| Age | Commit message (Collapse) | Author |
|---|
|
DISTCHECK_CONFIGURE_FLAGS is an automake variable and is not properly
detected when it is inside configure.ac
|
|
Signed-off-by: Richard Yao <ryao@gentoo.org>
|
|
Commit ff2c503df091e6e4e9ab48cdb6df6ec8b7b525d0 introduced accept4() into
udev, which broke compatibility with kernels older than Linux 2.6.32 (or
Linux 2.6.36 on ARM). The purpose of accept4() is to permit
O_NONBLOCK and O_CLOEXEC to be specified at the accept() call site
while previously, they had to be set using fcntl() because Linux does
not inherit them.
Since accept4() increases the minimum kernel version, we add a fallback
path for situations in which it is unavailable.
Reported-by: Stephen Klimaszewski <steev@gentoo.org>
Signed-off-by: Richard Yao <ryao@gentoo.org>
|
|
Module loading support was unconditionally disabled in the initial
autotools tests. I had missed this in my testing because I still had the
80-drivers.rules file from udev 171 on my system, which masked this
issue by invoking modprobe explicitly. This patch corrects the issue.
Signed-off-by: Richard Yao <ryao@gentoo.org>
|
|
The use of the eudev version in pkgconfig files caused build failures,
so we switch to the udev version for compatibility purposes.
Resolves gentoo/eudev#22
Signed-off-by: Richard Yao <ryao@gentoo.org>
|
|
|
|
|
|
This commit makes sure that src/gudev/gudevprivate.h is included
in the distribution, fixing a failure in building gudev with
`make distcheck`. There's also some code cleanup and migration
of all configurable paths to the main configure.ac file.
Signed-off-by: Anthony G. Basile <blueness@gentoo.org>
|
|
This commit the related issues of building gudev with/out
introspection. It draws on suggestions from nvinson in
https://github.com/gentoo/eudev/pull/20
Signed-off-by: Anthony G. Basile <blueness@gentoo.org>
|
|
Signed-off-by: Anthony G. Basile <blueness@gentoo.org>
|
|
We move the test-udev.c and test-libudev.c from src/test to test.
This corrects the a problem with hard coded relative paths finding
the test/test/sys directory created by test/udev-test.pl.
This commit draws heavily from nvinson patch in
https://github.com/gentoo/eudev/pull/20
Signed-off-by: Anthony G. Basile <blueness@gentoo.org>
|
|
This commit is a continuation of the previous one in which all the configured
paths obtained in configure.ac are propagated to the Makefile.am and .c files
via AM_CPPFLAGS of the form -DUDEV_CONF_FILE=\"$(udevconffile)\". This should
address the issue in
https://github.com/gentoo/eudev/issues/17
Signed-off-by: Anthony G. Basile <blueness@gentoo.org>
|
|
The configuration of the installation paths for various components
was scattered between the main configure.ac file and the various
Makefile.am's. These components are: udev config file, hwdb, keymaps
and force-release keymaps and the rules. This commit consolidates
them all into one point in configure.ac and anticipates the inclusion
of new AM_CPPFLAGS of the form -DHWDB_BIN=@udevhwdbdir@ as upstream
has done, so it is easier to address issues like:
https://github.com/gentoo/eudev/issues/17
Signed-off-by: Anthony G. Basile <blueness@gentoo.org>
|
|
The original checks for python were to build some man pages and
for systemd-python. These checks are no longer relevant. There
still is a dependency on python for tests but this is checked by
test/rules-test.sh which tests for python or skips. Also, there
is a dependency on perl for both tests and hwdb. These checks can
be reintroduced at a future date when re revisit both.
Signed-off-by: Anthony G. Basile <blueness@gentoo.org>
|
|
The GTK_DOC_CHECK macro includes AC_ARG_ENABLE/AS_HELP_STRING etc
for --enable-gtk-doc and there is no need for a second check.
One caveat should be noted: GTK_DOC_CHECK must begin the line for
gtkdocize to recognize its presense in configure.ac.
Signed-off-by: Anthony G. Basile <blueness@gentoo.org>
|
|
Inspired by e30431623a7d871da123cc37055ac49abf2c20ea from systemd.
Signed-off-by: Richard Yao <ryao@gentoo.org>
|
|
We support module loading through modprobe when libkmod is neither
available nor wanted.
Signed-off-by: Richard Yao <ryao@gentoo.org>
|
|
This opens the door to using disabling the use of kmod from autotools.
