@xoofx True, generic should be fine.
I remember my blocker with .NET was not pure covariance/contravariance in methods (always possible to generate a thunk/trampoline method to transform ptr) but array/data covariance (i.e. ITest[] to object[]).
I wish they didn't copy Java for that ;)
@xoofx Tried fat pointers in another project but hit 2 potential issues:
1. cast & co/contravariance (unless all pointers are fat)
2. if half-copied with non-cooperative GC it could cause issues (some CPU have atomic instr)
Probably OK in #starklang though since you have full control!
@xoofx Good luck!
This part of the engine is definitely worthy of a story/article, it was really interesting experimentation and research.
Let me know if you need any help/review!
I will be at the GDC from Tuesday to Thursday.
If you want to meet me to discuss #Xenko, don't hesitate to reach out!
No booth, but I will be strolling the exhibition floor with my #Xenko t-shirt!
Thanks @TRBPlay for graciously sponsoring flight tickets, GDC pass and t-shirts!
I have joined the race for the .NET Foundation Board!
https://t.co/JKSfiM9jKV
I believe #gamedev field and high-performance .NET should be well represented within the foundation.
Quick background: I am lead developer of #Xenko opensource .NET Game Engine
The wait is over...
Xenko 3.0 is now out, released under the permissive MIT license (incl. editor) and now turned into a community project!
More details at https://t.co/L6Dnj3jhj2
Any help welcome!
#xenko#gamedev#indiedev#FOSS#opensource#dotnet