- Closes engines, pools and scheduler a previous server left open when the next server starts
- Flushes them at JVM exit once the stopping server thread has ended
- Corrects the scheduler shutdown comment
- Treats only loading saved biome information as an engine failure
- Leaves a container without loot and warns once when its saved biome cannot be recovered
- Keeps a failed chunk unetched and retries it without stopping the rest of the level's update pass
- Ends a level's pass only when its engine or mantle is closed
- Covers the deferred, permanent and isolated update paths through updateChunk and the server tick
- Waits for the destination chunk without a wall-clock deadline so a cold first chunk no longer fails the teleport
- Reports the failure cause, an unloaded dimension, an offline player or a refused teleport separately
- Adds the new messages to every locale
- Defers a chunk update pass while saved biome information is loading and retries it on a later pass
- Forgets only the updates already applied so a retry never refills a chest
- Propagates engine failures from loot application instead of marking the chunk done
- Stops downloads, protocol, pregen, tools and services at server stopping
- Keeps engines, mantle, generation history, the generation pool and burst pools open while the loader drains chunk generation
- Closes each engine when its level unloads and sweeps the rest at server stopped
- Logs a failed level unload instead of aborting the loader's level close loop
- GenerationClosedException covers closed history routers, stages, saved biome runtimes and admissions
- It also covers ownership stores, hydrology caches and engine runtimes
- GenerationFailures tells engine, lifecycle, executor and JVM failures from content failures
- Engine failures rethrow from object, cave object, procedural, floating, upper and decorator catch sites
- Biome layer, noise bound, rotation and mantle write catch sites rethrow them too
- Only content failures are skipped
- A source plan is never cached or persisted without an object a healthy engine places
- A transient engine failure no longer turns imported features off
- Matches the volmlib$ handler names of the client world-open and world-type mixins
- Keys expected mixins by their mixin config file and reports the target class name
- Verifies every expected handler is an injector declared by its mixin class
- Requires every registered volmlib mixin to be audited or deliberately unaudited
- The modded structure-state biome lookup resolves from the natural true biome stream, the same as the Bukkit ring search
- Stronghold ring placement no longer plans hydrology tiles around its 128 positions, so chunk structure starts stop waiting on river plans out to ~24k blocks
- Ring lookups are cached per runtime and cleared with the other biome caches
- The modded chunk generator hands VolmLib the bound loader for chunk-generation creature checks
- Each loader build tests its spawn position check: Forge and NeoForge fire MobSpawnEvent.PositionCheck and honor a denial, Fabric applies the vanilla rule and obstruction checks
- The forced datapack, datapack commands, and world check render biomes and dimension types with the 26.3 data fixer and pack format 121
- Modded tests use the 26.3 feature, placement, spawner, sign, and biome resolver APIs
- The water spawn test binds a blocks_motion fixture so solid waterlogged blocks still obstruct spawns
- Fabric, Forge, and NeoForge build against Minecraft 26.3 with Fabric Loader 0.19.5, Forge 26.3-66.0.8, and NeoForge 26.3.0.33-beta
- Fabric API modules follow fabric-api 0.161.0+26.3, ForgeGradle is 7.0.40, and ModDevGradle is 2.0.147
- Loader metadata requires Fabric Loader 0.19.5, Forge 66, NeoForge 26.3, and Minecraft 26.3
- The mod resource pack declares formats 97.1 through 121.0
- The v26_3_R1 headless sources no longer unpack a version overlay
- The README lists the 26.3 loader targets
- Modded platform gets registry access from the modded server factory
- Engine bootstrap and tile parity test get tile readers from the modded server factory
- Registry access tests build through the modded server factory
- Language asset test locates the directory through the English file
- Structure template test reads the mixin class bytes instead of loading it
- Water spawn test initializes block state caches before running
- Logging policy check runs after the native sources are unpacked
- Native class boundary check runs after classes compile
- Modded jars ship pack.mcmeta for client resources and mod data
- Modded jars bundle jlibdeflate and its native libraries
- Artifact check requires pack.mcmeta and the jlibdeflate entries
- Artifact check rejects client language resources without pack.mcmeta
- Test loads the pack metadata against the current Minecraft formats
- Modded platform hands the running server to registry access
- Tile reader is built from the running server
- Registry test covers startup, reload and shutdown of dynamic registries
- Every getChunk(..).use() site calls Mantle.useChunk, which cannot fail because a save closed the chunk in between
- Covered-chunk cleanup, custom block and static object layers take their chunk through useChunk
- Discovered flag writes go through Mantle.flag instead of an unpinned chunk
- Test mantles stub useChunk the way the real one behaves
- Index planned tiles by occupied chunk; empty chunks skip composition, tile residency and cave-cell lookups
- Plan the whole pregen area nearest first with several roots, past the tile budget when plans persist
- Restore persisted tiles without a planning root permit
- Never memoize answers the server thread gave before its hydrology plan landed
- Semantic journal and saved-biome appends write without forcing; a shared flusher forces touched files within 2s outside every store lock
- GenerationHistory.sync forces pending history at world save, engine shutdown, pregen completion and before activation promotion
- A failed deferred force poisons the store the same way an uncertain append did
- Cache the active epoch pack root instead of regex-validating it on every stage open
- Tests pin append-without-force, sync barriers, failure poisoning and the save, shutdown and pregen hooks
- Packages regrouped by domain: generation, world, pack, studio, structure, platform, command, localization, diagnostics
- Forked VolmLib utilities dropped for the shared library: noise, stream, interpolation, hunk, matter
- Engine-coupled stream decorators moved to generation.stream.GenerationStreams
- Hunk section splits now run on the Iris burst pool through HunkBurst
- Artifact entry lists, purity allowlists and shrink rules follow the new paths
Surface hydrology policy split, native terrain height cache, watched-edit
hotload cancellation, world-save requestSave with tracked ownership flush,
background-task drain that follows saves admitted mid-drain, peaceful spawn
guard, cached-completion terminal state, IO drains off the owner thread.
