[3540][UI][core] Fix interface not being updated in thinclient

[DjLegolas]

When changing the interface in both GTK and ConsoleUI, we call
directly to deluge.common.is_interface, which checks the interface on
the local machine and not on the daemon's machine.

The WebUI does not seem to have any validation on input.

closes: https://dev.deluge-torrent.org/ticket/3540

[cas--]

Ah yes I encountered this issue!

Will need to handle backward compatibility in the client for daemons that don't have that method exported

[DjLegolas]

depends on: #427

[cas--]

I have some thoughts on this change... 😄

[cas--]

Firstly this could be more readable with an intermediary variable and secondly the fallback should be True not False since we cannot check the validity of the interface. Also I don't think the version should specify patch number since this new core function in theory should bump the minor version.

Suggested change

	if (
	client.is_daemon_version_equal_or_greater('2.1.1')
	and client.core.is_valid_interface(listen_interface)
	or not listen_interface
	):
	listen_interface_valid=(
	client.core.is_valid_interface(listen_interface)
	if client.is_daemon_version_equal_or_greater('2.2')
	else True
	)
	if listen_interface_valid or not listen_interface:

Oh and I just noticed that the call to client.core method will return a deferred so is always True so this will need a callback to get the actual result. That would lead us to look at the alternative option of not using the client daemon version check and instead handle the deferred fail for that method call. It might not be as explicit as the version check however...

[cas--]

I have concerns about the naming of this method and the underlying problem we are solving.

In the first instance valid seems redundant but can understand clarification as a core method. Even if we were to use it I would suggest using it as a suffix rather than prefix is_interface_valid.

However I am leaning towards either core.has_interface or core.has_interface_name with perhaps more weight to the latter since common.is_interface seems to be my source of the contention. In addition the check if an IP is valid doesn't require a core method call and we don't actually check that the IP is assigned to one of the core host interfaces.

Perhaps core.has_interface will suffice as a core method with a future check that IP is assigned to an interface.

My final concern is that should we prevent the user from assigning an 'invalid' interface, e.g. interface is down briefly, and instead show a UI warning... That does increase the scope of this change so will need to revisit that