Search Issue Tracker
Fixed in 2020.1
[Android] Shadergraph blackboard variables are mixed when building to Android devices
Steps to reproduce:
1. Open the attached project "ShadergraphQuestTest.zip"
2. Build it on a mobile device
3. Notice the Sphere in the Scene becomes red or brown without a hole in the middle
Expected results: The Sphere is Rendered Green with a hole in the middle
Actual results: The Sphere is Rendered brown or red without a hole in the middle
Affects GLES2/3. Works OK on Vulkan.
Reproducible with: 2019.3.0f2, 2020.1.0a16
VLNQA00170, Google Pixel 2 (Pixel 2), Android 10, CPU: Snapdragon 835 MSM8998, GPU: Adreno (TM) 540
VLNQA00098, Lge LG G4 (LG-H815), Android 6.0, CPU: Snapdragon 808 MSM8992, GPU: Adreno (TM) 418
VLNQA00270, Samsung Galaxy S10e (SM-G970F), Android 9, CPU: Exynos 9 9820, GPU: Mali-G76
VLNQA00152, Huawei P9 (EVA-L19), Android 7.0, CPU: HiSilicon Kirin 955, GPU: Mali-T880
VLNQA00128, Samsung Galaxy Note8 (SM-N950F), Android 9, CPU: Exynos 9 Octa 8895, GPU: Mali-G71
Note: ShaderGraph brakes in 2019.2 and lower. Screenshots from StandAlone and from Android are in the description. The green SPhere is the StandAlone build and it is the expected results. IL2cpp makes the sphere be Rendered red more frequently while Mono makes brown more frequent color
All about bugs
View bugs we have successfully reproduced, and vote for the bugs you want to see fixed most urgently.
- [2D Path] Box selection for objects inheriting GenericScriptablePath is offset on the Y-axis
- [Linux] Unity licensing command on Linux doesn't complete when using -batchmode and -nographics
- SerializedReference GameObject property gets set to null after entering Play mode
- Profiler - RawFrameDataIterator ThreadID will always return 0 for profiler frame data loaded from .data files
- Crash on scripting_field_get_type when a Script has List<> of LayoutKind.Explicit struct with [FieldOffset(1)] ScriptableObject