mirrors/PocketMine-MP
1
0
Fork 0
mirror of https://github.com/pmmp/PocketMine-MP.git synced 2025-05-21 21:14:39 +00:00
Code Issues Projects Releases Wiki Activity
10,114 Commits 16 Branches 441 Tags
Commit Graph

8868 Commits

Author SHA1 Message Date
Dylan K. Taylor
49d3a42120 phpstan: make EntityEvent generic, fix a bunch of 'actual-problems' ignored errors 2020-06-21 00:27:32 +01:00
Dylan K. Taylor
c523595e85 Rewrite TeleportCommand (sadly I can't make this commit any smaller) 
this pile of shit was overdue a rewrite. The new version is much easier to understand.
 2020-06-21 00:04:18 +01:00
Dylan K. Taylor
7c7e4f2093 WhitelistCommand: fixed silence on unknown subcommand 2020-06-20 23:05:39 +01:00
Dylan K. Taylor
88c1014f03 TimingsCommand: fixed silence when using a nonexisting subcommand 2020-06-20 23:04:08 +01:00
Dylan K. Taylor
e105578be0 LegacySkinAdapter: an extra check for resourcePatch (fixes 3 phpstan explicitMixed errors) 2020-06-18 20:08:38 +01:00
Matt
c601352777
Fixed wrong meta value when pick cake block. (#3593) 2020-06-18 11:14:57 +01:00
Dylan K. Taylor
77c71e22b2 SkinAdapter::fromSkinData() may now throw InvalidSkinException 
fixes a rogue TODO in LegacySkinAdapter and invalid skins maybe showing up as Steve instead of getting kicked off the server
 2020-06-17 21:44:22 +01:00
Dylan K. Taylor
1c13ba5656 Avoid parameter ordering bugs during packet decoding 
A PhpStorm refactor could have side effects on code that directly reads stuff from the packet input stream in the arguments block, because those calls will get moved into a different order if the constructor gets refactored. This would, obviously, break packet decoding, so that's something we should avoid and really should not encourage.
 2020-06-17 21:01:01 +01:00
Dylan K. Taylor
f970be0e4d SkinImage: 128x64 is not a valid classic skin size 
MC itself doesn't accept classic skins of this size.
 2020-06-17 20:31:28 +01:00
Dylan K. Taylor
11a3f9f1b9 VerifyLoginTask: fast-fail by checking header x5u before verifying signature 
this is less costly, although it doesn't make any difference except in invalid cases.
 2020-06-17 17:52:19 +01:00
Dylan K. Taylor
09771849ae VerifyLoginTask optimisation: do not copy the entire LoginPacket into the worker thread 
this is especially bad considering the fact that the cached buffer is copied. That said, it's only a few kilobytes so it's not a huge problem, but nonetheless...
 2020-06-17 17:46:22 +01:00
Dylan K. Taylor
57a310230a ScoreboardIdentityPacketEntry: added missing field default 2020-06-15 23:59:53 +01:00
Dylan K. Taylor
130c55d9f1 EntityLink: remove rogue default ctor parameters 2020-06-15 23:58:11 +01:00
Dylan K. Taylor
2712befa82 SkinData: fixed capeImage type violation (doesn't accept null) 2020-06-15 23:51:48 +01:00
Dylan K. Taylor
a4e250a3e6 TextFormat: improved exception messages for PCRE failures 2020-06-15 23:43:01 +01:00
Dylan K. Taylor
23b97d8e2d TextFormat: wrap all preg_replace() usages in a type-safe exception-throwing version 
fixes 3 phpstan level 8 errors
 2020-06-15 23:31:46 +01:00
Dylan K. Taylor
b0b1b29de4 Chunk: specify list<int> for heightMap 2020-06-15 23:07:35 +01:00
Dylan K. Taylor
f3063e797f ChunkUtils: provide a more explicit parameter type 2020-06-15 23:05:52 +01:00
Dylan K. Taylor
8dcc88712c ChunkUtils: fixed phpstan level 7 type inference error 2020-06-15 23:04:12 +01:00
Dylan K. Taylor
04191ec44a Rail: specify type for local static variable 
fixes a phpstan level 7 error
phpstan doesn't make any assumptions about local static variable types because analysing them would require too much work, apparently.
 2020-06-15 22:47:20 +01:00
Dylan K. Taylor
62ea7c93a9 added a dedicated InvalidSkinException 2020-06-15 21:48:24 +01:00
Dylan K. Taylor
cf06b5b8cf Player: explicitly check for false return of dataPacket() 
fixes a phpstan error on l7
this won't ever actually be a problem, but this isn't obvious from the type system.
