David Wong
David Wong
This was an attempt to cleanly separate the circuit compilation with the typechecking pass, even for builtins. As you can see the typecheck for function calls happen here: https://github.com/zksecurity/noname/blob/main/src/type_checker/checker.rs#L817 and...
oh I see :)) time for a non-beta release :D ?
I exported that function: ```rust pub fn init_rust_panic_hook() { panic::set_hook(Box::new(|_| { let bt = Backtrace::new(); println!("Rust custom panic hook: {:?}", bt); })); } ``` and I'm initializing it in Ocaml:...
removing the use of `Backtrace` and simply printing the argument to the closure actually improved things. my theory is that the panic raised by OCaml is also caught by my...
adding a backtrace to an error type (with thiserror) also creates issues. I've posted on ocamlforum as well: https://discuss.ocaml.org/t/catching-panics-in-rust/11730
BTW, to contrast here how Rust display errors, I find them quite clear due to the spacing, arrows, underlines, the colors, and the hints on how to act on the...
maybe this issue should be about converting the current `examples/const.no`: ```rust const player_one = 1; const player_two = 2; fn main(pub player: Field) -> Field { assert_eq(player_one, player); let next_player...
^ still remains to add it to the prove/verify CLI
^ yeah not a good idea IMO. I think having two functions could help: ```rust fn init_analyze(nast: NAST, is_lib: bool) -> Result; fn analyze(self: TAST, nast: NAST, is_lib: bool) ->...
@katat we can close this no?