Studio pregen at scale: detached historical runtimes keep the configured
noise cache instead of the Studio-wide cache; retired runtimes are no longer
pinned by the volumetric fragment filter's per-thread scratch; pregen mantle
cleanup and native region flushes run on per-run threads instead of the
server thread and the shared Iris IO pool (the flush on the IO pool starved
tectonic plate loads and wedged every Paper worker in Mantle.getChunks);
background-task cancellation claims tasks atomically.
- Resolve IrisTerrainNormalActuator by keeping master's primitive riverWaterSurfaceStream lookup and primarySurfaceLayerOrNull accessors while retaining the terrain column, span, layer surface and ceiling locals that write air through terrain3D gaps
- Resolve HydrologyFootprintCompiler by keeping the MaterializedSurface planned-surface record with terrain ownership beside master's package-private SurfaceRasterIndex, which retains its ownsTerrain override
- Resolve HydrologyCaveContainmentPlanner by dropping the nested RecordingCaveVoxelView, SurfaceObservations and ViewObservations types that master promoted to their own files
- Port the surface-observation ownership check into CaveSurfaceObservations so cached cave validations replay both height and ownership observations through Observation.matches
- Resolve PackExportClosureTest by keeping master's retirement of the source-text packager guard while retaining the snippet export tests, and restore the Files and Path imports those tests use
- Resolve IrisBiomeLayerSlopeSourceTest by folding the lower-ledge slope palette case into master's lockLayers parameterization
- Recapture the studio schema digest for the terrain3D biome property, its IrisTerrain3D definition, the terrain-3d snippet enumeration and the reworded upperDimensionCarving description, verified by reconstructing the previous digest from the stripped schema
- Add an optional per-biome `terrain3D` profile with density and fissure noise styles, amplitude, horizontal and vertical feature scales, crack depth, width and scale, slope and fluid-clearance gates, and a seed offset, each range-checked before use
- Add a terrain runtime that samples the signed density and fissure fields on a four-block lattice around the biome's generated height and resolves each column into solid spans, so surfaces gain projecting rock, undercuts, ledges and narrow cracks that fade out on flat ground and near the fluid line
- Remove unsupported rock by flood-filling span components across neighbouring columns and dropping any fragment of 512 blocks or fewer, while larger connected masses are kept
- Report the shaped column top as the natural terrain height, keep the unshaped generator height available as a separate stream, and measure surface slope between neighbouring ledge floors instead of the flat heightmap
- Write air through the gaps in the normal terrain actuator, generate a fresh surface palette for every ledge floor, and paint ceiling layers beneath undercuts
- Carry each layer's own column through the dimension stack and upper dimension so stack top heights, solidity and surface lookups follow the spans, and each exposed face draws its palette from its own source height and slope
- Sample a layer's slope condition once per column at the surface being generated, and gate surface decorators on the slope of the ledge they sit on rather than the column heightmap
- Decorate every exposed ledge floor and the ceiling above it during the decorant pass
- Treat shaped openings as surface rather than cave when resolving biomes, skip them during cave carving, suppress cave zone markers inside them, and resolve their boundary biome from the surface biome stream
- Include shaped openings in mantle carve queries and carved columns, fall back to terrain solidity for surface-solid checks across stack, upper and normal terrain, and answer both queries from the island mask when placing objects on floating islands
- Let the hydrology planned surface report whether it owns a column, so the cave voxel view consults natural terrain solidity where hydrology does not own the terrain, and split replayed plan observations into height and ownership records
- Validate `terrain3D` on biomes and `snippet/terrain-3d` snippets as a blocking pack error, rejecting unknown fields, wrong types, out-of-range numbers, unknown noise styles, style nesting past 32 levels and snippet references that escape the pack
- Copy the whole snippet tree during modded studio pack export through one shared routine instead of only `snippet/style`, so terrain and nested style snippets survive export
- Add a terrain probe and Gradle task that generates real pack chunks, compares predicted column geometry against the actual blocks, requires retained overhang coverage for named biomes, reports detached-component topology, and writes cross-section images with a JSON summary
- Enumerate image-mapped regions and skip compat-excluded ones when collecting biomes for stacked and upper dimension terrain
- Move art.arcane.iris.testsupport into core/src/testFixtures and apply java-test-fixtures to core
