Bukkit

@MCTylerPVP

That seems overly optimistic towards the non-free plugins!

Seriously, it's certainly not "everything", though fight and ("emo-") fly are pretty severe at present. We'll cut down fly for sure, because that's our main field after all. Fighting will need ProtocolLib and will follow soon, though we can't really prevent bots.

@martini002

It "will be" (mostly) solved, once we are at it. We've just worked around a nasty type of false positive. While such always adds some danger for more exploits (...), we'll be on topic with fight checks and flying exploits from then on.

Unless it is very simple to fix, we usually try to remove false positives first, then tackle the exploits. Currently it's a little bit lengthy but we'll catch up soon.

@martini002

@PowerHustlers

It'll certainly be solved for most (possibly will need ProtocolLib though, at least for fighting)..

@Glitchblade

With "rough" compatibility we usually describe, that testing may not have been very intense (thus also beta) and that we just tried to tackle the most basic and/or severe issues, e.g. support for new types of blocks.

Essentially the beta release (3.12.0) is way outdated for 1.8. Instead the latest development builds are crafted towards a release and should be preferred for 1.8, at least until next release.

In general we keep some serious backwards compatibility, so having the latest (stable) release usually means improvements. There are rare cases when using the latest NCP on a much older version of Spigot/CraftBukkit will be more lenient, e.g. using the sprint key with flying, faster block breaking times. I'd also call the compatibility with 1.4.5-R1.0 "rough", because we hardly test that part ever.

@JoinMice

该许可“nocheatplus.checks.fight”允许绕过相关的战斗所有检查。如果你想有一个球员用一个特殊的程序，而不是一个服务器端插件飞，你应该给他们的许可“nocheatplus.checks.movi​​ng.survivalfly”。那么“creativefly”检查将监视飞行速度。 （使用谷歌翻译）

@Youtube_thejust219

NCP does not alter blocking in any way. That might be a server side bug (is your server up to date?), or it might be a plugin messing with stuff.

Do you know what it is triggered by?--

Please keep to the ticket, if it is that issue.

@Sikatsuu

Supporting changing blocks is quite complicated in nature. After the next release there is a chance we will newly evaluate how complicated it might be or if there is a simplified model we could implement, always having in mind not allowing to fly and such. Essentially on server side, you get situations where players move through blocks, hover in the air, take 1-block distances in almost one step, and similar.

Edit: Pistons are covered in Ticket 364.

@martini002

Good, thank you :).

@Bobcat00

We are heading for a release, currently trying to iron out a range of false positives, which show "best" with snow layers of differing height, we will also add a replacement for fight.speed using ProtocolLib.

The builds since 823 might have introduced new side issues with survivalfly, but they also do contain at least these improvements:

• Better attribute support, more precision (e.g. not allowing twice speed with speed potions).
• Better support for slowness potions, including level -100 or -200 (don't remember right now :p).
• Instant ladder.
• Hardening nofall (prevents generating artificial damage to lift off or glide).
• Prevent abuse of no-pvp regions to let NCP assume velocity being dealt.
• Hardening passable (fixed one potential exploit).
• In 1.8 the "step height" was increased from 0.5 to 0.6, enabling players to walk onto beds.
• Prevent Fastconsume from "endless" violation loops triggered by a false positive.
• Various fixes in the moving logic (sprinting, ...).

Since false positives count most for most people, it's probably rising and falling with the moving checks. I'd assume, unless for a snow-layers based server, you might be better off with the latest dev build until the next release. Can you test it with typical setups and a couple of players?

@File14

That would indeed be interesting, also setting probabilities for cancel (or commands). Currently it's not supported, i assume it's been like that with the original NoCheat, can't be entirely sure though.

A couple of optimizations rely on cancelling and it's some question how/where exactly to work a similar solution or if to just make delay+repeat possible everywhere (might then include an overhaul of some chunk of the internal bookkeeping of violations for performance reasons). So this isn't as quick to change as it might seem at first glance.