mirrors/PocketMine-MP
1
0
Fork 0
mirror of https://github.com/pmmp/PocketMine-MP.git synced 2025-08-28 05:55:12 +00:00
Code Issues Projects Releases Wiki Activity
12,583 Commits 20 Branches 447 Tags
Commit Graph

97 Commits

Author SHA1 Message Date
Dylan K. Taylor
e09d78238f Removed EmptySubChunk and SubChunkInterface 2020-10-31 23:10:31 +00:00
Dylan K. Taylor
f50be1ba63 Chunk: Use getSubChunkChecked() in getFullBlock() 2020-10-31 23:06:12 +00:00
Dylan K. Taylor
01001dca74 Use Chunk::getSubChunkChecked() in places where we know that an invalid coordinate cannot be requested 2020-10-31 22:48:41 +00:00
Dylan K. Taylor
b270029161 Rename Chunk::getWritableSubChunk() -> Chunk::getSubChunkChecked() 
this is not specific to 'writable', it's just an opt-in to checked bounds so that an EmptySubChunk will never be received.
 2020-10-31 22:46:33 +00:00
Dylan K. Taylor
0cbc5c9a4a region: harden handling of ByteArrayTag 
previously this would just explode if the wrong length of data was given.
 2020-10-31 15:37:06 +00:00
Dylan K. Taylor
3e1263eb79 Chunk: remove all proxy APIs to lighting information 
these aren't used by internals and they shouldn't be used by plugins either.
 2020-10-29 12:11:06 +00:00
Dylan K. Taylor
1859dac789 Implemented self-contained (pass 1) chunk relighting 
this doesn't handle propagating light across chunk borders yet, since that's much more complex to implement.
 2020-10-27 18:29:32 +00:00
Dylan K. Taylor
8bab9cc108 RegionWorldProvider: use morton2d directly instead of abusing chunkHash() 
while these currently do the same thing, it's very confusing and creates an unnecessary dependency on World.
 2020-10-26 15:34:32 +00:00
Dylan K. Taylor
d6ed542fcd Merge remote-tracking branch 'origin/stable' into master 
# Conflicts:
#	composer.json
#	composer.lock
#	src/CrashDump.php
#	src/pocketmine/Server.php
#	src/pocketmine/level/format/io/region/McRegion.php
#	tests/phpstan/configs/check-explicit-mixed-baseline.neon
#	tests/phpstan/configs/l7-baseline.neon
#	tests/phpstan/configs/l8-baseline.neon
#	tests/travis/setup-php.yml
 2020-10-26 16:18:53 +00:00
Dylan K. Taylor
4c0c2ebd24 CS cleanup 2020-10-26 15:56:30 +00:00
Dylan K. Taylor
9191e75392 LevelDB: quick and dirty hack for air with bad metadata 
this needs a proper solution, but this is a pressing issue that can't wait.
 2020-10-03 16:41:26 +01:00
Dylan K. Taylor
75e3a0aa0f Merge branch 'stable' into master 
# Conflicts:
#	composer.lock
#	resources/vanilla
#	src/command/defaults/StatusCommand.php
#	src/pocketmine/entity/Entity.php
#	src/pocketmine/level/format/SubChunk.php
 2020-09-26 13:27:08 +01:00
Dylan K. Taylor
89cce4c749 performance: only calculate light for chunks inside ticking areas 
this produces a major performance improvement for large render distances, and reduces the impact of lighting calculation to zero on servers which have random blockupdates turned off.
 2020-09-26 13:13:12 +01:00
Dylan K. Taylor
03de2bcc67 Chunk: simplify heightmap calculation 2020-09-08 23:03:52 +01:00
Dylan K. Taylor
0fd3d91038 LightArray: hide constants ZERO and FIFTEEN from the API 
this makes it easier to implement this in C++ with the same API. Since the C++ version doesn't use strings, these constants aren't needed anyway.
 2020-09-08 15:13:58 +01:00
Dylan K. Taylor
01f8116cdd Fix some of the implicit immutability issues of EmptySubChunk 
it's useful to have an immutable stub around for the sake of feeding back dummy read values, but for write values it has to barf instead of being quiet.
