terasology / health Goto Github PK

Some event should be provided for other systems to heal an entity to full health.
Currently, other systems increase the health value by huge amount to do the same. Providing with an event for it will improve intersystem interaction, as they will not have to directly access HealthComponent.

Currently, the RegenComponent has many logic bits in it. ECS asks for components with only data, and logic should be kept away for the system to handle.

When a block is damaged, and the game is reloaded, the block remains at same "damage state".

The cause might be that scheduler information is lost on leaving the game.

Possible solution: Add a postBegin() to schedule regen

In a multiplayer setting, when the client is damaged, the visual damage indicator is shown on the host instead of the client. This happens irrespective of whether the damage inflicted comes from the host, another client or other origins such as falling.

If a player drinks a regen potion, it activates regen by ActivateRegenEvent(). But the regen component, and hence the regeneration information is lost if the player is at full health.

This is a mystery to me 🔮 According to the logs, the ComponentSystemManager is trying to load a CoreHealthSystem as a system, but complains that it is not a subclass of ComponentSystem. The mystery part - this class does not (no longer) exist, not in this module nor in any other module 👀

Could this be a reflection issue?

22:59:45.674 [main] ERROR o.t.e.core.ComponentSystemManager - Cannot load CoreHealthSystem, must be a subclass of ComponentSystem

The package names currently used in this module make it difficult to distinguish which classes are in this module and which are from engine or elsewhere.

Recommend all packages here descend from a single org.terasology.health (or org.terasology.module.health?) instead of org.terasology.logic & rendering.

Events need to have a default constructor (no arguments) if they shall be sent over the network. Currently, some of the events violate this requirement.

22:59:44.119 [main] ERROR o.t.e.e.metadata.EventMetadata - Event 'Health:OnRestoredEvent' is a network event but lacks a default constructor - will not be replicated
22:59:44.120 [main] ERROR o.t.e.e.metadata.EventMetadata - Event 'Health:MaxHealthChangedEvent' is a network event but lacks a default constructor - will not be replicated
22:59:44.121 [main] ERROR o.t.e.e.metadata.EventMetadata - Event 'Health:OnDamagedEvent' is a network event but lacks a default constructor - will not be replicated

The API sections in the module docs are currently leading to

The links in question are

• http://jenkins.terasology.io/teraorg/job/Terasology/job/Modules/job/H/job/Health/job/master/javadoc/overview-summary.html
• http://jenkins.terasology.io/teraorg/job/Terasology/job/Modules/job/H/job/Health/job/develop/javadoc/overview-summary.html

This is due to the switch to a newer jenkins (http://newerjenkins.terasology.io instead of https://jenkins.terasology.io).

The new javadocs will be available on

• http://newerjenkins.terasology.io/job/Terasology/job/Modules/job/H/job/Health/job/develop/...
• http://newerjenkins.terasology.io/job/Terasology/job/Modules/job/H/job/Health/job/master/...

once they ran for the first time which they did not do yet.

Blocks #104

Problem

Module docs contain references to Java implementation details, such as specific components, systems, events, or methods. Somebody reading the docs and wanting to learn more about these components, systems, events, or methods would need to search for them in the JavaDocs (provided the link works again in the first place, see #101 ), which can be tedious depending on the module size.

Suggestion

Link all component, system, event, and method references to their respective JavaDoc.

Depends on #101

It would be better if baseRegen with 0 value is not added to the scheduler.

Probably related to two paths hitting the same map in the regeneration scheduler

Scenario: Server/client hosting a BuilderSampleGameplay world, 2nd client joins. Beat randomly on a few blocks and observe yep they don't regen right now, that's another known issue.

Host beat on 2nd client via fisticuffs for a few seconds, observed regen happen fine for 2nd client.

