The bug
Using malformed block states or biome ids for the superflat presets crashes the game.
Crash report
19w12b\: [^crash-2019-03-24_19.34.55-client.txt]
Description: mouseClicked event handler
n: Non [a-z0-9/._-] character in path of location: minecraft:dirt,minecraft:grass_block
at qq.<init>(SourceFile:38)
at qq.<init>(SourceFile:43)
at cef.b(SourceFile:372)
at cxo.a(SourceFile:63)
at cyf.b(SourceFile:167)
at cve.onPress(SourceFile:20)
at cve.mouseClicked(SourceFile:45)
at cvz.mouseClicked(SourceFile:45)
at cub.b(SourceFile:86)
at cyi.wrapScreenError(SourceFile:450)
at cub.a(SourceFile:86)
at org.lwjgl.glfw.GLFWMouseButtonCallbackI.callback(GLFWMouseButtonCallbackI.java:36)
at org.lwjgl.system.JNI.invokeV(Native Method)
at org.lwjgl.glfw.GLFW.glfwWaitEventsTimeout(GLFW.java:3129)
at csu.c(SourceFile:292)
at cua.b(SourceFile:1044)
at cua.d(SourceFile:993)
at cua.b(SourceFile:412)
at net.minecraft.client.main.Main.main(SourceFile:154)
How to reproduce
Try to use the following superflat preset
minecraft:bedrock;minecraft:dirt,minecraft:grass_block;minecraft:plains;plains
Linked issues
is duplicated by
relates to
Attachments
Comments


Um, just to clarify, this is happening in a regularly formed world I made before 1.10.. So my issue isn't about superflat. Or am I misunderstanding the title?
Affects for 18w31a.
Attached an updated crash report
[media]for the "malformed biome ids" section of this report.
Seems to be fixed as of 1.13.2-pre2. If a malformed block state is entered in the superflat customization, Minecraft resets to the "Classic Flat" preset. If a malformed biome ID is entered, it is detected by the game and removed. The game doesn't crash anymore.
Not fixed and updated the preset.
Can confirm for 18w43c. Pasted preset crashes the game.

Confirmed for 1.14-prerelease 5.
minecraft:bedrock;minecraft:plains;village;minecraft:plains
does not crash; the game just removes the second "minecraft:plains". However,
minecraft:bedrock;minecraft:pl,ains;village
(note the comma in "minecraft:pl,ains") still crashes.
For this type of error, is it possible to improve the logging to indicate which region file and chunk was the source of the bad entry for the crash?

Fixed in 19w46b
Confirmed for 1.13-pre5