Attached screenshot and minimal world repro. If you stand at the location indicated in the screenshot and save/reload the world, you will hit the bug.
Root cause has been narrowed down to:
Item frame with some contained item at X=0 / Z=0 in chunk
Mounted on block in adjacent chunk
Behind that block, a comparator is sensing the item frame
Stand in a chunk adjacent to the chunk containing the item frame, save and load the world, and all❓ entities in the chunk with the item frame will be duplicated a huge number of times.
From MC-56077:
This is an item duplication glitch that involves the new comparator measuring item frame feature in snapshot 14w18.
How to Replicate This Bug:
1.) Find the east/west edge of a chunk. (Preferably using F3)
2.) Place any redstone current-carrying block on the west side of the chunk edge.
3.) Place an item frame on the block, facing east.
4.) Place a comparator next the block, pointing the opposite direction that the item frame is facing, so that the comparator is in a different chunk than the item frame.
5.) Put any block you want to duplicate into the item frame.
6.) Quit and re-join the world. You will now have a stack of item frames and the item you put in the frame.
Related issues
is duplicated by
relates to
Attachments
Comments


Confirmed.

Also attaching log file, which has some nice long stack overflow exceptions that were ignored when loading. Last level load in the log is related to the screenshot.

Duplication causes exception (or exception causes duplication):
[07:21:34 INFO]: Client> [07:21:30] [Server thread/ERROR]: Couldn't load chunk
[07:21:34 INFO]: Client> s: Exception getting block type in world
[07:21:34 INFO]: Client> at tj.f(SourceFile:1168) ~[14w11b.jar:?]
[07:21:34 INFO]: Client> at tp.a(SourceFile:158) ~[14w11b.jar:?]
[07:21:34 INFO]: Client> at ave.a(SourceFile:291) ~[14w11b.jar:?]
...
[07:21:34 INFO]: Client> at ny.d(SourceFile:117) ~[14w11b.jar:?]
[07:21:34 INFO]: Client> at akh.b(SourceFile:287) ~[14w11b.jar:?]
[07:21:34 INFO]: Client> at akh.f(SourceFile:282) ~[14w11b.jar:?]
[07:21:34 INFO]: Client> at akh.c(SourceFile:218) ~[14w11b.jar:?]
[07:21:34 INFO]: Client> at akh.a(SourceFile:2360) ~[14w11b.jar:?]
[07:21:34 INFO]: Client> Caused by: java.lang.StackOverflowError
[07:21:34 INFO]: Client> at java.lang.Package.<init>(Unknown Source) ~[?:1.7.0_51]
[07:21:34 INFO]: Client> at java.lang.Package.<init>(Unknown Source) ~[?:1.7.0_51]
[07:21:34 INFO]: Client> at java.lang.Package$1.run(Unknown Source) ~[?:1.7.0_51]
[07:21:34 INFO]: Client> at java.lang.Package$1.run(Unknown Source) ~[?:1.7.0_51]
[07:21:34 INFO]: Client> at java.security.AccessController.doPrivileged(Native Method) ~[?:1.7.0_51]
[07:21:34 INFO]: Client> at java.lang.Package.defineSystemPackage(Unknown Source) ~[?:1.7.0_51]
[07:21:34 INFO]: Client> at java.lang.Package.getSystemPackage(Unknown Source) ~[?:1.7.0_51]
[07:21:34 INFO]: Client> at java.lang.Package.getPackage(Unknown Source) ~[?:1.7.0_51]
[07:21:34 INFO]: Client> at java.lang.Class.getPackage(Unknown Source) ~[?:1.7.0_51]
[07:21:34 INFO]: Client> at org.apache.logging.log4j.core.impl.ThrowableProxy.resolvePackageElement(ThrowableProxy.java:437) ~[log4j-core-2.0-beta9.jar:2.0-beta9]
[07:21:34 INFO]: Client> at org.apache.logging.log4j.core.impl.ThrowableProxy.resolvePackageData(ThrowableProxy.java:395) ~[log4j-core-2.0-beta9.jar:2.0-beta9]
[07:21:34 INFO]: Client> at org.apache.logging.log4j.core.impl.ThrowableProxy.<init>(ThrowableProxy.java:115) ~[log4j-core-2.0-beta9.jar:2.0-beta9]
[07:21:34 INFO]: Client> at org.apache.logging.log4j.core.impl.ThrowableProxy.<init>(ThrowableProxy.java:98) ~[log4j-core-2.0-beta9.jar:2.0-beta9]
[07:21:34 INFO]: Client> at org.apache.logging.log4j.core.impl.Log4jLogEvent.<init>(Log4jLogEvent.java:114) ~[log4j-core-2.0-beta9.jar:2.0-beta9]
[07:21:34 INFO]: Client> at org.apache.logging.log4j.core.impl.Log4jLogEvent.<init>(Log4jLogEvent.java:90) ~[log4j-core-2.0-beta9.jar:2.0-beta9]
[07:21:34 INFO]: Client> at org.apache.logging.log4j.core.impl.DefaultLogEventFactory.createEvent(DefaultLogEventFactory.java:49) ~[log4j-core-2.0-beta9.jar:2.0-beta9]
[07:21:34 INFO]: Client> at org.apache.logging.log4j.core.config.LoggerConfig.log(LoggerConfig.java:365) ~[log4j-core-2.0-beta9.jar:2.0-beta9]
[07:21:34 INFO]: Client> at org.apache.logging.log4j.core.Logger.log(Logger.java:110) ~[log4j-core-2.0-beta9.jar:2.0-beta9]
[07:21:34 INFO]: Client> at org.apache.logging.log4j.spi.AbstractLogger.error(AbstractLogger.java:609) ~[log4j-api-2.0-beta9.jar:2.0-beta9]
[07:21:34 INFO]: Client> at ny.f(SourceFile:139) ~[14w11b.jar:?]
[07:21:34 INFO]: Client> at ny.c(SourceFile:81) ~[14w11b.jar:?]
[07:21:34 INFO]: Client> at ny.d(SourceFile:117) ~[14w11b.jar:?]
[07:21:34 INFO]: Client> at akh.b(SourceFile:287) ~[14w11b.jar:?]
...
[07:21:34 INFO]: Client> at akh.x(SourceFile:2377) ~[14w11b.jar:?]
[07:21:34 INFO]: Client> at akh.c(SourceFile:2398) ~[14w11b.jar:?]
[07:21:34 INFO]: Client> at anw.g(SourceFile:168) ~[14w11b.jar:?]
[07:21:34 INFO]: Client> at ann.g(SourceFile:131) ~[14w11b.jar:?]
[07:21:34 INFO]: Client> at ann.i(SourceFile:101) ~[14w11b.jar:?]
[07:21:34 INFO]: Client> at ann.b(SourceFile:197) ~[14w11b.jar:?]
[07:21:34 INFO]: Client> at anw.a(SourceFile:133) ~[14w11b.jar:?]
[07:21:34 INFO]: Client> at akh.f(SourceFile:2821) ~[14w11b.jar:?]

