ChanServ changed the topic of #nixos-systemd to: NixOS <3 systemd | https://jitsi.nixcon.net/systemd | Next meeting 08.12.2020 14:00 UTC (every two weeks)
<Mic92> which systemd version?
<hexa-> nixpkgs master
<hexa-> since a rebuild 3 days ago
pbb has quit [Quit: http://quassel-irc.org - Chat comfortably. Anywhere.]
pbb has joined #nixos-systemd
<{^_^}> Alexays/Waybar#994 (by rxguy, 6 weeks ago, closed): Crash on start - 'std::runtime_error' - Wrong size of RFKILL event
<V> unsure if related
<{^_^}> systemd/systemd#18679 (by keszybz, 2 weeks ago, merged): struct rfkill_event read/write size leniency
<{^_^}> systemd/systemd#18677 (by keszybz, 2 weeks ago, closed): systemd-rfkill: Couldn't write rfkill event structure, too short.
<V> ah, yes
<V> each time you read from /dev/rfkill, it produces a new rfkill object, which means that reading only the v1 fields is safe and will not cause you to read the start of the v2 fields on the next read. what systemd were doing here was definitely incorrect
<V> (source: have written rfkill bindings)
<hexa-> V++
<{^_^}> V's karma got increased to 13
<andi-> V: isn't that a user space breakage and should be reported upstream?
<andi-> or are readers supposed to ignore new trailing fields
<V> andi-: the latter
<V> There's the _V1 constant so readers know how many bytes they're dealing with
<V> Presumably there will be a _V2 constant, so you can go if (nread >= _V2) { v2 fields }
<V> if (< _V1) shouldn't happen