Race-condition causing duplicate actors to run in parallel?

[sandreae]

This occurred when running flathub beta version of Reflection. Not sure if it's related but there were two windows running at the time, although only one network instance should be shared between them.

2026-01-20T12:28:32.173073Z  WARN p2panda_net_next::actors::streams::eventually_consistent: sync manager failed with reason: ActorAlreadyRegistered(
    "net.sync.session.2+f93a56",
) actor_id=0.7365 topic=323331f6e8

Looking at the logs it appears that the session error caused the sync manager to fail, which is not desirable behaviour.

[adzialocha]

With the new version of p2panda-net that shouldn't be a problem (actors are not named and can not collide), but I'm still curious if this is maybe a pointer at a bug in Reflection (maybe we move the issue there?)