MovementController.java

- Endless movement while standing still [quick]fix

- Critical issues fixed

* Cannot use 2.5.x or newer jython with an outdated jython-engine

* Removing entries in direct iterations guarantees the subsequent call to entry.getNext() return null, thus invalidating the whole iteration after first removal

GPL...

* http://www.l2jfree-users.com/l2jfreeusers/posts/list/15/324.page

* thx heX1r0

* fix for airships

* NPE fix

Refactoring actors to correct folders, rename some instances

* reduced delay between arrival and beginning of the action

* NPE fix in sorting

GameServer

* decreasing selector thread latency

L2PcInstance

* offensive skills can be cast even if it gives charges and it has reached the limit

MovementController

* decreasing latency

ExclusiveTask

* just for divided entries in stats

* just a new admin command to get every skill under ID 2000 (so the skills that players usually have)

* it stores into db, so it could take time (~1k skill...)

L2PcInstance, RequestGMCommand, GMViewSkillInfo, SkillList

* sorting skill list sent to client, to get the same order regardless the db/learn order

* showing every skill to GMs - except if it's the SkillList packet, and subclass is active (let subclasses be normal..)

* sorting skills after loading, before actually adding it to the map - to spare some cpu cycles later :)

* skills basicly sorted by id, except the normally "non-visible" skills, which goes to the end

GameServer, GameTimeController, L2BoatInstance, L2Character, MovementController

* dividing movement code from game time calculation - only tick calculation left in GameTimeController

* calling EVT_ARRIVED_REVALIDATE from MovementController regularly, instead of the current solution, where it was executed one by one

* fixing a bug -> happened if the calculation was so long, that the position update was delayed a lot

L2Character

* just some fix for auto-attack after skill usage -> isAutoAttackable() should be called on the new target, not the old one

ThreadPoolManager

* re-added a simple ThreadPoolExecutor for instant tasks -> it don't have to wrap the Runnable into a Future<?> objects, so it's faster, and generates less garbage

* some tweak on shutdown -> it don't have to wait necessarily till timeout expires, instead of it returns immediately, when all worker thread is idle

ServerBasePacket

* fix for previous commits