Fix weighted baked models not respecting children render types
Allow fluid container model to use base texture as particle
Fix inverted behavior in composite model building. Fixes#8871
* Add getter for the component names in an unbaked geometry
* Fix render type hint not being copied in BlockGeometryBakingContext
* Ensure BlockRenderDispatches's renderSingleBlock uses the correct buffer
* Make ShulkerBoxBlockEntity#canPlaceItemThroughFace delegate to Item#canFitInsideContainerItems.
* Switch to using Or and add comment.
* Switch Or to And.
* Fix door datagenerator
Fix datagenerator for door blocks. Successor to #8687, addresses comments made there about statement complexity.
* Fix extra space around parameter
Fix extra space before comma around a parameter.
* dimension specific tag fix
* remove forge:is_beach cause vanilla has it already
* remove forge tags for new 1.19 vanilla tags (savanna, beach, overworld, end)
Co-authored-by: Flemmli97 <Flemmli97@users.noreply.github.com>
* convert GLM serializer class to codec
* cleanup
* GLM list needs to be sorted
* datagen
* simplify serialization
* fix test mods (oops)
* properly use suppliers for codec as they are registry obj
* Add an easy way to get the NamedGuiOverlay from a vanilla overlay
* Fix static member ordering crash in UnitTextureAtlasSprite
* Allow boss bar rendering to be cancelled
* Make fluid container datagen use the new name
* Add a classpath transformer discoverer to load custom transformation services from the classpath
* Update ClasspathTransformerDiscoverer to 1.18
* Update license year
* Update license header
* Fix the other license headers
* Update ClasspathTransformerDiscoverer to 1.19
* Handle modded packets on the network thread
- On the server we simply need to remove the call to
ensureRunningOnSameThread.
- On the client side, we now handle the packet at the very start of the
call. We make sure we're running from a network thread to prevent
calling the handling code twice.
While this does mean we no longer call .release(), in practice this
doesn't cause any leaks as ClientboundCustomPayloadPacket releases
for us.
* Clarify behaviour a little in the documentation
* Javadoc formatting
* Add a helper method for handling packets on the main thread
Also rename the network thread one. Should make it clearer the expected
behaviour of the two, and make it clearer there's a potentially breaking
change.
* Add back consumer() methods
Also document EventNetworkChannel, to clarify the thread behaviour
there.
* Add since = "1.19" to deprecated annotations
* Make the resource lookups cached.
* Include configurability and handle patch cleanup.
* Document and comment the cache manager.
* Make thread selection configurable.
* Implement a configurable loading mechanic that falls back to default behaviour when the config is not bound yet.
* Use boolean supplier and fix wildcard import.
* Clean up the VPR since this is more elegant.
* Clean up the VPR since this is more elegant.
* Address review comments.
* Address more review comments.
* Fix formatting on `getSource`
* Address comments by ichtt
* Adapt to pups requests.
* Stupid idea.
* Attempt this again with a copy on write list.
* Fix a concurrency and loading issue.
* Fix#8813
Checks if the paths are valid resource paths.
* Move the new methods on vanilla Patch.
LexManos