Latest Apple clang (v13) defines __cplusplus=199711 by default, but
protobuf requires at least 201103 (c++11)
Backported patch to fix c++ detection:
30fe936a88
Signed-off-by: Sergey V. Lobanov <sergey@lobanov.in>
tcsh requires host-tool (gethost) to generate tc.defs.c file.
To build this tool on macos, it is required to run ./configure
script wihout target vars/flags, but this package is not
splitted to host and target build.
This patch splits build to host and target builds to compile
host-tool(gethost) first, then compile tsch.
Due to lack of support of separate building host and target,
OpenWrt Makefile is modified to do it in the order below:
1. build host-tool (gethost)
2. build sh.err.h header for target
3. build tc.const.h header for target
4. copy gethost tool from host staging dir to target build dir
5. touch (make -t) gethost to protect it from recompiling
6. build tsch for target
It is required to do step 2 and 3 because 'make -t' generates
empty files but these files required to build tcsh
Patch 020-cross.patch has been removed due to not required
anymore (gethost build is separated and not affected by
target configure flags and vars)
Signed-off-by: Sergey V. Lobanov <sergey@lobanov.in>
./configure script detects macos specific system headers
(IOKit/serial/ioss.h and sys/ttycom.h) that are not available
during compile time. There is no way to pass ac_cv_* vars to
./configure script due to perl wrappers
To fix this issue, fake(empty) headers provided during compile
time if build host is MacOS
Signed-off-by: Sergey V. Lobanov <sergey@lobanov.in>
setup.py detects macos (darwin) and adds -flat_namespace flag. This
flag is not compatible with GCC that is used to compile target.
This patch patch disables darwin detection
Signed-off-by: Sergey V. Lobanov <sergey@lobanov.in>
Version 1.0.1 brought the following changes:
[v1.0.1] - 2021-11-26
Primarily fixes a few issues in the kernel module that were found
during a quick review from Russell King:
https://lore.kernel.org/netdev/YYPThd7aX+TBWslz@shell.armlinux.org.uk/https://lore.kernel.org/netdev/YYPU1gOvUPa00JWg@shell.armlinux.org.uk/
- mdio: The mvls subcommand now supports flushing the ATU
- mdio-netlink: Plug some glaring holes around integer overflows of
the PC.
- mdio-netlink: Release reference to MDIO bus after a transaction
completes.
So, update to the latest version and switch the kernel module back
to fetching tarballs like the userspace tool does.
Signed-off-by: Robert Marko <robimarko@gmail.com>
lpc21isp Makefile detects Darwin and defines __APPLE__ that is not
required for cross-compile build for OpenWrt
This patch sets OSTYPE="Linux" due to OpenWrt is always Linux
Signed-off-by: Sergey V. Lobanov <sergey@lobanov.in>
luaossl detects OS and changes compilation flags depends on OS.
If Darwin is detected then it adds GCC non-compatible flags.
OpenWrt is always Linux so build OS detection is disabled
Signed-off-by: Sergey V. Lobanov <sergey@lobanov.in>
hwinfo host-compile builds hwinfo entirely that is not really
required. It is required to build only some host tools to
generate ids and isdn related C-code
This patch changes host build to build only required host tools
Added patch to remove malloc.h using as deprecated, not required
and not supported on macos system (stdlib.h has required functions
on Linux and MacOS systems)
Signed-off-by: Sergey V. Lobanov <sergey@lobanov.in>
neon detects Darwin on target build and fails due to darwin-specific
build behaviour. OS detection is disable (as non required) via
ne_cv_os_uname=Linux as a part of CONFIGURE_VARS
Signed-off-by: Sergey V. Lobanov <sergey@lobanov.in>
Qemu version updated to 6.2.0, patch set refreshed for it.
Options --disable-jemalloc --disable-tcmalloc was replaced
by --enable-malloc=CHOICE, defaults to system.
Libudev search was moved from configure to meson.build, and now it's
not so easy to disable it. Even though --disable-mpath present.
Delete patches 0008-falloc and 0009-fs - they're already in 6.2.0.