 2020-06-15 21:24:19 +01:00
Dylan K. Taylor
a8ec51daac Player: do not assign maybe-false result of json_encode() to ModalFormRequestPacket 
fixes a phpstan error on l7
 2020-06-15 21:22:33 +01:00
Dylan K. Taylor
6a7b77fee2 first look at making region writes reuse old space 2020-06-15 18:36:54 +01:00
Dylan K. Taylor
b902f9ded0 RegionLoader: fixed 2 phpstan level 7 errors (eof reading region header) 2020-06-15 15:09:37 +01:00
Dylan K. Taylor
9bb8a8f761 RegionLoader: added utility function getProportionUnusedSpace() 2020-06-15 13:49:03 +01:00
Dylan K. Taylor
63b14a083c RegionLoader: added utility function generateSectorMap() 
this proved very useful while debugging some internal issues.
 2020-06-15 13:48:17 +01:00
Dylan K. Taylor
627a7c951a RegionLoader: added missing const import 2020-06-15 13:46:03 +01:00
Dylan K. Taylor
bb2685ca65 RegionLocationTableEntry: cap firstSector at 16777216 
this is the biggest sector start that the location table can represent, due to the binary format. Larger values than this will overflow and cause corruption.
This provides an effective limit of 64 GB on region files.
 2020-06-15 13:23:08 +01:00
Dylan K. Taylor
d38709a7ae RegionLoader: remove unused variable 2020-06-15 12:26:20 +01:00
Dylan K. Taylor
b559a65346 RegionLoader: account for possible corrupted header pointing to itself 2020-06-15 12:13:42 +01:00
Dylan K. Taylor
b92a2ded8a RegionLoader: check for zero sector count when loading location table 
implementations shouldn't be writing location entries that have an offset but zero sectors, but just in case they do, we need to be aware of it.
 2020-06-15 12:08:55 +01:00
Dylan K. Taylor
22f25dfbdb RegionLocationTableEntry: require sector count to be at least 1 2020-06-15 12:05:48 +01:00
Dylan K. Taylor
6bf840c72e RegionLoader: use actual null instead of zeroed entry for non-allocated chunks 
this forces the code to be properly aware of non-allocated chunks, because it'll crash with NPE if it isn't.
 2020-06-15 12:02:03 +01:00
Dylan K. Taylor
745be19a56 RegionLoader: fixed regions ballooning when writing chunks to the end of file 
we already have a region growth problem due to the lack of garbage collection, but this bug was making it worse. If the region already contained 1024 allocated chunks, 4MB of file space would get wasted before the next chunk would be appended to the file.
 2020-06-14 23:40:33 +01:00
Dylan K. Taylor
e05bee5ffb RegionLoader: do a full check for chunk overlaps during initial load 2020-06-14 22:39:01 +01:00
Dylan K. Taylor
d8d994351b phpstan 0.12.29 2020-06-14 16:25:55 +01:00
Dylan K. Taylor
0029efa370 Server: add getPlayerDataPath(), reduce logic duplication 2020-06-14 12:40:24 +01:00
Dylan K. Taylor
df13e967fd imports cleanup 2020-06-14 10:27:15 +01:00
Dylan K. Taylor
097c260dbb Eradicate all usages of strtoupper() 
strtoupper() is an evil function whose behaviour depends on the system locale. mb_strtoupper() has more consistent behaviour.
 2020-06-13 19:47:00 +01:00
Dylan K. Taylor
710c162604 QueryRegenerateEvent: fixed possible type violation on listPlugins 2020-06-10 12:11:39 +01:00
Dylan K. Taylor
409c8c1703 TimingsCommand: workaround a PHPStan type specifying bug 2020-06-10 12:11:10 +01:00
Dylan K. Taylor
376926c700 TimingsCommand: fix missing type information for async task local storage 2020-06-10 12:10:33 +01:00
Dylan K. Taylor
c3fabe833e FileWriteTask: mark as deprecated 2020-06-10 11:22:18 +01:00
Dylan K. Taylor
3e09ff5350 EnchantTable: fix formatting issue [ci skip] 2020-06-10 10:54:22 +01:00
Dylan K. Taylor
7255065106 LevelDB: stop passing false to places where it's not expected 2020-06-10 10:45:54 +01:00
Dylan K. Taylor
76f1add3b3 Timezone: return false if date_parse() fails 
I hate this, but I don't want to change it to throw right now because it'll create a bunch of extra work.
 2020-06-10 10:34:34 +01:00
Dylan K. Taylor
fcc9e62c65 CommandEnum: specify enumValues type 
phpstan 0.12.26 starts reporting errors about the result of array_search() being given to some constructor or another because of the lack of key type specification.
 2020-06-10 10:33:35 +01:00
Dylan K. Taylor
3d50aafcc4 ShapedRecipe: remove a curly-brace array-access that nothing noticed 2020-06-06 11:12:45 +01:00
Dylan K. Taylor
a27b29897c TimingsCommand: tell operators what's going on for audit purposes 
previously nobody except the person who was managing timings would know that timings was running, being pasted or whatever else. Since timings can impact performance (and, for example, block the main thread when writing timings to a file), access to it should be logged so that server owners know what's going on.
 2020-06-05 11:12:36 +01:00