Sway: Remove extraneous properties from IPC command get_outputs

[jplatte]

for a setup with two monitors:

$ swaymsg -t get_outputs | grep focused
     "focused": false,
     "focused": false,

$ swaymsg -t get_version
sway version 1.0-alpha.5-364-g15dadaaa (Sep 18 2018, branch 'makepkg')

[RyanDwyer]

This is intended. focused is only true for the single focused node (not the ancestors), so it can never be true on an output.

The i3 docs (https://i3wm.org/docs/ipc.html) confirm this:

... will result in eventually reaching the one node with focused set to true

And looking at the i3 code, it's only set for one node (focused is a global variable): https://github.com/i3/i3/blob/ac88d06632b6235e51d89806b7889ba5bfb9eb78/src/ipc.c#L420

[jplatte]

Okay. Then that field being there at all is a bit misleading though. I just checked and in i3, the field is not part of the get_outputs IPC command output.

[RyanDwyer]

Confirmed. An example of what they return is:

{
    "name": "LVDS-1",
    "active": true,
    "primary": true,
    "rect": {
      "x": 0,
      "y": 0,
      "width": 1920,
      "height": 1080
    },
    "current_workspace": "3"
},

I'd support removing irrelevant elements from our get_outputs reply.

[jplatte]

Also, as an aside: My main goal was to query the name of the currently focused output. I managed to do that after reading up on jq, which I had seen used for filtering out information from i3's ipc previously. This is what I came up with, for anyone looking to do the same thing:

swaymsg -t get_tree | jq -r '.nodes[] | select(recurse(.nodes[], .floating_nodes[]) | .focused).name'

[her001]

May I ask what other properties besides "focused" are considered extraneous?

[ianyfan]

Probably anything that isn't in the i3 reply.

[progandy]

Not everything that isn't in i3 is extraneous. Some of the properties have to be included by sway to handle output configuration that is done by xrandr in Xorg.

[Moelf]

$ swaymsg -t get_outputs | grep focused
     "focused": true,
     "focused": false,

looks like now we have some useful outputs?

[ammgws]

@ianyfan It looks like this can be closed now.

[ianyfan]

I don't think so. Run swaymsg -t get_outputs -r in both sway and i3 and compare them.

[ammgws]

Ah OK I had conflated the issue of focused always returning false with the original issue of non-conformance with the i3 IPC.