(SAFETY COMMIT)
Largely breaking change.
* Interfaces in front of data types (and 'managers'), some interfaces
removed.
* Data and configuration fetching.
* Check activation checking (config flags, isEnabled, hasBypass).
* CheckType (activation checks, factories removed).
* Lots of collateral signature changes, including IPlayerData.
The (I)WorldDataManager stores per-world data (+ per world per check
type).
* Raw configurations.
* Typical flags: check activation, debug, lag adaption.
* Generic data, such as check configurations or per world check data.
The (I)PlayerDataManager stores per player data.
* Check Data.
* Typical flags: debug
* Exemption
* Check data (and config cache).
* Further mappings and later OfflinePlayerData.
* The registration interface will allow defining, how instances are
handled for registered types (factory, proxy, what on world change, what
on logout, global removal handler, per player removal handler).
(I)PlayerData is intended to be/become the central access point.
* External interface is IPlayerData now.
* Per player debug flags, exemptions.
* Fetching configuration and data: local cache, relaying fetching to
registered factories and proxy-registries/storage (e.g. fetching
configuration from per world storage).
Other fixes/changes:
(+) Extend the debug player command (set true/false, reset to world
default, arbitrary check types).
(+) PlayerData maintains a currentWorldIdentifier (to be used instead of
ChatData in future).
(+) The WorldConfigProvider getAll implementation returns a
LinkedHashSet now, avoiding duplicates.
(+) Move DefaultGenericInstanceRegistry to NCPCore.
(+) Thread-safety considerations for DefaultGenericInstanceRegistry.
(+) Don't log errors on hasBypass checking. TBD: Instead intercept
during listener methods (or even as a feature within the listener node:
e.g. @ThreadContext(primaryThread=true, skipOffContext=true,
cancelOffContext=true).
(+) Add fight.wrongturn permissions to plugin.yml.
(+) Missing GPLv3 headers.
Broken/Missing:
* WorldData inheritance from default: propagate all changes done
directly to the default config to children (all worlds that don't have
an explicit world_config.yml set) - possibly add an OverrideState or
similar, (NONE, FROM_DEFAULT, EXPLICIT) and don't override EXPLICIT if
coming from the default. Calling override on the default WorldData is
not to be confused with calling override for WorldDataManager (override
for all worlds as EXPLICIT).
* Organize overriding for special circumstances (version dependent
activation and the like). Might want to add registered override
handlers to be called on reload automatically.
* Store generic per check type per world data in the WorldDataManager,
such as configurations and per-world check data. TBD: Factories, cleanup
(!).
* Most efficient referencing (IWorldCheckTypeNode, IHandle<something>?).
* All the registry stuff (see PlayerData).
* Use interfaces for auto registry (and a flag within
RegistrationContext?) - world unload, world change, player join / leave.
* (Data expiration handling including transition to IOfflinePlayerData,
because now data is a little heavier.)
* Further details.
* Store the per check type flags within PlayerData/PlayerCheckTypeTree.
Access methods within PlayerData.
* (More) thread-safe access, with the twist that alterations are mostly
done within the thread-context (primary thread vs. asynchronous).
(+) Mimic legacy behavior, by non-nested entries. However untouched
nested entries are possible, creating ExemptionContext instances with
other negative ids.
(-) No ExemptionRegistry is implemented. This just aims at replacing the
internals, without altering the (legacy) behavior.
* Each asynchronous permission check yields an update request anyway,
thus frequent bulk update requests have been removed.
* With join and world change, lazy permission updating is requested.
Breaking:
* Remove CheckConfig.getCachePermissions().
The original idea had been to just count an abstract value for load, so
the number of permission checks per tick stays limited, however that
doesn't scale with the CPU power and other.
This doesn't account for lag, but keeps the impact down slightly. Not
sure this is really useful.
(+) Only run set back for online players. Quitting has to be handled
differently.
(+) Comments.
* Remove the old per-world permission cache.
* Use the new permission caching feature.
* Always use permission subscriptions.
