Itay Shoshani [Sat, 9 Nov 2024 22:30:40 +0000 (00:30 +0200)]
openthread-br: bump source version to 2024-11-20 and add missing dependencies
Bumping ot-br-posix that openthread-br is based on to latest code.
Due to some major changes required some more compilation flags, and in order to avoid the ot-br agent had to add some dependencies as well
Dependecies:
1. kmod-tun: otherwise the otbr fails to load and crashed immidiedtly
Compilation flags:
1. OTBR_NAT64, OT_NAT64_BORDER_ROUTING set to OFF: a temporary workaround to a compilation problem with openwrt, could be reverted once the issue here is fixed: openthread/ot-br-posix#2606
2. OT_TARGET_OPENWRT: wasn't assumed from OTBR_OPENWRT in some targets
Dirk Brenken [Sun, 24 Nov 2024 14:44:20 +0000 (15:44 +0100)]
banIP: release 1.0.1-1
* optimized procd settings for better performance
* made the log monitor working again (even on master with apk migration issues)
* reworked the fetch autodetection function (still broken in master due to apk migration)
When the syslog-ng installed it records all messages to /var/log/messages.
This makes the default OpenWrt's ubox /sbin/logread useless and return nothing.
The logread script was added to the syslog-ng as a shim.
It's based on `tail /var/log/messages` and repeats basic options that the ubox logread have.
The script is installed instead of the /sbin/logread and this is causes an error.
Instead install it to /usr/libexec/logread.sh and specify as an alternative /sbin/logread.
The ubox logread also needs to be made as alternative in /usr/libexec/logread-ubox.
So we need to give it a separate name.
The most simple and clear would be logread-shell or just logread.sh.
Suggested-by: Andreas Gnau <redacted> Signed-off-by: Sergey Ponomarev <redacted>
Dirk Brenken [Sat, 23 Nov 2024 21:15:00 +0000 (22:15 +0100)]
adblock: release 4.2.3-1
* optimized procd settings for better performance
* reworked autodetection functions (still broken in master due to apk migration)
* made the tld function optional, set 'adb_tld' accordingly (enabled by default)
* reworked count function
* various code improvements
This release marks the transition of Node.js 22.x into Long Term Support (LTS) with the codename 'Jod'. The 22.x release line now moves into "Active LTS" and will remain so until October 2025. After that time, it will move into "Maintenance" until end of life in April 2027.
Daniel Golle [Fri, 22 Nov 2024 17:56:57 +0000 (17:56 +0000)]
gst1-libav: remove obsolete configuration options
Since the build was switched to meson a couple of years ago gst1-libav
simply uses libffmpeg and the codec configuration previous passed via
CONFIGURE_ARGS is completely ignored. This is misleading as all those
options are still present in menuconfig but don't have any effect.
Remove all dead code from build recipe.
Fixes: de426d431 ("gst1-libav: update to 1.17.1") Signed-off-by: Daniel Golle <redacted>
John Audia [Fri, 22 Nov 2024 15:59:59 +0000 (10:59 -0500)]
ncdu: update to 1.21
Release history:
* Perform tilde expansion on paths in the config file (from 2.7)
* Fix JSON import of escaped UTF-16 surrogate pairs (from 2.7)
* Fix displaying and exporting zero values when extended info is not available (from 2.6)
* Fix JSON export and import of the “other filesystem” flag (from 2.5)
Martin Pecka [Mon, 8 Jan 2024 00:24:07 +0000 (01:24 +0100)]
strongswan: Add support for EAP-MSCHAPv2 authentication
Support for EAP-MSCHAPv2 authentication scheme is added.
Different from the previously supported schemes, this one is
usually asymmetric in the way that server auth method (pubkey) is
different from the client auth method (eap-mschapv2).
The code handles this asymmetry automatically.
A new UCI config section mschapv2_secrets is added where the user
can specify the EAP identities and their passwords that are
accepted by the server. AFAIK, there is no way to select which
EAP IDs should be accepted by which remote, except setting
`eap_id` to something different than `%any`. But `eap_id`
does not support template matching, so either only a single
identity or all can be configured for one remote. This is why
the EAP identities are not subsections of remotes, but are
a standalone section.
Signed-off-by: Martin Pecka <redacted> Signed-off-by: Martin Pecka <redacted>
Martin Pecka [Sun, 7 Jan 2024 22:23:57 +0000 (23:23 +0100)]
strongswan: Fix pools to be only generated once
Before this commit, if a user configures multiple remotes in UCI,
each remote generates one output section of pools.
This doesn't hurt because swanctl just merges all of them,
but it is apparently not needed to have N copies of the same.
This commit changes the behavior to only create one pools
section at the end of the generated swanctl config.
Signed-off-by: Martin Pecka <redacted> Signed-off-by: Martin Pecka <redacted>
Daniel Golle [Sat, 16 Nov 2024 17:36:04 +0000 (17:36 +0000)]
mptcpd: update to version 0.13
Changes since v0.12: 7fb6416 tests: fix segfault with no /etc/protocols (#286) 0dddc42 README: Update branch references in badge URLs 5eea3f4 github: Update branch names for actions b102523 scripts: get debug info for bug reports 334bb9c Make listening socket creation optional. (#297) a4e7b65 {etc,man}: improve check_route address notification flag documentation (#301) ffa276f Support ELL 0.68 l_netlink_message API. (#303) 269f10a gh: daily validation of the ELL compatibility (#308) 824ff8c Support listener events (#304) 5d9f33a Some clean-ups (#305) 05a08bf lib: switch to v4.0.1: only added items (#306) 3702487 scripts: debug: also print system info (#307) 5d8db15 ell: only include 'ell/ell.h' header (#309) 772afe9 AUTHORS: Update e-mail addresses. 0c36dba mptcpd 0.13
Drop backported patches which are part of the release.