2nd client beat on host for several more seconds, observed regen starting fine but then crashed the server:

21:15:18.188 [main] ERROR o.t.e.e.internal.PojoEntityManager - Adding a component (class org.terasology.logic.health.RegenComponent) over an existing component for entity 31
21:15:18.188 [main] WARN  o.t.logic.health.RegenComponent - Regen added baseRegen
21:15:18.188 [main] WARN  o.t.logic.health.RegenComponent - Regen added baseRegen
21:15:18.241 [main] WARN  o.t.logic.health.RegenComponent - regenerating 3
21:15:18.241 [main] ERROR o.terasology.engine.TerasologyEngine - Uncaught exception, attempting clean game shutdown
java.util.ConcurrentModificationException: null
	at java.util.TreeMap$PrivateEntryIterator.nextEntry(TreeMap.java:1211)
	at java.util.TreeMap$EntryIterator.next(TreeMap.java:1247)
	at java.util.TreeMap$EntryIterator.next(TreeMap.java:1242)
	at com.google.common.collect.AbstractMapBasedMultimap$Itr.next(AbstractMapBasedMultimap.java:1177)
	at org.terasology.logic.health.RegenAuthoritySystem.regenerate(RegenAuthoritySystem.java:102)
	at org.terasology.logic.health.RegenAuthoritySystem.invokeRegenOperations(RegenAuthoritySystem.java:98)
	at org.terasology.logic.health.RegenAuthoritySystem.update(RegenAuthoritySystem.java:70)
	at org.terasology.engine.modes.StateIngame.update(StateIngame.java:171)
	at org.terasology.engine.TerasologyEngine.tick(TerasologyEngine.java:458)
	at org.terasology.engine.TerasologyEngine.mainLoop(TerasologyEngine.java:421)
	at org.terasology.engine.TerasologyEngine.run(TerasologyEngine.java:397)
	at org.terasology.engine.Terasology.main(Terasology.java:156)
21:15:18.241 [main] INFO  o.terasology.engine.TerasologyEngine - Shutting down Terasology...
21:15:18.303 [main] INFO  o.t.p.i.ReadWriteStorageManager - Saving - Creating game snapshot
21:15:18.352 [main] INFO  o.t.p.i.ReadWriteStorageManager - Saving - Snapshot created: Writing phase starts
21:15:18.363 [main] INFO  o.t.n.internal.NetworkSystemImpl - Client disconnected: Cerv2
21:15:18.365 [main] INFO  o.t.logic.console.ConsoleImpl - [NOTIFICATION] Player "Cerv2" has left the game
21:15:18.365 [main] INFO  o.t.n.internal.NetworkSystemImpl - Client disconnected: Cerv1
21:15:18.366 [main] INFO  o.t.logic.console.ConsoleImpl - [NOTIFICATION] Player "Cerv1" has left the game
21:15:18.366 [main] INFO  o.t.n.internal.NetworkSystemImpl - Network shutdown

Suspect something relatively simple is wrong in scheduler management land :-)

Converting regenEndTime map to sorted map removes the need for soonestEndTime and can improve performance during removal of completed regen.

Context

The Health module doc mentions the DamageResistComponent that allows configuring damage types an entity will not be damaged by. Most basic damage types are provided by the Terasology engine code base. There is a minor note on damage types in the Asset System Tutorial. Like all other prefabs, damage types can be modified on module level using Deltas & Overrides. New damage types can further be provided as module assets.

Task

Extend the module docs with information on how to add more damage types.

Open Questions

Should this information be in module-level docs or rather in the engine wiki and only linked to in the Health module docs?

In #84 we introduced the notion of Duration and Instant with corresponding TypeHandlers. As the logs indicate, defining those in the module is not allowed ("Denied access to class").

22:59:42.851 [Thread-3] ERROR o.t.g.m.s.JavaModuleClassLoader - Denied access to class (not allowed with this module's permissions): org.terasology.persistence.typeHandling.TypeHandler
22:59:42.853 [Thread-3] ERROR o.t.g.m.s.JavaModuleClassLoader - Denied access to class (not allowed with this module's permissions): org.terasology.persistence.typeHandling.TypeHandler