Merge tuinity (#6413)
This PR contains all of Tuinity's patches. Very notable ones are:
- Highly optimised collisions
- Optimised entity lookups by bounding box (Mojang made regressions in 1.17, this brings it back to 1.16)
- Starlight https://github.com/PaperMC/Starlight
- Rewritten dataconverter system https://github.com/PaperMC/DataConverter
- Random block ticking optimisation (wrongly dropped from Paper 1.17)
- Chunk ticking optimisations
- Anything else I've forgotten in the 60 or so patches
If you are a previous Tuinity user, your config will not migrate. You must do it yourself. The config options have simply been moved into paper.yml, so it will be an easy migration. However, please note that the chunk loading options in tuinity.yml are NOT compatible with the options in paper.yml.
* Port tuinity, initial patchset
* Update gradle to 7.2
jmp said it fixes rebuildpatches not working for me. it fucking better
* Completely clean apply
* Remove tuinity config, add per player api patch
* Remove paper reobf mappings patch
* Properly update gradlew
* Force clean rebuild
* Mark fixups
Comments and ATs still need to be done
* grep -r "Tuinity"
* Fixup
* Ensure gameprofile lastaccess is written only under the state lock
* update URL for dataconverter
* Only clean rebuild tuinity patches
might fix merge conflicts
* Use UTF-8 for gradlew
* Clean rb patches again
* Convert block ids used as item ids
Neither the converters of pre 1.13 nor DFU handled these cases,
as by the time they were written the game at the time didn't
consider these ids valid - they would be air. Because of this,
some worlds have logspam since only DataConverter (not DFU or
legacy converters) will warn when an invalid id has been
seen.
While quite a few do need to now be considered as air, quite a lot
do not. So it makes sense to add conversion for these items, instead
of simply suppressing or ignoring the logs. I've now added id -> string conversion
for all block ids that could be used as items that existed in the game
before 1.7.10 (I have no interest in tracking down the
exact version block ids stopped working) that were on
https://minecraft-ids.grahamedgecombe.com/
Items that did not directly convert to new items will
be instead converted to air: stems, wheat crops, piston head,
tripwire wire block
* Fix LightPopulated parsing in V1466
The DFU code was checking if the number existed, not if it
didn't exist. I misread the original code.
* Always parse protochunk light sources unless it is marked as non-lit
Chunks not marked as lit will always go through the light engine,
so they should always have their block sources parsed.
* Update custom names to JSON for players
Missed this fix from CB, as it was inside
the DataFixers class.
I decided to double check all of the CB changes again:
DataFixers.java was the only area I missed, as I had inspected all
datafixer diffs and implemented them all into DataConverter. I also
checked Bootstrap.java again, and re-evaluated their changes. I had
previously done this, but determined that they were all bad.
The changes to make standing_sign block map to oak_sign block in
V1450 is bad, because that's not the item id V1450 accepts. Only
in 1.14 did oak_sign even exist, and as expected there is a converter
to rename all existing sign items/blocks.
The fix to register the portal block under id 1440 is useless, as
the flattenning logic will default to the lowest registered id - which
is the exact blockstate that CB registers into 1440. So it just
doesn't do anything.
The extra item ids in the id -> string converter are already added,
but I found this from EMC originally.
The change for the spawn egg id 23 -> Arrow is just wrong,
that id DOES correspond to TippedArrow, NOT Arrow. As
expected, the spawn egg already has a dedicated mapping for
Arrow, which is id 10 - which was Arrow's entity id.
I also ported a fix for the cooked_fished id update. This doesn't
really matter since there is already a dataconverter to fix this,
but the game didn't accept cooked_fished at the time. So I see
no harm.
* Review all converters and walkers
- Refactor V99 to have helper methods for defining entity/tile
entity types
- Automatically namespace all ids that should be namespaced.
While vanilla never saved non-namespaced data for things that
are namespaced, plugins/users might have.
- Synchronised the identity ensure map in HelperBlockFlatteningV1450
- Code style consistency
- Add missing log warning in V102 for ITEM_NAME type conversion
- Use getBoolean instead of getByte
- Use ConverterAbstractEntityRename for V143 TippedArrow -> Arrow
rename, as it will affect ENTITY_NAME type
- Always set isVillager to false in V502 for Zombie
- Register V808's converter under subversion 1 like DFU
- Register a breakpoint for V1.17.1. In the future, all final
versions of major releases will have a breakpoint so that
the work required to determine if a converter needs a breakpoint
is minimal
- Validate that a dataconverter is only registered for a version
that is registered
- ConverterFlattenTileEntity is actually ConverterFlattenEntity
It even registered the converters under TILE_ENTITY, instead of
ENTITY.
- Fix id comparison in V1492 STRUCTURE_FEATURE renamer
- Use ConverterAbstractStatsRename for V1510 stats renamer
At the time I had written that class, the abstract renamer didn't
exist.
- Ensure OwnerUUID is at least set to empty string in
V1904 if the ocelot is converted to a cat (this is
likely so that it retains a collar)
- Use generic read/write for Records in V1946
Records is actually a list, not a map. So reading map was
invalid.
* Always set light to zero when propagating decrease
This fixes an almost infinite loop where light values
would be spam queued on a very small subset on blocks.
This also likely fixes the memory issues people were
seeing.
* re-organize patches
* Apply and fix conflicts
* Revert some patches
getChunkAt retains chunks so that plugins don't spam loads
revert mc-4 fix will remain unless issues pop up
* Shuffle iterated chunks if per player is not enabled
Can help with some mob spawning stacking up at locations
* Make per player default, migrate all configs
* Adjust comments in fixups
* Rework config for player chunk loader
Old config is not compatible. Move all configs to be
under `settings` in paper.yml
The player chunk loader has been modified to
less aggressively load chunks, but to send
chunks at higher rates compared to tuinity. There are
new config entries to tune this behavior.
* Add back old constructor to CompressionEncoder/Decoder (fixes
Tuinity #358)
* Raise chunk loading default limits
* Reduce worldgen thread workers for lower core count cpus
* Raise limits for chunk loading config
Also place it under `chunk-loading`
* Disable max chunk send rate by default
* Fix conflicts and rebuild patches
* Drop default send rate again
Appears to be still causing problems for no known reason
* Raise chunk send limits to 100 per player
While a low limit fixes ping issues for some people, most people
do not suffer from this issue and thus should not suffer from
an extremely slow load-in rate.
* Rebase part 1
Autosquash the fixups
* Move not implemented up
* Fixup mc-dev fixes
Missed this one
* Rebase per player viewdistance api into the original api patch
* Remove old light engine patch part 1
The prioritisation must be kept from it, so that part
has been rebased into the priority patch.
Part 2 will deal with rebasing all of the patches _after_
* Rebase remaining patches for old light patch removal
* Remove other mid tick patch
* Remove Optimize-PlayerChunkMap-memory-use-for-visibleChunks.patch
Replaced by `Do not copy visible chunks`
* Revert AT for Vec3i setX/Y/Z
The class is immutable. set should not be exposed
* Remove old IntegerUtil class
* Replace old CraftChunk#getEntities patch
* Remove import for SWMRNibbleArray in ChunkAccess
* Finished merge checklist
* Remove ensureTickThread impl in urgency patch
Co-authored-by: Spottedleaf <Spottedleaf@users.noreply.github.com>
Co-authored-by: Jason Penilla <11360596+jpenilla@users.noreply.github.com>
2021-08-31 11:02:11 +00:00
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Spottedleaf <Spottedleaf@users.noreply.github.com>
Date: Thu, 17 Jun 2021 19:55:02 -0700
Subject: [PATCH] Rewrite entity bounding box lookup calls
For whatever reason, Mojang thought it was OK to make this system
scale logn relative to the number of entity sections loaded.
On top of that, they do a hashtable lookup per section - before
this was just a basic array access.
This patch brings back entity slices for lookup only.
diff --git a/src/main/java/io/papermc/paper/world/ChunkEntitySlices.java b/src/main/java/io/papermc/paper/world/ChunkEntitySlices.java
new file mode 100644
index 0000000000000000000000000000000000000000..47b5f75d9f27cf3ab947fd1f69cbd609fb9f2749
--- /dev/null
+++ b/src/main/java/io/papermc/paper/world/ChunkEntitySlices.java
@@ -0,0 +1,500 @@
+package io.papermc.paper.world;
+
+import com.destroystokyo.paper.util.maplist.EntityList;
+import it.unimi.dsi.fastutil.objects.Reference2ObjectMap;
+import it.unimi.dsi.fastutil.objects.Reference2ObjectOpenHashMap;
+import net.minecraft.server.level.ChunkHolder;
+import net.minecraft.server.level.ServerLevel;
+import net.minecraft.util.Mth;
+import net.minecraft.world.entity.Entity;
+import net.minecraft.world.entity.EntityType;
+import net.minecraft.world.entity.boss.EnderDragonPart;
+import net.minecraft.world.entity.boss.enderdragon.EnderDragon;
+import net.minecraft.world.phys.AABB;
+import java.util.Arrays;
+import java.util.Iterator;
+import java.util.List;
+import java.util.function.Predicate;
+
+public final class ChunkEntitySlices {
+
+ protected final int minSection;
+ protected final int maxSection;
+ protected final int chunkX;
+ protected final int chunkZ;
+ protected final ServerLevel world;
+
+ protected final EntityCollectionBySection allEntities;
+ protected final EntityCollectionBySection hardCollidingEntities;
+ protected final Reference2ObjectOpenHashMap<Class<? extends Entity>, EntityCollectionBySection> entitiesByClass;
+ protected final EntityList entities = new EntityList();
+
+ public ChunkHolder.FullChunkStatus status;
+
+ // TODO implement container search optimisations
+
+ public ChunkEntitySlices(final ServerLevel world, final int chunkX, final int chunkZ, final ChunkHolder.FullChunkStatus status,
+ final int minSection, final int maxSection) { // inclusive, inclusive
+ this.minSection = minSection;
+ this.maxSection = maxSection;
+ this.chunkX = chunkX;
+ this.chunkZ = chunkZ;
+ this.world = world;
+
+ this.allEntities = new EntityCollectionBySection(this);
+ this.hardCollidingEntities = new EntityCollectionBySection(this);
+ this.entitiesByClass = new Reference2ObjectOpenHashMap<>();
+
+ this.status = status;
+ }
+
+ // Paper start - optimise CraftChunk#getEntities
+ public org.bukkit.entity.Entity[] getChunkEntities() {
+ List<org.bukkit.entity.Entity> ret = new java.util.ArrayList<>();
+ final Entity[] entities = this.entities.getRawData();
+ for (int i = 0, size = Math.min(entities.length, this.entities.size()); i < size; ++i) {
+ final Entity entity = entities[i];
+ if (entity == null) {
+ continue;
+ }
+ final org.bukkit.entity.Entity bukkit = entity.getBukkitEntity();
+ if (bukkit != null && bukkit.isValid()) {
+ ret.add(bukkit);
+ }
+ }
+
+ return ret.toArray(new org.bukkit.entity.Entity[0]);
+ }
+ // Paper end - optimise CraftChunk#getEntities
+
+ public boolean isEmpty() {
+ return this.entities.size() == 0;
+ }
+
+ private void updateTicketLevels() {
+ final Entity[] entities = this.entities.getRawData();
+ for (int i = 0, size = Math.min(entities.length, this.entities.size()); i < size; ++i) {
+ final Entity entity = entities[i];
+ entity.chunkStatus = this.status;
+ }
+ }
+
+ public synchronized void updateStatus(final ChunkHolder.FullChunkStatus status) {
+ this.status = status;
+ this.updateTicketLevels();
+ }
+
+ public synchronized void addEntity(final Entity entity, final int chunkSection) {
+ if (!this.entities.add(entity)) {
+ return;
+ }
+ entity.chunkStatus = this.status;
+ final int sectionIndex = chunkSection - this.minSection;
+
+ this.allEntities.addEntity(entity, sectionIndex);
+
+ if (entity.hardCollides()) {
+ this.hardCollidingEntities.addEntity(entity, sectionIndex);
+ }
+
+ for (final Iterator<Reference2ObjectMap.Entry<Class<? extends Entity>, EntityCollectionBySection>> iterator =
+ this.entitiesByClass.reference2ObjectEntrySet().fastIterator(); iterator.hasNext();) {
+ final Reference2ObjectMap.Entry<Class<? extends Entity>, EntityCollectionBySection> entry = iterator.next();
+
+ if (entry.getKey().isInstance(entity)) {
+ entry.getValue().addEntity(entity, sectionIndex);
+ }
+ }
+ }
+
+ public synchronized void removeEntity(final Entity entity, final int chunkSection) {
+ if (!this.entities.remove(entity)) {
+ return;
+ }
+ entity.chunkStatus = ChunkHolder.FullChunkStatus.INACCESSIBLE;
+ final int sectionIndex = chunkSection - this.minSection;
+
+ this.allEntities.removeEntity(entity, sectionIndex);
+
+ if (entity.hardCollides()) {
+ this.hardCollidingEntities.removeEntity(entity, sectionIndex);
+ }
+
+ for (final Iterator<Reference2ObjectMap.Entry<Class<? extends Entity>, EntityCollectionBySection>> iterator =
+ this.entitiesByClass.reference2ObjectEntrySet().fastIterator(); iterator.hasNext();) {
+ final Reference2ObjectMap.Entry<Class<? extends Entity>, EntityCollectionBySection> entry = iterator.next();
+
+ if (entry.getKey().isInstance(entity)) {
+ entry.getValue().removeEntity(entity, sectionIndex);
+ }
+ }
+ }
+
+ public void getHardCollidingEntities(final Entity except, final AABB box, final List<Entity> into, final Predicate<? super Entity> predicate) {
+ this.hardCollidingEntities.getEntities(except, box, into, predicate);
+ }
+
+ public void getEntities(final Entity except, final AABB box, final List<Entity> into, final Predicate<? super Entity> predicate) {
+ this.allEntities.getEntitiesWithEnderDragonParts(except, box, into, predicate);
+ }
+
+ public <T extends Entity> void getEntities(final EntityType<?> type, final AABB box, final List<? super T> into,
+ final Predicate<? super T> predicate) {
+ this.allEntities.getEntities(type, box, (List)into, (Predicate)predicate);
+ }
+
+ protected EntityCollectionBySection initClass(final Class<? extends Entity> clazz) {
+ final EntityCollectionBySection ret = new EntityCollectionBySection(this);
+
+ for (int sectionIndex = 0; sectionIndex < this.allEntities.entitiesBySection.length; ++sectionIndex) {
+ final BasicEntityList<Entity> sectionEntities = this.allEntities.entitiesBySection[sectionIndex];
+ if (sectionEntities == null) {
+ continue;
+ }
+
+ final Entity[] storage = sectionEntities.storage;
+
+ for (int i = 0, len = Math.min(storage.length, sectionEntities.size()); i < len; ++i) {
+ final Entity entity = storage[i];
+
+ if (clazz.isInstance(entity)) {
+ ret.addEntity(entity, sectionIndex);
+ }
+ }
+ }
+
+ return ret;
+ }
+
+ public <T extends Entity> void getEntities(final Class<? extends T> clazz, final Entity except, final AABB box, final List<? super T> into,
+ final Predicate<? super T> predicate) {
+ EntityCollectionBySection collection = this.entitiesByClass.get(clazz);
+ if (collection != null) {
+ collection.getEntitiesWithEnderDragonParts(except, clazz, box, (List)into, (Predicate)predicate);
+ } else {
+ synchronized (this) {
+ this.entitiesByClass.putIfAbsent(clazz, collection = this.initClass(clazz));
+ }
+ collection.getEntitiesWithEnderDragonParts(except, clazz, box, (List)into, (Predicate)predicate);
+ }
+ }
+
+ public synchronized void updateEntity(final Entity entity) {
+ /*// TODO
+ if (prev aabb != entity.getBoundingBox()) {
+ this.entityMap.delete(entity, prev aabb);
+ this.entityMap.insert(entity, prev aabb = entity.getBoundingBox());
+ }*/
+ }
+
+ protected static final class BasicEntityList<E extends Entity> {
+
+ protected static final Entity[] EMPTY = new Entity[0];
+ protected static final int DEFAULT_CAPACITY = 4;
+
+ protected E[] storage;
+ protected int size;
+
+ public BasicEntityList() {
+ this(0);
+ }
+
+ public BasicEntityList(final int cap) {
+ this.storage = (E[])(cap <= 0 ? EMPTY : new Entity[cap]);
+ }
+
+ public boolean isEmpty() {
+ return this.size == 0;
+ }
+
+ public int size() {
+ return this.size;
+ }
+
+ private void resize() {
+ if (this.storage == EMPTY) {
+ this.storage = (E[])new Entity[DEFAULT_CAPACITY];
+ } else {
+ this.storage = Arrays.copyOf(this.storage, this.storage.length * 2);
+ }
+ }
+
+ public void add(final E entity) {
+ final int idx = this.size++;
+ if (idx >= this.storage.length) {
+ this.resize();
+ this.storage[idx] = entity;
+ } else {
+ this.storage[idx] = entity;
+ }
+ }
+
+ public int indexOf(final E entity) {
+ final E[] storage = this.storage;
+
+ for (int i = 0, len = Math.min(this.storage.length, this.size); i < len; ++i) {
+ if (storage[i] == entity) {
+ return i;
+ }
+ }
+
+ return -1;
+ }
+
+ public boolean remove(final E entity) {
+ final int idx = this.indexOf(entity);
+ if (idx == -1) {
+ return false;
+ }
+
+ final int size = --this.size;
+ final E[] storage = this.storage;
+ if (idx != size) {
+ System.arraycopy(storage, idx + 1, storage, idx, size - idx);
+ }
+
+ storage[size] = null;
+
+ return true;
+ }
+
+ public boolean has(final E entity) {
+ return this.indexOf(entity) != -1;
+ }
+ }
+
+ protected static final class EntityCollectionBySection {
+
+ protected final ChunkEntitySlices manager;
+ protected final long[] nonEmptyBitset;
+ protected final BasicEntityList<Entity>[] entitiesBySection;
+ protected int count;
+
+ public EntityCollectionBySection(final ChunkEntitySlices manager) {
+ this.manager = manager;
+
+ final int sectionCount = manager.maxSection - manager.minSection + 1;
+
+ this.nonEmptyBitset = new long[(sectionCount + (Long.SIZE - 1)) >>> 6]; // (sectionCount + (Long.SIZE - 1)) / Long.SIZE
+ this.entitiesBySection = new BasicEntityList[sectionCount];
+ }
+
+ public void addEntity(final Entity entity, final int sectionIndex) {
+ BasicEntityList<Entity> list = this.entitiesBySection[sectionIndex];
+
+ if (list != null && list.has(entity)) {
+ return;
+ }
+
+ if (list == null) {
+ this.entitiesBySection[sectionIndex] = list = new BasicEntityList<>();
+ this.nonEmptyBitset[sectionIndex >>> 6] |= (1L << (sectionIndex & (Long.SIZE - 1)));
+ }
+
+ list.add(entity);
+ ++this.count;
+ }
+
+ public void removeEntity(final Entity entity, final int sectionIndex) {
+ final BasicEntityList<Entity> list = this.entitiesBySection[sectionIndex];
+
+ if (list == null || !list.remove(entity)) {
+ return;
+ }
+
+ --this.count;
+
+ if (list.isEmpty()) {
+ this.entitiesBySection[sectionIndex] = null;
+ this.nonEmptyBitset[sectionIndex >>> 6] ^= (1L << (sectionIndex & (Long.SIZE - 1)));
+ }
+ }
+
+ public void getEntities(final Entity except, final AABB box, final List<Entity> into, final Predicate<? super Entity> predicate) {
+ if (this.count == 0) {
+ return;
+ }
+
+ final int minSection = this.manager.minSection;
+ final int maxSection = this.manager.maxSection;
+
+ final int min = Mth.clamp(Mth.floor(box.minY - 2.0) >> 4, minSection, maxSection);
+ final int max = Mth.clamp(Mth.floor(box.maxY + 2.0) >> 4, minSection, maxSection);
+
+ // TODO use the bitset
+
+ final BasicEntityList<Entity>[] entitiesBySection = this.entitiesBySection;
+
+ for (int section = min; section <= max; ++section) {
+ final BasicEntityList<Entity> list = entitiesBySection[section - minSection];
+
+ if (list == null) {
+ continue;
+ }
+
+ final Entity[] storage = list.storage;
+
+ for (int i = 0, len = Math.min(storage.length, list.size()); i < len; ++i) {
+ final Entity entity = storage[i];
+
+ if (entity == null || entity == except || !entity.getBoundingBox().intersects(box)) {
+ continue;
+ }
+
+ if (predicate != null && !predicate.test(entity)) {
+ continue;
+ }
+
+ into.add(entity);
+ }
+ }
+ }
+
+ public void getEntitiesWithEnderDragonParts(final Entity except, final AABB box, final List<Entity> into,
+ final Predicate<? super Entity> predicate) {
+ if (this.count == 0) {
+ return;
+ }
+
+ final int minSection = this.manager.minSection;
+ final int maxSection = this.manager.maxSection;
+
+ final int min = Mth.clamp(Mth.floor(box.minY - 2.0) >> 4, minSection, maxSection);
+ final int max = Mth.clamp(Mth.floor(box.maxY + 2.0) >> 4, minSection, maxSection);
+
+ // TODO use the bitset
+
+ final BasicEntityList<Entity>[] entitiesBySection = this.entitiesBySection;
+
+ for (int section = min; section <= max; ++section) {
+ final BasicEntityList<Entity> list = entitiesBySection[section - minSection];
+
+ if (list == null) {
+ continue;
+ }
+
+ final Entity[] storage = list.storage;
+
+ for (int i = 0, len = Math.min(storage.length, list.size()); i < len; ++i) {
+ final Entity entity = storage[i];
+
+ if (entity == null || entity == except || !entity.getBoundingBox().intersects(box)) {
+ continue;
+ }
+
+ if (predicate == null || predicate.test(entity)) {
+ into.add(entity);
+ } // else: continue to test the ender dragon parts
+
+ if (entity instanceof EnderDragon) {
+ for (final EnderDragonPart part : ((EnderDragon)entity).subEntities) {
+ if (part == except || !part.getBoundingBox().intersects(box)) {
+ continue;
+ }
+
+ if (predicate != null && !predicate.test(part)) {
+ continue;
+ }
+
+ into.add(part);
+ }
+ }
+ }
+ }
+ }
+
+ public void getEntitiesWithEnderDragonParts(final Entity except, final Class<?> clazz, final AABB box, final List<Entity> into,
+ final Predicate<? super Entity> predicate) {
+ if (this.count == 0) {
+ return;
+ }
+
+ final int minSection = this.manager.minSection;
+ final int maxSection = this.manager.maxSection;
+
+ final int min = Mth.clamp(Mth.floor(box.minY - 2.0) >> 4, minSection, maxSection);
+ final int max = Mth.clamp(Mth.floor(box.maxY + 2.0) >> 4, minSection, maxSection);
+
+ // TODO use the bitset
+
+ final BasicEntityList<Entity>[] entitiesBySection = this.entitiesBySection;
+
+ for (int section = min; section <= max; ++section) {
+ final BasicEntityList<Entity> list = entitiesBySection[section - minSection];
+
+ if (list == null) {
+ continue;
+ }
+
+ final Entity[] storage = list.storage;
+
+ for (int i = 0, len = Math.min(storage.length, list.size()); i < len; ++i) {
+ final Entity entity = storage[i];
+
+ if (entity == null || entity == except || !entity.getBoundingBox().intersects(box)) {
+ continue;
+ }
+
+ if (predicate == null || predicate.test(entity)) {
+ into.add(entity);
+ } // else: continue to test the ender dragon parts
+
+ if (entity instanceof EnderDragon) {
+ for (final EnderDragonPart part : ((EnderDragon)entity).subEntities) {
+ if (part == except || !part.getBoundingBox().intersects(box) || !clazz.isInstance(part)) {
+ continue;
+ }
+
+ if (predicate != null && !predicate.test(part)) {
+ continue;
+ }
+
+ into.add(part);
+ }
+ }
+ }
+ }
+ }
+
+ public <T extends Entity> void getEntities(final EntityType<?> type, final AABB box, final List<? super T> into,
+ final Predicate<? super T> predicate) {
+ if (this.count == 0) {
+ return;
+ }
+
+ final int minSection = this.manager.minSection;
+ final int maxSection = this.manager.maxSection;
+
+ final int min = Mth.clamp(Mth.floor(box.minY - 2.0) >> 4, minSection, maxSection);
+ final int max = Mth.clamp(Mth.floor(box.maxY + 2.0) >> 4, minSection, maxSection);
+
+ // TODO use the bitset
+
+ final BasicEntityList<Entity>[] entitiesBySection = this.entitiesBySection;
+
+ for (int section = min; section <= max; ++section) {
+ final BasicEntityList<Entity> list = entitiesBySection[section - minSection];
+
+ if (list == null) {
+ continue;
+ }
+
+ final Entity[] storage = list.storage;
+
+ for (int i = 0, len = Math.min(storage.length, list.size()); i < len; ++i) {
+ final Entity entity = storage[i];
+
+ if (entity == null || (type != null && entity.getType() != type) || !entity.getBoundingBox().intersects(box)) {
+ continue;
+ }
+
+ if (predicate != null && !predicate.test((T)entity)) {
+ continue;
+ }
+
+ into.add((T)entity);
+ }
+ }
+ }
+ }
+}
diff --git a/src/main/java/io/papermc/paper/world/EntitySliceManager.java b/src/main/java/io/papermc/paper/world/EntitySliceManager.java
new file mode 100644
index 0000000000000000000000000000000000000000..3ba094e640d7fe7803e2bbdab8ff3beb6f50e8a0
--- /dev/null
+++ b/src/main/java/io/papermc/paper/world/EntitySliceManager.java
@@ -0,0 +1,391 @@
+package io.papermc.paper.world;
+
+import io.papermc.paper.util.CoordinateUtils;
+import io.papermc.paper.util.WorldUtil;
+import it.unimi.dsi.fastutil.longs.Long2ObjectOpenHashMap;
+import net.minecraft.core.BlockPos;
+import net.minecraft.server.level.ChunkHolder;
+import net.minecraft.server.level.ServerLevel;
+import net.minecraft.util.Mth;
+import net.minecraft.world.entity.Entity;
+import net.minecraft.world.entity.EntityType;
+import net.minecraft.world.phys.AABB;
+import java.util.List;
+import java.util.concurrent.locks.StampedLock;
+import java.util.function.Predicate;
+
+public final class EntitySliceManager {
+
+ protected static final int REGION_SHIFT = 5;
+ protected static final int REGION_MASK = (1 << REGION_SHIFT) - 1;
+ protected static final int REGION_SIZE = 1 << REGION_SHIFT;
+
+ public final ServerLevel world;
+
+ private final StampedLock stateLock = new StampedLock();
+ protected final Long2ObjectOpenHashMap<ChunkSlicesRegion> regions = new Long2ObjectOpenHashMap<>(64, 0.7f);
+
+ private final int minSection; // inclusive
+ private final int maxSection; // inclusive
+
+ protected final Long2ObjectOpenHashMap<ChunkHolder.FullChunkStatus> statusMap = new Long2ObjectOpenHashMap<>();
+ {
+ this.statusMap.defaultReturnValue(ChunkHolder.FullChunkStatus.INACCESSIBLE);
+ }
+
+ public EntitySliceManager(final ServerLevel world) {
+ this.world = world;
+ this.minSection = WorldUtil.getMinSection(world);
+ this.maxSection = WorldUtil.getMaxSection(world);
+ }
+
+ public void chunkStatusChange(final int x, final int z, final ChunkHolder.FullChunkStatus newStatus) {
+ if (newStatus == ChunkHolder.FullChunkStatus.INACCESSIBLE) {
+ this.statusMap.remove(CoordinateUtils.getChunkKey(x, z));
+ } else {
+ this.statusMap.put(CoordinateUtils.getChunkKey(x, z), newStatus);
+ final ChunkEntitySlices slices = this.getChunk(x, z);
+ if (slices != null) {
+ slices.updateStatus(newStatus);
+ }
+ }
+ }
+
+ public synchronized void addEntity(final Entity entity) {
+ final BlockPos pos = entity.blockPosition();
+ final int sectionX = pos.getX() >> 4;
+ final int sectionY = Mth.clamp(pos.getY() >> 4, this.minSection, this.maxSection);
+ final int sectionZ = pos.getZ() >> 4;
+ final ChunkEntitySlices slices = this.getOrCreateChunk(sectionX, sectionZ);
+ slices.addEntity(entity, sectionY);
+
+ entity.sectionX = sectionX;
+ entity.sectionY = sectionY;
+ entity.sectionZ = sectionZ;
+ }
+
+ public synchronized void removeEntity(final Entity entity) {
+ final ChunkEntitySlices slices = this.getChunk(entity.sectionX, entity.sectionZ);
+ slices.removeEntity(entity, entity.sectionY);
+ if (slices.isEmpty()) {
+ this.removeChunk(entity.sectionX, entity.sectionZ);
+ }
+ }
+
+ public void moveEntity(final Entity entity) {
+ final BlockPos newPos = entity.blockPosition();
+ final int newSectionX = newPos.getX() >> 4;
+ final int newSectionY = Mth.clamp(newPos.getY() >> 4, this.minSection, this.maxSection);
+ final int newSectionZ = newPos.getZ() >> 4;
+
+ if (newSectionX == entity.sectionX && newSectionY == entity.sectionY && newSectionZ == entity.sectionZ) {
+ return;
+ }
+
+ synchronized (this) {
+ // are we changing chunks?
