JSON signs with scoreboard objectives erase their block data upon chunk load

Hi, I've tried on my map with this command :

blockdata ~ ~ ~2 {Text1:"[\"\",{\"text\":\"hi\"}]",Text2:"[\"\",{\"score\":{\"name\":\"@p\",\"objective\":\"hi\"}}]"}

and the system is working, I don't find the problem, sorry.

Here is the link of my map download : http://www.mediafire.com/download/6sghiati11s31sy/Test_Bug.zip

Good bye.

@unknown Le problème est: Si tu redémarrer la map le texte sur la pancarte/panneau est disparu }=)
(The problem is: If you restart the map/world, the text on the sign is gone.)

Can confirm that, and on triggering the CB the text and score on the sign is back.

This has nothing to do with scoreboard objectives. The problem comes from the missing Text3 and Text4 tags. See this comment on MC-87143

@unknown That's what I suspected myself, I was in that discussion and not only I begged Mr. Broes to change that..
That's why I tested a sign (incl. scoreboard data) WITH Text3 + Text4, but it still didn't work.
Either I messed up the command, or line 3 and 4 not being referred to in the command is not the problem here.

Yep, I confirmed it again, I referenced line3 and 4 of the sign and the sign was still blank on relog.
If someone could look over my command I used, if I messed it up, I'm not doing so well with strict JSON:

setblock ~ ~1 ~ standing_sign 0 replace {Text1:"{\"text\":\"\"}",Text2:"[\"\",{\"score\":{\"name\":\"@p\",\"objective\":\"hi\"}}]",Text3:"{\"text\":\"\"}",Text4:"{\"text\":\"\"}"}

Yeah, confirmed, the launcher even print an error :

[Server thread/ERROR]: Failed to load data for block entity Sign
java.lang.NullPointerException
	at aqn$1.h(SourceFile:99) ~[1.9-pre4.jar:?]
	at ex.a(SourceFile:38) ~[1.9-pre4.jar:?]
	at ev.a(SourceFile:37) ~[1.9-pre4.jar:?]
	at ev.a(SourceFile:63) ~[1.9-pre4.jar:?]
	at aqn.a(SourceFile:107) ~[1.9-pre4.jar:?]
	at apv.a(SourceFile:108) [1.9-pre4.jar:?]
	at ass.a(SourceFile:306) [1.9-pre4.jar:?]
	at ass.a(SourceFile:83) [1.9-pre4.jar:?]
	at ass.a(SourceFile:70) [1.9-pre4.jar:?]
	at lo.f(SourceFile:124) [1.9-pre4.jar:?]
	at lo.c(SourceFile:76) [1.9-pre4.jar:?]
	at lo.d(SourceFile:91) [1.9-pre4.jar:?]
	at net.minecraft.server.MinecraftServer.l(SourceFile:325) [1.9-pre4.jar:?]
	at byn.a(SourceFile:114) [1.9-pre4.jar:?]
	at byn.j(SourceFile:130) [1.9-pre4.jar:?]
	at net.minecraft.server.MinecraftServer.run(SourceFile:427) [1.9-pre4.jar:?]
	at java.lang.Thread.run(Thread.java:745) [?:1.8.0_66]

Probably caused by this method in the aht.class (probably World.java):

public MinecraftServer u()
  {
    return null;
  }

That being said, I don't know if the changes to resolve MC-87143 could in theory affect other sign-related bugs..
So at the very least, going by what Marcono said here, quote:

it might be worth a look into it, unless it's completely out of the question it can have anything to do with it .

Thanks for making me aware, I should look more often into my Game Output ;-;