There's still some issues with LightArray which I don't currently have a solution for, but I'm thinking about separating light storage from chunks anyway.
 2020-09-07 14:43:26 +01:00
Dylan K. Taylor
5fbc842f7a LightArray: do not accept NULL in the constructor 
it makes more sense to pass LightArray::ZERO or just use LightArray::fill(0) if a zeroed light array is desired.
 2020-09-06 16:18:47 +01:00
Dylan K. Taylor
6b6f77f8af Chunk: improved heightmap calculation performance 
recalculateHeightMapColumn is stateless, so it can't make any assumptions about which subchunks to check for blocks. However, in most the average case (6 allocated subchunks), this causes 2500+ useless SubChunk->getHighestBlockAt() calls (10 per column). Since we're calculating in bulk, we can figure out which subchunks are empty one time and ignore them for all 256 columns.
In the average case, this produced a 50-60% performance improvement for heightmap calculation (~1.1 ms -> 0.5 ms).
In extreme cases where the height is extremely varied, this produces no observable performance benefit, but for most cases with flattish terrain, it's an improvement.
It can likely be further improved, but further performance improvements are outside the scope of this commit and will likely result in more complexity increases.
 2020-09-05 00:16:07 +01:00
Dylan K. Taylor
e1816bd415 SubChunkInterface: clarify documentation of isEmptyFast() 2020-09-04 22:10:42 +01:00
Dylan K. Taylor
a68b0b1f93 LightArray: make collectGarbage() more preload-friendly 
opcache preloading doesn't store non-class constants, but it does store class constants. Class constants that reference non-class constants are pre-resolved into values before storing in opcache SHM, so class constants are OK to use, but non-class constants will come back as undefined.
 2020-09-04 17:23:23 +01:00
Dylan K. Taylor
b81cc671e9 Major performance improvement to basic sky light calculation 
this was degraded whenever it was I decided to make chunks always be allocated. This commit uses a fast path for light filling in subchunks which are completely clear of the heightmap, which returns the performance back to its old fast levels.
 2020-09-04 15:51:33 +01:00
Dylan K. Taylor
279abb871d Remove all usages of CompoundTag->hasTag() 
in pretty much every case, these usages really wanted to read the tag's contents anyway, which can be combined with a getTag() and instanceof call for more concise and static analysis friendly code.
In the few cases where the tag contents wasn't needed, it still wanted to check the type, which, again, can be done in a more static analysis friendly way by just using getTag() and instanceof.
 2020-07-10 21:01:43 +01:00
Dylan K. Taylor
a5d77d5106 LevelDB: remove unused function 2020-07-08 23:24:54 +01:00
Dylan K. Taylor
0188323d74 fixed a bunch of NBT-related phpstan level 8 errors 2020-07-05 21:01:13 +01:00
Dylan K. Taylor
a2677eba02 Merge commit 'f3063e797fab56ff8d0213c8e5a3904085a6eb79' 
# Conflicts:
#	resources/vanilla
#	src/pocketmine/level/format/io/ChunkUtils.php
 2020-06-17 20:43:30 +01:00
Dylan K. Taylor
d2130265f8 Merge commit '8dcc88712c6b8a4d4a0c6be2f6b908ae85378209' 
# Conflicts:
#	resources/vanilla
#	src/pocketmine/level/format/io/ChunkUtils.php
#	tests/phpstan/configs/l7-baseline.neon
 2020-06-17 20:42:28 +01:00
Dylan K. Taylor
506f98efc4 Eliminate usages of BaseNbtSerializer->readCompressed() and BaseNbtSerializer->writeCompressed() 2020-06-17 14:31:56 +01:00
Dylan K. Taylor