Pitfalls:
* Permission subscriptions might fail under certain conditions
(legacy?).
Only count the fall height below the set back y from lift-off (rather).
Breaking:
* Adjust method signatures and public visibility / interface for NoFall.
Issues with bunny hopping remain (...), but this might help with
elevators.
Use of velocity entries has been made more strict (directly use, remove
previously queued ones).
(+) Extend/alter slightly, e.g. added a regex example.
Replacement characters are not needed, and default values won't keep
reappearing.
Tests pending, e.g. dump with the upcoming registry log.
Benefits:
* Improves performance, where permission lookup has major impact, with
timeout based lookup, static permissions (skip permission check
entirely), and world/offline based invalidation. (Once fully
implemented.)
* Hopefully more efficient: use Bukkit Permission for faster defaults.
* (Allows control over how which permission is to be
updated/invalidated, which is useful per se.)
Risks:
* Complex changes yield bugs.
* Other plugins depending on NCP might break.
* Cache incoherence might happen (permissions are changed dynamically +-
unintended malconfiguration, or in case of bugs).
* (Endless loops certainly have all been fixed.)
Breaking:
* Lots of more or less internal API has been changed or removed: Check,
CheckType, CheckUtils, TickTask, ...
* Permission checking behavior has been altered.
Rough points:
* Implement a permission cache within PlayerData.
* Remove the player tasks and permission updates in favour of handling
those within DataManager and PlayerData.
* Adjust everything else to it (partly TBD).
* Updating sets of permissions (e.g. for CHAT) is done more lazily now,
i.e. one per 10 ticks). An actual permission check would still yield an
update next tick (asynchronous).
* Fixed/extended random spots (DualCollection, MiniListener registration
support, StringUtil).
Missing:
* Basic implementation
* Cleanup after logout (stages: 1. non-essential like permissions,
2. unrecoverable like set-back location, 3. complete data removal).
* Coverage
* Might have missed spots.
* NoCheatPlus.nameSetPerms should be replaced by caching + default
config for world-wise updating.
* Command permissions are always checked. At least for players,
cache based lookup should get implemented.
* More unit tests.
* Extended configurability: Per-world settings/policies.
* Efficiency
* Not all parts of the implementation are 100%/optimal yet.
Relatively cheap addition, done:
* Internal API within BlockProperties.
Missing:
* Support configuration: compatibility.blocks.breakingtime
* Support a command to update by the last n mined blocks (!). Needs
storing all side conditions within BlockBreakKey and convenience methods
for getBlockBreakKey + use internally.
The old ListenerManager is removed, new system in place. Removed
doesManageListeners(). (The new system isn't that new, but it's been
fixed and adapted to using RegistrationOrder.)
New
* Register all Bukkit events via the new EventRegistryBukkit.
* This way listeners can be ordered by numeric priority and tags
(regular expressions for beforeTag and afterTag).
* Unregistering listeners is possible (a listener node stays in the
Bukkit registry, but only one per event).
* It's possible to add listeners with minimal impact (MiniListener).
* The registry registers by event class 'hard' i.e., no relations
between already registered classes are checked.
* Order isn't necessarily stable nor even reproducible for randomized
start conditions with the same elements.
Point
* Compatibility hooks can easily place event listeners
before/after/between NCP default listeners, without resorting to tricks
like 'load-before'.
* Future registry of NCP itself: unregistering listeners is necessary
for runtime-removable checks, order is useful if not necessary, to be
able to add check listeners at any point of time.
Breaks:
* Anything relying on the previous (optional) managelisteners feature.
Missing:
* Lots of testing/inspection.
* Ensure all NCP listeners are coming with name/tag at least.
* Provide meaningful tags/RegistrationOrder for fine grained access
(e.g. after feature.moving but before feature.inventory).
* Change cncp to hard depend on NCP and use listener priorities.
* The annotations made for RegistrationOrder are defined in there.
* Add RegisterMethodWithOrder (method) for event handlers.
* Add RegisterEventsWithOrder (type) for default order with listeners.
