* wip backport of custom uniforms
* Functioning custom uniforms maybe?
* Workon separating dynamic/non-dynamic uniforms and making variables available to custom uniforms (currently the JVM dies horribly)
* wip
* It no longer crashes horrifically
* star imports
* Initial MemoryStack backport
* Remove stubs, for now
Will re-add as needed
* Implement mallocInt
* Replace some buffer allocation
* Add mallocFloat
* Replace more uses of direct buffer allocation
* Reorder methods
Fixes a clinit crash and makes it easier to compare to upstream
* Remove another direct buffer creation
* Store graphics vendor
* AMD driver workarounds
* Test TRANSFORM_BIT
* Test GL_TEXTURE_BIT
* Fix handling of bound textures on glDeleteTextures
* Switch up the LOGGER in RenderSystem so we don't trigger a Forge Configuration load; AMD driver default matrix mode workaround
* Switch BatchingFontRenderer to use the gl methods via GLSM to better respect cache bypassing
Fix a few instances of GLSM not respecting the bypass
* Start with "runningSplash" true always to disable the cache until loaded into the world
---------
Co-authored-by: Caedis <[email protected]>
* More Caching
* Pull TextureTracker & TextureInfoCache out of Iris and into GLSM
* Cache more TextureInfo state
* GLSM: glTexParameter, glGetTexParameter(i|f), glGetTexLevelParameteri
* Move IrisRenderSystem -> GLSM.RenderSystem
* Ensure DSA works properly with GLSM
* Further updates to DSA for texParameter/getTexParameter
* A few more cached fields for TextureInfo
* Properly handle deleteTextures
* Ensure anyone crazy enough to use an Int/FloatBuffer for single value params get redirected
* Add optional runtime disabling of GLSM
* Switch checkGLError from an inject to a redirect, hopefully fewer allocations
* Fewer allocations for clearBufferState
* Microoptimization for (setup|clear)BufferState