c618932d25 Eliminate usages of BinaryStream->setBuffer() and BinaryStream->reset() 2020-06-17 13:49:43 +01:00
Dylan K. Taylor
7d73630fb7 Merge branch 'stable' 
# Conflicts:
#	resources/vanilla
#	src/world/format/io/region/RegionGarbageMap.php
#	src/world/format/io/region/RegionLoader.php
#	tests/phpstan/configs/l7-baseline.neon
#	tests/phpunit/world/format/io/region/RegionLocationTableEntryTest.php
 2020-06-15 20:10:15 +01:00
Dylan K. Taylor
0d186fe42c Merge commit '725506510' 
# Conflicts:
#	resources/vanilla
#	src/pocketmine/level/format/io/leveldb/LevelDB.php
 2020-06-13 11:55:10 +01:00
Dylan K. Taylor
c93038f574 Inject WorldProviderManager to WorldManager's constructor, no longer singleton 2020-05-24 19:22:04 +01:00
Dylan K. Taylor
437e4d75ab WorldProviderManager: use $this-> instead of self:: 2020-05-24 19:16:57 +01:00
Dylan K. Taylor
b05fab3e3c FormatConverter: do not hardcode progress update interval 2020-05-24 14:30:59 +01:00
Dylan K. Taylor
c95951479c FormatConverter: beware paths with \ on linux 2020-05-24 14:28:07 +01:00
Dylan K. Taylor
c9af5ce7a9 Convert GeneratorManager to singleton 2020-05-23 10:13:03 +01:00
Dylan K. Taylor
640428c415 Convert WorldProviderManager to singleton 2020-05-23 10:02:09 +01:00
Dylan K. Taylor
b3454b3488 BiomeArray: check lower bounds in ::idx() as well as upper bounds 2020-05-16 18:42:47 +01:00
Dylan K. Taylor
ae6a7b7cc9 imports cleanup 2020-05-16 18:42:08 +01:00
Dylan K. Taylor
88715c7055 extracted a HeightArray type from Chunk 2020-05-16 18:41:27 +01:00
Dylan K. Taylor
a31240f60b FastChunkSerializer: added constants for internal flags 2020-05-16 18:06:00 +01:00
Dylan K. Taylor
82d361d75f extract a BiomeArray unit from Chunk 
this now also properly validates data read from disk.
 2020-05-16 17:36:22 +01:00
Dylan K. Taylor
45f9c61d0f forward-port of 2f47597d75b514944a2f3f5d293f2fb72dd7fb56 as best it fits 2020-05-13 13:28:35 +01:00
Dylan K. Taylor
ac5cf2443e convert TileFactory to singleton 2020-04-26 01:11:30 +01:00
Dylan K. Taylor
7d9df6af6f Convert EntityFactory to singleton 2020-04-24 22:43:02 +01:00
Dylan K. Taylor
3a42c21cc1 wrap up block_id_map in a class 2020-04-24 21:28:27 +01:00
Dylan K. Taylor
accc0da0cb Chunk no longer depends on BlockFactory 
really this light population crap shouldn't be in the chunk to begin with, but that's a bit more complicated.
 2020-04-23 21:31:28 +01:00
Dylan K. Taylor
163c3855eb Merge branch 'next-minor' 
# Conflicts:
#	resources/vanilla
#	src/plugin/PluginBase.php
#	src/plugin/PluginDescription.php
#	src/pocketmine/Player.php
#	src/pocketmine/network/rcon/RCON.php
#	src/pocketmine/network/rcon/RCONInstance.php
#	src/pocketmine/scheduler/AsyncTask.php
#	src/pocketmine/tile/Spawnable.php
#	src/scheduler/AsyncPool.php
#	src/utils/Config.php
#	src/utils/Timezone.php
#	src/utils/UUID.php
#	src/utils/Utils.php
#	src/world/format/io/region/RegionLoader.php
 2020-04-19 11:13:41 +01:00
Dylan K. Taylor
ca909ebc1d Merge branch 'next-minor' 2020-03-10 13:05:37 +00:00
Dylan K. Taylor
995309424e updated pocketmine/nbt dependency 
this is going to need work on exception handling, but right now it's so inconsistent that it doesn't matter anyway.
 2020-03-04 17:53:37 +00:00
Dylan K. Taylor
a633e415ef FastChunkSerializer: use machine endianness for pack() (thanks @Frago9876543210) 
this is faster due to not having to reverse bytes. Since we don't use this format for persistence, it's OK to use machine byte order for this.
 2020-02-25 15:08:53 +00:00