Clarification on the state of the Bukkit project.

I have been following the CraftBukkit Twitter for updates, and the most recent Tweet @CraftBukkit by EvilSeph has me a bit confused.

This seems to suggest that a stable Bukkit build may not be possible at all. Are these "issues" in Minecraft simply bugs/things that will be worked out in a future update? Or has Mojang put a stop to the development of Bukkit by intentionally changing the framework of Minecraft?

I just don't want to be mislead by the idea that MC 1.7 could be the end of Bukkit.

 

That's quite a leap from "We will have dev builds." to "It's the end of Bukkit."

 

Well, I was more concerned with the second part of the tweet, and the "issues we can't fix".

Are these issues that can't be fixed for the time being?... or that may never be fixable? These are rumors that are floating around my server community. If a stable version may never be possible, I think it would be safe to say it would be the demise of Bukkit.

 

It's simply stating that minecraft itself has issues... And apparently bukkit cannot fix these well enough without compromising functionality. The issues are enough bukkit doesn't want to tag it as a stable :/ this doesn't mean it won't work or that minecraft won't fix it in the future.

^^^my guess xD. Admittedly it is a little vague and has had me curious as well since I read it

 

Sounds for me like: "There is a Bug Mojang needs to fix."

 

There's always a Bug Mojang needs to fix

 

This is what I have taken it to mean, and that it's something that will be worked out. Has this sort of thing happened in the past? I guess I would still like to get some clarification on this from someone who is involved in Bukkit development.

 

My guess is that there are some pretty nasty bugs after integrating the bukkit hooks in that fixing with the obfuscation would be terribly hard

 

Don't worry.

I found this tweet also confusing, and talked with EvilSeph and TnT on IRC. Bukkit is fine, there are just some issues with 1.7.2... so when dev build emerges and there are issues... it's not on bukkit. This is just a clarification that bukkit is still trying to uphold their quality standards, however, there might be glaring Mojang bugs in a dev build that are not their fault.

That's what I gathered at least.

 

This is most likely in reference to bugs like this one: https://mojang.atlassian.net/browse/MC-17630

These are bugs in the code that Bukkit most likely can't fix because it is a code optimization issue thing that is at the core of how Minecraft works. There are several bugs like this right now that are a pain in everyone's backsides. Take a look at the top 10 most viewed and voted on bugs, and I think you will get a good idea of what EvilSeph is talking about.

 

That's an old ticket.

Anyway, I don't understand why the Bukkit folks can't implement a fix for Mojang errors and be ahead of the game for a change.

 

Exactly. It has been open since June 13 of this year. Issue has yet to be resolved. It's been reported to be better in 1.7 but still not completely fixed, hence why it is still open. These are game-crippling bugs that Mojang needs to fix, and won't be fixed in a bukkit implementation until Mojang fixes them most likely.

 

Since that bug has been there for 5 months and people are using Minecraft just fine since then, I wouldn't call it "game-crippling".

 

Did you even read the bug and the huge comment thread? Single player users don't usually have problems (although if you read the comment thread there are many who report they can generate the problem in SP as well). It is mostly a problem for server admins who run SMP servers with multiple players. If you are playing on a server that essentially runs in creative mode with no mobs spawning, then yea you won't see this issue. But both Dinnerbone and Jeb have acknowledged the problem, as it has effected many high-profile players who livestream and that kind of stuff. It causes server crippling lag with SMP, to the point where the server will eventually die because the CPU will eventually get pegged at 100% and TPS dips to very low levels. Even with just a few players, you can easily kill a server if they are all in different chunks with zombies locking onto villagers or players with no route to them.

I have had nights turned off on my server since June, because of this very bug. Enabling nights cripples the server in no time with even a couple of players online. If you haven't experienced this bug, then good for you. But for MANY server admins, this is GAME CRIPPLING and the only way to fix it is to disable nights or remove zombies from the game. This problem exists in almost all server implementations, except for servers that use Forge, who introduced a weird workaround. I'm certain these are the kinds of bugs that EvilSeph is referencing.

 

I know about the problem. I run an SMP server with multiple players, and this hasn't been a problem for us.