Signed-off-by: Richard Yao <ryao@gentoo.org>
|
|
The keys/keymap test preformed by src/keymap/check-keymaps.sh must
find both keys.txt and the keymap directory. When building out of
the source tree, eg when doing `make distcheck`, these are located
at ${top_builddir}/src/keymap/keys.txt and ${top_srcdir}/keymap/,
respectively. This patch fixes the build so that these are now
correctly found.
Signed-off-by: Anthony G. Basile <blueness@gentoo.org>
|
|
Signed-off-by: Anthony G. Basile <blueness@gentoo.org>
|
|
${rootlibdir}/udev instead of ${root}/lib/udev
|
|
This commit reintroduces code to
1) build src/keymap
2) test keymap/Makefile.am that it has all the key maps listed
3) test that all the key names in keymap/* are in <linux/input.h>
4) do a syntax check on rules/*.rules which now includes
95-keymap.rules and 95-keyboard-force-release.rules
For #4, the regex expressions in rule-syntax-check.py had to be
updated. They do not allow trailing comments via # comment, and
I did not include that. A rule in 95-keymap.rules had to have its
comment moved from the end to another line.
Signed-off-by: Anthony G. Basile <blueness@gentoo.org>
|
|
The build failure that prompted its removal has been resolved.
Signed-off-by: Richard Yao <ryao@gentoo.org>
|
|
not available
|
|
|
|
|
|
|
|
|
|
|
|
Changes to rules were introduced by
7c2dee4a4d7f1b264031daaee786a8fe429884e1 while builtin-blkid support was
introduced in other commits. The removal of systemd resulted in this
code causing linker errors. This code adds complexity with no clear
benefit, so we remove it.
Signed-off-by: Richard Yao <ryao@gentoo.org>
|
|
Changes to rules were introduced by
06316d9f1a91b4d3efdb7402e72498cb3deb1806 while kmod support was
introduced in other commits. A ton of commits were made involving kmod
and it is quite clear that it is broken, so we remove it.
This changes our rules to depend on modprobe. As long as the modprobe
binary is in /, and not /usr, udev module loading should function
properly.
Signed-off-by: Richard Yao <ryao@gentoo.org>
|
|
|
|
|
|
|
|
|
|
The structure of the source tree is basically correct and this is
about as far as we can go without hacking at the C code.
|
|
The original Makefile.am was drawn to the top level. This commit
breaks it out into the various directories with SUBDIRS connecting
them. This makes each directory easier to maintain.
|
|
This is the first pass attempting to keep as much of the build system
as is necessary for only udev from the fork. Emphasis was given to
configure.ac. Gutting had to be done to Makefile.am but this needs
work to be broken out into SUBDIR Makefile.am which each address those
pieces.
|
|
glibc moved clock_* functions from librt to the core libc. As a result,
clock_gettime is no more a suitable symbol to use when finding librt.
Look for mq_open instead.
Reference:
http://www.sourceware.org/git/gitweb.cgi?p=glibc.git&h=6e6249d0b461b952d0f544792372663feb6d792a
Fixes a FTBFS in Fedora Rawhide.
|
|
This is the usual setup, where pythonX.Y and pythonX.Y-config go
together. Using python-config with python3 will only lead to
confusion.
--libs is changed to --ldflags, since the latter also includes other
required flags like -L<dir>.
The tests for HAVE_PYTHON and HAVE_PYTHON_DEVEL are separated. It is
possible to have python development libraries installed without the
binary (or to want to build python modules without using python during
the build).
A line is added to the output, to show what flags will be used for
python.
|
|
|
|
|
|
|
|
|
|
than bus
This should make session termination more reliable, as D-Bus doesn't
have to be around anymore for this to succeed.
|
|
This also enables time-based rotation (but not vacuuming) after 1month,
so that not more one month of journal is lost at a time per vacuuming.
|
|
Being able to be explicit about Python support (in addition to the
default of auto-detecting it) and acting upon the result, specifying
it as an option gains us more control about both dependencies and
the resulting build.
Furthermore, relying purely on auto-detection can lead to problems for
source-based distros. E. g. systemd being built before *both* 32-bit &
64-bit ABIs are installed will lead to build failures as systemd's
build system will pick up either 32-/64-bit Python, conclude both are
available and fail if that's not the case.
|
|
AC_CHECK_FUNCS may be successful, even though name_to_handle_at and
'struct file_handle' are not available.
|
|
|
|
|