[10:27:35] [Server thread/ERROR]: Failed to load data for block entity Sign
java.lang.NullPointerException
	at aqn$1.h(SourceFile:99) ~[1.9-pre4.jar:?]
	at ex.a(SourceFile:38) ~[1.9-pre4.jar:?]
	at ev.a(SourceFile:37) ~[1.9-pre4.jar:?]
	at ev.a(SourceFile:63) ~[1.9-pre4.jar:?]
	at aqn.a(SourceFile:107) ~[1.9-pre4.jar:?]
	at apv.a(SourceFile:108) [1.9-pre4.jar:?]
	at ass.a(SourceFile:306) [1.9-pre4.jar:?]
	at ass.a(SourceFile:83) [1.9-pre4.jar:?]
	at ass.a(SourceFile:70) [1.9-pre4.jar:?]
	at lo.f(SourceFile:124) [1.9-pre4.jar:?]
	at lo.c(SourceFile:76) [1.9-pre4.jar:?]
	at lo.d(SourceFile:91) [1.9-pre4.jar:?]
	at net.minecraft.server.MinecraftServer.l(SourceFile:325) [1.9-pre4.jar:?]
	at byn.a(SourceFile:114) [1.9-pre4.jar:?]
	at byn.j(SourceFile:130) [1.9-pre4.jar:?]
	at net.minecraft.server.MinecraftServer.run(SourceFile:427) [1.9-pre4.jar:?]
	at java.lang.Thread.run(Unknown Source) [?:1.8.0_40]

I've the same error.

This affects 1.8.9 worlds when upgrading, yes?

@unknown I went with 1.8.9 into the testworld with that scoreboard-sign and can confirm the scoreboard updates with no problem in 1.8.9, and on relog the sign is not blank.

I used the command

setblock ~ ~1 ~ standing_sign 0 replace {Text1:"{\"text\":\"\"}",Text2:"[\"\",{\"score\":{\"name\":\"@p\",\"objective\":\"hi\"}}]",Text3:"{\"text\":\"\"}",Text4:"{\"text\":\"\"}"}

(of course including the scoreboard before, as provided in the OP)

When opening this testworld incl that sign in 1.9 pre-4 the sign turns blank, so i can confirm that.
As I also referred line 3 and 4 in that sign, it cannot go blank due to that (but, again, I don't know what Mr. Broes did in total as for the changes on signs).

I can't find this in the comments this far, so I thought I'd add it in a quick note: This happens whenever the chunk unloads and re-loads. So it doesn't require the world to be closed and re-opened, it just requires you to leave the area and come back.

@unknown Indeed, confirmed, the sign's text gets also erased if the chunk unloads.
It outputs a slightly different error now than before:

[14:39:38] [Server thread/ERROR]: Failed to load data for block entity Sign
java.lang.NullPointerException
	at aqn$1.h(SourceFile:99) ~[1.9.jar:?]
	at ex.a(SourceFile:38) ~[1.9.jar:?]
	at ev.a(SourceFile:37) ~[1.9.jar:?]
	at ev.a(SourceFile:63) ~[1.9.jar:?]
	at aqn.a(SourceFile:107) ~[1.9.jar:?]
	at apv.a(SourceFile:108) [1.9.jar:?]
	at ass.a(SourceFile:306) [1.9.jar:?]
	at ass.a(SourceFile:83) [1.9.jar:?]
	at ass.a(SourceFile:70) [1.9.jar:?]
	at lo.f(SourceFile:124) [1.9.jar:?]
	at lo.c(SourceFile:76) [1.9.jar:?]
	at lu.<init>(SourceFile:40) [1.9.jar:?]
	at lv.c(SourceFile:195) [1.9.jar:?]
	at lv.c(SourceFile:287) [1.9.jar:?]
	at mm.d(SourceFile:334) [1.9.jar:?]
	at mb.a(SourceFile:508) [1.9.jar:?]
	at it.a(SourceFile:126) [1.9.jar:?]
	at it$b.a(SourceFile:18) [1.9.jar:?]
	at fh$1.run(SourceFile:13) [1.9.jar:?]
	at java.util.concurrent.Executors$RunnableAdapter.call(Unknown Source) [?:1.8.0_40]
	at java.util.concurrent.FutureTask.run(Unknown Source) [?:1.8.0_40]
	at g.a(SourceFile:45) [1.9.jar:?]
	at net.minecraft.server.MinecraftServer.D(SourceFile:597) [1.9.jar:?]
	at net.minecraft.server.MinecraftServer.C(SourceFile:553) [1.9.jar:?]
	at byn.C(SourceFile:154) [1.9.jar:?]
	at net.minecraft.server.MinecraftServer.run(SourceFile:457) [1.9.jar:?]
	at java.lang.Thread.run(Unknown Source) [?:1.8.0_40]

