Search Issue Tracker
Fixed in 2019.3
Fixed in 2018.4, 2019.2
Inconsistent results when casting a negative float to uint between ARM and x86
Steps to reproduce:
1. Open and run QA-Supplied project.
2. Notice the 3rd field from the top is set as uint maximum (4294967295).
3. Build and run.
4. Notice the same field.
Expected: 3rd field from the top is set as uint maximum.
Actual: 3rd field from the top is set as 0.
Reproduced in: 2017.4.30f1, 2018.4.5f1, 2019.1.13f1, 2019.2.0f1, 2019.3.0a11
Note: not reproducible on Windows.
Reproducible on Mono, IL2CPP.
Reproducible on iPhone 8 iOS 11.0 2019.3.0a11
VLNQA00031, Lge Nexus 5 (Nexus 5), Android 6.0.1, CPU: Snapdragon 800 MSM8974, GPU: Adreno (TM) 330
VLNQA00002, Samsung Galaxy S5 Neo (SM-G903F), Android 6.0.1, CPU: Exynos 7 Octa 7580, GPU: Mali-T720
VLNQA00146, Htc 10 (HTC 10), Android 7.0, CPU: Snapdragon 820 MSM8996, GPU: Adreno (TM) 530
VLNQA00096, Samsung Galaxy S8 (SM-G950F), Android 8.0.0, CPU: Exynos 9 Octa 8895, GPU: Mali-G71
VLNQA00268, Samsung Galaxy S10+ (SM-G975F), Android 9, CPU: Exynos 9 9820, GPU: Mali-G76
All about bugs
View bugs we have successfully reproduced, and vote for the bugs you want to see fixed most urgently.
- [iOS] No fallback for Light Probe Proxy Volume for devices which do not support 32bit floating point textures
- Layout is not valid unless the GameView is resized
- [iOS] Directional lightmap compression causes lightmap artifacts
- NullReferenceExceptions in PreprocessReflectionProbeData
- [2d][Sprite Import] Specific Sprite seems to be cutoff when added into the scene