common/stringtriebuilder.cpp:388: more undefined references to `std::type_info::operator==(std::type_info const&) const' follow
| collect2: error: ld returned 1 exit status
Keep using vendored ICU on mips since this breaks build for gjs
introspection code which is run under qemu-user
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Neither gpsd nor significant part of gps-utils depend on python. However
installing gps-utils package would require installing python, which adds
about 17 MB to the image size. Split away gps-utils-python to allow
using gpsd/gps-utils in the systems with limited resources (or in the
initramfs).
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
* Disable RUN_GPERF_IF_PRESENT. Otherwise cmake will look for gperf. For
target, it will not find it and proceeds to use the pregenerated
files. However, for native it finds `/usr/bin/gperf`, but then tries
to run `gperf` instead, which fails since it is not in hosttools.
* Disable ENABLE_CPP for native. Otherwise it needs iconv, which it
cannot find.
Signed-off-by: Peter Kjellerstedt <peter.kjellerstedt@axis.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
gcc-12 has a better static analyser and finds additional use-after-free
case, addressed by removing the double free
Signed-off-by: Khem Raj <raj.khem@gmail.com>
If one adds some configuration that disables the creation of the Python
files, e.g., minimal=yes, then the installation fails when it tries to
copy the non-existent files. Avoid this by checking for their existence
before copying them.
Signed-off-by: Peter Kjellerstedt <peter.kjellerstedt@axis.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Fixes build with gcc-12
intl/icu/source/common/stringtriebuilder.cpp:388: more undefined references to `std::type_info::operator==(std::type_info const&) const' follow
collect2: error: ld returned 1 exit status
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Martin Jansa
Richard Neill
Nicolas Dechesne
Gianfranco Costamagna
Michael Opdenacker
Denys Dmytriyenko
Mingli Yu
Jiaqing Zhao
Khem Raj
Armin Kuster
Willy Tu
wangmy
Peter Kjellerstedt
Ben Fekih, Hichem
Dmitry Baryshkov