FastAsyncWorldEdit

FastAsyncWorldEdit

152k Downloads

Cannot load from char array because "getSection" is null

BoomEaro opened this issue ยท 5 comments

commented

Server Implementation

Paper

Server Version

1.16.5

Describe the bug

Randomly, when inserting the schematic, an error occurs that seems to be related to lighting processing.
I'm using WorldEdit API to insert the schematic and seem to be doing everything right using a single thread pool that does something like this:

try (EditSession es = WorldEdit.getInstance().newEditSession(w)) {
                Operation op = new ClipboardHolder(cb)
                        .createPaste(es)
                        .to(BlockVector3.at(loc.getX(), loc.getY(), loc.getZ()))
                        .ignoreAirBlocks(false)
                        .copyEntities(false)
                        .copyBiomes(false)
                        .build();

                Operations.complete(op);
            }

And everything works, however, from time to time, the schematic is inserted incorrectly and an error occurs.

To Reproduce

The error happens randomly, there is no way to repeat it all the time.

Expected behaviour

Correct insertion of schematics without errors

Screenshots / Videos

No response

Error log (if applicable)

https://pastebin.com/z5XhDgZd

Fawe Debugpaste

https://athion.net/ISPaster/paste/view/509a03a0c2c9480bb04f91fcc4920819

Fawe Version

2.0.2-SNAPSHOT-84;faa2477

Checklist

Anything else?

No response

commented

Likely due to synchronisation issues with the layer or getChunk being trimmed

commented

I saw that in your debugpaste there is a stacktrace before the one you mentioned above. Failed to locate relight method in LightEngineThreaded on Tuinity. Is everything up to date? which could be the issue of this problem. This would need a verification.

commented

It's nothing to do with out of date server versions^ as I said most likely due to a rare and slight synchronisation issue. It's somewhat related to other rare issues where changes to the world are not necessarily captured, and typically occur when the queue.target-size config value is hit

commented

Most likely fixed by the recent processSet changes/improvement

commented

It wasn't because I just reproduced, but hopefully #2771 will do something