Terasology: Crash report for CoreSampleGameplay after multiple play sessions in one game launch

[brylie]

What you were trying to do

Try out the CoreSampleGameplay

What actually happened

Got a NullPointerException

How to reproduce

1. Start Terasology
2. Start new CoreSampleGameplay single player game

Log details and game version

https://pastebin.com/vSbG7efe

Game version: Omega

Computer details

Kubuntu 18.04

[Cervator]

Thanks again for submitting :-)

This one is a little tricky as it seems like the log file has several rounds of action in it, meaning you might have hit an issue that wouldn't get hit if doing it as the sole action when starting the game fresh.

Specifically 00:30:51.772 [New I/O worker #3] INFO o.t.n.i.ClientConnectionHandler - Downloading core:2.0.0 (1.92 MB,1 modules remain) sticks out. The shipped game was a develop Omega zip with v2.0.1-SNAPSHOT, meaning the game is trying to downgrade an embedded module. Maybe we should add either a warning or block downgrading modules, at least those shipped with the game and thus tied to the engine which isn't version tweakable at runtime.

There are also a couple 00:36:02.076 [Timer-1] ERROR o.t.n.i.ClientConnectionHandler - Server timeout threshold of 10000 ms exceeded. although those probably matter less. But when you've tried joining a server whichever modules the server runs is what the game will try to load up.

I figure in the end you tried starting a CoreSampleGameplay session, but it crashed on start-up? There isn't a good error at the very end of the log file, which is a pity. I can start the game fresh and run CoreSampleGameplay just fine, so the original issue is likely not a problem, but a more tricky multi-session game setup workflow can lead to corrupt game state _is_ - and I'm pretty sure I've hit something similar myself.

I'm going to rename this a bit to try to cover hitting an odd error like this after entering and leaving multiple game worlds in one execution of the game then see if we can get some more details on that case 👍

Edit: Marginally related to #2753 which could be helped by better error handling when modules at funny versions end up being downloaded, and #3131 which also is about better dependency feedback

[brylie]

Thanks for the follow-up.

Looking forward to the upcoming playtesting session! ☺

[Cervator]

Going to close this as tentatively fixed - we just resolved for sure a _very_ similar sounding issue with #3903 and while its origin is dated after this issue was submitted I suspect it could have been the same bug surviving a refactor in the area.

Hope you're doing well @brylie, if you run into this again by chance sometime let us know! Fix is in develop now, should in in the builds soon.