Signed-off-by: Vladimir Ermakov <vooon341@gmail.com>
During update we use sudo to start an instance. It probably could be
workarounded, but given the MariaDB size, dependency on sudo shouldn't
be an issue.
Signed-off-by: Michal Hrusecky <michal.hrusecky@turris.com>
If you are migrating to MariaDB package, you might have old datadir
stored somewhere using different setup with different users. If you
trust us enough to enable autoupgrade, you probably trust us enough to
chown your datadir as well. This can prevent some potential issues.
Signed-off-by: Michal Hrusecky <michal.hrusecky@turris.com>
The runtime testing always ran on master branch aka snapshots since the
branch wasn't passed over to the container execution!
Signed-off-by: Paul Spooren <mail@aparcar.org>
* Updating i2pd package to 2.40.0
* Rewrite Makefile
* Remove usage of PKG_INSTALL (package's make install)
* Rewrite init.rc configuration and script
* Remove '--service' option from init, which only sets datadir to /var/lib/i2pd
* Use '--datadir' option in init, otherwise datadir changing via uci is not works
* Update patch for i2pd.conf
Signed-off-by: R4SAS I2P <r4sas@i2pmail.org>
When ModemManager is started on boot we may end up with hotplug events
reported directly to the daemon, plus some others already cached in
the cache file before the daemon was started.
If the cached events correspond to the same device that is still
notifying ports directly, we may end up with a modem object created
before the cached events have been emitted, so the modem may not
handle all control/data ports it should.
E.g.:
- modem detected
- hotplug event for wwan0 port, cached as MM not running
- hotplug event for cdc-wdm0 port, cached as MM not running
- hotplug event for ttyUSB0, cached as MM not running
- MM starts
- hotplug event for ttyUSB1, directly processed as MM is running
- hotplug event for ttyUSB2, directly processed as MM is running
- modem object created with ttyUSB1 and ttyUSB2
- 2s after MM starts, cached events for wwan0, cdc-wdm0 and ttyUSB0
happen, but are ignored because the modem object has already been
created
MM expects that ports of the same device are reported with less than
1500ms in between ports. In other words, if ports are reported more
than 1500ms after the last reported port, they may get ignored.
If we remove the 2s timeout, the report of the cached events will
happen as soon as MM starts, which makes it much more likely to happen
in the timeslot that MM expects for ports of the same device reported.
The logic is still not perfect, and we may also need to increase that
1500ms timeout inside MM, but removing the 2s timeout right away here
makes sense.
This 2s timeout was introduced along with the new wrapper launcher for
the daemon, it didn't exist before.
Signed-off-by: Aleksander Morgado <aleksander@aleksander.es>
* replace 020-fix-ss-path.patch with an upstream patch 020-remove-which-dep.patch
* remove 030-remove-pynotify2-dep.patch and do not install aa-notify
* simplify 060-openwrt-dnsmasq-profile.patch
* add two upstream patches for Python 3.10 compatibility
* add an upstream patch to move from deprecated distutils which broke
the build with Python 3.10+
Signed-off-by: Michal Vasilek <michal.vasilek@nic.cz>
This backports a patch from upstream gpgme to fix compilation with glibc 2.34.
It fixes the following build problem:
posix-io.c: In function '_gpgme_io_spawn':
posix-io.c:577:23: error: void value not ignored as it ought to be
577 | while ((i = closefrom (fd)) && errno == EINTR)
| ^
make[5]: *** [Makefile:947: posix-io.lo] Error 1
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Sergey V. Lobanov
Lucian Cristian
Josef Schlehofer
Robert Marko
Espen Jürgensen
Nick Hainke
Kerma Gérald
Florian Eckert
Rosen Penev
Sebastian Kemper
Vladimir Ermakov
Othmar Truniger
Michael Heimpold
Jeffery To
Glen Huang
Claudio Marelli
Michal Hrusecky
Paul Spooren
R4SAS I2P
Aleksander Morgado
Michal Vasilek
Hauke Mehrtens
James White