mul_assign not implemented

[DavePearce]

In the full traces being checked using the -N switch, there is a panic arising:

panicked at src/column.rs:132:53:

This points to this piece of code:

pub(crate) fn mul_assign(&mut self, other: &Value) {
        match (self, other) {
            (Value::BigInt(ref mut i1), Value::BigInt(ref i2)) => *i1 *= i2,
            (Value::BigInt(x), Value::Native(y)) => todo!("{} *= {}", x, y), // <---- HERE
            (Value::BigInt(_), Value::ExoNative(_)) => todo!(),
            (Value::Native(_), Value::BigInt(_)) => todo!(),
            (Value::Native(ref mut f1), Value::Native(ref f2)) => f1.mul_assign(f2),
            (Value::Native(_), Value::ExoNative(_)) => todo!(),
            (Value::ExoNative(_), Value::BigInt(_)) => todo!(),
            (Value::ExoNative(_), Value::Native(_)) => todo!(),
            (Value::ExoNative(_), Value::ExoNative(_)) => todo!(),
        }
    }

Relevant part of the backtrace:

  3: corset::column::Value::mul_assign
             at ./src/column.rs:134:53
   4: corset::check::check_lookup::pseudo_rlc
             at ./src/check.rs:344:13
   5: corset::check::check_lookup
             at ./src/check.rs:389:37
   6: corset::check::check::{{closure}}
             at ./src/check.rs:531:41

So, that looks like a Value::BigInt is creeping in somehow without being converted to a Value::Native.

[DavePearce]

There is some interplay with expansion here. For example, running with -N / -eN / -eeN gives the above whilst -eeeN / -eeeeN does not.