Error with the simulation

[OscarLopezCentenera]

I've been encountering an issue related to simulation value ranges. I'm sharing a link to a Rust HDL project that implements an 8-bit signed adder (Signed<8>), whose output is a 9-bit signed value (Signed<9>).

When running the simulation in Rust, I populate the test_cases variable with input values ranging from -127 to 127. This works correctly. However, if I use the full expected range for a Signed<8> — from -128 to 127 — the simulation fails specifically when testing the -128 case, resulting in an "overflowed" error.

This is not the first time I have experienced this issue, and it seems consistent when dealing with the value -128. If I adjust the input range to start at -127, which appears to be the default behavior, everything works correctly.

I would like to know whether this is due to a limitation of the tool — meaning that it does not support handling -128 correctly — or if it is simply an unimplemented feature in the current version of the simulation framework.

https://drive.google.com/drive/folders/1rBo9JDkt8n50aqYVuEfK90CKt5o6wpWF?usp=sharing

The code

let mut sim = Simulation::<Sumador>::new();
sim.add_testbench(move |mut ep| {
   
    let mut x = ep.init()?;

   //here, change (-127..128) to (-128..128) ant its fails
    let test_cases: Vec<(i32, i32)> = (-127..128).flat_map(|a| (-127..128).map(move |b| ((a,b)))).collect();

    for &(a, b) in test_cases.iter() {
        x.a.next = Signed::<8>::from(a as i64);
        x.b.next = Signed::<8>::from(b as i64);
        let x_clone = x.clone();
        let x = ep.wait(1, x_clone)?;
        //println!("a= {}, b= {}, sum= {}", x.a.val().bigint(), x.b.val().bigint(), x.sum.val().bigint());
        let expected_sum = a + b;
        sim_assert_eq!(ep, x.sum.val(), Signed::<9>::from(expected_sum as i64), x);
    }
    ep.done(x)?;
    Ok(())
});