packages, branch master

shunt: ignore unspecified netifd nexthop

2026-10-06T19:49:15Z

netifd always emits a nexthop for its routes, the unspecified address for a gateway-less one, e.g. wg0 with route_allowed_ips. Taken as gateway, IPv4 treats 0.0.0.0 as none, but the kernel rejects "via ::" with EINVAL, so the IPv6 policy route was never installed and marked traffic fell through to main, where the interface metrics decided (IPv6 leak via WAN). Co-authored-by: Claude Signed-off-by: Dirk Brenken

mpd: match the version in the binary for the CI tests

2026-10-06T19:26:57Z

mpd 0.24 aborts in main() when /proc/self/status reports more than one thread, before it even parses the command line (src/Main.cxx). qemu-user always reports several, so "mpd --version" cannot run on the emulated CI targets and the generic version check fails there. Add test-version.sh matching the version string compiled into the binary, and drop the two unusable "mpd --version" calls from test.sh, whose failures it silently ignored anyway. Signed-off-by: Alexandru Ardelean

mpd: update to 0.24.15

2026-10-06T19:26:57Z

Refresh PKG_HASH and point PKG_SOURCE_URL at the 0.24 release directory. 0.24 removed the SoundCloud plugin and replaced yajl with nlohmann/json, so -Dsoundcloud= and -Dyajl= are no longer meson options and yajl leaves DEPENDS. Pin -Dnlohmann_json=disabled rather than leave the new option at auto: its only users here, qobuz and snapcast, are both disabled. 0.24 also dropped the Boost dependency: meson no longer looks for it and the one remaining include is a header-only lround fallback for a libstdc++ built without C99 math. Nothing links a boost library, so drop +boost too. Drop 010-fmt.patch. It was added in 0.23.16 to fix compilation against fmt 11.1, and 0.24 rewrote Response.hxx, dropping the FMT_VERSION ladder the patch edited. src/client/Response.cxx compiles against the feed's fmt 12.1.0 unpatched. Signed-off-by: Alexandru Ardelean

mpd: fix the pulseaudio rpath-link guard

2026-10-06T19:26:57Z

-Dpulse= is derived with $(if $(CONFIG_PACKAGE_audio-support)), true for =m as well, while the rpath-link that lets the linker find libpulsecommon was guarded by ifeq(...,y). Selecting audio-support as a module built pulse support without the flag, leaving every pa_* symbol undefined. Test the same way the option does. Signed-off-by: Alexandru Ardelean

travelmate: minor cp fixes

2026-10-06T19:11:36Z

- handle captive portals failing after the redirect, fixes #30709 - Strip the port from the effective url host and, for a consistent foreign host comparison, from the probe host as well Signed-off-by: Dirk Brenken

ripe-atlas: fit get_key to the installed variant

2026-10-06T18:18:20Z

With only ripe-atlas-common installed, get_key told users to start the probe to create the key, but the init creates it only once ripe-atlas-probe or ripe-atlas-anchor is installed. For an anchor, it pointed to the registration form of software probes. Say which package is missing, and point an anchor to the anchor application. Fixes: 21b1fefb4172 ("ripe-atlas: add get_key and probeid to init script") Signed-off-by: Josef Schlehofer

ripe-atlas: provide atlas-probe from ripe-atlas-probe

2026-10-06T18:18:20Z

ripe-atlas-common provides atlas-probe at version 5130, newer than the atlas-probe 2.6.3 it is meant to replace, so apk replaces atlas-probe with ripe-atlas-common whenever only that package is installed, for example when LuCI installs ripe-atlas-common or ripe-atlas-anchor. atlas-sw-probe then keeps running without its measurement programs, and its key is not taken over. Provide atlas-probe from ripe-atlas-probe instead, which also replaces atlas-sw-probe and takes over its key. Fixes: b77264c9226c ("ripe-atlas: add new package") Signed-off-by: Josef Schlehofer

realtek-poe: increase version to v1.3.2

2026-10-06T18:17:21Z

This picks up a memory leak fix, some ubus fixes, and a couple of Zyxel device quirks. Signed-off-by: Martin Kennedy

strongswan: stage the migration in the LuCI session through ubus

2026-10-06T07:53:01Z

When LuCI passes a session id, migrate-ipsec-to-swanctl moves its uci delta into /tmp/run/rpcd/uci-, rpcd's private save directory of that session. rpcd stops naming this directory after the session id, as the id is a bearer credential that leaked through directory listings, so the migrated configuration would no longer show up as pending changes in LuCI. Stage the migration through rpcd's uci ubus API instead: revert the session's ipsec changes, delete all sections and add the sections of the migrated configuration. Then revert the changes of the uci command line tool, which the old code moved away. This works with old and new rpcd versions. If staging fails, both are reverted and the script exits with an error. Fixes: d5f33e6410cf ("strongswan: support optional UCI session staging in migrate-ipsec-to-swanctl") Assisted-by: Claude:claude-opus-5-5 Signed-off-by: Hauke Mehrtens

tio: use Lua 5.4

2026-10-06T07:40:22Z

tio embeds Lua for its scripting support, which is not optional. It looks Lua up through pkg-config as lua-5.4, lua-5.3, lua-5.2, lua-5.1 and lua, in this order, with no option to select the version. The package depends on lua (5.1), which tio has found as lua.pc so far. Since lua5.4 stages lua-5.4.pc, tio finds Lua 5.4 first and links liblua5.4 without depending on it. Depend on liblua5.4, the version tio looks up first, instead of lua, which is the documented preferred version anyway. Only the library is needed; tio does not run the lua interpreter. Signed-off-by: Mirko Vogt