+ if (newSectionX != entity.sectionX || newSectionZ != entity.sectionZ) {
+ final ChunkEntitySlices slices = this.getOrCreateChunk(newSectionX, newSectionZ);
+ final ChunkEntitySlices old = this.getChunk(entity.sectionX, entity.sectionZ);
+ synchronized (old) {
+ old.removeEntity(entity, entity.sectionY);
+ if (old.isEmpty()) {
+ this.removeChunk(entity.sectionX, entity.sectionZ);
+ }
+ }
+
+ synchronized (slices) {
+ slices.addEntity(entity, newSectionY);
+
+ entity.sectionX = newSectionX;
+ entity.sectionY = newSectionY;
+ entity.sectionZ = newSectionZ;
+ }
+ } else {
+ final ChunkEntitySlices slices = this.getChunk(newSectionX, newSectionZ);
+ // same chunk
+ synchronized (slices) {
+ slices.removeEntity(entity, entity.sectionY);
+ slices.addEntity(entity, newSectionY);
+ }
+ entity.sectionY = newSectionY;
+ }
+ }
+
+ }
+
+ public void getEntities(final Entity except, final AABB box, final List<Entity> into, final Predicate<? super Entity> predicate) {
+ final int minChunkX = (Mth.floor(box.minX) - 2) >> 4;
+ final int minChunkZ = (Mth.floor(box.minZ) - 2) >> 4;
+ final int maxChunkX = (Mth.floor(box.maxX) + 2) >> 4;
+ final int maxChunkZ = (Mth.floor(box.maxZ) + 2) >> 4;
+
+ final int minRegionX = minChunkX >> REGION_SHIFT;
+ final int minRegionZ = minChunkZ >> REGION_SHIFT;
+ final int maxRegionX = maxChunkX >> REGION_SHIFT;
+ final int maxRegionZ = maxChunkZ >> REGION_SHIFT;
+
+ for (int currRegionZ = minRegionZ; currRegionZ <= maxRegionZ; ++currRegionZ) {
+ final int minZ = currRegionZ == minRegionZ ? minChunkZ & REGION_MASK : 0;
+ final int maxZ = currRegionZ == maxRegionZ ? maxChunkZ & REGION_MASK : REGION_MASK;
+
+ for (int currRegionX = minRegionX; currRegionX <= maxRegionX; ++currRegionX) {
+ final ChunkSlicesRegion region = this.getRegion(currRegionX, currRegionZ);
+
+ if (region == null) {
+ continue;
+ }
+
+ final int minX = currRegionX == minRegionX ? minChunkX & REGION_MASK : 0;
+ final int maxX = currRegionX == maxRegionX ? maxChunkX & REGION_MASK : REGION_MASK;
+
+ for (int currZ = minZ; currZ <= maxZ; ++currZ) {
+ for (int currX = minX; currX <= maxX; ++currX) {
+ final ChunkEntitySlices chunk = region.get(currX | (currZ << REGION_SHIFT));
+ if (chunk == null || !chunk.status.isOrAfter(ChunkHolder.FullChunkStatus.BORDER)) {
+ continue;
+ }
+
+ chunk.getEntities(except, box, into, predicate);
+ }
+ }
+ }
+ }
+ }
+
+ public void getHardCollidingEntities(final Entity except, final AABB box, final List<Entity> into, final Predicate<? super Entity> predicate) {
+ final int minChunkX = (Mth.floor(box.minX) - 2) >> 4;
+ final int minChunkZ = (Mth.floor(box.minZ) - 2) >> 4;
+ final int maxChunkX = (Mth.floor(box.maxX) + 2) >> 4;
+ final int maxChunkZ = (Mth.floor(box.maxZ) + 2) >> 4;
+
+ final int minRegionX = minChunkX >> REGION_SHIFT;
+ final int minRegionZ = minChunkZ >> REGION_SHIFT;
+ final int maxRegionX = maxChunkX >> REGION_SHIFT;
+ final int maxRegionZ = maxChunkZ >> REGION_SHIFT;
+
+ for (int currRegionZ = minRegionZ; currRegionZ <= maxRegionZ; ++currRegionZ) {
+ final int minZ = currRegionZ == minRegionZ ? minChunkZ & REGION_MASK : 0;
+ final int maxZ = currRegionZ == maxRegionZ ? maxChunkZ & REGION_MASK : REGION_MASK;
+
+ for (int currRegionX = minRegionX; currRegionX <= maxRegionX; ++currRegionX) {
+ final ChunkSlicesRegion region = this.getRegion(currRegionX, currRegionZ);
+
+ if (region == null) {
+ continue;
+ }
+
+ final int minX = currRegionX == minRegionX ? minChunkX & REGION_MASK : 0;
+ final int maxX = currRegionX == maxRegionX ? maxChunkX & REGION_MASK : REGION_MASK;
+
+ for (int currZ = minZ; currZ <= maxZ; ++currZ) {
+ for (int currX = minX; currX <= maxX; ++currX) {
+ final ChunkEntitySlices chunk = region.get(currX | (currZ << REGION_SHIFT));
+ if (chunk == null || !chunk.status.isOrAfter(ChunkHolder.FullChunkStatus.BORDER)) {
+ continue;
+ }
+
+ chunk.getHardCollidingEntities(except, box, into, predicate);
+ }
+ }
+ }
+ }
+ }
+
+ public <T extends Entity> void getEntities(final EntityType<?> type, final AABB box, final List<? super T> into,
+ final Predicate<? super T> predicate) {
+ final int minChunkX = (Mth.floor(box.minX) - 2) >> 4;
+ final int minChunkZ = (Mth.floor(box.minZ) - 2) >> 4;
+ final int maxChunkX = (Mth.floor(box.maxX) + 2) >> 4;
+ final int maxChunkZ = (Mth.floor(box.maxZ) + 2) >> 4;
+
+ final int minRegionX = minChunkX >> REGION_SHIFT;
+ final int minRegionZ = minChunkZ >> REGION_SHIFT;
+ final int maxRegionX = maxChunkX >> REGION_SHIFT;
+ final int maxRegionZ = maxChunkZ >> REGION_SHIFT;
+
+ for (int currRegionZ = minRegionZ; currRegionZ <= maxRegionZ; ++currRegionZ) {
+ final int minZ = currRegionZ == minRegionZ ? minChunkZ & REGION_MASK : 0;
+ final int maxZ = currRegionZ == maxRegionZ ? maxChunkZ & REGION_MASK : REGION_MASK;
+
+ for (int currRegionX = minRegionX; currRegionX <= maxRegionX; ++currRegionX) {
+ final ChunkSlicesRegion region = this.getRegion(currRegionX, currRegionZ);
+
+ if (region == null) {
+ continue;
+ }
+
+ final int minX = currRegionX == minRegionX ? minChunkX & REGION_MASK : 0;
+ final int maxX = currRegionX == maxRegionX ? maxChunkX & REGION_MASK : REGION_MASK;
+
+ for (int currZ = minZ; currZ <= maxZ; ++currZ) {
+ for (int currX = minX; currX <= maxX; ++currX) {
+ final ChunkEntitySlices chunk = region.get(currX | (currZ << REGION_SHIFT));
+ if (chunk == null || !chunk.status.isOrAfter(ChunkHolder.FullChunkStatus.BORDER)) {
+ continue;
+ }
+
+ chunk.getEntities(type, box, (List)into, (Predicate)predicate);
+ }
+ }
+ }
+ }
+ }
+
+ public <T extends Entity> void getEntities(final Class<? extends T> clazz, final Entity except, final AABB box, final List<? super T> into,
+ final Predicate<? super T> predicate) {
+ final int minChunkX = (Mth.floor(box.minX) - 2) >> 4;
+ final int minChunkZ = (Mth.floor(box.minZ) - 2) >> 4;
+ final int maxChunkX = (Mth.floor(box.maxX) + 2) >> 4;
+ final int maxChunkZ = (Mth.floor(box.maxZ) + 2) >> 4;
+
+ final int minRegionX = minChunkX >> REGION_SHIFT;
+ final int minRegionZ = minChunkZ >> REGION_SHIFT;
+ final int maxRegionX = maxChunkX >> REGION_SHIFT;
+ final int maxRegionZ = maxChunkZ >> REGION_SHIFT;
+
+ for (int currRegionZ = minRegionZ; currRegionZ <= maxRegionZ; ++currRegionZ) {
+ final int minZ = currRegionZ == minRegionZ ? minChunkZ & REGION_MASK : 0;
+ final int maxZ = currRegionZ == maxRegionZ ? maxChunkZ & REGION_MASK : REGION_MASK;
+
+ for (int currRegionX = minRegionX; currRegionX <= maxRegionX; ++currRegionX) {
+ final ChunkSlicesRegion region = this.getRegion(currRegionX, currRegionZ);
+
+ if (region == null) {
+ continue;
+ }
+
+ final int minX = currRegionX == minRegionX ? minChunkX & REGION_MASK : 0;
+ final int maxX = currRegionX == maxRegionX ? maxChunkX & REGION_MASK : REGION_MASK;
+
+ for (int currZ = minZ; currZ <= maxZ; ++currZ) {
+ for (int currX = minX; currX <= maxX; ++currX) {
+ final ChunkEntitySlices chunk = region.get(currX | (currZ << REGION_SHIFT));
+ if (chunk == null || !chunk.status.isOrAfter(ChunkHolder.FullChunkStatus.BORDER)) {
+ continue;
+ }
+
+ chunk.getEntities(clazz, except, box, into, predicate);
+ }
+ }
+ }
+ }
+ }
+
+ public ChunkEntitySlices getChunk(final int chunkX, final int chunkZ) {
+ final ChunkSlicesRegion region = this.getRegion(chunkX >> REGION_SHIFT, chunkZ >> REGION_SHIFT);
+ if (region == null) {
+ return null;
+ }
+
+ return region.get((chunkX & REGION_MASK) | ((chunkZ & REGION_MASK) << REGION_SHIFT));
+ }
+
+ public ChunkEntitySlices getOrCreateChunk(final int chunkX, final int chunkZ) {
+ final ChunkSlicesRegion region = this.getRegion(chunkX >> REGION_SHIFT, chunkZ >> REGION_SHIFT);
+ ChunkEntitySlices ret;
+ if (region == null || (ret = region.get((chunkX & REGION_MASK) | ((chunkZ & REGION_MASK) << REGION_SHIFT))) == null) {
+ ret = new ChunkEntitySlices(this.world, chunkX, chunkZ, this.statusMap.get(CoordinateUtils.getChunkKey(chunkX, chunkZ)),
+ WorldUtil.getMinSection(this.world), WorldUtil.getMaxSection(this.world));
+
+ this.addChunk(chunkX, chunkZ, ret);
+
+ return ret;
+ }
+
+ return ret;
+ }
+
+ public ChunkSlicesRegion getRegion(final int regionX, final int regionZ) {
+ final long key = CoordinateUtils.getChunkKey(regionX, regionZ);
+ final long attempt = this.stateLock.tryOptimisticRead();
+ if (attempt != 0L) {
+ try {
+ final ChunkSlicesRegion ret = this.regions.get(key);
+
+ if (this.stateLock.validate(attempt)) {
+ return ret;
+ }
+ } catch (final Error error) {
+ throw error;
+ } catch (final Throwable thr) {
+ // ignore
+ }
+ }
+
+ this.stateLock.readLock();
+ try {
+ return this.regions.get(key);
+ } finally {
+ this.stateLock.tryUnlockRead();
+ }
+ }
+
+ public synchronized void removeChunk(final int chunkX, final int chunkZ) {
+ final long key = CoordinateUtils.getChunkKey(chunkX >> REGION_SHIFT, chunkZ >> REGION_SHIFT);
+ final int relIndex = (chunkX & REGION_MASK) | ((chunkZ & REGION_MASK) << REGION_SHIFT);
+
+ final ChunkSlicesRegion region = this.regions.get(key);
+ final int remaining = region.remove(relIndex);
+
+ if (remaining == 0) {
+ this.stateLock.writeLock();
+ try {
+ this.regions.remove(key);
+ } finally {
+ this.stateLock.tryUnlockWrite();
+ }
+ }
+ }
+
+ public synchronized void addChunk(final int chunkX, final int chunkZ, final ChunkEntitySlices slices) {
+ final long key = CoordinateUtils.getChunkKey(chunkX >> REGION_SHIFT, chunkZ >> REGION_SHIFT);
+ final int relIndex = (chunkX & REGION_MASK) | ((chunkZ & REGION_MASK) << REGION_SHIFT);
+
+ ChunkSlicesRegion region = this.regions.get(key);
+ if (region != null) {
+ region.add(relIndex, slices);
+ } else {
+ region = new ChunkSlicesRegion();
+ region.add(relIndex, slices);
+ this.stateLock.writeLock();
+ try {
+ this.regions.put(key, region);
+ } finally {
+ this.stateLock.tryUnlockWrite();
+ }
+ }
+ }
+
+ public static final class ChunkSlicesRegion {
+
+ protected final ChunkEntitySlices[] slices = new ChunkEntitySlices[REGION_SIZE * REGION_SIZE];
+ protected int sliceCount;
+
+ public ChunkEntitySlices get(final int index) {
+ return this.slices[index];
+ }
+
+ public int remove(final int index) {
+ final ChunkEntitySlices slices = this.slices[index];
+ if (slices == null) {
+ throw new IllegalStateException();
+ }
+
+ this.slices[index] = null;
+
+ return --this.sliceCount;
+ }
+
+ public void add(final int index, final ChunkEntitySlices slices) {
+ final ChunkEntitySlices curr = this.slices[index];
+ if (curr != null) {
+ throw new IllegalStateException();
+ }
+
+ this.slices[index] = slices;
+
+ ++this.sliceCount;
+ }
+ }
+}
diff --git a/src/main/java/net/minecraft/server/level/ServerLevel.java b/src/main/java/net/minecraft/server/level/ServerLevel.java
2021-11-15 19:01:19 +00:00
index ac3ef76f0db5a2d0f02b5e48a8bf511d9b0104d4..5c588c39de11bbabdc2f50ef4204007c622fdc6a 100644
Merge tuinity (#6413)
This PR contains all of Tuinity's patches. Very notable ones are:
- Highly optimised collisions
- Optimised entity lookups by bounding box (Mojang made regressions in 1.17, this brings it back to 1.16)
- Starlight https://github.com/PaperMC/Starlight
- Rewritten dataconverter system https://github.com/PaperMC/DataConverter
- Random block ticking optimisation (wrongly dropped from Paper 1.17)
- Chunk ticking optimisations
- Anything else I've forgotten in the 60 or so patches
If you are a previous Tuinity user, your config will not migrate. You must do it yourself. The config options have simply been moved into paper.yml, so it will be an easy migration. However, please note that the chunk loading options in tuinity.yml are NOT compatible with the options in paper.yml.
* Port tuinity, initial patchset
* Update gradle to 7.2
jmp said it fixes rebuildpatches not working for me. it fucking better
* Completely clean apply
* Remove tuinity config, add per player api patch
* Remove paper reobf mappings patch
* Properly update gradlew
* Force clean rebuild
* Mark fixups
Comments and ATs still need to be done
* grep -r "Tuinity"
* Fixup
* Ensure gameprofile lastaccess is written only under the state lock
* update URL for dataconverter
* Only clean rebuild tuinity patches
might fix merge conflicts
* Use UTF-8 for gradlew
* Clean rb patches again
* Convert block ids used as item ids
Neither the converters of pre 1.13 nor DFU handled these cases,
as by the time they were written the game at the time didn't
consider these ids valid - they would be air. Because of this,
some worlds have logspam since only DataConverter (not DFU or
legacy converters) will warn when an invalid id has been
seen.
While quite a few do need to now be considered as air, quite a lot
do not. So it makes sense to add conversion for these items, instead
of simply suppressing or ignoring the logs. I've now added id -> string conversion
for all block ids that could be used as items that existed in the game
before 1.7.10 (I have no interest in tracking down the
exact version block ids stopped working) that were on
https://minecraft-ids.grahamedgecombe.com/
Items that did not directly convert to new items will
be instead converted to air: stems, wheat crops, piston head,
tripwire wire block
* Fix LightPopulated parsing in V1466
The DFU code was checking if the number existed, not if it
didn't exist. I misread the original code.
* Always parse protochunk light sources unless it is marked as non-lit
Chunks not marked as lit will always go through the light engine,
so they should always have their block sources parsed.
* Update custom names to JSON for players
Missed this fix from CB, as it was inside
the DataFixers class.
I decided to double check all of the CB changes again:
DataFixers.java was the only area I missed, as I had inspected all
datafixer diffs and implemented them all into DataConverter. I also
checked Bootstrap.java again, and re-evaluated their changes. I had
previously done this, but determined that they were all bad.
The changes to make standing_sign block map to oak_sign block in
V1450 is bad, because that's not the item id V1450 accepts. Only
in 1.14 did oak_sign even exist, and as expected there is a converter
to rename all existing sign items/blocks.
The fix to register the portal block under id 1440 is useless, as
the flattenning logic will default to the lowest registered id - which
is the exact blockstate that CB registers into 1440. So it just
doesn't do anything.
The extra item ids in the id -> string converter are already added,
but I found this from EMC originally.
The change for the spawn egg id 23 -> Arrow is just wrong,
that id DOES correspond to TippedArrow, NOT Arrow. As
expected, the spawn egg already has a dedicated mapping for
Arrow, which is id 10 - which was Arrow's entity id.
I also ported a fix for the cooked_fished id update. This doesn't
really matter since there is already a dataconverter to fix this,
but the game didn't accept cooked_fished at the time. So I see
no harm.
* Review all converters and walkers
- Refactor V99 to have helper methods for defining entity/tile
entity types
- Automatically namespace all ids that should be namespaced.
While vanilla never saved non-namespaced data for things that
are namespaced, plugins/users might have.
- Synchronised the identity ensure map in HelperBlockFlatteningV1450
- Code style consistency
- Add missing log warning in V102 for ITEM_NAME type conversion
- Use getBoolean instead of getByte
- Use ConverterAbstractEntityRename for V143 TippedArrow -> Arrow
rename, as it will affect ENTITY_NAME type
- Always set isVillager to false in V502 for Zombie
- Register V808's converter under subversion 1 like DFU
- Register a breakpoint for V1.17.1. In the future, all final
versions of major releases will have a breakpoint so that
the work required to determine if a converter needs a breakpoint
is minimal
- Validate that a dataconverter is only registered for a version
that is registered
- ConverterFlattenTileEntity is actually ConverterFlattenEntity
It even registered the converters under TILE_ENTITY, instead of
ENTITY.
- Fix id comparison in V1492 STRUCTURE_FEATURE renamer
- Use ConverterAbstractStatsRename for V1510 stats renamer
At the time I had written that class, the abstract renamer didn't
exist.
- Ensure OwnerUUID is at least set to empty string in
V1904 if the ocelot is converted to a cat (this is
likely so that it retains a collar)
- Use generic read/write for Records in V1946
Records is actually a list, not a map. So reading map was
invalid.
* Always set light to zero when propagating decrease
This fixes an almost infinite loop where light values
would be spam queued on a very small subset on blocks.
This also likely fixes the memory issues people were
seeing.
* re-organize patches
* Apply and fix conflicts
* Revert some patches
getChunkAt retains chunks so that plugins don't spam loads
revert mc-4 fix will remain unless issues pop up
* Shuffle iterated chunks if per player is not enabled
Can help with some mob spawning stacking up at locations
* Make per player default, migrate all configs
* Adjust comments in fixups
* Rework config for player chunk loader
Old config is not compatible. Move all configs to be
under `settings` in paper.yml
The player chunk loader has been modified to
less aggressively load chunks, but to send
chunks at higher rates compared to tuinity. There are
new config entries to tune this behavior.
* Add back old constructor to CompressionEncoder/Decoder (fixes
Tuinity #358)
* Raise chunk loading default limits
* Reduce worldgen thread workers for lower core count cpus
* Raise limits for chunk loading config
Also place it under `chunk-loading`
* Disable max chunk send rate by default
* Fix conflicts and rebuild patches
* Drop default send rate again
Appears to be still causing problems for no known reason
* Raise chunk send limits to 100 per player
While a low limit fixes ping issues for some people, most people
do not suffer from this issue and thus should not suffer from
an extremely slow load-in rate.