The following error occurs when CONFIG_USE_APK is set.
```
ln /mnt/gl-mt6000-main/openwrt/build_dir/target-aarch64_cortex-a53_musl/git-2.46.2/.pkgdir/git/usr/lib/git-core/git /mnt/gl-mt6000-main/openwrt/build_dir/target-aarch64_cortex-a53_musl/git-2.46.2/.pkgdir/git/usr/bin/git
ln /mnt/gl-mt6000-main/openwrt/build_dir/target-aarch64_cortex-a53_musl/git-2.46.2/.pkgdir/git/usr/lib/git-core/git-shell /mnt/gl-mt6000-main/openwrt/build_dir/target-aarch64_cortex-a53_musl/git-2.46.2/.pkgdir/git/usr/bin/git-shell
touch /mnt/gl-mt6000-main/openwrt/build_dir/target-aarch64_cortex-a53_musl/git-2.46.2/.pkgdir/git.installed
mkdir -p /mnt/gl-mt6000-main/openwrt/staging_dir/target-aarch64_cortex-a53_musl/root-mediatek/stamp
SHELL= flock /mnt/gl-mt6000-main/openwrt/tmp/.root-copy.flock -c 'cp -fpR /mnt/gl-mt6000-main/openwrt/build_dir/target-aarch64_cortex-a53_musl/git-2.46.2/.pkgdir/git/. /mnt/gl-mt6000-main/openwrt/staging_dir/target-aarch64_cortex-a53_musl/root-mediatek/'
cp: '/mnt/gl-mt6000-main/openwrt/build_dir/target-aarch64_cortex-a53_musl/git-2.46.2/.pkgdir/git/./usr/bin/git' and '/mnt/gl-mt6000-main/openwrt/staging_dir/target-aarch64_cortex-a53_musl/root-mediatek/./usr/bin/git' are the same file
cp: '/mnt/gl-mt6000-main/openwrt/build_dir/target-aarch64_cortex-a53_musl/git-2.46.2/.pkgdir/git/./usr/bin/git-shell' and '/mnt/gl-mt6000-main/openwrt/staging_dir/target-aarch64_cortex-a53_musl/root-mediatek/./usr/bin/git-shell' are the same file
make[2]: *** [Makefile:167: /mnt/gl-mt6000-main/openwrt/staging_dir/target-aarch64_cortex-a53_musl/root-mediatek/stamp/.git_installed] Error 1
```
Use $(LN) instead of ln and use relative paths.
Daniel Golle [Thu, 14 Nov 2024 03:45:57 +0000 (03:45 +0000)]
xdg-dbus-proxy: add package
xdg-dbus-proxy is a filtering proxy for D-Bus connections. It is used to
allow partial access to D-Bus form sandboxed processes, eg. when using
bubblewrap.
Florian Eckert [Wed, 6 Nov 2024 13:28:57 +0000 (14:28 +0100)]
modemmanager: check during SIM unlocking whether initialization has been completed
After the SIM has been successfully unlocked, it is initialized. This can
take longer on some modems, so we must wait until the modem is ready to
execute the next commands.
Otherwise the modem cannot be enabled and aborts with the following
error message:
error: couldn't enable the modem: 'GDBus.Error:org.freedesktop.ModemManager1.Error.Core.WrongState: modem in initializing state'
Florian Eckert [Tue, 5 Nov 2024 07:37:02 +0000 (08:37 +0100)]
modemmanager: swap plmn and technology set command in the proto
In the current implementation of the modemmanager 'proto', if a 'plmn' and a
'technology' is configured, the 'plmn' is set first and then second the
'technology' on a 'proto' setup.
However, this is problematic if a 'technology' has already been set in an
earlier run. It is possible, that this previously set 'technology' is not
available at the current location, as the modem remembers the setting.
To fix this, first set the technology and then the plmn.
modemmanager: add log file if syslog is set to DEBUG
If the ModemManager is started with debug, all outputs are written to the
system log. To simplify debugging, a logging file is now created under
'/var/log/mm.log' in this case. This simplifies error analysis.
Marius Dinu [Sun, 17 Apr 2022 10:23:51 +0000 (13:23 +0300)]
btrfs-progs: add "btrfs device scan" before overlay/extroot
If users choose to build OpenWrt with btrfs included
and want to use a btrfs RAID for overlay/extroot,
then devices need to be scanned *before* mounting overlay/extroot.
If not, btrfs won't find all RAID drives and fail to mount.
This commit:
- creates a duplicate (symlink) of the btrfs scan script so that it runs
both before and after overlay/extroot mount,
- changes the scan command to not depend on blkid (-d),
- outputs the scan results to kernel log to show what's being detected,
as the system logger is not yet running.
Daniel Golle [Tue, 5 Nov 2024 03:20:57 +0000 (03:20 +0000)]
gst1-plugins-bad: package play and player libraries
Create packages for the libgstplay and libgstplayer.
Introduce InstallDev to allow using and detecting libraries which are
part of the 'bad' set by other packages.