Sulkta-OSS/dynmap-neoforge
2
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Make sure chunk unload queue is processed quickly enough on MCPC

Browse source
This commit is contained in:
Mike Primm 2013-03-18 20:30:13 -05:00
parent bee16ef331
commit 0d25fd2edc
6 changed files with 28 additions and 13 deletions
Download patch file Download diff file Expand all files Collapse all files

5
src/main/java/org/dynmap/bukkit/NewMapChunkCache.java
View file

@ -935,16 +935,13 @@ public class NewMapChunkCache implements MapChunkCache {
/* If wasn't loaded before, we need to do unload */
if (!wasLoaded) {
chunks_read++;
/* It looks like bukkit "leaks" entities - they don't get removed from the world-level table
* when chunks are unloaded but not saved - removing them seems to do the trick */
helper.removeEntitiesFromChunk(c);
/* Since we only remember ones we loaded, and we're synchronous, no player has
* moved, so it must be safe (also prevent chunk leak, which appears to happen
* because isChunkInUse defined "in use" as being within 256 blocks of a player,
* while the actual in-use chunk area for a player where the chunks are managed
* by the MC base server is 21x21 (or about a 160 block radius).
* Also, if we did generate it, need to save it */
helper.unloadChunkNoSave(w, chunk.x, chunk.z);
helper.unloadChunkNoSave(w, c, chunk.x, chunk.z);
}
else if (isunloadpending) { /* Else, if loaded and unload is pending */
w.unloadChunkRequest(chunk.x, chunk.z); /* Request new unload */