Some projects have been removing their setup.py shims under the
assumption that they are not needed due to setuptools being usable
via pep517, however this is not actually the case as it is not
possible to bootstrap setuptools without setup.py style builds as
setuptools has a dependency on wheel when being used via pep517
which is itself not installable via pep517 as that would create a
dependency cycle between wheel and setuptools(which means wheel
itself must use a legacy setup.py build/installation for itself).
In addition our pep517 toolchain itself requires setuptools for
building/installing dependencies, although it would appear these
dependencies will eventually move to flit. Until this is resolved
we must use setup.py style builds for setuptools packages.
Since it is not yet possible to convert setup.py based setuptools
builds to pep517 based setuptools builds we simply need to execute
the setuptools shim code directly if the setup.py file is missing.
See:
https://setuptools.pypa.io/en/latest/setuptools.html#setup-cfg-only-projects
Signed-off-by: James Hilliard <james.hilliard1@gmail.com>
Tested-by: Yegor Yefremov <yegorslists@googlemail.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Set CBOR2_BUILD_C_EXTENSION=1 to ensure c extensions are always built.
Signed-off-by: James Hilliard <james.hilliard1@gmail.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Add host-pkgconf dependency to fix the following build failure raised
since commit 4e3be3ae9d:
run pkg_config fail: "Failed to run `\"/nvmedata/autobuild/instance-6/output-1/per-package/host-rust/host/bin/pkg-config\" \"--libs\" \"--cflags\" \"openssl\"`: No such file or directory (os error 2)"
Fixes:
- http://autobuild.buildroot.org/results/b046523960079cbf85931b8f67c3b98dd07fbbda
Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
numactl unconditionally uses __atomic_fetch_and resulting in the
following build failure on architectures that need libatomic to provide
atomic intrinsics (e.g. microblaze) since commit
4ed540ddf5:
/nvmedata/autobuild/instance-5/output-1/host/lib/gcc/microblaze-buildroot-linux-uclibc/10.3.0/../../../../microblaze-buildroot-linux-uclibc/bin/ld: ./.libs/libnuma.a(libnuma.o): in function `numa_node_to_cpus_v1':
(.text+0x2a34): undefined reference to `__atomic_fetch_and_1'
Fixes:
- http://autobuild.buildroot.org/results/e225cb83dae390d9dc543d4da85c52180efbd40a
Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
The xradio driver is broken for kernel configurations with disabled
power management core functionality (CONFIG_PM). The xradio package
makefile attempts to enable all the required kernel configuration
options, including CONFIG_PM. However certain architectures have
no support for CONFIG_PM. Those are or1k, nios2, csky, and more.
One possible fix would be to exclude those architectures in the
package Config.in. On the other hand, the xradio driver is rarely
used: orangepi_zero is the only such board in Buildroot. So it
makes sense to keep it simple and enable xradio only for ARM.
Any new xradio users will be considered on a case by case basis.
Fixes:
- http://autobuild.buildroot.net/results/2ad11cd7a9ca6b694854730e46260a944e8c4bac/
Signed-off-by: Sergey Matyukevich <geomatsi@gmail.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Fix the following build failure with libressl raised since bump to
version 3.5.2 in commit 8b216927db:
capture_openssl.c: In function 'P_hash':
capture_openssl.c:101:18: error: storage size of 'hm' isn't known
101 | HMAC_CTX hm;
| ^~
Fixes:
- http://autobuild.buildroot.org/results/f1e8cdb3ac35a30055ab79d41e6cc038e5339c37
Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Fix the following build failure with libressl raised since bump to
version 3.5.2 in commit 8b216927db:
pkcs11h-openssl.c: In function 'DSA_meth_set1_name':
pkcs11h-openssl.c:239:41: error: invalid use of incomplete typedef 'DSA_METHOD' {aka 'struct dsa_method'}
239 | rv = _pkcs11h_mem_strdup ((void *)&meth->name, name);
| ^~
Fixes:
- http://autobuild.buildroot.org/results/9b0d3bf7d97696c7be6de1724daaba196626b865
Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
5.15.x is getting quite old, so default to 5.17.x instead.
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
The 5.16.x series is now EOL upstream, so drop the linux-headers option and
add legacy handling for it.
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
As reported by [1], the lxc test is broken since lxc >= 4.0.11.
A patch was added to lxc 4.0.11 to use the new mount api for devpts
setup [2] but the fall back code doesn't work when this new mount
API is not supported. This API was added in kernel 5.6.
