do not always recalculate existing multiblocks. Only do so,
if the adjacent block is either part of the multiblock's bounding
box, or if the adjacent block would be a valid part of the multiblock.
* Deduplicate crafting patterns internals
This should save a bit memory for duplicate patterns as well as speed up
loading new patterns on reloading chunks.
* Encode recipe id in pattern
* Use the recipe id directly for looking it up
* Slightly refactored how patterns encode whether they are crafting patterns or not.
* Renamed PatternHelper.
* Added APIs to encode patterns. (#4463)
* Removed ICraftingPatternItem.
* Renamed isCrafting to isCraftable
Co-authored-by: shartte <shartte@users.noreply.github.com>
Replaced fuzzy ints with a well defined enum
Moved API related cell interfaces to a new subpackage
Introduced a new client helper method to register a cell texture for a storage channel
Added a method to obtain the storage channel for a drive or chest slot
Added a new cell state for cells with stored items but neither types nor completely full
This better indicates power issue with dynamic consumption like
extracting/inserting items into a network. E.g. the terminals will
shutdown, no light/channel/etc indicators.
Due to the networks trying to extract from the largest possible energy
storage first, the network buffer will be used as last option after
energy cells or controllers.
Potentially needs evaluation in case of adhoc networks with a high
transfer rate locally. Increasing the local buffer might be an idea or
have it scale with the amount of local gridnodes. As long as it does not
invalidates energy cells.
* Updated Grid API
Added Nonull/Nullable/Nonnegative where applicable.
Deprecated unused methods
Changed concrete classes with interfaces (HashMap -> Map)
IGridTickable will no longer accept null as TickingRequest. Do not
implement it, if it should not tick.
* Renamed parameter to match conventions
* Adds a debug generator for ForgeEnergy.
This adds a simple debug block to test external energy injection via
ForgeEnergy and not rely solely on creative cells.
The energy has two functions. The first is acting as an infinite
ForgEnergy battery, this is unused by AE2 itself as nothing pulls
energy.
The second one is injecting energy every tick into any adjacent block
accepting ForgeEnergy.
The base generation is 8 per tick, but this is increased by the power of
adjacent TileEnergyGenerators.
* Remove obsolete api and fix some warning
* Move MEMonitorHandler to internal code and fix some warnings
* Rename exceptions to conform to naming scheme
* Use itemstack as itemdef
* HIGH_TAG/LOW_TAG should be compared both directions
* Remove getTagCompound
* Make Itemlist implementation independent
* Cache item id for performance reasons
* Add preconditions to saveguard against external meddling
* Chache itemDamage
* Remove IAEStackSearchKey for now, rename getDisplayStack
* Further improvments for the enerygrid.
Simplified quartz fiber.
Removed different Actionable code paths and replaced with a unified
approach.
Dropped lastProvider/Requester, it simply iterates over all, but still
drops them when necessary.
* Refactored the BaseActionSource
It now uses an interface `IActionSource` instead of a concrete class and further
subclasses.
Instead of relying on a specific class for a certain action type, it now
uses methods with Optional as return values to determine a player or
machine issuing an action. Refer to the JavaDocs for the exact behaviour.
* `IActionHost` no longer extends `IGridHost`
It never used the additional functionality and if needed the `IGridNode`
will also provide a reference to the corresponding `IGridHost`.
Due to most crafting related GUIs being hardcoded to `IGridHost`,
they no longer work when `IActionHost` is not extending it. Actually
`IActionHost` is the better solution for it, as it prevents us from looking
the grid up via the `IGridHost` and potentially finding a wrong grid.
* Interfaces now only lock the currently worked slot.
This allows `DualityInterface#onChangeInventory` to update slots of the same interface, should they be the source for a requested item and therefore updating their working set accordingly to queue further crafting requests for unfulfilled stack requirements.
There is currently no difference between an external grid and the local
machines trying to extract power from the grid. Thus preventing any
external grid from stealing power also does prevent machines from
extracting, if there is no other provider available.
Also a small change to use a priorityqueue with the lowest percentage
stored first. This can better utilize the upper limit by reaching it
faster and exiting earlier.
* Performance improvements for the energygrid
Reworked the old recursive approach to a queue based loop.
Extract will try to prefer the next grid with the hightest amount of
stored energy.
Inject will try to prefer the grid with the lowest percentage stored.
Other operations are first come, first serve.
* Added a local buffer storage to EnergyGrid
This replaces the old not really working buffer with a special
IAEPowerStorage acting as buffer/proxy for the local energy demand as
well as temporary overflow should something provide more energy than
requested.
Currently it set to hold a maximum of 200 AE (+ optional overflow until
consumed). It will only be used locally, no other grid can use it to
avoid starving the neighbor grids before finding a energy cell.
* Fixes IExternalPowerSink
All implementations currently depend on the network demand being a valid
source, which might not be true.
Further it can cause the sink to iterate the network twice (demand and
inject) and both again for simulate and modulate.
Also it did not return the actual leftover amount instead of relying on
the demand matching it.
* Minor fixes related to removing nodes from a grid.
The grid did remove IStackWatcherHost not IEnergyWatcherHost, this was
fine for AE2 as only level emitters use it and they implement both.
But not for potential addons.
Also they would potentually not being removed as the are indexed by the
gridnodes not the machine.
Fixes#1004
Sebastian Hartte