ModernFix

Changes since 3.5.2

• Use half fix for packet leak - embeddedt
• Synchronize LivingEntityRenderer layer list - embeddedt
• Fix classloading hack not working anymore - embeddedt
• Fix search tree provider crash - embeddedt
• Use less invasive mixin on Blocks.rebuildCache() - embeddedt
• Add JEI plugin to fabric.mod.json - embeddedt
• Move search tree replacement to common module - embeddedt
• Fix workaround message printing twice - embeddedt
• Move search tree replacement to common, support both modloaders - embeddedt
• Workaround for Forge issue #9505 - embeddedt
• Fix compile error - embeddedt
• Remove blockstate compression - embeddedt
• Fix performance issue when loading large NBT maps - embeddedt
• Add support for Fabric datagen from runClient - embeddedt
• Invalidate pack caches directly instead of iterating resource manager pack list - embeddedt
• Fix resource pack cache not invalidating on F3+T - embeddedt
• Update patch for 1.19 - embeddedt
• Add button to title screen for datagenning inside runClient - embeddedt
• Move iteration later in texture atlas patch - embeddedt
• Shorten structure hashes to fix issues on Windows - embeddedt
• Optimize server thread aliveness check - embeddedt
• More Fabric dynamic model loading improvements - embeddedt
• Fix texture reload failing on second try - embeddedt
• Add more paths to texture search - embeddedt
• Enable GitHub Gradle cache - embeddedt
• Fix Continuity emissives - embeddedt
• Make RenderType EqualsStrategy more predictable - embeddedt
• Disable faster texture loading when Stitch is present - embeddedt
• Adjust faster texture loading for FAPI compat - embeddedt
• Update incompatible mod disabling list - embeddedt
• Update patch for 1.19 - embeddedt
• Avoid running Fabric loading hook except for extra models - embeddedt
• Minor optimization when turning MRL into blockstate list - embeddedt
• Set the missing model early if needed - embeddedt
• Handle bakedTopLevelModels not being a dynamic provider yet - embeddedt
• Rewrite Fabric dynamic model loader to not load all models at startup - embeddedt