Daniel Rudolf

Thanks for the fast response :+1: I honestly don't see an issue with DBus' uint32 IDs: Since the ID is provided by the notification server, the ID is guaranteed to...

It took me quite some time to find the cause of my issue: It seems like that one can't close a notification using another DBus bus if the bus that...

It's definitely an upstream issue that was introduced to workaround misbehaving clients... :unamused: See https://gitlab.gnome.org/GNOME/gnome-shell/-/issues/8072

> Is it to interact with notifications that we sent by a previous run of the app? Or is it rather about having multiple processes running at the same time...

Windows doesn't have a similar event (I checked). Can't say anything about OS X because I neither understand their overly complex API, nor do I have an Apple device to...

> Do you think there is any value in distinguishing between expired and programatically closed notifications? We could. It's been five months, so I've forgotten a lot of my original...

Yes. I assume you ask because I forgot to remove `timeout` from the examples? I just fixed that, sorry. Since I was over it I also now fully unified the...

PR is up-to-date with `main` now (and thus all todos done) and ready to review + merge :+1: After this has been merged I'll look into getting #200 back on...

> I am not very familiar with immutabledict. Is there a possible Python stdlib solution instead? For example MappingProxyType? `MappingProxyType` has a rather poor API not even closely similar to...

Done. Since you haven't had other comments about #208 I've updated this PR in anticipation of #208 as well. Meaning: You must merge #208 first, then re-run CI (mypy fails...