* Rebase part 1
Autosquash the fixups
* Move not implemented up
* Fixup mc-dev fixes
Missed this one
* Rebase per player viewdistance api into the original api patch
* Remove old light engine patch part 1
The prioritisation must be kept from it, so that part
has been rebased into the priority patch.
Part 2 will deal with rebasing all of the patches _after_
* Rebase remaining patches for old light patch removal
* Remove other mid tick patch
* Remove Optimize-PlayerChunkMap-memory-use-for-visibleChunks.patch
Replaced by `Do not copy visible chunks`
* Revert AT for Vec3i setX/Y/Z
The class is immutable. set should not be exposed
* Remove old IntegerUtil class
* Replace old CraftChunk#getEntities patch
* Remove import for SWMRNibbleArray in ChunkAccess
* Finished merge checklist
* Remove ensureTickThread impl in urgency patch
Co-authored-by: Spottedleaf <Spottedleaf@users.noreply.github.com>
Co-authored-by: Jason Penilla <11360596+jpenilla@users.noreply.github.com>
2021-08-31 11:02:11 +00:00
--- a/src/main/java/net/minecraft/server/level/ServerLevel.java
+++ b/src/main/java/net/minecraft/server/level/ServerLevel.java
@@ -437,7 +437,7 @@ public class ServerLevel extends Level implements WorldGenLevel {
DataFixer datafixer = minecraftserver.getFixerUpper();
EntityPersistentStorage<Entity> entitypersistentstorage = new EntityStorage(this, new File(convertable_conversionsession.getDimensionPath(resourcekey), "entities"), datafixer, flag2, minecraftserver);
- this.entityManager = new PersistentEntitySectionManager<>(Entity.class, new ServerLevel.EntityCallbacks(), entitypersistentstorage);
+ this.entityManager = new PersistentEntitySectionManager<>(Entity.class, new ServerLevel.EntityCallbacks(), entitypersistentstorage, this.entitySliceManager); // Paper
StructureManager definedstructuremanager = minecraftserver.getStructureManager();
int j = this.spigotConfig.viewDistance; // Spigot
PersistentEntitySectionManager persistententitysectionmanager = this.entityManager;
diff --git a/src/main/java/net/minecraft/server/level/WorldGenRegion.java b/src/main/java/net/minecraft/server/level/WorldGenRegion.java
index 0f6b534a4c789a2f09f6c4624e5d58b99c7ed0e6..21d1e0c9c471e9e556b5bd70166a769b46105c7a 100644
--- a/src/main/java/net/minecraft/server/level/WorldGenRegion.java
+++ b/src/main/java/net/minecraft/server/level/WorldGenRegion.java
@@ -77,6 +77,23 @@ public class WorldGenRegion implements WorldGenLevel {
@Nullable
private Supplier<String> currentlyGenerating;
+ // Paper start
+ // No-op, this class doesn't provide entity access
+ @Override
+ public List<Entity> getHardCollidingEntities(Entity except, AABB box, Predicate<? super Entity> predicate) {
+ return Collections.emptyList();
+ }
+
+ @Override
+ public void getEntities(Entity except, AABB box, Predicate<? super Entity> predicate, List<Entity> into) {}
+
+ @Override
+ public void getHardCollidingEntities(Entity except, AABB box, Predicate<? super Entity> predicate, List<Entity> into) {}
+
+ @Override
+ public <T> void getEntitiesByClass(Class<? extends T> clazz, Entity except, AABB box, List<? super T> into, Predicate<? super T> predicate) {}
+ // Paper end
+
public WorldGenRegion(ServerLevel world, List<ChunkAccess> list, ChunkStatus chunkstatus, int i) {
this.generatingStatus = chunkstatus;
this.writeRadiusCutoff = i;
diff --git a/src/main/java/net/minecraft/world/entity/Entity.java b/src/main/java/net/minecraft/world/entity/Entity.java
2021-11-15 19:01:19 +00:00
index e9eacba3816ff3b76c51e9d3011c739b24864355..b7b802053c48c740161747c89cc55ade80094cb7 100644
Merge tuinity (#6413)
This PR contains all of Tuinity's patches. Very notable ones are:
- Highly optimised collisions
- Optimised entity lookups by bounding box (Mojang made regressions in 1.17, this brings it back to 1.16)
- Starlight https://github.com/PaperMC/Starlight
- Rewritten dataconverter system https://github.com/PaperMC/DataConverter
- Random block ticking optimisation (wrongly dropped from Paper 1.17)
- Chunk ticking optimisations
- Anything else I've forgotten in the 60 or so patches
If you are a previous Tuinity user, your config will not migrate. You must do it yourself. The config options have simply been moved into paper.yml, so it will be an easy migration. However, please note that the chunk loading options in tuinity.yml are NOT compatible with the options in paper.yml.
* Port tuinity, initial patchset
* Update gradle to 7.2
jmp said it fixes rebuildpatches not working for me. it fucking better
* Completely clean apply
* Remove tuinity config, add per player api patch
* Remove paper reobf mappings patch
* Properly update gradlew
* Force clean rebuild
* Mark fixups
Comments and ATs still need to be done
* grep -r "Tuinity"
* Fixup
* Ensure gameprofile lastaccess is written only under the state lock
* update URL for dataconverter
* Only clean rebuild tuinity patches
might fix merge conflicts
* Use UTF-8 for gradlew
* Clean rb patches again
* Convert block ids used as item ids
Neither the converters of pre 1.13 nor DFU handled these cases,
as by the time they were written the game at the time didn't
consider these ids valid - they would be air. Because of this,
some worlds have logspam since only DataConverter (not DFU or
legacy converters) will warn when an invalid id has been
seen.
While quite a few do need to now be considered as air, quite a lot
do not. So it makes sense to add conversion for these items, instead
of simply suppressing or ignoring the logs. I've now added id -> string conversion
for all block ids that could be used as items that existed in the game
before 1.7.10 (I have no interest in tracking down the
exact version block ids stopped working) that were on
https://minecraft-ids.grahamedgecombe.com/
Items that did not directly convert to new items will
be instead converted to air: stems, wheat crops, piston head,
tripwire wire block
* Fix LightPopulated parsing in V1466
The DFU code was checking if the number existed, not if it
didn't exist. I misread the original code.
* Always parse protochunk light sources unless it is marked as non-lit
Chunks not marked as lit will always go through the light engine,
so they should always have their block sources parsed.
* Update custom names to JSON for players
Missed this fix from CB, as it was inside
the DataFixers class.
I decided to double check all of the CB changes again:
DataFixers.java was the only area I missed, as I had inspected all
datafixer diffs and implemented them all into DataConverter. I also
checked Bootstrap.java again, and re-evaluated their changes. I had
previously done this, but determined that they were all bad.
The changes to make standing_sign block map to oak_sign block in
V1450 is bad, because that's not the item id V1450 accepts. Only
in 1.14 did oak_sign even exist, and as expected there is a converter
to rename all existing sign items/blocks.
The fix to register the portal block under id 1440 is useless, as
the flattenning logic will default to the lowest registered id - which
is the exact blockstate that CB registers into 1440. So it just
doesn't do anything.
The extra item ids in the id -> string converter are already added,
but I found this from EMC originally.
The change for the spawn egg id 23 -> Arrow is just wrong,
that id DOES correspond to TippedArrow, NOT Arrow. As
expected, the spawn egg already has a dedicated mapping for
Arrow, which is id 10 - which was Arrow's entity id.
I also ported a fix for the cooked_fished id update. This doesn't
really matter since there is already a dataconverter to fix this,
but the game didn't accept cooked_fished at the time. So I see
no harm.
* Review all converters and walkers
- Refactor V99 to have helper methods for defining entity/tile
entity types
- Automatically namespace all ids that should be namespaced.
While vanilla never saved non-namespaced data for things that
are namespaced, plugins/users might have.
- Synchronised the identity ensure map in HelperBlockFlatteningV1450
- Code style consistency
- Add missing log warning in V102 for ITEM_NAME type conversion
- Use getBoolean instead of getByte
- Use ConverterAbstractEntityRename for V143 TippedArrow -> Arrow
rename, as it will affect ENTITY_NAME type
- Always set isVillager to false in V502 for Zombie
- Register V808's converter under subversion 1 like DFU
- Register a breakpoint for V1.17.1. In the future, all final
versions of major releases will have a breakpoint so that
the work required to determine if a converter needs a breakpoint
is minimal
- Validate that a dataconverter is only registered for a version
that is registered
- ConverterFlattenTileEntity is actually ConverterFlattenEntity
It even registered the converters under TILE_ENTITY, instead of
ENTITY.
- Fix id comparison in V1492 STRUCTURE_FEATURE renamer
- Use ConverterAbstractStatsRename for V1510 stats renamer
At the time I had written that class, the abstract renamer didn't
exist.
- Ensure OwnerUUID is at least set to empty string in
V1904 if the ocelot is converted to a cat (this is
likely so that it retains a collar)
- Use generic read/write for Records in V1946
Records is actually a list, not a map. So reading map was
invalid.
* Always set light to zero when propagating decrease
This fixes an almost infinite loop where light values
would be spam queued on a very small subset on blocks.
This also likely fixes the memory issues people were
seeing.
* re-organize patches
* Apply and fix conflicts
* Revert some patches
getChunkAt retains chunks so that plugins don't spam loads
revert mc-4 fix will remain unless issues pop up
* Shuffle iterated chunks if per player is not enabled
Can help with some mob spawning stacking up at locations
* Make per player default, migrate all configs
* Adjust comments in fixups
* Rework config for player chunk loader
Old config is not compatible. Move all configs to be
under `settings` in paper.yml
The player chunk loader has been modified to
less aggressively load chunks, but to send
chunks at higher rates compared to tuinity. There are
new config entries to tune this behavior.
* Add back old constructor to CompressionEncoder/Decoder (fixes
Tuinity #358)
* Raise chunk loading default limits
* Reduce worldgen thread workers for lower core count cpus
* Raise limits for chunk loading config
Also place it under `chunk-loading`
* Disable max chunk send rate by default
* Fix conflicts and rebuild patches
* Drop default send rate again
Appears to be still causing problems for no known reason
* Raise chunk send limits to 100 per player
While a low limit fixes ping issues for some people, most people
do not suffer from this issue and thus should not suffer from
an extremely slow load-in rate.
* Rebase part 1
Autosquash the fixups
* Move not implemented up
* Fixup mc-dev fixes
Missed this one
* Rebase per player viewdistance api into the original api patch
* Remove old light engine patch part 1
The prioritisation must be kept from it, so that part
has been rebased into the priority patch.
Part 2 will deal with rebasing all of the patches _after_
* Rebase remaining patches for old light patch removal
* Remove other mid tick patch
* Remove Optimize-PlayerChunkMap-memory-use-for-visibleChunks.patch
Replaced by `Do not copy visible chunks`
* Revert AT for Vec3i setX/Y/Z
The class is immutable. set should not be exposed
* Remove old IntegerUtil class
* Replace old CraftChunk#getEntities patch
* Remove import for SWMRNibbleArray in ChunkAccess
* Finished merge checklist
* Remove ensureTickThread impl in urgency patch
Co-authored-by: Spottedleaf <Spottedleaf@users.noreply.github.com>
Co-authored-by: Jason Penilla <11360596+jpenilla@users.noreply.github.com>
2021-08-31 11:02:11 +00:00
--- a/src/main/java/net/minecraft/world/entity/Entity.java
+++ b/src/main/java/net/minecraft/world/entity/Entity.java
@@ -417,6 +417,56 @@ public abstract class Entity implements Nameable, EntityAccess, CommandSource, n
}
// Paper end - make end portalling safe
+ // Paper start
+ /**
+ * Overriding this field will cause memory leaks.
+ */
+ private final boolean hardCollides;
+
+ private static final java.util.Map<Class<? extends Entity>, Boolean> cachedOverrides = java.util.Collections.synchronizedMap(new java.util.WeakHashMap<>());
+ {
+ /* // Goodbye, broken on reobf...
+ Boolean hardCollides = cachedOverrides.get(this.getClass());
+ if (hardCollides == null) {
+ try {
+ java.lang.reflect.Method getHardCollisionBoxEntityMethod = Entity.class.getMethod("canCollideWith", Entity.class);
+ java.lang.reflect.Method hasHardCollisionBoxMethod = Entity.class.getMethod("canBeCollidedWith");
+ if (!this.getClass().getMethod(hasHardCollisionBoxMethod.getName(), hasHardCollisionBoxMethod.getParameterTypes()).equals(hasHardCollisionBoxMethod)
+ || !this.getClass().getMethod(getHardCollisionBoxEntityMethod.getName(), getHardCollisionBoxEntityMethod.getParameterTypes()).equals(getHardCollisionBoxEntityMethod)) {
+ hardCollides = Boolean.TRUE;
+ } else {
+ hardCollides = Boolean.FALSE;
+ }
+ cachedOverrides.put(this.getClass(), hardCollides);
+ }
+ catch (ThreadDeath thr) { throw thr; }
+ catch (Throwable thr) {
+ // shouldn't happen, just explode
+ throw new RuntimeException(thr);
+ }
+ } */
+ this.hardCollides = this instanceof Boat
+ || this instanceof net.minecraft.world.entity.monster.Shulker
+ || this instanceof net.minecraft.world.entity.vehicle.AbstractMinecart
+ || this.shouldHardCollide();
+ }
+
+ // plugins can override
+ protected boolean shouldHardCollide() {
+ return false;
+ }
+
+ public final boolean hardCollides() {
+ return this.hardCollides;
+ }
+
+ public net.minecraft.server.level.ChunkHolder.FullChunkStatus chunkStatus;
+
+ public int sectionX = Integer.MIN_VALUE;
+ public int sectionY = Integer.MIN_VALUE;
+ public int sectionZ = Integer.MIN_VALUE;
+ // Paper end
+
public Entity(EntityType<?> type, Level world) {
this.id = Entity.ENTITY_COUNTER.incrementAndGet();
this.passengers = ImmutableList.of();
@@ -2280,11 +2330,11 @@ public abstract class Entity implements Nameable, EntityAccess, CommandSource, n
return InteractionResult.PASS;
}
- public boolean canCollideWith(Entity other) {
+ public boolean canCollideWith(Entity other) { // Paper - diff on change, hard colliding entities override this - TODO CHECK ON UPDATE - AbstractMinecart/Boat override
return other.canBeCollidedWith() && !this.isPassengerOfSameVehicle(other);
}
- public boolean canBeCollidedWith() {
+ public boolean canBeCollidedWith() { // Paper - diff on change, hard colliding entities override this TODO CHECK ON UPDATE - Boat/Shulker override
return false;
}
diff --git a/src/main/java/net/minecraft/world/level/EntityGetter.java b/src/main/java/net/minecraft/world/level/EntityGetter.java
index 325e244c46ec208a2e7e18d71ccbbfcc25fc1bce..94130509e3a7980c378cc95c46821cf0fc753ce6 100644
--- a/src/main/java/net/minecraft/world/level/EntityGetter.java
+++ b/src/main/java/net/minecraft/world/level/EntityGetter.java
@@ -18,6 +18,18 @@ import net.minecraft.world.phys.shapes.Shapes;
import net.minecraft.world.phys.shapes.VoxelShape;
public interface EntityGetter {
+
+ // Paper start
+ List<Entity> getHardCollidingEntities(Entity except, AABB box, Predicate<? super Entity> predicate);
+
+ void getEntities(Entity except, AABB box, Predicate<? super Entity> predicate, List<Entity> into);
+
+ void getHardCollidingEntities(Entity except, AABB box, Predicate<? super Entity> predicate, List<Entity> into);
+
+ <T> void getEntitiesByClass(Class<? extends T> clazz, Entity except, final AABB box, List<? super T> into,
+ Predicate<? super T> predicate);
+ // Paper end
+
List<Entity> getEntities(@Nullable Entity except, AABB box, Predicate<? super Entity> predicate);
<T extends Entity> List<T> getEntities(EntityTypeTest<Entity, T> filter, AABB box, Predicate<? super T> predicate);
@@ -55,8 +67,8 @@ public interface EntityGetter {
return Stream.empty();
} else {
AABB aABB = box.inflate(1.0E-7D);
- return this.getEntities(entity, aABB, predicate.and((entityx) -> {
- if (entityx.getBoundingBox().intersects(aABB)) {
+ Predicate<Entity> hardCollides = (entityx) -> { // Paper - optimise entity hard collisions
+ if (true || entityx.getBoundingBox().intersects(aABB)) { // Paper - always true
if (entity == null) {
if (entityx.canBeCollidedWith()) {
return true;
@@ -67,7 +79,11 @@ public interface EntityGetter {
}
return false;
- })).stream().map(Entity::getBoundingBox).map(Shapes::create);
+ }; // Paper start - optimise entity hard collisions
+ predicate = predicate == null ? hardCollides : hardCollides.and(predicate);
+ return (entity != null && entity.hardCollides() ? this.getEntities(entity, aABB, predicate) : this.getHardCollidingEntities(entity, aABB, predicate))
+ .stream().map(Entity::getBoundingBox).map(Shapes::create);
+ // Paper end - optimise entity hard collisions
}
}
diff --git a/src/main/java/net/minecraft/world/level/Level.java b/src/main/java/net/minecraft/world/level/Level.java
2021-11-15 19:01:19 +00:00
index d0cfa782498eca8a13ce7d7f28d5e55ae4e402fa..1dd2e968bde16da2d2da63ca3c30515e1fd5b620 100644
Merge tuinity (#6413)
This PR contains all of Tuinity's patches. Very notable ones are:
- Highly optimised collisions
- Optimised entity lookups by bounding box (Mojang made regressions in 1.17, this brings it back to 1.16)
- Starlight https://github.com/PaperMC/Starlight
- Rewritten dataconverter system https://github.com/PaperMC/DataConverter
- Random block ticking optimisation (wrongly dropped from Paper 1.17)
- Chunk ticking optimisations
- Anything else I've forgotten in the 60 or so patches
If you are a previous Tuinity user, your config will not migrate. You must do it yourself. The config options have simply been moved into paper.yml, so it will be an easy migration. However, please note that the chunk loading options in tuinity.yml are NOT compatible with the options in paper.yml.
* Port tuinity, initial patchset
* Update gradle to 7.2
jmp said it fixes rebuildpatches not working for me. it fucking better
* Completely clean apply
* Remove tuinity config, add per player api patch
* Remove paper reobf mappings patch
* Properly update gradlew
* Force clean rebuild
* Mark fixups
Comments and ATs still need to be done
* grep -r "Tuinity"
* Fixup
* Ensure gameprofile lastaccess is written only under the state lock
* update URL for dataconverter
* Only clean rebuild tuinity patches
might fix merge conflicts
* Use UTF-8 for gradlew
* Clean rb patches again
* Convert block ids used as item ids
Neither the converters of pre 1.13 nor DFU handled these cases,
as by the time they were written the game at the time didn't
consider these ids valid - they would be air. Because of this,
some worlds have logspam since only DataConverter (not DFU or
legacy converters) will warn when an invalid id has been
seen.
While quite a few do need to now be considered as air, quite a lot
do not. So it makes sense to add conversion for these items, instead
of simply suppressing or ignoring the logs. I've now added id -> string conversion
for all block ids that could be used as items that existed in the game
before 1.7.10 (I have no interest in tracking down the
exact version block ids stopped working) that were on
https://minecraft-ids.grahamedgecombe.com/
Items that did not directly convert to new items will
be instead converted to air: stems, wheat crops, piston head,
tripwire wire block
* Fix LightPopulated parsing in V1466
The DFU code was checking if the number existed, not if it
didn't exist. I misread the original code.
* Always parse protochunk light sources unless it is marked as non-lit
Chunks not marked as lit will always go through the light engine,
so they should always have their block sources parsed.
* Update custom names to JSON for players
Missed this fix from CB, as it was inside
the DataFixers class.
I decided to double check all of the CB changes again:
DataFixers.java was the only area I missed, as I had inspected all
datafixer diffs and implemented them all into DataConverter. I also
checked Bootstrap.java again, and re-evaluated their changes. I had
previously done this, but determined that they were all bad.
The changes to make standing_sign block map to oak_sign block in
V1450 is bad, because that's not the item id V1450 accepts. Only
in 1.14 did oak_sign even exist, and as expected there is a converter
to rename all existing sign items/blocks.
The fix to register the portal block under id 1440 is useless, as
the flattenning logic will default to the lowest registered id - which
is the exact blockstate that CB registers into 1440. So it just
doesn't do anything.
The extra item ids in the id -> string converter are already added,
but I found this from EMC originally.
The change for the spawn egg id 23 -> Arrow is just wrong,
that id DOES correspond to TippedArrow, NOT Arrow. As
expected, the spawn egg already has a dedicated mapping for
Arrow, which is id 10 - which was Arrow's entity id.
I also ported a fix for the cooked_fished id update. This doesn't
really matter since there is already a dataconverter to fix this,
but the game didn't accept cooked_fished at the time. So I see
no harm.
* Review all converters and walkers
- Refactor V99 to have helper methods for defining entity/tile
entity types
- Automatically namespace all ids that should be namespaced.
While vanilla never saved non-namespaced data for things that
are namespaced, plugins/users might have.
- Synchronised the identity ensure map in HelperBlockFlatteningV1450
- Code style consistency
- Add missing log warning in V102 for ITEM_NAME type conversion
- Use getBoolean instead of getByte
- Use ConverterAbstractEntityRename for V143 TippedArrow -> Arrow
rename, as it will affect ENTITY_NAME type
- Always set isVillager to false in V502 for Zombie
- Register V808's converter under subversion 1 like DFU
- Register a breakpoint for V1.17.1. In the future, all final
versions of major releases will have a breakpoint so that
the work required to determine if a converter needs a breakpoint
is minimal
- Validate that a dataconverter is only registered for a version
that is registered
- ConverterFlattenTileEntity is actually ConverterFlattenEntity
It even registered the converters under TILE_ENTITY, instead of
ENTITY.
- Fix id comparison in V1492 STRUCTURE_FEATURE renamer
- Use ConverterAbstractStatsRename for V1510 stats renamer
At the time I had written that class, the abstract renamer didn't
exist.
- Ensure OwnerUUID is at least set to empty string in
V1904 if the ocelot is converted to a cat (this is
likely so that it retains a collar)
- Use generic read/write for Records in V1946
Records is actually a list, not a map. So reading map was
invalid.
* Always set light to zero when propagating decrease
This fixes an almost infinite loop where light values
would be spam queued on a very small subset on blocks.
This also likely fixes the memory issues people were
seeing.
* re-organize patches
* Apply and fix conflicts
* Revert some patches
getChunkAt retains chunks so that plugins don't spam loads
revert mc-4 fix will remain unless issues pop up
* Shuffle iterated chunks if per player is not enabled
Can help with some mob spawning stacking up at locations
* Make per player default, migrate all configs
* Adjust comments in fixups
* Rework config for player chunk loader
Old config is not compatible. Move all configs to be
under `settings` in paper.yml
The player chunk loader has been modified to
less aggressively load chunks, but to send
chunks at higher rates compared to tuinity. There are
new config entries to tune this behavior.
* Add back old constructor to CompressionEncoder/Decoder (fixes
Tuinity #358)
* Raise chunk loading default limits
* Reduce worldgen thread workers for lower core count cpus
* Raise limits for chunk loading config
Also place it under `chunk-loading`
* Disable max chunk send rate by default
* Fix conflicts and rebuild patches
* Drop default send rate again
Appears to be still causing problems for no known reason
* Raise chunk send limits to 100 per player
While a low limit fixes ping issues for some people, most people
do not suffer from this issue and thus should not suffer from
an extremely slow load-in rate.
* Rebase part 1
Autosquash the fixups
* Move not implemented up
* Fixup mc-dev fixes
Missed this one
* Rebase per player viewdistance api into the original api patch
* Remove old light engine patch part 1
The prioritisation must be kept from it, so that part
has been rebased into the priority patch.
