Linus Lewandowski

It's hard to say if it's a good idea or not. In most cases, the DBus Introspection XML is a part of a pre-existing standard (like MPRIS), and the programmer...

OK, so now back to your pull request: 1. @signalled is a bad idea, as not every call to the setter will change the value (and therefore should emit the...

@signalled - right, I've forgotten about this annotation. So this is necessary, but I still don't think that it should send the signal automatically. And probably should be used on...

Objects implementing multiple interfaces are a bit of a complicated case. There are cases like MPRIS2, where all methods are grouped into 4 interfaces, and 3 of them are extensions;...

(Of course for methods on the default interface, @interface is not required.)

About typed_signal - objects can send signals, and currently that's being defined with ``` python from pydbus.generic import signal class X: SignalName = signal() ``` To autogenerate XML for them,...

That's why I haven't suggested typed_signal as a decorator :D I didn't write "@" before its name :D Probably it should be a class that inherits pydbus.generic.signal, and has a...

(GitHub collapsed one of two TypedSignal-related comments as outdated, you might need to click "Show outdated")

I haven't written any code for fd passing yet, and looks like GDBus does not support them transparently, but requires using g_dbus_message_get_unix_fd_list - like in https://git.gnome.org/browse/glib/tree/gio/tests/gdbus-example-unix-fd-client.c#n27 I guess it's necessary...

- Unpack 's' to unicodes on Python 2? No idea. I guess it's simpler to have a single codebase when assuming that unicodes don't exist.