10:19 <krobelus> I'm trying to understand whether terminals (and service managers) should be sending SIGKILL. I suppose it is what most users expect. But I feel like I'd personally prefer "no implicit SIGKILL" so I can find and fix misbehaving applications. I suppose 965d8a3a (terminal: don't get stuck waiting for misbehaving slaves to terminate, 2019-11-02) might be motivated by the server/client scenario, didn't

10:19 <krobelus> test that yet.

10:20 <krobelus> I think SIGTERM/SIGKILL will also kill disowned processes, which seems weird.. probably irrelevant in practice since it shouldn't normally happen

10:36 <bookworm> krobelus: sigkill doesn't kill process that are disowned, you decouple the sessions

10:39 <bookworm> now, if the terminal goes away, that might mean that the process in question immediately gets a fatal signal when it tries to read to it, or more likely write to it, but that's another question