You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
32-bit MinGW g++ doesn't call the correct overload for the new operator when a type is 16 bytes aligned. (jrouwe#1614)
It uses the non-aligned version, which on 32 bit platforms usually returns an 8 byte aligned block. We therefore default to 16 byte aligned allocations when the regular new operator is used.
See: godotengine/godot#105455 (comment)
Copy file name to clipboardExpand all lines: Docs/ReleaseNotes.md
+1Lines changed: 1 addition & 0 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -11,6 +11,7 @@ For breaking API changes see [this document](https://github.com/jrouwe/JoltPhysi
11
11
12
12
### Bug Fixes
13
13
14
+
* 32-bit MinGW g++ doesn't call the correct overload for the new operator when a type is 16 bytes aligned. This could cause unaligned read access violations.
14
15
* Fixed compiling in double precision and fixed issues with floating point contraction that caused unit test failures on LoongArch architecture.
15
16
* Added an epsilon to the `CastRay` / `CastShape` early out condition to avoid dividing by a very small number and overflowing to INF. This can cause a float overflow exception.
16
17
* Fixed Samples requiring Vulkan extension `VK_EXT_device_address_binding_report` without checking if it is available.
0 commit comments