I haven't noticed anything weird or different.

Elaborated.

I guess it would be nice to split the arguments and default them to INVALID_ENTITY and INVALID_PLAYER.

SE.

There was a reason, IIRC @Itms discussed that with @Krinkle on IRC (quite) some time ago.

It keeps setting undefined players to their defaults.
When no victory condition is defined on game setup opening, it is also not possible to tick the other two conquests, until one changes the victory conditions.
It used to display "Endless game" when not victory conditions were ticket.

Rebased.

Everything is an object.

Ah, I get this now :)

Added a consequence.

So that's basically the attackerData-diff?

Rebased, comment.

Nice POC, but unwanted, also not needed ;)

I can't reveal the map when in the game?
Conquest is enabled even if it was unticked.
Cheats disabled in-game while enabled according to the settings.
Relic settings are not used/remembered in the game.
World population idem.

Some things I found:

• It is now possible to select a biome for a random random map.
• Selecting the Conquest VC should disable Conquest Structures/Units.
• I can't start the match:

ERROR: Errors executing script event "Press"
ERROR: JavaScript error: gamesettings/attributes/PlayerAI.js line 23
attribs.settings.PlayerData[i] is undefined
  LegacySerialize@gamesettings/attributes/PlayerAI.js:23:5
  LegacySerialize@gamesettings/GameSettings.js:32:16
  saveFile@gui/gamesetup/Controls/GameSettingsFile.js:48:51
  onLaunchGame@gui/gamesetup/Controls/GameSettingsControl.js:125:26
  launchGame@gui/gamesetup/Controls/StartGameControl.js:35:4
  onPress@gui/gamesetup/Pages/GameSetupPage/Panels/Buttons/StartGameButton.js:60:46

Thanks :)

Rebased.

Rebased.

Rebased.

In rP24979#49386, @wraitii wrote:

Hm, I would say not right now, but we probably want to consider starting the tutorial campaign instead.

Yeah, that sounds good.

Other things to keep in mind:

• This breaks victory conditions?
• Diplomacy panel still shows the removed player.

This doesn't seem to always work?
At least not on Acropolis bay ^^

Rebased.

Rebased.

Fix trigger maps.

Those get replaced by the global entity renamed message in GarrisonHolder.

Update some comments.

@Nescio I created a ticket for it: #6083.

Sounds like a globalscript, the cheats (and their actions) in JSON/JS files in data. Nice, doable, but out of scope for this patch, IMHO.

Too large of an endeavour right now.

More elaborate.

Okay, so the FSM already enters the next state before finishing processing the message. That means that this.order needs to be set when the order is accepted, and it cannot be set when we know the order has been accepted, which was my plan.

This patch fixes the problem and I will commit it. The more elaborate and correct solution should be done, but I regard that more as a feature.

(Assuming the idea is to give more information than: "Oh no, something broke in your 161-file commit.".)

It still errors, but now gives also a more informative error :)

Smells like D3284?

That is exactly what the code does now :)

I guess.

Ah, yeah, I experienced that one when testing this ;)

Fix @param on Starting to collect.

Comments.

Yeah, you can purge them. It's kinda guessing which map has the walls and which map doesn't.

Correct!
I guess I only tested walls on random maps.

Add ToDo.

Also, if a domestic animal will be hit by something that cannot slaughter (e.g. splash or a champion), it will flee.

We don't need the renamed parameter in the GarrisonedUnitsChanged-message anymore.

Delete holder after renaming.

Rebased.

In D3607#157232, @wraitii wrote:

It's probably better than SVN, but it's not optimal :p

Please explain?

Fixed tests and renaming turrets.

Rebased.

Rebased.

Since D3233 now kills animals in one go when slaughtered, we can fix this.

Rebased.

Need to fix turrets switching points due to UnitAI garrisoning on its own.
And tests reveal some other bug.