14:45:27] [Server thread/ERROR]: Failed to load data for block entity Sign
java.lang.NullPointerException
	at aqn$1.h(SourceFile:99) ~[1.9.jar:?]
	at ex.a(SourceFile:38) ~[1.9.jar:?]
	at ev.a(SourceFile:37) ~[1.9.jar:?]
	at ev.a(SourceFile:63) ~[1.9.jar:?]
	at aqn.a(SourceFile:107) ~[1.9.jar:?]
	at apv.a(SourceFile:108) [1.9.jar:?]
	at ass.a(SourceFile:306) [1.9.jar:?]
	at ass.a(SourceFile:83) [1.9.jar:?]
	at ass.a(SourceFile:70) [1.9.jar:?]
	at lo.f(SourceFile:124) [1.9.jar:?]
	at lo.c(SourceFile:76) [1.9.jar:?]
	at lo.d(SourceFile:91) [1.9.jar:?]
	at aht.a(SourceFile:297) [1.9.jar:?]
	at aht.a(SourceFile:883) [1.9.jar:?]
	at lp.a(SourceFile:934) [1.9.jar:?]
	at mm.c(SourceFile:329) [1.9.jar:?]
	at mm.a(SourceFile:160) [1.9.jar:?]
	at md.b(SourceFile:116) [1.9.jar:?]
	at md.c(SourceFile:62) [1.9.jar:?]
	at ek.a(SourceFile:232) [1.9.jar:?]
	at ma.c(SourceFile:187) [1.9.jar:?]
	at net.minecraft.server.MinecraftServer.D(SourceFile:643) [1.9.jar:?]
	at net.minecraft.server.MinecraftServer.C(SourceFile:553) [1.9.jar:?]
	at byn.C(SourceFile:154) [1.9.jar:?]
	at net.minecraft.server.MinecraftServer.run(SourceFile:457) [1.9.jar:?]
	at java.lang.Thread.run(Unknown Source) [?:1.8.0_40]

Please link to this comment in the description

The following is based on decompiled version of Minecraft 1.9 using MCP 9.24 beta. All method and class names are the names used in the decompiled version.

The reason why this happens is that on loading a tile entity the public static TileEntity createTileEntity(MinecraftServer server, NBTTagCompound compound) method of the net.minecraft.tileentity.TileEntity class is called. If the created tile entity is not null, it is added to the world.
The problem is that the methods adding the sign to the world set the position and the world however the creating method reads the NBT data of the sign. When the NBT data of the sign is read, its text gets updated, for scores this requires the server to get the scoreboard. The server is received by calling the public MinecraftServer getServer() method of an anonymous class of the net.minecraft.command.ICommandSender interface the net.minecraft.tileentity.TileEntitySign class creates. This method gets the server from the world field of the sign which is at this point not set, which creates a java.lang.NullPointerException.

This can be fixed by having the public static TileEntity createTileEntity(MinecraftServer server, NBTTagCompound compound) method of the net.minecraft.tileentity.TileEntity class set the world and position instead of the public void addTileEntity(BlockPos pos, TileEntity tileEntityIn) method of the net.minecraft.world.chunk.Chunk class.

I've tried the following:

In *.mca-Savegame and as result of an empty blockdata request (/blockdata x y z {}), I've got in 1.8 and 1.9:

Text1:{"extra":["hi"],"text":""}
Text2:{"extra":[{"score":{"name":"playername","objective":"hi","value":"100"}}],"text":""}

The behavior in 1.8 was: No score-update nor selector-update on chunk update or reload.
The savegame and blockdata contains a field "value" with the value at the creation time.
Therefore for loading the sign, the scoreboard is not required. The "value"-field is simply showen.

Please fix this to the way it works in 1.8. I want to hold "old" scores in signs. These should be persistent, because the objective will change during playing the map, and the old score should be displayed.

Also, in 1.9 the value-data is updated while cloning a sign. In 1.8, the data remains the old score (from "value").

If you want so - to force reload the displayed score - you could send a blockdata-update (like in 1.8). Or add a tag like "resolved" for books to get both behaviors.