i3 does not correctly set *DESKTOP* atoms

[qvint]

i3 does not correctly set _NET_DESKTOP_NAMES, _NET_NUMBER_OF_DESKTOPS and _NET_WM_DESKTOP atoms if the last window of the inactive workspace is closed. Here are the steps to reproduce the error:

1. Go to the empty workspace #1 and open a terminal on it.
2. Do the same thing with the empty workspace #3.
3. Then go to the empty workspace #2 and open a terminal again. Type in it: sleep 5; exit
4. Switch to the workspace #1 and wait until the workspace #2 disappears from the status line.
5. And for now, atom values are:

   $ xprop -root
   <...>
   _NET_DESKTOP_NAMES(UTF8_STRING) = "1", "2", "3"
   _NET_NUMBER_OF_DESKTOPS(CARDINAL) = 3
   <...>
   

   while the status line shows only #1 and #3. Expected values are:

   _NET_DESKTOP_NAMES(UTF8_STRING) = "1", "3"
   _NET_NUMBER_OF_DESKTOPS(CARDINAL) = 2
   

6. Now switch to the workspace #3 and do xprop | grep _NET_WM_DESKTOP with the terminal window #3 itself. It will log 2, while we are in the second existent workspace (value 1 is expected in this case).

Output of i3 --moreversion 2>&- || i3 --version:

Binary i3 version:  4.14.1 (2017-09-24) © 2009 Michael Stapelberg and contributors
Running i3 version: 4.14.1 (2017-09-24) (pid 11740)bort…)
Loaded i3 config: /home/user/.config/i3/config (<...>)

The i3 binary you just called: /usr/bin/i3
The i3 binary you are running: i3

[i3bot]

I don’t see a link to logs.i3wm.org. Did you follow https://i3wm.org/docs/debugging.html? (In case you actually provided a link to a logfile, please ignore me.)

[Airblader]

Thanks for reporting! The fix should be pretty simple, we just have to call the corresponding ewmh_update_* functions in some place we missed so far, apparently (for anyone reading who wants to take a shot here – it's a good bug to get started with!)

[Riyyi]

@op8867555's commit fixed the issue partly, but when I have a sticky floating window opened, the issue persists.
Im using i3 version 4.16.1-169-g60a6d149, for reference.

[Airblader]

Can you please open a new issue with reproduction steps for this case?