Fix thread-safety issues around armok mode logic

[lethosor]

Followup to feedback in #4674.

I verified that dfhack.internal.setArmokTools() and dfhack.setMortalMode() can still be called from the lua interpreter without deadlocking. I haven't seen meaningful changes occur from calling them, though, so my testing may not be exhaustive.

[myk002]

quick way to test:

1. disable mortal mode
2. go to the main dwarfmode map and hit Ctrl-Shift-T to bring up gui/teleport (which should come up)
3. enable mortal mode
4. go back to the map and hit Ctrl-Shift-T again. nothing should happen

you can see more info about the change in processing by running this before doing the testing:

debugfilter set debug core keybinding

[myk002]

also, to test setArmokTools:

1. enable mortal mode
2. on main dwarfmode map, hit Ctrl-Shift-T and fail to bring up gui/teleport
3. exit back to title screen
4. edit hack/docs/docs/tools/gui/teleport.txt and take armok out of the tags
5. load up a world, Ctrl-Shift-T should now launch gui/teleport even with mortal mode enabled

[lethosor]

I was trying to filter out teleport in gui/launcher, with no luck. Didn't try loading a fort.

[myk002]

that shouldn't be affected by this change, and should continue to work.

However, I think I see the issue that you might have been running into -- the tag filters in gui/launcher are not reset when you change the mortal mode config setting in gui/control-panel. I'll see if I can make that experience a little more smooth.

vidcap-2024-06-06_15.51.52.mp4

[myk002]

fixed in #4680 and DFHack/scripts#1153

[lethosor]

That would explain what I was seeing, thanks.