Force immediate [endlevel]

[WhiteWolf]

Hello,

Currently I have a weapon special which might trigger a <Lua> Repeated [endlevel] execution, ignoring warning.
The weapon special has area damage which can kill. If there are two hero units adjacent to this attacker, area damage might kill an adjacent hero (first [endlevel]) however, the game doesn't end immediately, the remaining strikes are still resolved for some reason*. Alas, the targeted hero might also die in resolving this fight, which leads to a second [endlevel].
The scenario ends in defeat all right, but said error message is displayed, which is not nice to leave around for the user to see.

*This is the part where I think it should be possible to intervene. Is there a way to make the [endlevel] immediatly end the level, and not resolve the rest of the strikes?

Thanks

[josteph]

*This is the part where I think it should be possible to intervene. Is there a way to make the [endlevel] immediatly end the level, and not resolve the rest of the strikes?

I'm surprised to discover that's actually possible:

Code: Select all

diff --git a/data/scenario-test.cfg b/data/scenario-test.cfg
index 3ced99d9094..767350eb92a 100644
--- a/data/scenario-test.cfg
+++ b/data/scenario-test.cfg
@@ -710,6 +710,14 @@ Xu, Xu, Qxu, Qxu, Ql, Ql, Ql, Xu, Xu, Xu, Xu, Xu, Xu, Xu, Xu, Gg, Gg, Gg, Gg, Gg
         {MODIFY_UNIT (id=$second_unit.id) side 1}
     [/event]
     [event]
+        name=attacker hits
+        [cancel_action]
+        [/cancel_action]
+        [endlevel]
+            result=victory
+        [/endlevel]
+    [/event]
+    [event]
         name=prestart
         {VARIABLE teleports_on no}
     [/event]

Spoiler:

Calling cancel_action causes wml_aborted to be set in attack::fire_event, which terminates the battle.

[WhiteWolf]

Thanks, this seems to solve it fine I've added [cancel_action] after the [endlevel]'s for each last breath event of the heroes.
On the other hand, can this theoretically cause any other kind of bugs anywhere else? I'm guessing not, after all, it's already an endlevel in defeat at the death of a unit, which is pretty much immediate in a 'regular' fight, so I guess an 'extra' [cancel_action] won't discard anything necessary... but I'd like to be reassured

[josteph]

On the other hand, can this theoretically cause any other kind of bugs anywhere else? I'm guessing not, after all, it's already an endlevel in defeat at the death of a unit, which is pretty much immediate in a 'regular' fight, so I guess an 'extra' [cancel_action] won't discard anything necessary... but I'd like to be reassured

Indeed, on the one hand, battles can end before all strikes have taken place if one of the units dies partway through, so most code should be ready for that situation. However, normally, when a battle ends prematurely, one of the two combatants dies. That's not true in your case, and furthermore, the wiki does not document that battles can be terminated prematurely by using [cancel_action]. So, I think it's possible if you use this, you'll discover that some other addon breaks when a battle ends prematurely with both combatants still living.

But the code explicitly tests a flag that can only be set by the [cancel_action] facility, so I assume this trick is actually an intentional feature, even if it's not documented.