Clarification note
This bug seems to be caused by a JVM bug:
We've identified the big 1.13-pre4 crash as a bug in Java itself! We have a fix lined up, but it's always exciting to be reminded that the JRE isn't infallible.
The bug
When creating a new world on 1.13-pre4, the first time I loaded up a new world, the game hung up on the "loading world" screen. The second time, it loaded 25% of the spawn area. Closing the game and loading the world progresses the "preparing the spawn area" screen between zero and four percent each time. After ten times attempting to load a new world, I'm at 36%.
Linked issues
is duplicated by 22
MC-132069
PRE-4 Minecraft don't create new world
Resolved
MC-132075
Creating new world will never end
Resolved
MC-132083
Swimming animation is active on land. And new worlds don't load.
Resolved
MC-132094
Creating new world frozen
Resolved
MC-132103
Unable to create a new world
Resolved
Attachments
Comments 30
Duplicate of MC-132069 unless it is not reopened.
Happens to me as well. Odd.
Edit: ...and why do the most important things always break? How did they not notice this major bug sooner? Oh well. Back to 12.2 for now.
Just figured it out–this is a Java related issue. Running 1.8.0_25 (which was the version I ran in the first crash that came from the launcher) caused it to crash the game. Running an updated version which I got since then gave me no issues.
Sorry for any confusion.
EDIT: Dinnerbone on Twitter
EDIT 2: I'm pretty sure today's update to the launcher changed the built-in Java version (1.8.0_51), and fixed the issue for me..
The crash is still present in 18w43c, just occured a few times while generating chunks in a new singleplayer world. Not sure how to reproduce it consistently yet.
[media][media]OS: Windows 10 1803 (Build 17134.112)
Java: Bundled JRE (1.8.0_51)
JVM Arguments: Default + -Xms4G -Xmx4G
I have been having the issue in 18w43c as well. So far, all I've done in the version is fly around the world a lot, but the world continually fails to load in, freezing and loading a few chunks, then freezing and loading a few more, and every once in a while freezing and then loading in everything around me as expected. It continues this cycle regularly until eventually, around the time when I'd expect the world to load everything normally, the entire game simply crashes.
[media]
Edit: I tried running in Java 11 as well just to see if that would fix anything, but it didn't.
This crash is still present in 18w44a. I just experienced it on the first launch of a 1.13.1 world in 18w44a within minutes of the world version upgrade. I have attached my crash report (
[media]) that was generated after the crash. The crash occurred when I died and attempted to load the spawn chunks at a custom location set a couple thousand blocks from spawn- the world refused to render and within 30 seconds the game crashed. I have tried for hours to reproduce this crash, to no avail.
The crash report in 18w44a is identical to the ones posted by VideoklipBG (
[media]and
[media]). Although the file names are different, this is because of the obfuscation between versions since the line numbers match up perfectly. The line numbers in mine are all off a bit from the original crash reports.
The only thing that caught my eye when I was comparing the crash reports here was that mine referenced datafixer in one of the last few lines in the stacktrace. This led me to think that this was caused (and thus reproducible) by converting another 1.13.1 world. This does not seem to be the case at the moment.
Fry
It looks like something's broken in world generation.
I created two new worlds in SinglePlayer.
random seed
seed = "1.13-pre4"
In both cases MC stops running somewhere during loading the save game. {screenshots: taskmgr, loading screen}
Loading (a copy of) a save created in pre3 works OK.