Part 2 will deal with rebasing all of the patches _after_
* Rebase remaining patches for old light patch removal
* Remove other mid tick patch
* Remove Optimize-PlayerChunkMap-memory-use-for-visibleChunks.patch
Replaced by `Do not copy visible chunks`
* Revert AT for Vec3i setX/Y/Z
The class is immutable. set should not be exposed
* Remove old IntegerUtil class
* Replace old CraftChunk#getEntities patch
* Remove import for SWMRNibbleArray in ChunkAccess
* Finished merge checklist
* Remove ensureTickThread impl in urgency patch
Co-authored-by: Spottedleaf <Spottedleaf@users.noreply.github.com>
Co-authored-by: Jason Penilla <11360596+jpenilla@users.noreply.github.com>
2021-08-31 11:02:11 +00:00
--- a/src/main/java/net/minecraft/world/level/Level.java
+++ b/src/main/java/net/minecraft/world/level/Level.java
2021-10-05 03:14:11 +00:00
@@ -204,6 +204,50 @@ public abstract class Level implements LevelAccessor, AutoCloseable {
Merge tuinity (#6413)
This PR contains all of Tuinity's patches. Very notable ones are:
- Highly optimised collisions
- Optimised entity lookups by bounding box (Mojang made regressions in 1.17, this brings it back to 1.16)
- Starlight https://github.com/PaperMC/Starlight
- Rewritten dataconverter system https://github.com/PaperMC/DataConverter
- Random block ticking optimisation (wrongly dropped from Paper 1.17)
- Chunk ticking optimisations
- Anything else I've forgotten in the 60 or so patches
If you are a previous Tuinity user, your config will not migrate. You must do it yourself. The config options have simply been moved into paper.yml, so it will be an easy migration. However, please note that the chunk loading options in tuinity.yml are NOT compatible with the options in paper.yml.
* Port tuinity, initial patchset
* Update gradle to 7.2
jmp said it fixes rebuildpatches not working for me. it fucking better
* Completely clean apply
* Remove tuinity config, add per player api patch
* Remove paper reobf mappings patch
* Properly update gradlew
* Force clean rebuild
* Mark fixups
Comments and ATs still need to be done
* grep -r "Tuinity"
* Fixup
* Ensure gameprofile lastaccess is written only under the state lock
* update URL for dataconverter
* Only clean rebuild tuinity patches
might fix merge conflicts
* Use UTF-8 for gradlew
* Clean rb patches again
* Convert block ids used as item ids
Neither the converters of pre 1.13 nor DFU handled these cases,
as by the time they were written the game at the time didn't
consider these ids valid - they would be air. Because of this,
some worlds have logspam since only DataConverter (not DFU or
legacy converters) will warn when an invalid id has been
seen.
While quite a few do need to now be considered as air, quite a lot
do not. So it makes sense to add conversion for these items, instead
of simply suppressing or ignoring the logs. I've now added id -> string conversion
for all block ids that could be used as items that existed in the game
before 1.7.10 (I have no interest in tracking down the
exact version block ids stopped working) that were on
https://minecraft-ids.grahamedgecombe.com/
Items that did not directly convert to new items will
be instead converted to air: stems, wheat crops, piston head,
tripwire wire block
* Fix LightPopulated parsing in V1466
The DFU code was checking if the number existed, not if it
didn't exist. I misread the original code.
* Always parse protochunk light sources unless it is marked as non-lit
Chunks not marked as lit will always go through the light engine,
so they should always have their block sources parsed.
* Update custom names to JSON for players
Missed this fix from CB, as it was inside
the DataFixers class.
I decided to double check all of the CB changes again:
DataFixers.java was the only area I missed, as I had inspected all
datafixer diffs and implemented them all into DataConverter. I also
checked Bootstrap.java again, and re-evaluated their changes. I had
previously done this, but determined that they were all bad.
The changes to make standing_sign block map to oak_sign block in
V1450 is bad, because that's not the item id V1450 accepts. Only
in 1.14 did oak_sign even exist, and as expected there is a converter
to rename all existing sign items/blocks.
The fix to register the portal block under id 1440 is useless, as
the flattenning logic will default to the lowest registered id - which
is the exact blockstate that CB registers into 1440. So it just
doesn't do anything.
The extra item ids in the id -> string converter are already added,
but I found this from EMC originally.
The change for the spawn egg id 23 -> Arrow is just wrong,
that id DOES correspond to TippedArrow, NOT Arrow. As
expected, the spawn egg already has a dedicated mapping for
Arrow, which is id 10 - which was Arrow's entity id.
I also ported a fix for the cooked_fished id update. This doesn't
really matter since there is already a dataconverter to fix this,
but the game didn't accept cooked_fished at the time. So I see
no harm.
* Review all converters and walkers
- Refactor V99 to have helper methods for defining entity/tile
entity types
- Automatically namespace all ids that should be namespaced.
While vanilla never saved non-namespaced data for things that
are namespaced, plugins/users might have.
- Synchronised the identity ensure map in HelperBlockFlatteningV1450
- Code style consistency
- Add missing log warning in V102 for ITEM_NAME type conversion
- Use getBoolean instead of getByte
- Use ConverterAbstractEntityRename for V143 TippedArrow -> Arrow
rename, as it will affect ENTITY_NAME type
- Always set isVillager to false in V502 for Zombie
- Register V808's converter under subversion 1 like DFU
- Register a breakpoint for V1.17.1. In the future, all final
versions of major releases will have a breakpoint so that
the work required to determine if a converter needs a breakpoint
is minimal
- Validate that a dataconverter is only registered for a version
that is registered
- ConverterFlattenTileEntity is actually ConverterFlattenEntity
It even registered the converters under TILE_ENTITY, instead of
ENTITY.
- Fix id comparison in V1492 STRUCTURE_FEATURE renamer
- Use ConverterAbstractStatsRename for V1510 stats renamer
At the time I had written that class, the abstract renamer didn't
exist.
- Ensure OwnerUUID is at least set to empty string in
V1904 if the ocelot is converted to a cat (this is
likely so that it retains a collar)
- Use generic read/write for Records in V1946
Records is actually a list, not a map. So reading map was
invalid.
* Always set light to zero when propagating decrease
This fixes an almost infinite loop where light values
would be spam queued on a very small subset on blocks.
This also likely fixes the memory issues people were
seeing.
* re-organize patches
* Apply and fix conflicts
* Revert some patches
getChunkAt retains chunks so that plugins don't spam loads
revert mc-4 fix will remain unless issues pop up
* Shuffle iterated chunks if per player is not enabled
Can help with some mob spawning stacking up at locations
* Make per player default, migrate all configs
* Adjust comments in fixups
* Rework config for player chunk loader
Old config is not compatible. Move all configs to be
under `settings` in paper.yml
The player chunk loader has been modified to
less aggressively load chunks, but to send
chunks at higher rates compared to tuinity. There are
new config entries to tune this behavior.
* Add back old constructor to CompressionEncoder/Decoder (fixes
Tuinity #358)
* Raise chunk loading default limits
* Reduce worldgen thread workers for lower core count cpus
* Raise limits for chunk loading config
Also place it under `chunk-loading`
* Disable max chunk send rate by default
* Fix conflicts and rebuild patches
* Drop default send rate again
Appears to be still causing problems for no known reason
* Raise chunk send limits to 100 per player
While a low limit fixes ping issues for some people, most people
do not suffer from this issue and thus should not suffer from
an extremely slow load-in rate.
* Rebase part 1
Autosquash the fixups
* Move not implemented up
* Fixup mc-dev fixes
Missed this one
* Rebase per player viewdistance api into the original api patch
* Remove old light engine patch part 1
The prioritisation must be kept from it, so that part
has been rebased into the priority patch.
Part 2 will deal with rebasing all of the patches _after_
* Rebase remaining patches for old light patch removal
* Remove other mid tick patch
* Remove Optimize-PlayerChunkMap-memory-use-for-visibleChunks.patch
Replaced by `Do not copy visible chunks`
* Revert AT for Vec3i setX/Y/Z
The class is immutable. set should not be exposed
* Remove old IntegerUtil class
* Replace old CraftChunk#getEntities patch
* Remove import for SWMRNibbleArray in ChunkAccess
* Finished merge checklist
* Remove ensureTickThread impl in urgency patch
Co-authored-by: Spottedleaf <Spottedleaf@users.noreply.github.com>
Co-authored-by: Jason Penilla <11360596+jpenilla@users.noreply.github.com>
2021-08-31 11:02:11 +00:00
return this.typeKey;
}
+ // Paper start
+ protected final io.papermc.paper.world.EntitySliceManager entitySliceManager;
+
+ // Paper start - optimise CraftChunk#getEntities
+ public org.bukkit.entity.Entity[] getChunkEntities(int chunkX, int chunkZ) {
+ io.papermc.paper.world.ChunkEntitySlices slices = this.entitySliceManager.getChunk(chunkX, chunkZ);
+ if (slices == null) {
+ return new org.bukkit.entity.Entity[0];
+ }
+ return slices.getChunkEntities();
+ }
+ // Paper end - optimise CraftChunk#getEntities
+
+ @Override
+ public List<Entity> getHardCollidingEntities(Entity except, AABB box, Predicate<? super Entity> predicate) {
+ List<Entity> ret = new java.util.ArrayList<>();
+ this.entitySliceManager.getEntities(except, box, ret, predicate);
+ return ret;
+ }
+
+ @Override
+ public void getEntities(Entity except, AABB box, Predicate<? super Entity> predicate, List<Entity> into) {
+ this.entitySliceManager.getEntities(except, box, into, predicate);
+ }
+
+ @Override
+ public void getHardCollidingEntities(Entity except, AABB box, Predicate<? super Entity> predicate, List<Entity> into) {
+ this.entitySliceManager.getHardCollidingEntities(except, box, into, predicate);
+ }
+
+ @Override
+ public <T> void getEntitiesByClass(Class<? extends T> clazz, Entity except, final AABB box, List<? super T> into,
+ Predicate<? super T> predicate) {
+ this.entitySliceManager.getEntities((Class)clazz, except, box, (List)into, (Predicate)predicate);
+ }
+
+ @Override
+ public <T extends Entity> List<T> getEntitiesOfClass(Class<T> entityClass, AABB box, Predicate<? super T> predicate) {
+ List<T> ret = new java.util.ArrayList<>();
+ this.entitySliceManager.getEntities(entityClass, null, box, ret, predicate);
+ return ret;
+ }
+ // Paper end
+
protected Level(WritableLevelData worlddatamutable, ResourceKey<Level> resourcekey, final DimensionType dimensionmanager, Supplier<ProfilerFiller> supplier, boolean flag, boolean flag1, long i, org.bukkit.generator.ChunkGenerator gen, org.bukkit.generator.BiomeProvider biomeProvider, org.bukkit.World.Environment env, java.util.concurrent.Executor executor) { // Paper - Anti-Xray - Pass executor
this.spigotConfig = new org.spigotmc.SpigotWorldConfig(((net.minecraft.world.level.storage.PrimaryLevelData) worlddatamutable).getLevelName()); // Spigot
this.paperConfig = new com.destroystokyo.paper.PaperWorldConfig(((net.minecraft.world.level.storage.PrimaryLevelData) worlddatamutable).getLevelName(), this.spigotConfig); // Paper
2021-10-05 03:14:11 +00:00
@@ -281,6 +325,7 @@ public abstract class Level implements LevelAccessor, AutoCloseable {
Merge tuinity (#6413)
This PR contains all of Tuinity's patches. Very notable ones are:
- Highly optimised collisions
- Optimised entity lookups by bounding box (Mojang made regressions in 1.17, this brings it back to 1.16)
- Starlight https://github.com/PaperMC/Starlight
- Rewritten dataconverter system https://github.com/PaperMC/DataConverter
- Random block ticking optimisation (wrongly dropped from Paper 1.17)
- Chunk ticking optimisations
- Anything else I've forgotten in the 60 or so patches
If you are a previous Tuinity user, your config will not migrate. You must do it yourself. The config options have simply been moved into paper.yml, so it will be an easy migration. However, please note that the chunk loading options in tuinity.yml are NOT compatible with the options in paper.yml.
* Port tuinity, initial patchset
* Update gradle to 7.2
jmp said it fixes rebuildpatches not working for me. it fucking better
* Completely clean apply
* Remove tuinity config, add per player api patch
* Remove paper reobf mappings patch
* Properly update gradlew
* Force clean rebuild
* Mark fixups
Comments and ATs still need to be done
* grep -r "Tuinity"
* Fixup
* Ensure gameprofile lastaccess is written only under the state lock
* update URL for dataconverter
* Only clean rebuild tuinity patches
might fix merge conflicts
* Use UTF-8 for gradlew
* Clean rb patches again
* Convert block ids used as item ids
Neither the converters of pre 1.13 nor DFU handled these cases,
as by the time they were written the game at the time didn't
consider these ids valid - they would be air. Because of this,
some worlds have logspam since only DataConverter (not DFU or
legacy converters) will warn when an invalid id has been
seen.
While quite a few do need to now be considered as air, quite a lot
do not. So it makes sense to add conversion for these items, instead
of simply suppressing or ignoring the logs. I've now added id -> string conversion
for all block ids that could be used as items that existed in the game
before 1.7.10 (I have no interest in tracking down the
exact version block ids stopped working) that were on
https://minecraft-ids.grahamedgecombe.com/
Items that did not directly convert to new items will
be instead converted to air: stems, wheat crops, piston head,
tripwire wire block
* Fix LightPopulated parsing in V1466
The DFU code was checking if the number existed, not if it
didn't exist. I misread the original code.
* Always parse protochunk light sources unless it is marked as non-lit
Chunks not marked as lit will always go through the light engine,
so they should always have their block sources parsed.
* Update custom names to JSON for players
Missed this fix from CB, as it was inside
the DataFixers class.
I decided to double check all of the CB changes again:
DataFixers.java was the only area I missed, as I had inspected all
datafixer diffs and implemented them all into DataConverter. I also
checked Bootstrap.java again, and re-evaluated their changes. I had
previously done this, but determined that they were all bad.
The changes to make standing_sign block map to oak_sign block in
V1450 is bad, because that's not the item id V1450 accepts. Only
in 1.14 did oak_sign even exist, and as expected there is a converter
to rename all existing sign items/blocks.
The fix to register the portal block under id 1440 is useless, as
the flattenning logic will default to the lowest registered id - which
is the exact blockstate that CB registers into 1440. So it just
doesn't do anything.
The extra item ids in the id -> string converter are already added,
but I found this from EMC originally.
The change for the spawn egg id 23 -> Arrow is just wrong,
that id DOES correspond to TippedArrow, NOT Arrow. As
expected, the spawn egg already has a dedicated mapping for
Arrow, which is id 10 - which was Arrow's entity id.
I also ported a fix for the cooked_fished id update. This doesn't
really matter since there is already a dataconverter to fix this,
but the game didn't accept cooked_fished at the time. So I see
no harm.
* Review all converters and walkers
- Refactor V99 to have helper methods for defining entity/tile
entity types
- Automatically namespace all ids that should be namespaced.
While vanilla never saved non-namespaced data for things that
are namespaced, plugins/users might have.
- Synchronised the identity ensure map in HelperBlockFlatteningV1450
- Code style consistency
- Add missing log warning in V102 for ITEM_NAME type conversion
- Use getBoolean instead of getByte
- Use ConverterAbstractEntityRename for V143 TippedArrow -> Arrow
rename, as it will affect ENTITY_NAME type
- Always set isVillager to false in V502 for Zombie
- Register V808's converter under subversion 1 like DFU
- Register a breakpoint for V1.17.1. In the future, all final
versions of major releases will have a breakpoint so that
the work required to determine if a converter needs a breakpoint
is minimal
- Validate that a dataconverter is only registered for a version
that is registered
- ConverterFlattenTileEntity is actually ConverterFlattenEntity
It even registered the converters under TILE_ENTITY, instead of
ENTITY.
- Fix id comparison in V1492 STRUCTURE_FEATURE renamer
- Use ConverterAbstractStatsRename for V1510 stats renamer
At the time I had written that class, the abstract renamer didn't
exist.
- Ensure OwnerUUID is at least set to empty string in
V1904 if the ocelot is converted to a cat (this is
likely so that it retains a collar)
- Use generic read/write for Records in V1946
Records is actually a list, not a map. So reading map was
invalid.
* Always set light to zero when propagating decrease
This fixes an almost infinite loop where light values
would be spam queued on a very small subset on blocks.
This also likely fixes the memory issues people were
seeing.
* re-organize patches
* Apply and fix conflicts
* Revert some patches
getChunkAt retains chunks so that plugins don't spam loads
revert mc-4 fix will remain unless issues pop up
* Shuffle iterated chunks if per player is not enabled
Can help with some mob spawning stacking up at locations
* Make per player default, migrate all configs
* Adjust comments in fixups
* Rework config for player chunk loader
Old config is not compatible. Move all configs to be
under `settings` in paper.yml
The player chunk loader has been modified to
less aggressively load chunks, but to send
chunks at higher rates compared to tuinity. There are
new config entries to tune this behavior.
* Add back old constructor to CompressionEncoder/Decoder (fixes
Tuinity #358)
* Raise chunk loading default limits
* Reduce worldgen thread workers for lower core count cpus
* Raise limits for chunk loading config
Also place it under `chunk-loading`
* Disable max chunk send rate by default
* Fix conflicts and rebuild patches
* Drop default send rate again
Appears to be still causing problems for no known reason
* Raise chunk send limits to 100 per player
While a low limit fixes ping issues for some people, most people
do not suffer from this issue and thus should not suffer from
an extremely slow load-in rate.
* Rebase part 1
Autosquash the fixups
* Move not implemented up
* Fixup mc-dev fixes
Missed this one
* Rebase per player viewdistance api into the original api patch
* Remove old light engine patch part 1
The prioritisation must be kept from it, so that part
has been rebased into the priority patch.
Part 2 will deal with rebasing all of the patches _after_
* Rebase remaining patches for old light patch removal
* Remove other mid tick patch
* Remove Optimize-PlayerChunkMap-memory-use-for-visibleChunks.patch
Replaced by `Do not copy visible chunks`
* Revert AT for Vec3i setX/Y/Z
The class is immutable. set should not be exposed
* Remove old IntegerUtil class
* Replace old CraftChunk#getEntities patch
* Remove import for SWMRNibbleArray in ChunkAccess
* Finished merge checklist
* Remove ensureTickThread impl in urgency patch
Co-authored-by: Spottedleaf <Spottedleaf@users.noreply.github.com>
Co-authored-by: Jason Penilla <11360596+jpenilla@users.noreply.github.com>
2021-08-31 11:02:11 +00:00
this.chunkPacketBlockController = this.paperConfig.antiXray ?
new com.destroystokyo.paper.antixray.ChunkPacketBlockControllerAntiXray(this, executor)
: com.destroystokyo.paper.antixray.ChunkPacketBlockController.NO_OPERATION_INSTANCE; // Paper - Anti-Xray
+ this.entitySliceManager = new io.papermc.paper.world.EntitySliceManager((ServerLevel)this); // Paper
}
// Paper start
2021-11-15 19:01:19 +00:00
@@ -987,26 +1032,7 @@ public abstract class Level implements LevelAccessor, AutoCloseable {
Merge tuinity (#6413)
This PR contains all of Tuinity's patches. Very notable ones are:
- Highly optimised collisions
- Optimised entity lookups by bounding box (Mojang made regressions in 1.17, this brings it back to 1.16)
- Starlight https://github.com/PaperMC/Starlight
- Rewritten dataconverter system https://github.com/PaperMC/DataConverter
- Random block ticking optimisation (wrongly dropped from Paper 1.17)
- Chunk ticking optimisations
- Anything else I've forgotten in the 60 or so patches
If you are a previous Tuinity user, your config will not migrate. You must do it yourself. The config options have simply been moved into paper.yml, so it will be an easy migration. However, please note that the chunk loading options in tuinity.yml are NOT compatible with the options in paper.yml.
* Port tuinity, initial patchset
* Update gradle to 7.2
jmp said it fixes rebuildpatches not working for me. it fucking better
* Completely clean apply
* Remove tuinity config, add per player api patch
* Remove paper reobf mappings patch
* Properly update gradlew
* Force clean rebuild
* Mark fixups
Comments and ATs still need to be done
* grep -r "Tuinity"
* Fixup
* Ensure gameprofile lastaccess is written only under the state lock
* update URL for dataconverter
* Only clean rebuild tuinity patches
might fix merge conflicts
* Use UTF-8 for gradlew
* Clean rb patches again
* Convert block ids used as item ids
Neither the converters of pre 1.13 nor DFU handled these cases,
as by the time they were written the game at the time didn't
consider these ids valid - they would be air. Because of this,
some worlds have logspam since only DataConverter (not DFU or
legacy converters) will warn when an invalid id has been
seen.
While quite a few do need to now be considered as air, quite a lot
do not. So it makes sense to add conversion for these items, instead
of simply suppressing or ignoring the logs. I've now added id -> string conversion
for all block ids that could be used as items that existed in the game
before 1.7.10 (I have no interest in tracking down the
exact version block ids stopped working) that were on
https://minecraft-ids.grahamedgecombe.com/
Items that did not directly convert to new items will
be instead converted to air: stems, wheat crops, piston head,
tripwire wire block
* Fix LightPopulated parsing in V1466
The DFU code was checking if the number existed, not if it
didn't exist. I misread the original code.
* Always parse protochunk light sources unless it is marked as non-lit
Chunks not marked as lit will always go through the light engine,
so they should always have their block sources parsed.
* Update custom names to JSON for players
Missed this fix from CB, as it was inside
the DataFixers class.
I decided to double check all of the CB changes again:
DataFixers.java was the only area I missed, as I had inspected all
datafixer diffs and implemented them all into DataConverter. I also
checked Bootstrap.java again, and re-evaluated their changes. I had
previously done this, but determined that they were all bad.
The changes to make standing_sign block map to oak_sign block in
V1450 is bad, because that's not the item id V1450 accepts. Only
in 1.14 did oak_sign even exist, and as expected there is a converter
to rename all existing sign items/blocks.
The fix to register the portal block under id 1440 is useless, as
the flattenning logic will default to the lowest registered id - which
is the exact blockstate that CB registers into 1440. So it just
doesn't do anything.
The extra item ids in the id -> string converter are already added,
but I found this from EMC originally.
The change for the spawn egg id 23 -> Arrow is just wrong,
that id DOES correspond to TippedArrow, NOT Arrow. As
expected, the spawn egg already has a dedicated mapping for
Arrow, which is id 10 - which was Arrow's entity id.
I also ported a fix for the cooked_fished id update. This doesn't
really matter since there is already a dataconverter to fix this,
but the game didn't accept cooked_fished at the time. So I see
no harm.
* Review all converters and walkers
- Refactor V99 to have helper methods for defining entity/tile
entity types
- Automatically namespace all ids that should be namespaced.
While vanilla never saved non-namespaced data for things that
are namespaced, plugins/users might have.
- Synchronised the identity ensure map in HelperBlockFlatteningV1450
- Code style consistency
- Add missing log warning in V102 for ITEM_NAME type conversion
- Use getBoolean instead of getByte
- Use ConverterAbstractEntityRename for V143 TippedArrow -> Arrow
rename, as it will affect ENTITY_NAME type
- Always set isVillager to false in V502 for Zombie
- Register V808's converter under subversion 1 like DFU
- Register a breakpoint for V1.17.1. In the future, all final
versions of major releases will have a breakpoint so that
the work required to determine if a converter needs a breakpoint
is minimal
- Validate that a dataconverter is only registered for a version
that is registered
- ConverterFlattenTileEntity is actually ConverterFlattenEntity
It even registered the converters under TILE_ENTITY, instead of
ENTITY.
- Fix id comparison in V1492 STRUCTURE_FEATURE renamer
- Use ConverterAbstractStatsRename for V1510 stats renamer
At the time I had written that class, the abstract renamer didn't
exist.
- Ensure OwnerUUID is at least set to empty string in
V1904 if the ocelot is converted to a cat (this is
likely so that it retains a collar)
- Use generic read/write for Records in V1946
Records is actually a list, not a map. So reading map was
invalid.
* Always set light to zero when propagating decrease
This fixes an almost infinite loop where light values
would be spam queued on a very small subset on blocks.
This also likely fixes the memory issues people were
seeing.
* re-organize patches
* Apply and fix conflicts
* Revert some patches
getChunkAt retains chunks so that plugins don't spam loads
revert mc-4 fix will remain unless issues pop up
* Shuffle iterated chunks if per player is not enabled
Can help with some mob spawning stacking up at locations
* Make per player default, migrate all configs
* Adjust comments in fixups
* Rework config for player chunk loader
Old config is not compatible. Move all configs to be
under `settings` in paper.yml
The player chunk loader has been modified to
less aggressively load chunks, but to send
chunks at higher rates compared to tuinity. There are
new config entries to tune this behavior.
* Add back old constructor to CompressionEncoder/Decoder (fixes
Tuinity #358)
* Raise chunk loading default limits
* Reduce worldgen thread workers for lower core count cpus
* Raise limits for chunk loading config
Also place it under `chunk-loading`
* Disable max chunk send rate by default
* Fix conflicts and rebuild patches
* Drop default send rate again
Appears to be still causing problems for no known reason
* Raise chunk send limits to 100 per player
While a low limit fixes ping issues for some people, most people
do not suffer from this issue and thus should not suffer from
an extremely slow load-in rate.
* Rebase part 1
Autosquash the fixups
* Move not implemented up
* Fixup mc-dev fixes
Missed this one
* Rebase per player viewdistance api into the original api patch
* Remove old light engine patch part 1
The prioritisation must be kept from it, so that part
has been rebased into the priority patch.