 

I have posted this before mods, but just to inform:
Imagine developing Bukkit for 1.7 is like hey0's hMod being developed for alpha 1.2.0. :I

 

Well consider yourself lucky I guess. All of my servers are seeing this issue, along with the hundreds of other people who voted that ticket up and continue to monitor it, waiting for a fix from Mojang.

 

Maybe you have malicious players who intentionally surround themselves with lava in order to bog down your server.

 

Bobcat00
Zombies do lagg a server, there is no need to "malicious" players for them to generate lagg.

moldyspore
Thats a big issue, but most likely not the one that causes the Bukkit Team problems, as it's kinda old...

 

Lava? That is not even what that bug is about... It's about zombie pathfinding.

And I know that is not what my issue is, since I have confirmed it being that bug that causes the problem. I can reproduce it with only 2 people on, soon as zombies start locking onto a target they can't reach, server TPS starts dropping. Hence why I have nights shut off on my server. With nights off, we don't have an issue, because there aren't a million zombies spawning on the surface and locking onto villagers and players. Using a mod to remove zombies also works, but that would remove them from everywhere including underground and in caves, which seemed like overkill, so I decided just to turn off the daylight cycle with the new /gamerule command they added. The pathfinding issue is exacerbated by the the changes they made in 1.6 with zombie spawning. The issue has always existed, apparently, but it didn't show up until they increased the distance that zombies can lock on to targets, and increased the spawn rate of zombies.

Perhaps, but it is a bug that Bukkit won't be able to fix, it's something only Mojang can handle. And there are a ton of them. I know I won't consider my Bukkit-based server truly stable until Mojang issues a fix for at least the top 5 tracked issues right now (all of which have been around for a long time) and then Bukkit updates their server afterwards.

If EvilSeph isn't talking about these bugs, I wish he'd elaborate and let us know what he meant. If he isn't talking about bugs on the official tracker, then that tweet is pretty cryptic.

 

Surrounding yourself with lava is one way to cause the zombie pathfinding bug to cause problems. (Try it!) Maybe your players are intentionally blocking the zombies from getting to them, which then causes your server to lag. Maybe your players are doing this to cause the server to lag.

My players don't do this. If they did, I'd tell them to stop. And I've never seen a village where this is a problem (although the zombies end up killing all the villagers eventually). So this "game crippling" bug is not game crippling at all.

Anyway, I think EvilSeph was talking about something else entirely.

 

I know they aren't purposely using lava to cause the bug since I know most of the players on my servers personally and they would have no reason to cause lag on the server. It benefits nobody. Also, as I already said, the bug is present when just myself and 1 other person logs in and night's are turned on. Again, I have turned nights off, and since then there is no problem with this bug. That eliminates the possibility of it being a player-created problem, since the problem goes away with nights turned off, and immediately comes back when they are turned back on. If your theory were correct, then anyone who found a zombie spawner could cause the lag, and that simply hasn't been happening. Turning off nights is what most server admins have done that experience this issue. Either that or remove zombies altogether with a mod (read the comment thread on the bug ticket).

The bug presents itself in ANY village that is walled off, where zombies spawn naturally close enough to lock onto the villagers inside. As I said, it only happens when there is NO path to their target. Villagers that are not fenced or walled off, therefore, do not cause this problem. If zombies end up killing your villagers, that means there is at least ONE path for the zombies to get to them, and the bug never presents itself as visible lag.

Just because you have never experienced the bug does not mean it doesn't exist or that players are somehow maliciously creating lag on purpose. I find it strange why you continue to deny its existence when there is a bug ticket with a confirmed status on the Mojang bug tracker with hundreds up updates and reports from admins and players alike. As we are getting way off topic, I'm just going to leave it at that. Believe whatever you like.

 

I know it's a bug, and it really exists. I just disagree with your opinion that it's "server crippling". I also don't think that's what EvilSeph was referring to.

FWIW, none of the villages on my server are walled off. And the spawn area has walls but two entrances that are always open, so the zombies can always find a way in.

Blame it on Dinnerbone!