Fix AirLiquidPlace falses (and BadPacketsX false)

[Axionize]

The comments in the code explain everything very clearly and in detail.

This also indirectly fixes a BadPacketsX false which is triggered together when AirLiquidPlace fails in rapid succession multiple times in a tick.

    /* This check has been plagued by falses for ages, and I've finally figured it out.
    When breaking and placing on the same tick in the same tick, I believe the vanilla client always sends DIGGING ACTION packets first
    This check's falses all seem to stem from processing DiggingAction.START_DIGGING before PacketType.Play.Client.PLAYER_BLOCK_PLACEMENT in the same tick
    Since we process the break first, when we go to process the place it looks like the player placed against air in the async world
    We will often see:
        Async world updated: short_grass -> air at X: -32, Y: 69, Z: -240, tick 0, cause/source: DiggingAction.START_DIGGING
        AirLiquidPlace Check: Block state at X: -32, Y: 69, Z: -240 is air (valid: false), tick +0-1, cause/source: PacketType.Play.Client.PLAYER_BLOCK_PLACEMENT <---- previously falsed here
        Async world updated: air -> short_grass at X: -32, Y: 69, Z: -240, tick +3-4, cause: realtime task in applyBlockChanges(List<Vector3i> toApplyBlocks) source: PacketType.Play.Client.PONG
        Async world updated: short_grass -> air at X: -32, Y: 69, Z: -240, tick +0-1, cause: handleNettySyncTransaction(LatencyUtils.java:56) source: PacketType.Play.Client.PONG
    In addition, it is possible for:
        Async world updated: short_grass -> air at X: -49, Y: 69, Z: -190, tick 0, cause: realtime task in applyBlockChanges(List<Vector3i> toApplyBlocks) source: PacketType.Play.Client.PONG
        AirLiquidPlace Check: Block state at X: -49, Y: 69, Z: -190 is air (valid=false), tick 0 <--- false due to change from applyBlockChanges()
        Async world updated: grass_block[snowy=false] -> grass_block[snowy=false] at X: -49, Y: 69, Z: -189, tick 0, cause: handleNettySyncTransaction(LatencyUtils.java:56) source: PacketType.Play.Client.PONG
    And in even more rare cases:
        Async world updated: air -> air at X: -51, Y: 71, Z: -179, tick 0, cause: handleNettySyncTransaction(LatencyUtils.java:56) source: PacketType.Play.Client.PONG
        AirLiquidPlace Check: Block state at X: -49, Y: 70, Z: -180 is short_grass (valid=true), tick 0
        Async world updated: short_grass -> air at X: -49, Y: 70, Z: -180, tick 1, cause/source: DiggingAction.START_DIGGING <--- double dig here (see my BadPacketsX patch) this is legit behaviour. Can only be up to 2 in 1 tick though.
        Async world updated: air -> short_grass at X: -49, Y: 70, Z: -180, tick 1, cause/source: DiggingAction.START_DIGGING
        Async world updated: short_grass -> air at X: -51, Y: 70, Z: -179, tick 1, cause: realtime task in applyBlockChanges(List<Vector3i> toApplyBlocks) source: PacketType.Play.Client.PONG
        AirLiquidPlace Check: Block state at X: -49, Y: 70, Z: -179 is air (valid=false), tick 2 <--- falses here due to double dig if we only check the latest changed blockstate. We have to check all changes at the location in same tick.
        AirLiquidPlace Check: Block state at X: -49, Y: 70, Z: -179 is air (valid=true), tick 2
        All of which previously would've caused a false.
        To solve this we store recently changed blocks caused by DiggingAction.START_DIGGING (instant breaking) and check against the old block.
        Lots of other checks have similar issues, and with the new player.blockHistory we can patch those.
    So that's it right? It's unfalsable?
        Very close but not quite. Vanilla's client game desyncs, especially on a laggy connection where a player is breaking and placing grass 20 cps/sec in the same tick
        it is possible for short grass to be interacted with even if server-side the block is air much later, and it won't be accounted for because the modification isn't recent.
        This is incredibly rare, unreliable and is only triggerable if you intentionally want to false the check. Enough so that I consider a violation lvl of 2-3 to be perfect.
    */

EDIT: fixes #1508, #1403 and #1776

[SamB440]

CheckManagerlistener still has conflicts

[Axionize]

Yes I'll fix this when we resolve #1846 since it modifies similar lines.

[Axionize]

I've rebased it @SamB440

[Axionize]

I'll take care of this in a couple days; been pretty busy.

[Axionize]

I'll merge this myself, just need someone else to give it a review and make sure everything's ready.

[ManInMyVan]

Suggested change

// private final long time; // System time in milliseconds or nanoseconds for ordering

[ManInMyVan]

Suggested change