(kernel 5.5)
DEBUG conf - conf.c:lxc_setup_devpts_child:1682 - No new devpts instance will be
mounted since no pts devices are required
lxc-start lxc_iperf3 DEBUG conf - conf.c:lxc_setup_dev_console:1966 - Cleared
all (0) mounts from "/dev/console"
lxc-start lxc_iperf3 ERROR mount_utils - mount_utils.c:mount_at:661 - No such
file or directory - Failed to mount "/proc/self/fd/44" to "/proc/self/fd/43"
lxc-start lxc_iperf3 ERROR conf - conf.c:lxc_setup_dev_console:1988 - No such
file or directory - Failed to mount "10(/dev/pts/0)" on "43"
lxc-start lxc_iperf3 ERROR conf - conf.c:lxc_setup_console:2143 - No such file
or directory - Failed to setup console
(kernel 5.6)
lxc-start lxc_iperf3 TRACE mount_utils - mount_utils.c:can_use_mount_api:582 -
Kernel supports mount api
lxc-start lxc_iperf3 TRACE mount_utils - mount_utils.c:move_detached_mount:328
- Attach detached mount 45 to filesystem at 43
lxc-start lxc_iperf3 TRACE conf - conf.c:lxc_setup_dev_console:1990 - Setup
console "/dev/pts/0"
Bump the kernel to the current LTS 5.15.38 version that fully support the
mount API needed by lxc.
Fixes:
https://gitlab.com/buildroot.org/buildroot/-/jobs/2429013708
[1] http://lists.busybox.net/pipermail/buildroot/2022-January/635251.html
[2] be606e16fd
Signed-off-by: Romain Naour <romain.naour@gmail.com>
Cc: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
Switch U-Boot from vendor downstream version to upstream. Since the
upstream U-Boot uses binman to build the bootable binary (flash.bin) the
need for the iMX specifc prepare script is no longer required.
Signed-off-by: Heiko Thiery <heiko.thiery@gmail.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
Signed-off-by: Heiko Thiery <heiko.thiery@gmail.com>
[yann.morin.1998@free.fr: we now have 5.17 headers too]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
Add a dependency to sytemd if enabled so cmake can find libsystemd
and install the systemd service file.
Signed-off-by: Heiko Thiery <heiko.thiery@gmail.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
Drop patch that is now fixed upstream. Add a dependency to sytemd if
enabled so cmake can find libsystemd and install the systemd service file.
Signed-off-by: Heiko Thiery <heiko.thiery@gmail.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
Since the very beginning, libunistring was a mandatory dependency of
gnutls. However, it would use its internal copy if libunistring was not
selected. We never want that, so make libunistring an actual mandatory
dependency.
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
Version 3.7.4 added compression options with brotli, zlib and zstd.
These are automatically discovered, which makes their inclusion depend
on the build order. Therefore, explicitly enable/disable them.
Note that the configure help text says "--without-brotli" and
"--without-zstd", but the options are actually --without-libbrotli and
--without-libzstd. --without-zlib is correct in the help text.
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
The option is set a few lines below depending on
BR2_PACKAGE_GNUTLS_OPENSSL.
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
[yann.morin.1998@free.fr: move it all under a single conditional block]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
The pcre/libregex dependency was removed in version 3.7.3 with upstream
commit 26578b7d02c269ff1d34ff782d84c7667734d03d, which removed the
bundled libopts. Remove the pcre dependency and the relevant CONF_OPTS
handling.
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
gnutls's configure script has a weird approach where it first searches
for dependent libraries in the path specified by --prefix, before
searching in the default search path. Since we set --prefix to /usr,
and it doesn't take into account DESTDIR (which is anyway not set at
configure time), that means it will first search /usr/lib before
searching $(STAGING_DIR)/usr/lib.
Ideally, this would be fixed in the configure script itself. However,
the m4 file that does this is pretty complex, it's not immediately clear
where to add $DESTDIR. In addition it comes from gnulib which is a
somewhat annoying upstream.
Therefore, instead, bypass the prefix lookup with
--without-libfoo-prefix. Note that we could set
--with-libfoo-prefix=$(STAGING_DIR)/usr (the latter is already done for
librt and libpthread), but that's pretty pointless -
--without-libfoo-prefix in fact reverts to what should have been done in
the first place, i.e. use the toolchain search path.
Add --without-libfoo-prefix for all options defined in configure (found
with ./configure --help | grep without-.*-prefix). Most of these are
only used in tests (e.g. libcrypto) or even not at all (e.g. libiconv),
but it's fairly hard to discover this and to be sure that they are
indeed not needed, so better pass all of them.
Remove the now-redundant arguments for librt and libpthread.
Add a comment to remind people to revisit these when bumping the
version.
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Cc: Peter Korsgaard <peter@korsgaard.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
Not only is ninja much faster, it also avoids the following build
failure:
make[3]: printf: Argument list too long
This failure happens when the output directory is about 6 levels deep.
It's due to the make generator of gyp that uses some crazy shell
processing for splitting up paths on /. The ninja generator does no such
thing.
While we're at it, remove redundant parenthesis in
HOST_NODEJS_CONFIGURE_CMDS.
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
One per line, and alphabetical (was already the case for host).
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
[yann.morin.1998@free.fr: all host dependencies first]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
