targetworld issue

[mikeythemissile]

VoxelPort version: 2.046
Bukkit version: 1.3.2-R2.0
Client: Voxel Modpack version 9.1 R2

When executing the following command: "/vp targetworld ToSpace Space default", the following error is shown: "Invalid input! You entered something incorrectly. java.lang.NullPointerException " (Image link: http://puu.sh/1hYiq).
The console displays: "[INFO] mikeythemissile issued server command: /vp targetworld ToSpace Space default" (Image link: http://puu.sh/1hYkV)

The portal ToSpace exists, as does the world Space.

Regards,
Mike (mikeythemissile)

[mrnerdhair]

Something is going wrong in line 426 of PortManager.java. I suspect that the variable VoxelPort.s doesn't contain a valid reference to the Server object. Does an updated version of Bukkit help? If not, can you post the stack trace from the server log file?

[mikeythemissile]

The output from version is : This server is running CraftBukkit version git-Bukkit-1.4.7-R1.0-b2624jnks (MC: 1.4.7) (Implementing API version 1.4.7-R1.0), and the server has v2.046 of VoxelPort.
The error remains the same for the client, and the server log still only shows that the player executed the command (Image link: http://puu.sh/1ZKLO, note I tried the command twice)

[mrnerdhair]

It looks as if the error is intentionally swallowed because it was produced by a command; the logic being that that is probably indicative of a bad parameter. Unfortunately, that also makes it impossible to get the stack trace that will tell us exactly what the problem is.

Since I can't reproduce this on my own server directly, I need to get you to run a modified version of VoxelPort without the try/catch statement that swallows the error. That said, VoxelPort was already at version 2.046 when it was posted here at GitHub, so I'm not exactly sure which commit to build the modified version off of. That usually wouldn't matter much, but the package names were changed with commit fb7651a, and those are important.

If the command /vworld says "Herp derp" or the command "/help vp" says "Needs a description," I need you to upload the VoxelPort.java file you're using so I can be sure of the package names. Otherwise, tell me what the outputs are and I'll know which commit to base the debugging version off of.

[nristock]

I don't know why a good programmer would remove a try-catch block. I also don't know why a good programmer suppresses the stack trace, though. Instead of removing the statement simply add a printStackTrace one.

[mikeythemissile]

The output of those is commands is what you've written, the .jar I'm using is available here: http://dl.dropbox.com/u/23922866/VoxelPort.jar

[mrnerdhair]

Thanks for that file. It was important in figuring out which commit to use.

It turns out that you're using a snapshot of version 2.046, not an exact copy of the first version that bore that number, so I've prepared two modified JAR files: the first is as close to your current version as I can get, no bug fixes or different behavior or anything. The second has extra debugging code where I think the problem is, and is rigged to print the stack trace to the server log. The debugging code will only be triggered if you specify the "default" world type explicitly; leaving it off should produce the old behavior (though with a stack trace).

Please try this JAR first: https://www.dropbox.com/sh/l4be2iuhji8klxp/EGe7DfU4cf/VoxelPort-debug1.jar
Verify that it works on your server and displays the same bug.

Then try this JAR: https://www.dropbox.com/sh/l4be2iuhji8klxp/nNRO8tXQr-/VoxelPort-debug2.jar
Issue the same commands and it should provide debugging info. Please note that this JAR should not be considered "stable" for everyday usage; it's built specifically for this debugging task, and I haven't checked to make sure it won't accidentally eat your server in everyday usage. After running the test, you should probably switch back to your original JAR.

From the output this provides, I'll be able to determine where the problem is, and submit a patch to fix it if appropriate. Thanks for your cooperation!

[mikeythemissile]

The first debug .jar gives the same output as before, both for the client and in the server log.
However the second .jar doesn't provide any extra information, the server log only shows that the command was issued. Is there anything I need to change about the server, i.e. any extra configuration options?

[mrnerdhair]

That would be because I posted the wrong build product. My apologies for my noobish mistake. I should have noticed when I tested it the first time.

Here's the correct one: https://www.dropbox.com/sh/l4be2iuhji8klxp/rplXmw9EaK/VoxelPort-debug3.jar

Thanks for your patience.

[mikeythemissile]

Ok, when I try to run the same command, the following appears for the client: http://puu.sh/21uM9
And the server logs contains this: http://puu.sh/21uL5
Is there anything else you require?

[mrnerdhair]

Well, that's not what I expected. It's saying that the world named "Space" doesn't exist. Are you sure that there's a world by that name? Correct capitalization and everything? Does this issue recur with other worlds, and what plugins do you have installed on CraftBukkit?

[mikeythemissile]

The world "Space" definately exists, and when I try and set a portal from Space to World_2 (the default world), the same error is shown. I have the following plugins: SimpleBackup, bPermissions, DeathChest�, WorldEdit, �ChunkLoad� �VoxelGadget�, �VoxelSniper�, �LightDetector�, �CommandBook�, �CraftBukkitUpToDate�, �WirelessRedstone�, �SimpleCommandSigns�, �VoxelPort�, �dynmap�, �RedstoneCommandSigns�

[mrnerdhair]

I expected a plugin like Multiverse to be in that list. How do you access Space normally?

[mrnerdhair]

Also, can you post the entire server log file for an affected session?

[nristock]

I might be wrong, however, I guess either the world is not loaded when you issue the command or "Space" is just some kind of in-game alias.