- Consume core test fixtures from the bukkit plugin, nms, probe and the three loader adapters
- Delete the duplicate ProjectPaths copies in the bukkit-plugin and modded-common test trees
- Point ModdedMessageContributorTest at the shared ProjectPaths
- Guard CommandPackCompatTest, NativeTransitionColumnTest and StubPlatformStateTest with PlatformLeakGuard
- Release the probe runtime state after DimensionStackSuperflatProbeTest opens a workspace
- Keep core main classes ahead of the core jar on the test classpath so the bytecode gates still scan a directory
- Require the core source tree alongside the settings file when walking up for the repository root, so modded-common tests built inside a loader adapter no longer stop at that adapter's own settings file
- Add the ProjectPaths test helper that derives the module root from the test class code source and the repository root from the settings file above it
- Replace the working-directory-relative source and resource literals in thirty-five test classes with module and repository lookups
- Drop the bespoke source-root walk in PregenMantleCleanupContractTest
- Delete 22 test classes whose assertions only read production .java text and check call ordering or literal presence
- Remove 77 source-text assertion methods from 34 test classes that keep real behavioural coverage
- Replace the core process-stream source grep with an ASM gate over compiled core classes in CoreProcessStreamPurityGateTest
- Replace the packs-root source grep in PacksFolderContractTest with an ASM call-site gate over compiled core classes
- Replace the IrisData registered-root regex in PackTypeBukkitPurityGateTest with an ASM scan of the registerLoader class constants
- Trim six mixed tests down to their behavioural assertions and drop their source-text tails
- Drop the eight iris.*Source and three iris.moddedCommonSources test system properties now that nothing reads them
- Add a root moddedTest that runs adapters/modded-common through the Fabric build
- Make check depend on moddedTest, ordered ahead of the loader jars
- Gate the buildSrc jar on the buildSrc tests so they stop being skipped
- Drop an unreachable classpath guard from SchemaBuilderPlatformIsolationTest
- Replaced the nine static final PlatformBlockState fields in IrisDimensionStackActuator, IrisTerrainNormalActuator, IrisPerfectionModifier, IrisPostModifier, IrisBiomeLayerGenerator, BiomeBuffetGenerator, ChunkDataHunkHolder and EngineMantle with BoundBlockState holders that re-resolve when a different platform binds, costing one volatile read of the cached binding, one of the bound platform and a reference compare per read
- Turned the IrisPerfectionModifier ore-hiding table into a Map of BoundBlockState so the four base states follow the bound platform instead of the first one loaded
- Kept the last resolution in BoundBlockState while nothing is bound so in-flight generation on an unbinding server reads the same states it did before, matching what B.registries already does
- Added IrisPlatforms.getOrNull so the block-state holders read the binding once without an exception on a bare JVM
- Added PlatformStateRebindTest, which fails on the old code: it binds two mock platforms in turn and asserts ChunkDataHunkHolder and EngineMantle serve the second one, and rejects any static PlatformBlockState field left on the converted classes
- Indexed the .iob write palette by key so palette lookup is a hash probe rather than a scan of the palette list, cutting the 32767-state IrisObjectIoWriteLimitsTest from 4.9s to 0.7s while writing byte-identical files
- Built the palette once per write instead of once in validateWritable and again in the writer, and shared it between the plain and progress-reporting write paths
- Added IrisObjectIoPaletteTest, which pins the SHA-256 of a 30000-block, 3011-state object written by the previous implementation and asserts the palette keeps first-seen insertion order
- Added BukkitTestServer, a single mock Bukkit server installed once per JVM under a lock, stubbing logger, name, version, plugin manager and both createBlockData overloads, and failing loudly if the install does not take
- Moved IrisRuntimeSchedulerModeRoutingTest, WorldRuntimeControlServiceTimeLockTest, PreObjectMatterTest, CNGInjectorParityTest, BukkitSpiConformanceTest, MantleWriterPreObjectJournalTest, IrisFloatingChildBiomesCarvingResolutionTest and IrisDimensionCarvingResolverParityTest onto it, so whichever class runs first no longer decides whether createBlockData or getPluginManager is stubbed
- Extracted the KeyedBlockState test double out of IrisObjectIoWriteLimitsTest so the palette test can reuse it
- Replaced the IrisServices.clear() teardowns in nine test classes with a targeted remove of the PreservationRegistry each one registers, so a teardown no longer wipes a binding it did not install
- Registered a PreservationRegistry in PackDownloaderTest and StudioSVCWorldPackPublishTest, which reach IrisData.get and openRuntime and were silently getting loaderless IrisData whenever no other class had leaked one