Part 2 will deal with rebasing all of the patches _after_
* Rebase remaining patches for old light patch removal
* Remove other mid tick patch
* Remove Optimize-PlayerChunkMap-memory-use-for-visibleChunks.patch
Replaced by `Do not copy visible chunks`
* Revert AT for Vec3i setX/Y/Z
The class is immutable. set should not be exposed
* Remove old IntegerUtil class
* Replace old CraftChunk#getEntities patch
* Remove import for SWMRNibbleArray in ChunkAccess
* Finished merge checklist
* Remove ensureTickThread impl in urgency patch
Co-authored-by: Spottedleaf <Spottedleaf@users.noreply.github.com>
Co-authored-by: Jason Penilla <11360596+jpenilla@users.noreply.github.com>
2021-08-31 11:02:11 +00:00
public List<Entity> getEntities(@Nullable Entity except, AABB box, Predicate<? super Entity> predicate) {
this.getProfiler().incrementCounter("getEntities");
List<Entity> list = Lists.newArrayList();
-
- this.getEntities().get(box, (entity1) -> {
- if (entity1 != except && predicate.test(entity1)) {
- list.add(entity1);
- }
-
- if (entity1 instanceof EnderDragon) {
- EnderDragonPart[] aentitycomplexpart = ((EnderDragon) entity1).getSubEntities();
- int i = aentitycomplexpart.length;
-
- for (int j = 0; j < i; ++j) {
- EnderDragonPart entitycomplexpart = aentitycomplexpart[j];
-
- if (entity1 != except && predicate.test(entitycomplexpart)) {
- list.add(entitycomplexpart);
- }
- }
- }
-
- }, predicate == net.minecraft.world.entity.EntitySelector.CONTAINER_ENTITY_SELECTOR); // Paper
+ this.entitySliceManager.getEntities(except, box, list, predicate); // Paper - optimise this call
return list;
}
2021-11-15 19:01:19 +00:00
@@ -1015,26 +1041,22 @@ public abstract class Level implements LevelAccessor, AutoCloseable {
Merge tuinity (#6413)
This PR contains all of Tuinity's patches. Very notable ones are:
- Highly optimised collisions
- Optimised entity lookups by bounding box (Mojang made regressions in 1.17, this brings it back to 1.16)
- Starlight https://github.com/PaperMC/Starlight
- Rewritten dataconverter system https://github.com/PaperMC/DataConverter
- Random block ticking optimisation (wrongly dropped from Paper 1.17)
- Chunk ticking optimisations
- Anything else I've forgotten in the 60 or so patches
If you are a previous Tuinity user, your config will not migrate. You must do it yourself. The config options have simply been moved into paper.yml, so it will be an easy migration. However, please note that the chunk loading options in tuinity.yml are NOT compatible with the options in paper.yml.
* Port tuinity, initial patchset
* Update gradle to 7.2
jmp said it fixes rebuildpatches not working for me. it fucking better
* Completely clean apply
* Remove tuinity config, add per player api patch
* Remove paper reobf mappings patch
* Properly update gradlew
* Force clean rebuild
* Mark fixups
Comments and ATs still need to be done
* grep -r "Tuinity"
* Fixup
* Ensure gameprofile lastaccess is written only under the state lock
* update URL for dataconverter
* Only clean rebuild tuinity patches
might fix merge conflicts
* Use UTF-8 for gradlew
* Clean rb patches again
* Convert block ids used as item ids
Neither the converters of pre 1.13 nor DFU handled these cases,
as by the time they were written the game at the time didn't
consider these ids valid - they would be air. Because of this,
some worlds have logspam since only DataConverter (not DFU or
legacy converters) will warn when an invalid id has been
seen.
While quite a few do need to now be considered as air, quite a lot
do not. So it makes sense to add conversion for these items, instead
of simply suppressing or ignoring the logs. I've now added id -> string conversion
for all block ids that could be used as items that existed in the game
before 1.7.10 (I have no interest in tracking down the
exact version block ids stopped working) that were on
https://minecraft-ids.grahamedgecombe.com/
Items that did not directly convert to new items will
be instead converted to air: stems, wheat crops, piston head,
tripwire wire block
* Fix LightPopulated parsing in V1466
The DFU code was checking if the number existed, not if it
didn't exist. I misread the original code.
* Always parse protochunk light sources unless it is marked as non-lit
Chunks not marked as lit will always go through the light engine,
so they should always have their block sources parsed.
* Update custom names to JSON for players
Missed this fix from CB, as it was inside
the DataFixers class.
I decided to double check all of the CB changes again:
DataFixers.java was the only area I missed, as I had inspected all
datafixer diffs and implemented them all into DataConverter. I also
checked Bootstrap.java again, and re-evaluated their changes. I had
previously done this, but determined that they were all bad.
The changes to make standing_sign block map to oak_sign block in
V1450 is bad, because that's not the item id V1450 accepts. Only
in 1.14 did oak_sign even exist, and as expected there is a converter
to rename all existing sign items/blocks.
The fix to register the portal block under id 1440 is useless, as
the flattenning logic will default to the lowest registered id - which
is the exact blockstate that CB registers into 1440. So it just
doesn't do anything.
The extra item ids in the id -> string converter are already added,
but I found this from EMC originally.
The change for the spawn egg id 23 -> Arrow is just wrong,
that id DOES correspond to TippedArrow, NOT Arrow. As
expected, the spawn egg already has a dedicated mapping for
Arrow, which is id 10 - which was Arrow's entity id.
I also ported a fix for the cooked_fished id update. This doesn't
really matter since there is already a dataconverter to fix this,
but the game didn't accept cooked_fished at the time. So I see
no harm.
* Review all converters and walkers
- Refactor V99 to have helper methods for defining entity/tile
entity types
- Automatically namespace all ids that should be namespaced.
While vanilla never saved non-namespaced data for things that
are namespaced, plugins/users might have.
- Synchronised the identity ensure map in HelperBlockFlatteningV1450
- Code style consistency
- Add missing log warning in V102 for ITEM_NAME type conversion
- Use getBoolean instead of getByte
- Use ConverterAbstractEntityRename for V143 TippedArrow -> Arrow
rename, as it will affect ENTITY_NAME type
- Always set isVillager to false in V502 for Zombie
- Register V808's converter under subversion 1 like DFU
- Register a breakpoint for V1.17.1. In the future, all final
versions of major releases will have a breakpoint so that
the work required to determine if a converter needs a breakpoint
is minimal
- Validate that a dataconverter is only registered for a version
that is registered
- ConverterFlattenTileEntity is actually ConverterFlattenEntity
It even registered the converters under TILE_ENTITY, instead of
ENTITY.
- Fix id comparison in V1492 STRUCTURE_FEATURE renamer
- Use ConverterAbstractStatsRename for V1510 stats renamer
At the time I had written that class, the abstract renamer didn't
exist.
- Ensure OwnerUUID is at least set to empty string in
V1904 if the ocelot is converted to a cat (this is
likely so that it retains a collar)
- Use generic read/write for Records in V1946
Records is actually a list, not a map. So reading map was
invalid.
* Always set light to zero when propagating decrease
This fixes an almost infinite loop where light values
would be spam queued on a very small subset on blocks.
This also likely fixes the memory issues people were
seeing.
* re-organize patches
* Apply and fix conflicts
* Revert some patches
getChunkAt retains chunks so that plugins don't spam loads
revert mc-4 fix will remain unless issues pop up
* Shuffle iterated chunks if per player is not enabled
Can help with some mob spawning stacking up at locations
* Make per player default, migrate all configs
* Adjust comments in fixups
* Rework config for player chunk loader
Old config is not compatible. Move all configs to be
under `settings` in paper.yml
The player chunk loader has been modified to
less aggressively load chunks, but to send
chunks at higher rates compared to tuinity. There are
new config entries to tune this behavior.
* Add back old constructor to CompressionEncoder/Decoder (fixes
Tuinity #358)
* Raise chunk loading default limits
* Reduce worldgen thread workers for lower core count cpus
* Raise limits for chunk loading config
Also place it under `chunk-loading`
* Disable max chunk send rate by default
* Fix conflicts and rebuild patches
* Drop default send rate again
Appears to be still causing problems for no known reason
* Raise chunk send limits to 100 per player
While a low limit fixes ping issues for some people, most people
do not suffer from this issue and thus should not suffer from
an extremely slow load-in rate.
* Rebase part 1
Autosquash the fixups
* Move not implemented up
* Fixup mc-dev fixes
Missed this one
* Rebase per player viewdistance api into the original api patch
* Remove old light engine patch part 1
The prioritisation must be kept from it, so that part
has been rebased into the priority patch.
Part 2 will deal with rebasing all of the patches _after_
* Rebase remaining patches for old light patch removal
* Remove other mid tick patch
* Remove Optimize-PlayerChunkMap-memory-use-for-visibleChunks.patch
Replaced by `Do not copy visible chunks`
* Revert AT for Vec3i setX/Y/Z
The class is immutable. set should not be exposed
* Remove old IntegerUtil class
* Replace old CraftChunk#getEntities patch
* Remove import for SWMRNibbleArray in ChunkAccess
* Finished merge checklist
* Remove ensureTickThread impl in urgency patch
Co-authored-by: Spottedleaf <Spottedleaf@users.noreply.github.com>
Co-authored-by: Jason Penilla <11360596+jpenilla@users.noreply.github.com>
2021-08-31 11:02:11 +00:00
this.getProfiler().incrementCounter("getEntities");
List<T> list = Lists.newArrayList();
- this.getEntities().get(filter, box, (entity) -> {
- if (predicate.test(entity)) {
- list.add(entity);
- }
-
- if (entity instanceof EnderDragon) {
- EnderDragonPart[] aentitycomplexpart = ((EnderDragon) entity).getSubEntities();
- int i = aentitycomplexpart.length;
-
- for (int j = 0; j < i; ++j) {
- EnderDragonPart entitycomplexpart = aentitycomplexpart[j];
- T t0 = filter.tryCast(entitycomplexpart);
-
- if (t0 != null && predicate.test(t0)) {
- list.add(t0);
- }
- }
+ // Paper start - optimise this call
+ if (filter instanceof net.minecraft.world.entity.EntityType) {
+ this.entitySliceManager.getEntities((net.minecraft.world.entity.EntityType)filter, box, list, predicate);
+ } else {
+ Predicate<? super T> test = (obj) -> {
+ return filter.tryCast(obj) != null;
+ };
+ predicate = predicate == null ? test : test.and((Predicate)predicate);
+ Class base;
+ if (filter == null || (base = filter.getBaseClass()) == null || base == Entity.class) {
+ this.entitySliceManager.getEntities((Entity) null, box, (List)list, (Predicate)predicate);
+ } else {
+ this.entitySliceManager.getEntities(base, null, box, (List)list, (Predicate)predicate); // Paper - optimise this call
}
-
- });
+ }
+ // Paper end - optimise this call
return list;
}
diff --git a/src/main/java/net/minecraft/world/level/entity/PersistentEntitySectionManager.java b/src/main/java/net/minecraft/world/level/entity/PersistentEntitySectionManager.java
2021-11-05 14:08:24 +00:00
index e89396ea1d06f9e4a0a58d86cb7f9d857d50dc0a..976d206a17add01a31ae38b966913368cf386cb1 100644
Merge tuinity (#6413)
This PR contains all of Tuinity's patches. Very notable ones are:
- Highly optimised collisions
- Optimised entity lookups by bounding box (Mojang made regressions in 1.17, this brings it back to 1.16)
- Starlight https://github.com/PaperMC/Starlight
- Rewritten dataconverter system https://github.com/PaperMC/DataConverter
- Random block ticking optimisation (wrongly dropped from Paper 1.17)
- Chunk ticking optimisations
- Anything else I've forgotten in the 60 or so patches
If you are a previous Tuinity user, your config will not migrate. You must do it yourself. The config options have simply been moved into paper.yml, so it will be an easy migration. However, please note that the chunk loading options in tuinity.yml are NOT compatible with the options in paper.yml.
* Port tuinity, initial patchset
* Update gradle to 7.2
jmp said it fixes rebuildpatches not working for me. it fucking better
* Completely clean apply
* Remove tuinity config, add per player api patch
* Remove paper reobf mappings patch
* Properly update gradlew
* Force clean rebuild
* Mark fixups
Comments and ATs still need to be done
* grep -r "Tuinity"
* Fixup
* Ensure gameprofile lastaccess is written only under the state lock
* update URL for dataconverter
* Only clean rebuild tuinity patches
might fix merge conflicts
* Use UTF-8 for gradlew
* Clean rb patches again
* Convert block ids used as item ids
Neither the converters of pre 1.13 nor DFU handled these cases,
as by the time they were written the game at the time didn't
consider these ids valid - they would be air. Because of this,
some worlds have logspam since only DataConverter (not DFU or
legacy converters) will warn when an invalid id has been
seen.
While quite a few do need to now be considered as air, quite a lot
do not. So it makes sense to add conversion for these items, instead
of simply suppressing or ignoring the logs. I've now added id -> string conversion
for all block ids that could be used as items that existed in the game
before 1.7.10 (I have no interest in tracking down the
exact version block ids stopped working) that were on
https://minecraft-ids.grahamedgecombe.com/
Items that did not directly convert to new items will
be instead converted to air: stems, wheat crops, piston head,
tripwire wire block
* Fix LightPopulated parsing in V1466
The DFU code was checking if the number existed, not if it
didn't exist. I misread the original code.
* Always parse protochunk light sources unless it is marked as non-lit
Chunks not marked as lit will always go through the light engine,
so they should always have their block sources parsed.
* Update custom names to JSON for players
Missed this fix from CB, as it was inside
the DataFixers class.
I decided to double check all of the CB changes again:
DataFixers.java was the only area I missed, as I had inspected all
datafixer diffs and implemented them all into DataConverter. I also
checked Bootstrap.java again, and re-evaluated their changes. I had
previously done this, but determined that they were all bad.
The changes to make standing_sign block map to oak_sign block in
V1450 is bad, because that's not the item id V1450 accepts. Only
in 1.14 did oak_sign even exist, and as expected there is a converter
to rename all existing sign items/blocks.
The fix to register the portal block under id 1440 is useless, as
the flattenning logic will default to the lowest registered id - which
is the exact blockstate that CB registers into 1440. So it just
doesn't do anything.
The extra item ids in the id -> string converter are already added,
but I found this from EMC originally.
The change for the spawn egg id 23 -> Arrow is just wrong,
that id DOES correspond to TippedArrow, NOT Arrow. As
expected, the spawn egg already has a dedicated mapping for
Arrow, which is id 10 - which was Arrow's entity id.
I also ported a fix for the cooked_fished id update. This doesn't
really matter since there is already a dataconverter to fix this,
but the game didn't accept cooked_fished at the time. So I see
no harm.
* Review all converters and walkers
- Refactor V99 to have helper methods for defining entity/tile
entity types
- Automatically namespace all ids that should be namespaced.
While vanilla never saved non-namespaced data for things that
are namespaced, plugins/users might have.
- Synchronised the identity ensure map in HelperBlockFlatteningV1450
- Code style consistency
- Add missing log warning in V102 for ITEM_NAME type conversion
- Use getBoolean instead of getByte
- Use ConverterAbstractEntityRename for V143 TippedArrow -> Arrow
rename, as it will affect ENTITY_NAME type
- Always set isVillager to false in V502 for Zombie
- Register V808's converter under subversion 1 like DFU
- Register a breakpoint for V1.17.1. In the future, all final
versions of major releases will have a breakpoint so that
the work required to determine if a converter needs a breakpoint
is minimal
- Validate that a dataconverter is only registered for a version
that is registered
- ConverterFlattenTileEntity is actually ConverterFlattenEntity
It even registered the converters under TILE_ENTITY, instead of
ENTITY.
- Fix id comparison in V1492 STRUCTURE_FEATURE renamer
- Use ConverterAbstractStatsRename for V1510 stats renamer
At the time I had written that class, the abstract renamer didn't
exist.
- Ensure OwnerUUID is at least set to empty string in
V1904 if the ocelot is converted to a cat (this is
likely so that it retains a collar)
- Use generic read/write for Records in V1946
Records is actually a list, not a map. So reading map was
invalid.
* Always set light to zero when propagating decrease
This fixes an almost infinite loop where light values
would be spam queued on a very small subset on blocks.
This also likely fixes the memory issues people were
seeing.
* re-organize patches
* Apply and fix conflicts
* Revert some patches
getChunkAt retains chunks so that plugins don't spam loads
revert mc-4 fix will remain unless issues pop up
* Shuffle iterated chunks if per player is not enabled
Can help with some mob spawning stacking up at locations
* Make per player default, migrate all configs
* Adjust comments in fixups
* Rework config for player chunk loader
Old config is not compatible. Move all configs to be
under `settings` in paper.yml
The player chunk loader has been modified to
less aggressively load chunks, but to send
chunks at higher rates compared to tuinity. There are
new config entries to tune this behavior.
* Add back old constructor to CompressionEncoder/Decoder (fixes
Tuinity #358)
* Raise chunk loading default limits
* Reduce worldgen thread workers for lower core count cpus
* Raise limits for chunk loading config
Also place it under `chunk-loading`
* Disable max chunk send rate by default
* Fix conflicts and rebuild patches
* Drop default send rate again
Appears to be still causing problems for no known reason
* Raise chunk send limits to 100 per player
While a low limit fixes ping issues for some people, most people
do not suffer from this issue and thus should not suffer from
an extremely slow load-in rate.
* Rebase part 1
Autosquash the fixups
* Move not implemented up
* Fixup mc-dev fixes
Missed this one
* Rebase per player viewdistance api into the original api patch
* Remove old light engine patch part 1
The prioritisation must be kept from it, so that part
has been rebased into the priority patch.
Part 2 will deal with rebasing all of the patches _after_
* Rebase remaining patches for old light patch removal
* Remove other mid tick patch
* Remove Optimize-PlayerChunkMap-memory-use-for-visibleChunks.patch
Replaced by `Do not copy visible chunks`
* Revert AT for Vec3i setX/Y/Z
The class is immutable. set should not be exposed
* Remove old IntegerUtil class
* Replace old CraftChunk#getEntities patch
* Remove import for SWMRNibbleArray in ChunkAccess
* Finished merge checklist
* Remove ensureTickThread impl in urgency patch
Co-authored-by: Spottedleaf <Spottedleaf@users.noreply.github.com>
Co-authored-by: Jason Penilla <11360596+jpenilla@users.noreply.github.com>
2021-08-31 11:02:11 +00:00
--- a/src/main/java/net/minecraft/world/level/entity/PersistentEntitySectionManager.java
+++ b/src/main/java/net/minecraft/world/level/entity/PersistentEntitySectionManager.java
2021-09-01 12:03:36 +00:00
@@ -49,8 +49,10 @@ public class PersistentEntitySectionManager<T extends EntityAccess> implements A
Merge tuinity (#6413)
This PR contains all of Tuinity's patches. Very notable ones are:
- Highly optimised collisions
- Optimised entity lookups by bounding box (Mojang made regressions in 1.17, this brings it back to 1.16)
- Starlight https://github.com/PaperMC/Starlight
- Rewritten dataconverter system https://github.com/PaperMC/DataConverter
- Random block ticking optimisation (wrongly dropped from Paper 1.17)
- Chunk ticking optimisations
- Anything else I've forgotten in the 60 or so patches
If you are a previous Tuinity user, your config will not migrate. You must do it yourself. The config options have simply been moved into paper.yml, so it will be an easy migration. However, please note that the chunk loading options in tuinity.yml are NOT compatible with the options in paper.yml.
* Port tuinity, initial patchset
* Update gradle to 7.2
jmp said it fixes rebuildpatches not working for me. it fucking better
* Completely clean apply
* Remove tuinity config, add per player api patch
* Remove paper reobf mappings patch
* Properly update gradlew
* Force clean rebuild
* Mark fixups
Comments and ATs still need to be done
* grep -r "Tuinity"
* Fixup
* Ensure gameprofile lastaccess is written only under the state lock
* update URL for dataconverter
* Only clean rebuild tuinity patches
might fix merge conflicts
* Use UTF-8 for gradlew
* Clean rb patches again
* Convert block ids used as item ids
Neither the converters of pre 1.13 nor DFU handled these cases,
as by the time they were written the game at the time didn't
consider these ids valid - they would be air. Because of this,
some worlds have logspam since only DataConverter (not DFU or
legacy converters) will warn when an invalid id has been
seen.
While quite a few do need to now be considered as air, quite a lot
do not. So it makes sense to add conversion for these items, instead
of simply suppressing or ignoring the logs. I've now added id -> string conversion
for all block ids that could be used as items that existed in the game
before 1.7.10 (I have no interest in tracking down the
exact version block ids stopped working) that were on
https://minecraft-ids.grahamedgecombe.com/
Items that did not directly convert to new items will
be instead converted to air: stems, wheat crops, piston head,
tripwire wire block
* Fix LightPopulated parsing in V1466
The DFU code was checking if the number existed, not if it
didn't exist. I misread the original code.
* Always parse protochunk light sources unless it is marked as non-lit
Chunks not marked as lit will always go through the light engine,
so they should always have their block sources parsed.
* Update custom names to JSON for players
Missed this fix from CB, as it was inside
the DataFixers class.
I decided to double check all of the CB changes again:
DataFixers.java was the only area I missed, as I had inspected all
datafixer diffs and implemented them all into DataConverter. I also
checked Bootstrap.java again, and re-evaluated their changes. I had
previously done this, but determined that they were all bad.
The changes to make standing_sign block map to oak_sign block in
V1450 is bad, because that's not the item id V1450 accepts. Only
in 1.14 did oak_sign even exist, and as expected there is a converter
to rename all existing sign items/blocks.
The fix to register the portal block under id 1440 is useless, as
the flattenning logic will default to the lowest registered id - which
is the exact blockstate that CB registers into 1440. So it just
doesn't do anything.
The extra item ids in the id -> string converter are already added,
but I found this from EMC originally.
The change for the spawn egg id 23 -> Arrow is just wrong,
that id DOES correspond to TippedArrow, NOT Arrow. As
expected, the spawn egg already has a dedicated mapping for
Arrow, which is id 10 - which was Arrow's entity id.
I also ported a fix for the cooked_fished id update. This doesn't
really matter since there is already a dataconverter to fix this,
but the game didn't accept cooked_fished at the time. So I see
no harm.
* Review all converters and walkers
- Refactor V99 to have helper methods for defining entity/tile
entity types
- Automatically namespace all ids that should be namespaced.
While vanilla never saved non-namespaced data for things that
are namespaced, plugins/users might have.
- Synchronised the identity ensure map in HelperBlockFlatteningV1450
- Code style consistency
- Add missing log warning in V102 for ITEM_NAME type conversion
- Use getBoolean instead of getByte
- Use ConverterAbstractEntityRename for V143 TippedArrow -> Arrow
rename, as it will affect ENTITY_NAME type
- Always set isVillager to false in V502 for Zombie
- Register V808's converter under subversion 1 like DFU
- Register a breakpoint for V1.17.1. In the future, all final
versions of major releases will have a breakpoint so that
the work required to determine if a converter needs a breakpoint
is minimal
- Validate that a dataconverter is only registered for a version
that is registered
- ConverterFlattenTileEntity is actually ConverterFlattenEntity
It even registered the converters under TILE_ENTITY, instead of
ENTITY.
- Fix id comparison in V1492 STRUCTURE_FEATURE renamer
- Use ConverterAbstractStatsRename for V1510 stats renamer
At the time I had written that class, the abstract renamer didn't
exist.
- Ensure OwnerUUID is at least set to empty string in
V1904 if the ocelot is converted to a cat (this is
likely so that it retains a collar)
- Use generic read/write for Records in V1946
Records is actually a list, not a map. So reading map was
invalid.
* Always set light to zero when propagating decrease
This fixes an almost infinite loop where light values
would be spam queued on a very small subset on blocks.
This also likely fixes the memory issues people were
seeing.
* re-organize patches
* Apply and fix conflicts
* Revert some patches
getChunkAt retains chunks so that plugins don't spam loads
revert mc-4 fix will remain unless issues pop up
* Shuffle iterated chunks if per player is not enabled
Can help with some mob spawning stacking up at locations
* Make per player default, migrate all configs
* Adjust comments in fixups
* Rework config for player chunk loader
Old config is not compatible. Move all configs to be
under `settings` in paper.yml
The player chunk loader has been modified to
less aggressively load chunks, but to send
chunks at higher rates compared to tuinity. There are
new config entries to tune this behavior.
* Add back old constructor to CompressionEncoder/Decoder (fixes
Tuinity #358)
* Raise chunk loading default limits
* Reduce worldgen thread workers for lower core count cpus
* Raise limits for chunk loading config
Also place it under `chunk-loading`
* Disable max chunk send rate by default
* Fix conflicts and rebuild patches
* Drop default send rate again
Appears to be still causing problems for no known reason
* Raise chunk send limits to 100 per player
While a low limit fixes ping issues for some people, most people
do not suffer from this issue and thus should not suffer from
an extremely slow load-in rate.
