gingerBill
gingerBill
Well that's something I had never considered before, and is kind of crazy.
Can you try on the latest commit now to see if it is fixed?
I am struggling to replicate this.
This is two separate bugs. The first is a semantic checking bug. The second appears to be a code generation bug.
This isn't a bug per se, but rather just a possible stack overflow. So it's a question how would you warn about such a thing in the first place?
The current ABI for arm32 is mostly a placeholder, and will need correcting for aggregates. The problem is reverse engineering how LLVM does things more than anything. https://github.com/odin-lang/Odin/blob/master/src/llvm_abi.cpp#L1392-L1472
If you come from C, all parameters do this copy implicitly. ```c void foo(int x) { x = 123; // I can change it } ``` `x` in this case...
This isn't a bug and kind of the point of them. The original design rationale was for something like `os.args`. If you don't use `args`, it never got initialized, thus...
So to explain this "bug", this isn't one and it's actually extremely well defined behaviour to be this on purpose but of course this is subject to change. I wanted...
The problem is fundamentally to do with `when`.