* fix: adjust linked filter to be left-right and do not link to new forked instances
- Assume that child filters know about their own forks
- fixes#1874
* Remove appliesChunk|Layer for now
* feat: improve fawe limits (#2773)
- add FaweLimit implementations for increasing concurrency levels
- allow FaweLimit to perform processing (and forcefully disable as required) to capture [tile] entities
- Use `BlockVector3#set(Extent orDefault)` where appropriate to reduce block checks
- fixes#2679
- fixes#1874
(cherry picked from commit 6052fc3128)
* fix: actually apply a filter if applied from a full region chunk section
- cannot remember why I made this change in the first place in #2773 but this fixes edits having "empty middles"
- doesn't seem to have broken anything in testing
* Reduce limit to always atomic
- reset "origin"/relative X/Z when larger than short min/max value so we do not write incorrect positions
- history size can be larger than int max value
- fixes#2583
* Update to Sponge Schematic 3
Includes a major refactoring of how schematics are read.
(cherry picked from commit bd475b1d4acbcf2a95e5a8f3aee50d2fb2100ae8)
* Licenses lol
(cherry picked from commit a5ce8a47657aa987da8ca625cd658856d2eb3477)
* Fix imports
(cherry picked from commit e1892b7bd4ff0ca4592f8cb4e1b2d9363c4cd6ff)
* Update for final changes
(cherry picked from commit 2f6b50a4276b33b615d9dbc52e73e958308735f9)
* chore: ensure flushed clipboard in spongev2 writer
* feat: initial work on FastSchematicWriterV2
* fix: only write into palette once, write into data as varint
* chore: more work on FastSchematicWriterV3
* fix: make FastSchematicWriterV3 work
* fix/chore: write pos as doubles
* chore: start on reader (class exists at least)
* chore: replace while loop with simple if as char can be max 2 bytes
* chore/feat: more work on the fast v3 reader
* fix: offset must be inverted for origin
* chore: use the actual FileChannel for mark / reset (if present)
* chore: add null check again
* chore: buffer streams in isFormat check
* chore/feat: read schematic fully
* chore: don't hold a lazyreference (seems harder to gc with already computed value?)
* chore: remove debugs
* chore: optimize FastSchematicReaderV3
* chore: remove logger warn for now
* chore: constant not required anymore
* chore/feat: support non-file based inputstreams (using in memory LZ4 cache)
* chore: don't wrap streams unnecessary
* chore: cleanup
* chore: since comment for reader + writer
* chore: FAST_V3 not for testing anymore
* chore: update schematic and clipboard logic for linbus changes
* chore: undo format check on load
* fix: remove usages of old nbt types
* fix: use LinBus in FaweDelegateSchematicHandler
* fix: use ReaderUtil again
* chore: update supported schematic types for Arkitektonika
* chore: check for magic bytes in schematic (not tested yet)
* revert: magic bytes check
* fix: fix paletteAlreadyInitialized + biome placement on linear clipboards
* Update worldedit-core/src/main/java/com/fastasyncworldedit/core/extent/clipboard/io/FastSchematicWriterV3.java
---------
Co-authored-by: Octavia Togami <octavia.togami@gmail.com>
Co-authored-by: Hannes Greule <SirYwell@users.noreply.github.com>
* Add a BiomeCategories API (#2338)
* Add a BiomeCategories API
* licenses
* Use a supplier to retain the lazy-load & dynamicness of the existing system, but still retaining the inversion of control that this PR was intended to provide
* Minor fawe adapter cleanup
* Actually add the new files?
* Fixes
---------
Co-authored-by: Maddy Miller <mnmiller1@me.com>
- It's theoretically possible for the section FULL to return a null layer due to race condition with a trim operation
- Locally cache result and if null, recover
- I just had the error from #1592 again
- This seems to have stopped the error, but adding logging did not log, so possibly some bigger bytecode changes?
- Oh well