* Rebase part 1
Autosquash the fixups
* Move not implemented up
* Fixup mc-dev fixes
Missed this one
* Rebase per player viewdistance api into the original api patch
* Remove old light engine patch part 1
The prioritisation must be kept from it, so that part
has been rebased into the priority patch.
Part 2 will deal with rebasing all of the patches _after_
* Rebase remaining patches for old light patch removal
* Remove other mid tick patch
* Remove Optimize-PlayerChunkMap-memory-use-for-visibleChunks.patch
Replaced by `Do not copy visible chunks`
* Revert AT for Vec3i setX/Y/Z
The class is immutable. set should not be exposed
* Remove old IntegerUtil class
* Replace old CraftChunk#getEntities patch
* Remove import for SWMRNibbleArray in ChunkAccess
* Finished merge checklist
* Remove ensureTickThread impl in urgency patch
Co-authored-by: Spottedleaf <Spottedleaf@users.noreply.github.com>
Co-authored-by: Jason Penilla <11360596+jpenilla@users.noreply.github.com>
2021-08-31 11:02:11 +00:00
private final Long2ObjectMap<PersistentEntitySectionManager.ChunkLoadStatus> chunkLoadStatuses = new Long2ObjectOpenHashMap();
private final LongSet chunksToUnload = new LongOpenHashSet();
private final Queue<ChunkEntities<T>> loadingInbox = Queues.newConcurrentLinkedQueue();
+ public final io.papermc.paper.world.EntitySliceManager entitySliceManager; // Paper
- public PersistentEntitySectionManager(Class<T> entityClass, LevelCallback<T> handler, EntityPersistentStorage<T> dataAccess) {
+ public PersistentEntitySectionManager(Class<T> entityClass, LevelCallback<T> handler, EntityPersistentStorage<T> dataAccess, io.papermc.paper.world.EntitySliceManager entitySliceManager) { // Paper
+ this.entitySliceManager = entitySliceManager; // Paper
this.sectionStorage = new EntitySectionStorage<>(entityClass, this.chunkVisibility);
this.chunkVisibility.defaultReturnValue(Visibility.HIDDEN);
this.chunkLoadStatuses.defaultReturnValue(PersistentEntitySectionManager.ChunkLoadStatus.FRESH);
2021-11-05 14:08:24 +00:00
@@ -112,6 +114,7 @@ public class PersistentEntitySectionManager<T extends EntityAccess> implements A
Merge tuinity (#6413)
This PR contains all of Tuinity's patches. Very notable ones are:
- Highly optimised collisions
- Optimised entity lookups by bounding box (Mojang made regressions in 1.17, this brings it back to 1.16)
- Starlight https://github.com/PaperMC/Starlight
- Rewritten dataconverter system https://github.com/PaperMC/DataConverter
- Random block ticking optimisation (wrongly dropped from Paper 1.17)
- Chunk ticking optimisations
- Anything else I've forgotten in the 60 or so patches
If you are a previous Tuinity user, your config will not migrate. You must do it yourself. The config options have simply been moved into paper.yml, so it will be an easy migration. However, please note that the chunk loading options in tuinity.yml are NOT compatible with the options in paper.yml.
* Port tuinity, initial patchset
* Update gradle to 7.2
jmp said it fixes rebuildpatches not working for me. it fucking better
* Completely clean apply
* Remove tuinity config, add per player api patch
* Remove paper reobf mappings patch
* Properly update gradlew
* Force clean rebuild
* Mark fixups
Comments and ATs still need to be done
* grep -r "Tuinity"
* Fixup
* Ensure gameprofile lastaccess is written only under the state lock
* update URL for dataconverter
* Only clean rebuild tuinity patches
might fix merge conflicts
* Use UTF-8 for gradlew
* Clean rb patches again
* Convert block ids used as item ids
Neither the converters of pre 1.13 nor DFU handled these cases,
as by the time they were written the game at the time didn't
consider these ids valid - they would be air. Because of this,
some worlds have logspam since only DataConverter (not DFU or
legacy converters) will warn when an invalid id has been
seen.
While quite a few do need to now be considered as air, quite a lot
do not. So it makes sense to add conversion for these items, instead
of simply suppressing or ignoring the logs. I've now added id -> string conversion
for all block ids that could be used as items that existed in the game
before 1.7.10 (I have no interest in tracking down the
exact version block ids stopped working) that were on
https://minecraft-ids.grahamedgecombe.com/
Items that did not directly convert to new items will
be instead converted to air: stems, wheat crops, piston head,
tripwire wire block
* Fix LightPopulated parsing in V1466
The DFU code was checking if the number existed, not if it
didn't exist. I misread the original code.
* Always parse protochunk light sources unless it is marked as non-lit
Chunks not marked as lit will always go through the light engine,
so they should always have their block sources parsed.
* Update custom names to JSON for players
Missed this fix from CB, as it was inside
the DataFixers class.
I decided to double check all of the CB changes again:
DataFixers.java was the only area I missed, as I had inspected all
datafixer diffs and implemented them all into DataConverter. I also
checked Bootstrap.java again, and re-evaluated their changes. I had
previously done this, but determined that they were all bad.
The changes to make standing_sign block map to oak_sign block in
V1450 is bad, because that's not the item id V1450 accepts. Only
in 1.14 did oak_sign even exist, and as expected there is a converter
to rename all existing sign items/blocks.
The fix to register the portal block under id 1440 is useless, as
the flattenning logic will default to the lowest registered id - which
is the exact blockstate that CB registers into 1440. So it just
doesn't do anything.
The extra item ids in the id -> string converter are already added,
but I found this from EMC originally.
The change for the spawn egg id 23 -> Arrow is just wrong,
that id DOES correspond to TippedArrow, NOT Arrow. As
expected, the spawn egg already has a dedicated mapping for
Arrow, which is id 10 - which was Arrow's entity id.
I also ported a fix for the cooked_fished id update. This doesn't
really matter since there is already a dataconverter to fix this,
but the game didn't accept cooked_fished at the time. So I see
no harm.
* Review all converters and walkers
- Refactor V99 to have helper methods for defining entity/tile
entity types
- Automatically namespace all ids that should be namespaced.
While vanilla never saved non-namespaced data for things that
are namespaced, plugins/users might have.
- Synchronised the identity ensure map in HelperBlockFlatteningV1450
- Code style consistency
- Add missing log warning in V102 for ITEM_NAME type conversion
- Use getBoolean instead of getByte
- Use ConverterAbstractEntityRename for V143 TippedArrow -> Arrow
rename, as it will affect ENTITY_NAME type
- Always set isVillager to false in V502 for Zombie
- Register V808's converter under subversion 1 like DFU
- Register a breakpoint for V1.17.1. In the future, all final
versions of major releases will have a breakpoint so that
the work required to determine if a converter needs a breakpoint
is minimal
- Validate that a dataconverter is only registered for a version
that is registered
- ConverterFlattenTileEntity is actually ConverterFlattenEntity
It even registered the converters under TILE_ENTITY, instead of
ENTITY.
- Fix id comparison in V1492 STRUCTURE_FEATURE renamer
- Use ConverterAbstractStatsRename for V1510 stats renamer
At the time I had written that class, the abstract renamer didn't
exist.
- Ensure OwnerUUID is at least set to empty string in
V1904 if the ocelot is converted to a cat (this is
likely so that it retains a collar)
- Use generic read/write for Records in V1946
Records is actually a list, not a map. So reading map was
invalid.
* Always set light to zero when propagating decrease
This fixes an almost infinite loop where light values
would be spam queued on a very small subset on blocks.
This also likely fixes the memory issues people were
seeing.
* re-organize patches
* Apply and fix conflicts
* Revert some patches
getChunkAt retains chunks so that plugins don't spam loads
revert mc-4 fix will remain unless issues pop up
* Shuffle iterated chunks if per player is not enabled
Can help with some mob spawning stacking up at locations
* Make per player default, migrate all configs
* Adjust comments in fixups
* Rework config for player chunk loader
Old config is not compatible. Move all configs to be
under `settings` in paper.yml
The player chunk loader has been modified to
less aggressively load chunks, but to send
chunks at higher rates compared to tuinity. There are
new config entries to tune this behavior.
* Add back old constructor to CompressionEncoder/Decoder (fixes
Tuinity #358)
* Raise chunk loading default limits
* Reduce worldgen thread workers for lower core count cpus
* Raise limits for chunk loading config
Also place it under `chunk-loading`
* Disable max chunk send rate by default
* Fix conflicts and rebuild patches
* Drop default send rate again
Appears to be still causing problems for no known reason
* Raise chunk send limits to 100 per player
While a low limit fixes ping issues for some people, most people
do not suffer from this issue and thus should not suffer from
an extremely slow load-in rate.
* Rebase part 1
Autosquash the fixups
* Move not implemented up
* Fixup mc-dev fixes
Missed this one
* Rebase per player viewdistance api into the original api patch
* Remove old light engine patch part 1
The prioritisation must be kept from it, so that part
has been rebased into the priority patch.
Part 2 will deal with rebasing all of the patches _after_
* Rebase remaining patches for old light patch removal
* Remove other mid tick patch
* Remove Optimize-PlayerChunkMap-memory-use-for-visibleChunks.patch
Replaced by `Do not copy visible chunks`
* Revert AT for Vec3i setX/Y/Z
The class is immutable. set should not be exposed
* Remove old IntegerUtil class
* Replace old CraftChunk#getEntities patch
* Remove import for SWMRNibbleArray in ChunkAccess
* Finished merge checklist
* Remove ensureTickThread impl in urgency patch
Co-authored-by: Spottedleaf <Spottedleaf@users.noreply.github.com>
Co-authored-by: Jason Penilla <11360596+jpenilla@users.noreply.github.com>
2021-08-31 11:02:11 +00:00
EntitySection<T> entitysection = this.sectionStorage.getOrCreateSection(i);
entitysection.add(entity); // CraftBukkit - decompile error
+ this.entitySliceManager.addEntity((Entity)entity); // Paper
entity.setLevelCallback(new PersistentEntitySectionManager.Callback(entity, i, entitysection));
if (!existing) {
this.callbacks.onCreated(entity);
2021-11-05 14:08:24 +00:00
@@ -169,6 +172,7 @@ public class PersistentEntitySectionManager<T extends EntityAccess> implements A
Merge tuinity (#6413)
This PR contains all of Tuinity's patches. Very notable ones are:
- Highly optimised collisions
- Optimised entity lookups by bounding box (Mojang made regressions in 1.17, this brings it back to 1.16)
- Starlight https://github.com/PaperMC/Starlight
- Rewritten dataconverter system https://github.com/PaperMC/DataConverter
- Random block ticking optimisation (wrongly dropped from Paper 1.17)
- Chunk ticking optimisations
- Anything else I've forgotten in the 60 or so patches
If you are a previous Tuinity user, your config will not migrate. You must do it yourself. The config options have simply been moved into paper.yml, so it will be an easy migration. However, please note that the chunk loading options in tuinity.yml are NOT compatible with the options in paper.yml.
* Port tuinity, initial patchset
* Update gradle to 7.2
jmp said it fixes rebuildpatches not working for me. it fucking better
* Completely clean apply
* Remove tuinity config, add per player api patch
* Remove paper reobf mappings patch
* Properly update gradlew
* Force clean rebuild
* Mark fixups
Comments and ATs still need to be done
* grep -r "Tuinity"
* Fixup
* Ensure gameprofile lastaccess is written only under the state lock
* update URL for dataconverter
* Only clean rebuild tuinity patches
might fix merge conflicts
* Use UTF-8 for gradlew
* Clean rb patches again
* Convert block ids used as item ids
Neither the converters of pre 1.13 nor DFU handled these cases,
as by the time they were written the game at the time didn't
consider these ids valid - they would be air. Because of this,
some worlds have logspam since only DataConverter (not DFU or
legacy converters) will warn when an invalid id has been
seen.
While quite a few do need to now be considered as air, quite a lot
do not. So it makes sense to add conversion for these items, instead
of simply suppressing or ignoring the logs. I've now added id -> string conversion
for all block ids that could be used as items that existed in the game
before 1.7.10 (I have no interest in tracking down the
exact version block ids stopped working) that were on
https://minecraft-ids.grahamedgecombe.com/
Items that did not directly convert to new items will
be instead converted to air: stems, wheat crops, piston head,
tripwire wire block
* Fix LightPopulated parsing in V1466
The DFU code was checking if the number existed, not if it
didn't exist. I misread the original code.
* Always parse protochunk light sources unless it is marked as non-lit
Chunks not marked as lit will always go through the light engine,
so they should always have their block sources parsed.
* Update custom names to JSON for players
Missed this fix from CB, as it was inside
the DataFixers class.
I decided to double check all of the CB changes again:
DataFixers.java was the only area I missed, as I had inspected all
datafixer diffs and implemented them all into DataConverter. I also
checked Bootstrap.java again, and re-evaluated their changes. I had
previously done this, but determined that they were all bad.
The changes to make standing_sign block map to oak_sign block in
V1450 is bad, because that's not the item id V1450 accepts. Only
in 1.14 did oak_sign even exist, and as expected there is a converter
to rename all existing sign items/blocks.
The fix to register the portal block under id 1440 is useless, as
the flattenning logic will default to the lowest registered id - which
is the exact blockstate that CB registers into 1440. So it just
doesn't do anything.
The extra item ids in the id -> string converter are already added,
but I found this from EMC originally.
The change for the spawn egg id 23 -> Arrow is just wrong,
that id DOES correspond to TippedArrow, NOT Arrow. As
expected, the spawn egg already has a dedicated mapping for
Arrow, which is id 10 - which was Arrow's entity id.
I also ported a fix for the cooked_fished id update. This doesn't
really matter since there is already a dataconverter to fix this,
but the game didn't accept cooked_fished at the time. So I see
no harm.
* Review all converters and walkers
- Refactor V99 to have helper methods for defining entity/tile
entity types
- Automatically namespace all ids that should be namespaced.
While vanilla never saved non-namespaced data for things that
are namespaced, plugins/users might have.
- Synchronised the identity ensure map in HelperBlockFlatteningV1450
- Code style consistency
- Add missing log warning in V102 for ITEM_NAME type conversion
- Use getBoolean instead of getByte
- Use ConverterAbstractEntityRename for V143 TippedArrow -> Arrow
rename, as it will affect ENTITY_NAME type
- Always set isVillager to false in V502 for Zombie
- Register V808's converter under subversion 1 like DFU
- Register a breakpoint for V1.17.1. In the future, all final
versions of major releases will have a breakpoint so that
the work required to determine if a converter needs a breakpoint
is minimal
- Validate that a dataconverter is only registered for a version
that is registered
- ConverterFlattenTileEntity is actually ConverterFlattenEntity
It even registered the converters under TILE_ENTITY, instead of
ENTITY.
- Fix id comparison in V1492 STRUCTURE_FEATURE renamer
- Use ConverterAbstractStatsRename for V1510 stats renamer
At the time I had written that class, the abstract renamer didn't
exist.
- Ensure OwnerUUID is at least set to empty string in
V1904 if the ocelot is converted to a cat (this is
likely so that it retains a collar)
- Use generic read/write for Records in V1946
Records is actually a list, not a map. So reading map was
invalid.
* Always set light to zero when propagating decrease
This fixes an almost infinite loop where light values
would be spam queued on a very small subset on blocks.
This also likely fixes the memory issues people were
seeing.
* re-organize patches
* Apply and fix conflicts
* Revert some patches
getChunkAt retains chunks so that plugins don't spam loads
revert mc-4 fix will remain unless issues pop up
* Shuffle iterated chunks if per player is not enabled
Can help with some mob spawning stacking up at locations
* Make per player default, migrate all configs
* Adjust comments in fixups
* Rework config for player chunk loader
Old config is not compatible. Move all configs to be
under `settings` in paper.yml
The player chunk loader has been modified to
less aggressively load chunks, but to send
chunks at higher rates compared to tuinity. There are
new config entries to tune this behavior.
* Add back old constructor to CompressionEncoder/Decoder (fixes
Tuinity #358)
* Raise chunk loading default limits
* Reduce worldgen thread workers for lower core count cpus
* Raise limits for chunk loading config
Also place it under `chunk-loading`
* Disable max chunk send rate by default
* Fix conflicts and rebuild patches
* Drop default send rate again
Appears to be still causing problems for no known reason
* Raise chunk send limits to 100 per player
While a low limit fixes ping issues for some people, most people
do not suffer from this issue and thus should not suffer from
an extremely slow load-in rate.
* Rebase part 1
Autosquash the fixups
* Move not implemented up
* Fixup mc-dev fixes
Missed this one
* Rebase per player viewdistance api into the original api patch
* Remove old light engine patch part 1
The prioritisation must be kept from it, so that part
has been rebased into the priority patch.
Part 2 will deal with rebasing all of the patches _after_
* Rebase remaining patches for old light patch removal
* Remove other mid tick patch
* Remove Optimize-PlayerChunkMap-memory-use-for-visibleChunks.patch
Replaced by `Do not copy visible chunks`
* Revert AT for Vec3i setX/Y/Z
The class is immutable. set should not be exposed
* Remove old IntegerUtil class
* Replace old CraftChunk#getEntities patch
* Remove import for SWMRNibbleArray in ChunkAccess
* Finished merge checklist
* Remove ensureTickThread impl in urgency patch
Co-authored-by: Spottedleaf <Spottedleaf@users.noreply.github.com>
Co-authored-by: Jason Penilla <11360596+jpenilla@users.noreply.github.com>
2021-08-31 11:02:11 +00:00
io.papermc.paper.util.TickThread.ensureTickThread("Asynchronous chunk ticking status update"); // Paper
Visibility visibility = Visibility.fromFullChunkStatus(levelType);
+ this.entitySliceManager.chunkStatusChange(chunkPos.x, chunkPos.z, levelType); // Paper
this.updateChunkStatus(chunkPos, visibility);
}
2021-11-05 14:08:24 +00:00
@@ -461,6 +465,7 @@ public class PersistentEntitySectionManager<T extends EntityAccess> implements A
Merge tuinity (#6413)
This PR contains all of Tuinity's patches. Very notable ones are:
- Highly optimised collisions
- Optimised entity lookups by bounding box (Mojang made regressions in 1.17, this brings it back to 1.16)
- Starlight https://github.com/PaperMC/Starlight
- Rewritten dataconverter system https://github.com/PaperMC/DataConverter
- Random block ticking optimisation (wrongly dropped from Paper 1.17)
- Chunk ticking optimisations
- Anything else I've forgotten in the 60 or so patches
If you are a previous Tuinity user, your config will not migrate. You must do it yourself. The config options have simply been moved into paper.yml, so it will be an easy migration. However, please note that the chunk loading options in tuinity.yml are NOT compatible with the options in paper.yml.
* Port tuinity, initial patchset
* Update gradle to 7.2
jmp said it fixes rebuildpatches not working for me. it fucking better
* Completely clean apply
* Remove tuinity config, add per player api patch
* Remove paper reobf mappings patch
* Properly update gradlew
* Force clean rebuild
* Mark fixups
Comments and ATs still need to be done
* grep -r "Tuinity"
* Fixup
* Ensure gameprofile lastaccess is written only under the state lock
* update URL for dataconverter
* Only clean rebuild tuinity patches
might fix merge conflicts
* Use UTF-8 for gradlew
* Clean rb patches again
* Convert block ids used as item ids
Neither the converters of pre 1.13 nor DFU handled these cases,
as by the time they were written the game at the time didn't
consider these ids valid - they would be air. Because of this,
some worlds have logspam since only DataConverter (not DFU or
legacy converters) will warn when an invalid id has been
seen.
While quite a few do need to now be considered as air, quite a lot
do not. So it makes sense to add conversion for these items, instead
of simply suppressing or ignoring the logs. I've now added id -> string conversion
for all block ids that could be used as items that existed in the game
before 1.7.10 (I have no interest in tracking down the
exact version block ids stopped working) that were on
https://minecraft-ids.grahamedgecombe.com/
Items that did not directly convert to new items will
be instead converted to air: stems, wheat crops, piston head,
tripwire wire block
* Fix LightPopulated parsing in V1466
The DFU code was checking if the number existed, not if it
didn't exist. I misread the original code.
* Always parse protochunk light sources unless it is marked as non-lit
Chunks not marked as lit will always go through the light engine,
so they should always have their block sources parsed.
* Update custom names to JSON for players
Missed this fix from CB, as it was inside
the DataFixers class.
I decided to double check all of the CB changes again:
DataFixers.java was the only area I missed, as I had inspected all
datafixer diffs and implemented them all into DataConverter. I also
checked Bootstrap.java again, and re-evaluated their changes. I had
previously done this, but determined that they were all bad.
The changes to make standing_sign block map to oak_sign block in
V1450 is bad, because that's not the item id V1450 accepts. Only
in 1.14 did oak_sign even exist, and as expected there is a converter
to rename all existing sign items/blocks.
The fix to register the portal block under id 1440 is useless, as
the flattenning logic will default to the lowest registered id - which
is the exact blockstate that CB registers into 1440. So it just
doesn't do anything.
The extra item ids in the id -> string converter are already added,
but I found this from EMC originally.
The change for the spawn egg id 23 -> Arrow is just wrong,
that id DOES correspond to TippedArrow, NOT Arrow. As
expected, the spawn egg already has a dedicated mapping for
Arrow, which is id 10 - which was Arrow's entity id.
I also ported a fix for the cooked_fished id update. This doesn't
really matter since there is already a dataconverter to fix this,
but the game didn't accept cooked_fished at the time. So I see
no harm.
* Review all converters and walkers
- Refactor V99 to have helper methods for defining entity/tile
entity types
- Automatically namespace all ids that should be namespaced.
While vanilla never saved non-namespaced data for things that
are namespaced, plugins/users might have.
- Synchronised the identity ensure map in HelperBlockFlatteningV1450
- Code style consistency
- Add missing log warning in V102 for ITEM_NAME type conversion
- Use getBoolean instead of getByte
- Use ConverterAbstractEntityRename for V143 TippedArrow -> Arrow
rename, as it will affect ENTITY_NAME type
- Always set isVillager to false in V502 for Zombie
- Register V808's converter under subversion 1 like DFU
- Register a breakpoint for V1.17.1. In the future, all final
versions of major releases will have a breakpoint so that
the work required to determine if a converter needs a breakpoint
is minimal
- Validate that a dataconverter is only registered for a version
that is registered
- ConverterFlattenTileEntity is actually ConverterFlattenEntity
It even registered the converters under TILE_ENTITY, instead of
ENTITY.
- Fix id comparison in V1492 STRUCTURE_FEATURE renamer
- Use ConverterAbstractStatsRename for V1510 stats renamer
At the time I had written that class, the abstract renamer didn't
exist.
- Ensure OwnerUUID is at least set to empty string in
V1904 if the ocelot is converted to a cat (this is
likely so that it retains a collar)
- Use generic read/write for Records in V1946
Records is actually a list, not a map. So reading map was
invalid.
* Always set light to zero when propagating decrease
This fixes an almost infinite loop where light values
would be spam queued on a very small subset on blocks.
This also likely fixes the memory issues people were
seeing.
* re-organize patches
* Apply and fix conflicts
* Revert some patches
getChunkAt retains chunks so that plugins don't spam loads
revert mc-4 fix will remain unless issues pop up
* Shuffle iterated chunks if per player is not enabled
Can help with some mob spawning stacking up at locations
* Make per player default, migrate all configs
* Adjust comments in fixups
* Rework config for player chunk loader
Old config is not compatible. Move all configs to be
under `settings` in paper.yml
The player chunk loader has been modified to
less aggressively load chunks, but to send
chunks at higher rates compared to tuinity. There are
new config entries to tune this behavior.
* Add back old constructor to CompressionEncoder/Decoder (fixes
Tuinity #358)
* Raise chunk loading default limits
* Reduce worldgen thread workers for lower core count cpus
* Raise limits for chunk loading config
Also place it under `chunk-loading`
* Disable max chunk send rate by default
* Fix conflicts and rebuild patches
* Drop default send rate again
Appears to be still causing problems for no known reason
* Raise chunk send limits to 100 per player
While a low limit fixes ping issues for some people, most people
do not suffer from this issue and thus should not suffer from
an extremely slow load-in rate.
* Rebase part 1
Autosquash the fixups
* Move not implemented up
* Fixup mc-dev fixes
Missed this one
* Rebase per player viewdistance api into the original api patch
* Remove old light engine patch part 1
The prioritisation must be kept from it, so that part
has been rebased into the priority patch.