Duplication does not occur if comparator gets removed.

Still exists in latest snapshot. I noticed the duplication also occurs if you stand in the same chunk in the repro world; there may be some conditions to it as I didn't observe this in a previous repro environment.

I testet it and it still works in 14w20b.
The item frame has to be in a different chunk than the block it's hanging on in positive X or Z direction
Not only the item frame and its content get duplicated but also every other entity in the chunk the item frame is in.

Confirmed for 14w21a/b.

If you execute this bug with a villager you will get duplicates of the villager - the only way this seems to happen in the game.

Confirmed for 14w26a

confirmed for 14w26c

Confirmed for 14w27b

I should mention also that we originally discovered this issue accidentally in a multiplayer world: Besides the minimal repro, if you nether-portal in to an adjacent chunk in an area away from spawn (rather than just loading a single player world standing in an adjacent chunk) it also triggers the bug (and may crash the server, I think depending on the number of entities it results in)

Confirmed for 14w28a

Confirmed for 14w28b.

Confirmed for 14w29b.

Confirmed for 14w30a

Confirmed in 14w30b.

I cant run my snapshot server because of this and will not be able to update to 1.8 as this bug crashes online servers

Confirmed for 14w31a

Confirmed for 14w32a

I can also confirm for 14w32a. Screenshot attached.

Confirmed for 14w32c

Confirmed for 14w32d

Confirmed for 14w33a

Confirmed for 14w33b

Confirmed for 14w33c