Part 2 will deal with rebasing all of the patches _after_
* Rebase remaining patches for old light patch removal
* Remove other mid tick patch
* Remove Optimize-PlayerChunkMap-memory-use-for-visibleChunks.patch
Replaced by `Do not copy visible chunks`
* Revert AT for Vec3i setX/Y/Z
The class is immutable. set should not be exposed
* Remove old IntegerUtil class
* Replace old CraftChunk#getEntities patch
* Remove import for SWMRNibbleArray in ChunkAccess
* Finished merge checklist
* Remove ensureTickThread impl in urgency patch
Co-authored-by: Spottedleaf <Spottedleaf@users.noreply.github.com>
Co-authored-by: Jason Penilla <11360596+jpenilla@users.noreply.github.com>
2021-08-31 11:02:11 +00:00
long i = SectionPos.asLong(blockposition);
if (i != this.currentSectionKey) {
+ PersistentEntitySectionManager.this.entitySliceManager.moveEntity((Entity)this.entity); // Paper
Visibility visibility = this.currentSection.getStatus();
if (!this.currentSection.remove(this.entity)) {
2021-11-05 14:08:24 +00:00
@@ -509,6 +514,7 @@ public class PersistentEntitySectionManager<T extends EntityAccess> implements A
Merge tuinity (#6413)
This PR contains all of Tuinity's patches. Very notable ones are:
- Highly optimised collisions
- Optimised entity lookups by bounding box (Mojang made regressions in 1.17, this brings it back to 1.16)
- Starlight https://github.com/PaperMC/Starlight
- Rewritten dataconverter system https://github.com/PaperMC/DataConverter
- Random block ticking optimisation (wrongly dropped from Paper 1.17)
- Chunk ticking optimisations
- Anything else I've forgotten in the 60 or so patches
If you are a previous Tuinity user, your config will not migrate. You must do it yourself. The config options have simply been moved into paper.yml, so it will be an easy migration. However, please note that the chunk loading options in tuinity.yml are NOT compatible with the options in paper.yml.
* Port tuinity, initial patchset
* Update gradle to 7.2
jmp said it fixes rebuildpatches not working for me. it fucking better
* Completely clean apply
* Remove tuinity config, add per player api patch
* Remove paper reobf mappings patch
* Properly update gradlew
* Force clean rebuild
* Mark fixups
Comments and ATs still need to be done
* grep -r "Tuinity"
* Fixup
* Ensure gameprofile lastaccess is written only under the state lock
* update URL for dataconverter
* Only clean rebuild tuinity patches
might fix merge conflicts
* Use UTF-8 for gradlew
* Clean rb patches again
* Convert block ids used as item ids
Neither the converters of pre 1.13 nor DFU handled these cases,
as by the time they were written the game at the time didn't
consider these ids valid - they would be air. Because of this,
some worlds have logspam since only DataConverter (not DFU or
legacy converters) will warn when an invalid id has been
seen.
While quite a few do need to now be considered as air, quite a lot
do not. So it makes sense to add conversion for these items, instead
of simply suppressing or ignoring the logs. I've now added id -> string conversion
for all block ids that could be used as items that existed in the game
before 1.7.10 (I have no interest in tracking down the
exact version block ids stopped working) that were on
https://minecraft-ids.grahamedgecombe.com/
Items that did not directly convert to new items will
be instead converted to air: stems, wheat crops, piston head,
tripwire wire block
* Fix LightPopulated parsing in V1466
The DFU code was checking if the number existed, not if it
didn't exist. I misread the original code.
* Always parse protochunk light sources unless it is marked as non-lit
Chunks not marked as lit will always go through the light engine,
so they should always have their block sources parsed.
* Update custom names to JSON for players
Missed this fix from CB, as it was inside
the DataFixers class.
I decided to double check all of the CB changes again:
DataFixers.java was the only area I missed, as I had inspected all
datafixer diffs and implemented them all into DataConverter. I also
checked Bootstrap.java again, and re-evaluated their changes. I had
previously done this, but determined that they were all bad.
The changes to make standing_sign block map to oak_sign block in
V1450 is bad, because that's not the item id V1450 accepts. Only
in 1.14 did oak_sign even exist, and as expected there is a converter
to rename all existing sign items/blocks.
The fix to register the portal block under id 1440 is useless, as
the flattenning logic will default to the lowest registered id - which
is the exact blockstate that CB registers into 1440. So it just
doesn't do anything.
The extra item ids in the id -> string converter are already added,
but I found this from EMC originally.
The change for the spawn egg id 23 -> Arrow is just wrong,
that id DOES correspond to TippedArrow, NOT Arrow. As
expected, the spawn egg already has a dedicated mapping for
Arrow, which is id 10 - which was Arrow's entity id.
I also ported a fix for the cooked_fished id update. This doesn't
really matter since there is already a dataconverter to fix this,
but the game didn't accept cooked_fished at the time. So I see
no harm.
* Review all converters and walkers
- Refactor V99 to have helper methods for defining entity/tile
entity types
- Automatically namespace all ids that should be namespaced.
While vanilla never saved non-namespaced data for things that
are namespaced, plugins/users might have.
- Synchronised the identity ensure map in HelperBlockFlatteningV1450
- Code style consistency
- Add missing log warning in V102 for ITEM_NAME type conversion
- Use getBoolean instead of getByte
- Use ConverterAbstractEntityRename for V143 TippedArrow -> Arrow
rename, as it will affect ENTITY_NAME type
- Always set isVillager to false in V502 for Zombie
- Register V808's converter under subversion 1 like DFU
- Register a breakpoint for V1.17.1. In the future, all final
versions of major releases will have a breakpoint so that
the work required to determine if a converter needs a breakpoint
is minimal
- Validate that a dataconverter is only registered for a version
that is registered
- ConverterFlattenTileEntity is actually ConverterFlattenEntity
It even registered the converters under TILE_ENTITY, instead of
ENTITY.
- Fix id comparison in V1492 STRUCTURE_FEATURE renamer
- Use ConverterAbstractStatsRename for V1510 stats renamer
At the time I had written that class, the abstract renamer didn't
exist.
- Ensure OwnerUUID is at least set to empty string in
V1904 if the ocelot is converted to a cat (this is
likely so that it retains a collar)
- Use generic read/write for Records in V1946
Records is actually a list, not a map. So reading map was
invalid.
* Always set light to zero when propagating decrease
This fixes an almost infinite loop where light values
would be spam queued on a very small subset on blocks.
This also likely fixes the memory issues people were
seeing.
* re-organize patches
* Apply and fix conflicts
* Revert some patches
getChunkAt retains chunks so that plugins don't spam loads
revert mc-4 fix will remain unless issues pop up
* Shuffle iterated chunks if per player is not enabled
Can help with some mob spawning stacking up at locations
* Make per player default, migrate all configs
* Adjust comments in fixups
* Rework config for player chunk loader
Old config is not compatible. Move all configs to be
under `settings` in paper.yml
The player chunk loader has been modified to
less aggressively load chunks, but to send
chunks at higher rates compared to tuinity. There are
new config entries to tune this behavior.
* Add back old constructor to CompressionEncoder/Decoder (fixes
Tuinity #358)
* Raise chunk loading default limits
* Reduce worldgen thread workers for lower core count cpus
* Raise limits for chunk loading config
Also place it under `chunk-loading`
* Disable max chunk send rate by default
* Fix conflicts and rebuild patches
* Drop default send rate again
Appears to be still causing problems for no known reason
* Raise chunk send limits to 100 per player
While a low limit fixes ping issues for some people, most people
do not suffer from this issue and thus should not suffer from
an extremely slow load-in rate.
* Rebase part 1
Autosquash the fixups
* Move not implemented up
* Fixup mc-dev fixes
Missed this one
* Rebase per player viewdistance api into the original api patch
* Remove old light engine patch part 1
The prioritisation must be kept from it, so that part
has been rebased into the priority patch.
Part 2 will deal with rebasing all of the patches _after_
* Rebase remaining patches for old light patch removal
* Remove other mid tick patch
* Remove Optimize-PlayerChunkMap-memory-use-for-visibleChunks.patch
Replaced by `Do not copy visible chunks`
* Revert AT for Vec3i setX/Y/Z
The class is immutable. set should not be exposed
* Remove old IntegerUtil class
* Replace old CraftChunk#getEntities patch
* Remove import for SWMRNibbleArray in ChunkAccess
* Finished merge checklist
* Remove ensureTickThread impl in urgency patch
Co-authored-by: Spottedleaf <Spottedleaf@users.noreply.github.com>
Co-authored-by: Jason Penilla <11360596+jpenilla@users.noreply.github.com>
2021-08-31 11:02:11 +00:00
if (!this.currentSection.remove(this.entity)) {
PersistentEntitySectionManager.LOGGER.warn("Entity {} wasn't found in section {} (destroying due to {})", this.entity, SectionPos.of(this.currentSectionKey), reason);
}
+ PersistentEntitySectionManager.this.entitySliceManager.removeEntity((Entity)this.entity); // Paper
Visibility visibility = PersistentEntitySectionManager.getEffectiveStatus(this.entity, this.currentSection.getStatus());
diff --git a/src/main/java/org/bukkit/craftbukkit/CraftChunk.java b/src/main/java/org/bukkit/craftbukkit/CraftChunk.java
2021-11-05 14:08:24 +00:00
index 5088c84f6518cb686241b1db54faa8d813cb3eaa..07dfe575dea0f41a75bb9dad0ee2d541e983f6d7 100644
Merge tuinity (#6413)
This PR contains all of Tuinity's patches. Very notable ones are:
- Highly optimised collisions
- Optimised entity lookups by bounding box (Mojang made regressions in 1.17, this brings it back to 1.16)
- Starlight https://github.com/PaperMC/Starlight
- Rewritten dataconverter system https://github.com/PaperMC/DataConverter
- Random block ticking optimisation (wrongly dropped from Paper 1.17)
- Chunk ticking optimisations
- Anything else I've forgotten in the 60 or so patches
If you are a previous Tuinity user, your config will not migrate. You must do it yourself. The config options have simply been moved into paper.yml, so it will be an easy migration. However, please note that the chunk loading options in tuinity.yml are NOT compatible with the options in paper.yml.
* Port tuinity, initial patchset
* Update gradle to 7.2
jmp said it fixes rebuildpatches not working for me. it fucking better
* Completely clean apply
* Remove tuinity config, add per player api patch
* Remove paper reobf mappings patch
* Properly update gradlew
* Force clean rebuild
* Mark fixups
Comments and ATs still need to be done
* grep -r "Tuinity"
* Fixup
* Ensure gameprofile lastaccess is written only under the state lock
* update URL for dataconverter
* Only clean rebuild tuinity patches
might fix merge conflicts
* Use UTF-8 for gradlew
* Clean rb patches again
* Convert block ids used as item ids
Neither the converters of pre 1.13 nor DFU handled these cases,
as by the time they were written the game at the time didn't
consider these ids valid - they would be air. Because of this,
some worlds have logspam since only DataConverter (not DFU or
legacy converters) will warn when an invalid id has been
seen.
While quite a few do need to now be considered as air, quite a lot
do not. So it makes sense to add conversion for these items, instead
of simply suppressing or ignoring the logs. I've now added id -> string conversion
for all block ids that could be used as items that existed in the game
before 1.7.10 (I have no interest in tracking down the
exact version block ids stopped working) that were on
https://minecraft-ids.grahamedgecombe.com/
Items that did not directly convert to new items will
be instead converted to air: stems, wheat crops, piston head,
tripwire wire block
* Fix LightPopulated parsing in V1466
The DFU code was checking if the number existed, not if it
didn't exist. I misread the original code.
* Always parse protochunk light sources unless it is marked as non-lit
Chunks not marked as lit will always go through the light engine,
so they should always have their block sources parsed.
* Update custom names to JSON for players
Missed this fix from CB, as it was inside
the DataFixers class.
I decided to double check all of the CB changes again:
DataFixers.java was the only area I missed, as I had inspected all
datafixer diffs and implemented them all into DataConverter. I also
checked Bootstrap.java again, and re-evaluated their changes. I had
previously done this, but determined that they were all bad.
The changes to make standing_sign block map to oak_sign block in
V1450 is bad, because that's not the item id V1450 accepts. Only
in 1.14 did oak_sign even exist, and as expected there is a converter
to rename all existing sign items/blocks.
The fix to register the portal block under id 1440 is useless, as
the flattenning logic will default to the lowest registered id - which
is the exact blockstate that CB registers into 1440. So it just
doesn't do anything.
The extra item ids in the id -> string converter are already added,
but I found this from EMC originally.
The change for the spawn egg id 23 -> Arrow is just wrong,
that id DOES correspond to TippedArrow, NOT Arrow. As
expected, the spawn egg already has a dedicated mapping for
Arrow, which is id 10 - which was Arrow's entity id.
I also ported a fix for the cooked_fished id update. This doesn't
really matter since there is already a dataconverter to fix this,
but the game didn't accept cooked_fished at the time. So I see
no harm.
* Review all converters and walkers
- Refactor V99 to have helper methods for defining entity/tile
entity types
- Automatically namespace all ids that should be namespaced.
While vanilla never saved non-namespaced data for things that
are namespaced, plugins/users might have.
- Synchronised the identity ensure map in HelperBlockFlatteningV1450
- Code style consistency
- Add missing log warning in V102 for ITEM_NAME type conversion
- Use getBoolean instead of getByte
- Use ConverterAbstractEntityRename for V143 TippedArrow -> Arrow
rename, as it will affect ENTITY_NAME type
- Always set isVillager to false in V502 for Zombie
- Register V808's converter under subversion 1 like DFU
- Register a breakpoint for V1.17.1. In the future, all final
versions of major releases will have a breakpoint so that
the work required to determine if a converter needs a breakpoint
is minimal
- Validate that a dataconverter is only registered for a version
that is registered
- ConverterFlattenTileEntity is actually ConverterFlattenEntity
It even registered the converters under TILE_ENTITY, instead of
ENTITY.
- Fix id comparison in V1492 STRUCTURE_FEATURE renamer
- Use ConverterAbstractStatsRename for V1510 stats renamer
At the time I had written that class, the abstract renamer didn't
exist.
- Ensure OwnerUUID is at least set to empty string in
V1904 if the ocelot is converted to a cat (this is
likely so that it retains a collar)
- Use generic read/write for Records in V1946
Records is actually a list, not a map. So reading map was
invalid.
* Always set light to zero when propagating decrease
This fixes an almost infinite loop where light values
would be spam queued on a very small subset on blocks.
This also likely fixes the memory issues people were
seeing.
* re-organize patches
* Apply and fix conflicts
* Revert some patches
getChunkAt retains chunks so that plugins don't spam loads
revert mc-4 fix will remain unless issues pop up
* Shuffle iterated chunks if per player is not enabled
Can help with some mob spawning stacking up at locations
* Make per player default, migrate all configs
* Adjust comments in fixups
* Rework config for player chunk loader
Old config is not compatible. Move all configs to be
under `settings` in paper.yml
The player chunk loader has been modified to
less aggressively load chunks, but to send
chunks at higher rates compared to tuinity. There are
new config entries to tune this behavior.
* Add back old constructor to CompressionEncoder/Decoder (fixes
Tuinity #358)
* Raise chunk loading default limits
* Reduce worldgen thread workers for lower core count cpus
* Raise limits for chunk loading config
Also place it under `chunk-loading`
* Disable max chunk send rate by default
* Fix conflicts and rebuild patches
* Drop default send rate again
Appears to be still causing problems for no known reason
* Raise chunk send limits to 100 per player
While a low limit fixes ping issues for some people, most people
do not suffer from this issue and thus should not suffer from
an extremely slow load-in rate.
* Rebase part 1
Autosquash the fixups
* Move not implemented up
* Fixup mc-dev fixes
Missed this one
* Rebase per player viewdistance api into the original api patch
* Remove old light engine patch part 1
The prioritisation must be kept from it, so that part
has been rebased into the priority patch.
Part 2 will deal with rebasing all of the patches _after_
* Rebase remaining patches for old light patch removal
* Remove other mid tick patch
* Remove Optimize-PlayerChunkMap-memory-use-for-visibleChunks.patch
Replaced by `Do not copy visible chunks`
* Revert AT for Vec3i setX/Y/Z
The class is immutable. set should not be exposed
* Remove old IntegerUtil class
* Replace old CraftChunk#getEntities patch
* Remove import for SWMRNibbleArray in ChunkAccess
* Finished merge checklist
* Remove ensureTickThread impl in urgency patch
Co-authored-by: Spottedleaf <Spottedleaf@users.noreply.github.com>
Co-authored-by: Jason Penilla <11360596+jpenilla@users.noreply.github.com>
2021-08-31 11:02:11 +00:00
--- a/src/main/java/org/bukkit/craftbukkit/CraftChunk.java
+++ b/src/main/java/org/bukkit/craftbukkit/CraftChunk.java
2021-11-05 14:08:24 +00:00
@@ -131,9 +131,7 @@ public class CraftChunk implements Chunk {
2021-09-01 12:03:36 +00:00
long pair = ChunkPos.asLong(x, z);
if (entityManager.areEntitiesLoaded(pair)) { // PAIL rename isEntitiesLoaded
- return entityManager.getEntities(new ChunkPos(this.x, this.z)).stream()
- .map(net.minecraft.world.entity.Entity::getBukkitEntity)
- .filter(Objects::nonNull).toArray(Entity[]::new);
+ return getCraftWorld().getHandle().getChunkEntities(this.x, this.z); // Paper - optimise this
Merge tuinity (#6413)
This PR contains all of Tuinity's patches. Very notable ones are:
- Highly optimised collisions
- Optimised entity lookups by bounding box (Mojang made regressions in 1.17, this brings it back to 1.16)
- Starlight https://github.com/PaperMC/Starlight
- Rewritten dataconverter system https://github.com/PaperMC/DataConverter
- Random block ticking optimisation (wrongly dropped from Paper 1.17)
- Chunk ticking optimisations
- Anything else I've forgotten in the 60 or so patches
If you are a previous Tuinity user, your config will not migrate. You must do it yourself. The config options have simply been moved into paper.yml, so it will be an easy migration. However, please note that the chunk loading options in tuinity.yml are NOT compatible with the options in paper.yml.
* Port tuinity, initial patchset
* Update gradle to 7.2
jmp said it fixes rebuildpatches not working for me. it fucking better
* Completely clean apply
* Remove tuinity config, add per player api patch
* Remove paper reobf mappings patch
* Properly update gradlew
* Force clean rebuild
* Mark fixups
Comments and ATs still need to be done
* grep -r "Tuinity"
* Fixup
* Ensure gameprofile lastaccess is written only under the state lock
* update URL for dataconverter
* Only clean rebuild tuinity patches
might fix merge conflicts
* Use UTF-8 for gradlew
* Clean rb patches again
* Convert block ids used as item ids
Neither the converters of pre 1.13 nor DFU handled these cases,
as by the time they were written the game at the time didn't
consider these ids valid - they would be air. Because of this,
some worlds have logspam since only DataConverter (not DFU or
legacy converters) will warn when an invalid id has been
seen.
While quite a few do need to now be considered as air, quite a lot
do not. So it makes sense to add conversion for these items, instead
of simply suppressing or ignoring the logs. I've now added id -> string conversion
for all block ids that could be used as items that existed in the game
before 1.7.10 (I have no interest in tracking down the
exact version block ids stopped working) that were on
https://minecraft-ids.grahamedgecombe.com/
Items that did not directly convert to new items will
be instead converted to air: stems, wheat crops, piston head,
tripwire wire block
* Fix LightPopulated parsing in V1466
The DFU code was checking if the number existed, not if it
didn't exist. I misread the original code.
* Always parse protochunk light sources unless it is marked as non-lit
Chunks not marked as lit will always go through the light engine,
so they should always have their block sources parsed.
* Update custom names to JSON for players
Missed this fix from CB, as it was inside
the DataFixers class.
I decided to double check all of the CB changes again:
DataFixers.java was the only area I missed, as I had inspected all
datafixer diffs and implemented them all into DataConverter. I also
checked Bootstrap.java again, and re-evaluated their changes. I had
previously done this, but determined that they were all bad.
The changes to make standing_sign block map to oak_sign block in
V1450 is bad, because that's not the item id V1450 accepts. Only
in 1.14 did oak_sign even exist, and as expected there is a converter
to rename all existing sign items/blocks.
The fix to register the portal block under id 1440 is useless, as
the flattenning logic will default to the lowest registered id - which
is the exact blockstate that CB registers into 1440. So it just
doesn't do anything.
The extra item ids in the id -> string converter are already added,
but I found this from EMC originally.
The change for the spawn egg id 23 -> Arrow is just wrong,
that id DOES correspond to TippedArrow, NOT Arrow. As
expected, the spawn egg already has a dedicated mapping for
Arrow, which is id 10 - which was Arrow's entity id.
I also ported a fix for the cooked_fished id update. This doesn't
really matter since there is already a dataconverter to fix this,
but the game didn't accept cooked_fished at the time. So I see
no harm.
* Review all converters and walkers
- Refactor V99 to have helper methods for defining entity/tile
entity types
- Automatically namespace all ids that should be namespaced.
While vanilla never saved non-namespaced data for things that
are namespaced, plugins/users might have.
- Synchronised the identity ensure map in HelperBlockFlatteningV1450
- Code style consistency
- Add missing log warning in V102 for ITEM_NAME type conversion
- Use getBoolean instead of getByte
- Use ConverterAbstractEntityRename for V143 TippedArrow -> Arrow
rename, as it will affect ENTITY_NAME type
- Always set isVillager to false in V502 for Zombie
- Register V808's converter under subversion 1 like DFU
- Register a breakpoint for V1.17.1. In the future, all final
versions of major releases will have a breakpoint so that
the work required to determine if a converter needs a breakpoint
is minimal
- Validate that a dataconverter is only registered for a version
that is registered
- ConverterFlattenTileEntity is actually ConverterFlattenEntity
It even registered the converters under TILE_ENTITY, instead of
ENTITY.
- Fix id comparison in V1492 STRUCTURE_FEATURE renamer
- Use ConverterAbstractStatsRename for V1510 stats renamer
At the time I had written that class, the abstract renamer didn't
exist.
- Ensure OwnerUUID is at least set to empty string in
V1904 if the ocelot is converted to a cat (this is
likely so that it retains a collar)
- Use generic read/write for Records in V1946
Records is actually a list, not a map. So reading map was
invalid.
* Always set light to zero when propagating decrease
This fixes an almost infinite loop where light values
would be spam queued on a very small subset on blocks.
This also likely fixes the memory issues people were
seeing.
* re-organize patches
* Apply and fix conflicts
* Revert some patches
getChunkAt retains chunks so that plugins don't spam loads
revert mc-4 fix will remain unless issues pop up
* Shuffle iterated chunks if per player is not enabled
Can help with some mob spawning stacking up at locations
* Make per player default, migrate all configs
* Adjust comments in fixups
* Rework config for player chunk loader
Old config is not compatible. Move all configs to be
under `settings` in paper.yml
The player chunk loader has been modified to
less aggressively load chunks, but to send
chunks at higher rates compared to tuinity. There are
new config entries to tune this behavior.
* Add back old constructor to CompressionEncoder/Decoder (fixes
Tuinity #358)
* Raise chunk loading default limits
* Reduce worldgen thread workers for lower core count cpus
* Raise limits for chunk loading config
Also place it under `chunk-loading`
* Disable max chunk send rate by default
* Fix conflicts and rebuild patches
* Drop default send rate again
Appears to be still causing problems for no known reason
* Raise chunk send limits to 100 per player
While a low limit fixes ping issues for some people, most people
do not suffer from this issue and thus should not suffer from
an extremely slow load-in rate.
* Rebase part 1
Autosquash the fixups
* Move not implemented up
* Fixup mc-dev fixes
Missed this one
* Rebase per player viewdistance api into the original api patch
* Remove old light engine patch part 1
The prioritisation must be kept from it, so that part
has been rebased into the priority patch.
Part 2 will deal with rebasing all of the patches _after_
* Rebase remaining patches for old light patch removal
* Remove other mid tick patch
* Remove Optimize-PlayerChunkMap-memory-use-for-visibleChunks.patch
Replaced by `Do not copy visible chunks`
* Revert AT for Vec3i setX/Y/Z
The class is immutable. set should not be exposed
* Remove old IntegerUtil class
* Replace old CraftChunk#getEntities patch
* Remove import for SWMRNibbleArray in ChunkAccess
* Finished merge checklist
* Remove ensureTickThread impl in urgency patch
Co-authored-by: Spottedleaf <Spottedleaf@users.noreply.github.com>
Co-authored-by: Jason Penilla <11360596+jpenilla@users.noreply.github.com>
2021-08-31 11:02:11 +00:00
}
2021-09-01 12:03:36 +00:00
entityManager.ensureChunkQueuedForLoad(pair); // Start entity loading
2021-11-05 14:08:24 +00:00
@@ -169,9 +167,7 @@ public class CraftChunk implements Chunk {
}
}
2021-09-01 12:03:36 +00:00
- return entityManager.getEntities(new ChunkPos(this.x, this.z)).stream()
- .map(net.minecraft.world.entity.Entity::getBukkitEntity)
- .filter(Objects::nonNull).toArray(Entity[]::new);
+ return getCraftWorld().getHandle().getChunkEntities(this.x, this.z); // Paper - optimise this
Merge tuinity (#6413)
This PR contains all of Tuinity's patches. Very notable ones are:
- Highly optimised collisions
- Optimised entity lookups by bounding box (Mojang made regressions in 1.17, this brings it back to 1.16)
- Starlight https://github.com/PaperMC/Starlight
- Rewritten dataconverter system https://github.com/PaperMC/DataConverter
- Random block ticking optimisation (wrongly dropped from Paper 1.17)
- Chunk ticking optimisations
- Anything else I've forgotten in the 60 or so patches
If you are a previous Tuinity user, your config will not migrate. You must do it yourself. The config options have simply been moved into paper.yml, so it will be an easy migration. However, please note that the chunk loading options in tuinity.yml are NOT compatible with the options in paper.yml.
* Port tuinity, initial patchset
* Update gradle to 7.2
jmp said it fixes rebuildpatches not working for me. it fucking better
* Completely clean apply
* Remove tuinity config, add per player api patch
* Remove paper reobf mappings patch
* Properly update gradlew
* Force clean rebuild
* Mark fixups
Comments and ATs still need to be done
* grep -r "Tuinity"
* Fixup
* Ensure gameprofile lastaccess is written only under the state lock
* update URL for dataconverter
* Only clean rebuild tuinity patches
might fix merge conflicts
* Use UTF-8 for gradlew
* Clean rb patches again
* Convert block ids used as item ids
Neither the converters of pre 1.13 nor DFU handled these cases,
as by the time they were written the game at the time didn't
consider these ids valid - they would be air. Because of this,
some worlds have logspam since only DataConverter (not DFU or
legacy converters) will warn when an invalid id has been
seen.
While quite a few do need to now be considered as air, quite a lot
do not. So it makes sense to add conversion for these items, instead
of simply suppressing or ignoring the logs. I've now added id -> string conversion
for all block ids that could be used as items that existed in the game
before 1.7.10 (I have no interest in tracking down the
exact version block ids stopped working) that were on
https://minecraft-ids.grahamedgecombe.com/
Items that did not directly convert to new items will
be instead converted to air: stems, wheat crops, piston head,
tripwire wire block
* Fix LightPopulated parsing in V1466
The DFU code was checking if the number existed, not if it
didn't exist. I misread the original code.
* Always parse protochunk light sources unless it is marked as non-lit
Chunks not marked as lit will always go through the light engine,
so they should always have their block sources parsed.
* Update custom names to JSON for players
Missed this fix from CB, as it was inside
the DataFixers class.
I decided to double check all of the CB changes again:
DataFixers.java was the only area I missed, as I had inspected all
datafixer diffs and implemented them all into DataConverter. I also
checked Bootstrap.java again, and re-evaluated their changes. I had
previously done this, but determined that they were all bad.
The changes to make standing_sign block map to oak_sign block in
V1450 is bad, because that's not the item id V1450 accepts. Only
in 1.14 did oak_sign even exist, and as expected there is a converter
to rename all existing sign items/blocks.
The fix to register the portal block under id 1440 is useless, as
the flattenning logic will default to the lowest registered id - which
is the exact blockstate that CB registers into 1440. So it just
doesn't do anything.
The extra item ids in the id -> string converter are already added,
but I found this from EMC originally.
The change for the spawn egg id 23 -> Arrow is just wrong,
that id DOES correspond to TippedArrow, NOT Arrow. As
expected, the spawn egg already has a dedicated mapping for
Arrow, which is id 10 - which was Arrow's entity id.
I also ported a fix for the cooked_fished id update. This doesn't
really matter since there is already a dataconverter to fix this,
but the game didn't accept cooked_fished at the time. So I see
no harm.
* Review all converters and walkers
- Refactor V99 to have helper methods for defining entity/tile
entity types
- Automatically namespace all ids that should be namespaced.
While vanilla never saved non-namespaced data for things that
are namespaced, plugins/users might have.
- Synchronised the identity ensure map in HelperBlockFlatteningV1450
- Code style consistency
- Add missing log warning in V102 for ITEM_NAME type conversion
- Use getBoolean instead of getByte
- Use ConverterAbstractEntityRename for V143 TippedArrow -> Arrow
rename, as it will affect ENTITY_NAME type
- Always set isVillager to false in V502 for Zombie
- Register V808's converter under subversion 1 like DFU
- Register a breakpoint for V1.17.1. In the future, all final
versions of major releases will have a breakpoint so that
the work required to determine if a converter needs a breakpoint
is minimal
- Validate that a dataconverter is only registered for a version
that is registered
- ConverterFlattenTileEntity is actually ConverterFlattenEntity
It even registered the converters under TILE_ENTITY, instead of
ENTITY.
- Fix id comparison in V1492 STRUCTURE_FEATURE renamer
- Use ConverterAbstractStatsRename for V1510 stats renamer
At the time I had written that class, the abstract renamer didn't
exist.
- Ensure OwnerUUID is at least set to empty string in
V1904 if the ocelot is converted to a cat (this is
likely so that it retains a collar)
- Use generic read/write for Records in V1946
Records is actually a list, not a map. So reading map was
invalid.
* Always set light to zero when propagating decrease
This fixes an almost infinite loop where light values
would be spam queued on a very small subset on blocks.
This also likely fixes the memory issues people were
seeing.
* re-organize patches
* Apply and fix conflicts
* Revert some patches
getChunkAt retains chunks so that plugins don't spam loads
revert mc-4 fix will remain unless issues pop up
* Shuffle iterated chunks if per player is not enabled
Can help with some mob spawning stacking up at locations
* Make per player default, migrate all configs
* Adjust comments in fixups
* Rework config for player chunk loader
Old config is not compatible. Move all configs to be
under `settings` in paper.yml
The player chunk loader has been modified to
less aggressively load chunks, but to send
chunks at higher rates compared to tuinity. There are
new config entries to tune this behavior.
* Add back old constructor to CompressionEncoder/Decoder (fixes
Tuinity #358)
* Raise chunk loading default limits
* Reduce worldgen thread workers for lower core count cpus
* Raise limits for chunk loading config
Also place it under `chunk-loading`
* Disable max chunk send rate by default
* Fix conflicts and rebuild patches
* Drop default send rate again
Appears to be still causing problems for no known reason
* Raise chunk send limits to 100 per player
While a low limit fixes ping issues for some people, most people
do not suffer from this issue and thus should not suffer from
an extremely slow load-in rate.
* Rebase part 1
Autosquash the fixups
* Move not implemented up
* Fixup mc-dev fixes
Missed this one
* Rebase per player viewdistance api into the original api patch
* Remove old light engine patch part 1
The prioritisation must be kept from it, so that part
has been rebased into the priority patch.
Part 2 will deal with rebasing all of the patches _after_
* Rebase remaining patches for old light patch removal
* Remove other mid tick patch
* Remove Optimize-PlayerChunkMap-memory-use-for-visibleChunks.patch
Replaced by `Do not copy visible chunks`
* Revert AT for Vec3i setX/Y/Z
The class is immutable. set should not be exposed
* Remove old IntegerUtil class
* Replace old CraftChunk#getEntities patch
* Remove import for SWMRNibbleArray in ChunkAccess
* Finished merge checklist
* Remove ensureTickThread impl in urgency patch
Co-authored-by: Spottedleaf <Spottedleaf@users.noreply.github.com>
Co-authored-by: Jason Penilla <11360596+jpenilla@users.noreply.github.com>
2021-08-31 11:02:11 +00:00
}
@Override
diff --git a/src/main/java/org/bukkit/craftbukkit/util/DummyGeneratorAccess.java b/src/main/java/org/bukkit/craftbukkit/util/DummyGeneratorAccess.java
index c5d132dcf70a55de041c0187ff8fb889248bdfee..950d4381459d31d02acf55c4aef4f5e33367748b 100644
--- a/src/main/java/org/bukkit/craftbukkit/util/DummyGeneratorAccess.java
+++ b/src/main/java/org/bukkit/craftbukkit/util/DummyGeneratorAccess.java
@@ -254,4 +254,20 @@ public class DummyGeneratorAccess implements WorldGenLevel {
public Stream<? extends StructureStart<?>> startsForFeature(SectionPos pos, StructureFeature<?> feature) {
throw new UnsupportedOperationException("Not supported yet.");
}
+
+ // Paper start
+ @Override
+ public List<Entity> getHardCollidingEntities(Entity except, AABB box, Predicate<? super Entity> predicate) {
+ return java.util.Collections.emptyList();
+ }
+
+ @Override
+ public void getEntities(Entity except, AABB box, Predicate<? super Entity> predicate, List<Entity> into) {}
+
+ @Override
+ public void getHardCollidingEntities(Entity except, AABB box, Predicate<? super Entity> predicate, List<Entity> into) {}
+
+ @Override
+ public <T> void getEntitiesByClass(Class<? extends T> clazz, Entity except, AABB box, List<? super T> into, Predicate<? super T> predicate) {}
+ // Paper end
}
diff --git a/src/main/java/org/spigotmc/ActivationRange.java b/src/main/java/org/spigotmc/ActivationRange.java
2021-10-20 03:14:30 +00:00
index 85449fc6d19974622588e7f13e1dc78c8dffbeee..9c456cce42ef9d1654df9047d6fc1e0da13dc1c9 100644
Merge tuinity (#6413)
This PR contains all of Tuinity's patches. Very notable ones are:
- Highly optimised collisions
- Optimised entity lookups by bounding box (Mojang made regressions in 1.17, this brings it back to 1.16)
- Starlight https://github.com/PaperMC/Starlight
- Rewritten dataconverter system https://github.com/PaperMC/DataConverter
- Random block ticking optimisation (wrongly dropped from Paper 1.17)
- Chunk ticking optimisations
- Anything else I've forgotten in the 60 or so patches
If you are a previous Tuinity user, your config will not migrate. You must do it yourself. The config options have simply been moved into paper.yml, so it will be an easy migration. However, please note that the chunk loading options in tuinity.yml are NOT compatible with the options in paper.yml.
* Port tuinity, initial patchset
* Update gradle to 7.2
jmp said it fixes rebuildpatches not working for me. it fucking better
* Completely clean apply
* Remove tuinity config, add per player api patch
* Remove paper reobf mappings patch
* Properly update gradlew
* Force clean rebuild
* Mark fixups
Comments and ATs still need to be done
* grep -r "Tuinity"
* Fixup
* Ensure gameprofile lastaccess is written only under the state lock
* update URL for dataconverter
* Only clean rebuild tuinity patches
might fix merge conflicts
* Use UTF-8 for gradlew
* Clean rb patches again
* Convert block ids used as item ids
Neither the converters of pre 1.13 nor DFU handled these cases,
as by the time they were written the game at the time didn't
consider these ids valid - they would be air. Because of this,
some worlds have logspam since only DataConverter (not DFU or
legacy converters) will warn when an invalid id has been
seen.
While quite a few do need to now be considered as air, quite a lot
do not. So it makes sense to add conversion for these items, instead
of simply suppressing or ignoring the logs. I've now added id -> string conversion
for all block ids that could be used as items that existed in the game
before 1.7.10 (I have no interest in tracking down the
exact version block ids stopped working) that were on
https://minecraft-ids.grahamedgecombe.com/
Items that did not directly convert to new items will
be instead converted to air: stems, wheat crops, piston head,
tripwire wire block
* Fix LightPopulated parsing in V1466
The DFU code was checking if the number existed, not if it
didn't exist. I misread the original code.
* Always parse protochunk light sources unless it is marked as non-lit
Chunks not marked as lit will always go through the light engine,
so they should always have their block sources parsed.
* Update custom names to JSON for players
Missed this fix from CB, as it was inside
the DataFixers class.
I decided to double check all of the CB changes again:
DataFixers.java was the only area I missed, as I had inspected all
datafixer diffs and implemented them all into DataConverter. I also
checked Bootstrap.java again, and re-evaluated their changes. I had
previously done this, but determined that they were all bad.
The changes to make standing_sign block map to oak_sign block in
V1450 is bad, because that's not the item id V1450 accepts. Only
in 1.14 did oak_sign even exist, and as expected there is a converter
to rename all existing sign items/blocks.
The fix to register the portal block under id 1440 is useless, as
the flattenning logic will default to the lowest registered id - which
is the exact blockstate that CB registers into 1440. So it just
doesn't do anything.
The extra item ids in the id -> string converter are already added,
but I found this from EMC originally.
The change for the spawn egg id 23 -> Arrow is just wrong,
that id DOES correspond to TippedArrow, NOT Arrow. As
expected, the spawn egg already has a dedicated mapping for
Arrow, which is id 10 - which was Arrow's entity id.
I also ported a fix for the cooked_fished id update. This doesn't
really matter since there is already a dataconverter to fix this,
but the game didn't accept cooked_fished at the time. So I see
no harm.
* Review all converters and walkers
- Refactor V99 to have helper methods for defining entity/tile
entity types
- Automatically namespace all ids that should be namespaced.
While vanilla never saved non-namespaced data for things that
are namespaced, plugins/users might have.
- Synchronised the identity ensure map in HelperBlockFlatteningV1450
- Code style consistency
- Add missing log warning in V102 for ITEM_NAME type conversion
- Use getBoolean instead of getByte
- Use ConverterAbstractEntityRename for V143 TippedArrow -> Arrow
rename, as it will affect ENTITY_NAME type
- Always set isVillager to false in V502 for Zombie
- Register V808's converter under subversion 1 like DFU
- Register a breakpoint for V1.17.1. In the future, all final
versions of major releases will have a breakpoint so that
the work required to determine if a converter needs a breakpoint
is minimal
- Validate that a dataconverter is only registered for a version
that is registered
- ConverterFlattenTileEntity is actually ConverterFlattenEntity
It even registered the converters under TILE_ENTITY, instead of
ENTITY.
- Fix id comparison in V1492 STRUCTURE_FEATURE renamer
- Use ConverterAbstractStatsRename for V1510 stats renamer
At the time I had written that class, the abstract renamer didn't
exist.
- Ensure OwnerUUID is at least set to empty string in
V1904 if the ocelot is converted to a cat (this is
likely so that it retains a collar)
- Use generic read/write for Records in V1946
Records is actually a list, not a map. So reading map was
invalid.
* Always set light to zero when propagating decrease
This fixes an almost infinite loop where light values
would be spam queued on a very small subset on blocks.
This also likely fixes the memory issues people were
seeing.
* re-organize patches
* Apply and fix conflicts
* Revert some patches
getChunkAt retains chunks so that plugins don't spam loads
revert mc-4 fix will remain unless issues pop up
* Shuffle iterated chunks if per player is not enabled
Can help with some mob spawning stacking up at locations
* Make per player default, migrate all configs
* Adjust comments in fixups
* Rework config for player chunk loader
Old config is not compatible. Move all configs to be
under `settings` in paper.yml
The player chunk loader has been modified to
less aggressively load chunks, but to send
chunks at higher rates compared to tuinity. There are
new config entries to tune this behavior.
* Add back old constructor to CompressionEncoder/Decoder (fixes
Tuinity #358)
* Raise chunk loading default limits
* Reduce worldgen thread workers for lower core count cpus
* Raise limits for chunk loading config
Also place it under `chunk-loading`
* Disable max chunk send rate by default
* Fix conflicts and rebuild patches
* Drop default send rate again
Appears to be still causing problems for no known reason
* Raise chunk send limits to 100 per player
While a low limit fixes ping issues for some people, most people
do not suffer from this issue and thus should not suffer from
an extremely slow load-in rate.
* Rebase part 1
Autosquash the fixups
* Move not implemented up
* Fixup mc-dev fixes
Missed this one
* Rebase per player viewdistance api into the original api patch
* Remove old light engine patch part 1
The prioritisation must be kept from it, so that part
has been rebased into the priority patch.
Part 2 will deal with rebasing all of the patches _after_
* Rebase remaining patches for old light patch removal
* Remove other mid tick patch
* Remove Optimize-PlayerChunkMap-memory-use-for-visibleChunks.patch
Replaced by `Do not copy visible chunks`
* Revert AT for Vec3i setX/Y/Z
The class is immutable. set should not be exposed
* Remove old IntegerUtil class
* Replace old CraftChunk#getEntities patch
* Remove import for SWMRNibbleArray in ChunkAccess
* Finished merge checklist
* Remove ensureTickThread impl in urgency patch
Co-authored-by: Spottedleaf <Spottedleaf@users.noreply.github.com>
Co-authored-by: Jason Penilla <11360596+jpenilla@users.noreply.github.com>
2021-08-31 11:02:11 +00:00
--- a/src/main/java/org/spigotmc/ActivationRange.java
+++ b/src/main/java/org/spigotmc/ActivationRange.java
@@ -205,7 +205,13 @@ public class ActivationRange
2021-10-20 03:14:30 +00:00
ActivationType.VILLAGER.boundingBox = player.getBoundingBox().inflate( villagerActivationRange, 256, villagerActivationRange );
Merge tuinity (#6413)
This PR contains all of Tuinity's patches. Very notable ones are:
- Highly optimised collisions
- Optimised entity lookups by bounding box (Mojang made regressions in 1.17, this brings it back to 1.16)
- Starlight https://github.com/PaperMC/Starlight
- Rewritten dataconverter system https://github.com/PaperMC/DataConverter
- Random block ticking optimisation (wrongly dropped from Paper 1.17)
- Chunk ticking optimisations
- Anything else I've forgotten in the 60 or so patches
If you are a previous Tuinity user, your config will not migrate. You must do it yourself. The config options have simply been moved into paper.yml, so it will be an easy migration. However, please note that the chunk loading options in tuinity.yml are NOT compatible with the options in paper.yml.
* Port tuinity, initial patchset
* Update gradle to 7.2
jmp said it fixes rebuildpatches not working for me. it fucking better
* Completely clean apply
* Remove tuinity config, add per player api patch
* Remove paper reobf mappings patch
* Properly update gradlew
* Force clean rebuild
* Mark fixups
Comments and ATs still need to be done
* grep -r "Tuinity"
* Fixup
* Ensure gameprofile lastaccess is written only under the state lock
* update URL for dataconverter
* Only clean rebuild tuinity patches
might fix merge conflicts
* Use UTF-8 for gradlew
* Clean rb patches again
* Convert block ids used as item ids
Neither the converters of pre 1.13 nor DFU handled these cases,
as by the time they were written the game at the time didn't
consider these ids valid - they would be air. Because of this,
some worlds have logspam since only DataConverter (not DFU or
legacy converters) will warn when an invalid id has been
seen.
While quite a few do need to now be considered as air, quite a lot
do not. So it makes sense to add conversion for these items, instead
of simply suppressing or ignoring the logs. I've now added id -> string conversion
for all block ids that could be used as items that existed in the game
before 1.7.10 (I have no interest in tracking down the
exact version block ids stopped working) that were on
https://minecraft-ids.grahamedgecombe.com/
Items that did not directly convert to new items will
be instead converted to air: stems, wheat crops, piston head,
tripwire wire block
* Fix LightPopulated parsing in V1466
The DFU code was checking if the number existed, not if it
didn't exist. I misread the original code.
* Always parse protochunk light sources unless it is marked as non-lit
Chunks not marked as lit will always go through the light engine,
so they should always have their block sources parsed.
* Update custom names to JSON for players
Missed this fix from CB, as it was inside
the DataFixers class.
I decided to double check all of the CB changes again:
DataFixers.java was the only area I missed, as I had inspected all
datafixer diffs and implemented them all into DataConverter. I also
checked Bootstrap.java again, and re-evaluated their changes. I had
previously done this, but determined that they were all bad.
The changes to make standing_sign block map to oak_sign block in
V1450 is bad, because that's not the item id V1450 accepts. Only
in 1.14 did oak_sign even exist, and as expected there is a converter
to rename all existing sign items/blocks.
The fix to register the portal block under id 1440 is useless, as
the flattenning logic will default to the lowest registered id - which
is the exact blockstate that CB registers into 1440. So it just
doesn't do anything.
The extra item ids in the id -> string converter are already added,
but I found this from EMC originally.
The change for the spawn egg id 23 -> Arrow is just wrong,
that id DOES correspond to TippedArrow, NOT Arrow. As
expected, the spawn egg already has a dedicated mapping for
Arrow, which is id 10 - which was Arrow's entity id.
I also ported a fix for the cooked_fished id update. This doesn't
really matter since there is already a dataconverter to fix this,
but the game didn't accept cooked_fished at the time. So I see
no harm.
* Review all converters and walkers
- Refactor V99 to have helper methods for defining entity/tile
entity types
- Automatically namespace all ids that should be namespaced.
While vanilla never saved non-namespaced data for things that
are namespaced, plugins/users might have.
- Synchronised the identity ensure map in HelperBlockFlatteningV1450
- Code style consistency
- Add missing log warning in V102 for ITEM_NAME type conversion
- Use getBoolean instead of getByte
- Use ConverterAbstractEntityRename for V143 TippedArrow -> Arrow
rename, as it will affect ENTITY_NAME type
- Always set isVillager to false in V502 for Zombie
- Register V808's converter under subversion 1 like DFU
- Register a breakpoint for V1.17.1. In the future, all final
versions of major releases will have a breakpoint so that
the work required to determine if a converter needs a breakpoint
is minimal
- Validate that a dataconverter is only registered for a version
that is registered
- ConverterFlattenTileEntity is actually ConverterFlattenEntity
It even registered the converters under TILE_ENTITY, instead of
ENTITY.
- Fix id comparison in V1492 STRUCTURE_FEATURE renamer
- Use ConverterAbstractStatsRename for V1510 stats renamer
At the time I had written that class, the abstract renamer didn't
exist.
- Ensure OwnerUUID is at least set to empty string in
V1904 if the ocelot is converted to a cat (this is
likely so that it retains a collar)
- Use generic read/write for Records in V1946
Records is actually a list, not a map. So reading map was
invalid.
* Always set light to zero when propagating decrease
This fixes an almost infinite loop where light values
would be spam queued on a very small subset on blocks.
This also likely fixes the memory issues people were
seeing.
* re-organize patches
* Apply and fix conflicts
* Revert some patches
getChunkAt retains chunks so that plugins don't spam loads
revert mc-4 fix will remain unless issues pop up
* Shuffle iterated chunks if per player is not enabled
Can help with some mob spawning stacking up at locations
* Make per player default, migrate all configs
* Adjust comments in fixups
* Rework config for player chunk loader
Old config is not compatible. Move all configs to be
under `settings` in paper.yml
The player chunk loader has been modified to
less aggressively load chunks, but to send
chunks at higher rates compared to tuinity. There are
new config entries to tune this behavior.
* Add back old constructor to CompressionEncoder/Decoder (fixes
Tuinity #358)
* Raise chunk loading default limits
* Reduce worldgen thread workers for lower core count cpus
* Raise limits for chunk loading config
Also place it under `chunk-loading`
* Disable max chunk send rate by default
* Fix conflicts and rebuild patches
* Drop default send rate again
Appears to be still causing problems for no known reason
* Raise chunk send limits to 100 per player
While a low limit fixes ping issues for some people, most people
do not suffer from this issue and thus should not suffer from
an extremely slow load-in rate.
* Rebase part 1
Autosquash the fixups
* Move not implemented up
* Fixup mc-dev fixes
Missed this one
* Rebase per player viewdistance api into the original api patch
* Remove old light engine patch part 1
The prioritisation must be kept from it, so that part
has been rebased into the priority patch.
Part 2 will deal with rebasing all of the patches _after_
* Rebase remaining patches for old light patch removal
* Remove other mid tick patch
* Remove Optimize-PlayerChunkMap-memory-use-for-visibleChunks.patch
Replaced by `Do not copy visible chunks`
* Revert AT for Vec3i setX/Y/Z
The class is immutable. set should not be exposed
* Remove old IntegerUtil class
* Replace old CraftChunk#getEntities patch
* Remove import for SWMRNibbleArray in ChunkAccess
* Finished merge checklist
* Remove ensureTickThread impl in urgency patch
Co-authored-by: Spottedleaf <Spottedleaf@users.noreply.github.com>
Co-authored-by: Jason Penilla <11360596+jpenilla@users.noreply.github.com>
2021-08-31 11:02:11 +00:00
// Paper end
- world.getEntities().get(maxBB, ActivationRange::activateEntity);
+ // Paper start
+ java.util.List<Entity> entities = world.getEntities((Entity)null, maxBB, null);
+ for (int i = 0; i < entities.size(); i++) {
+ Entity entity = entities.get(i);
+ ActivationRange.activateEntity(entity);
+ }
+ // Paper end
}
MinecraftTimings.entityActivationCheckTimer.stopTiming();
}