guipublic
guipublic
If the array length is known at compile time, and is 0, then we can safely replace the `array[0]` by `{assert(false); 0 }` The assert being under the condition, the...
> Looks good but quite a few external repos are failing. May need to update them first or override CI. I updated the external repos in order to not cast...
> It doesn't look like we have any tests which enforce that this error is emitted. yes, I just added one.
Note that this is related to the question I asked in slack; we need first to clarify the expected behaviour that we want.
The behaviour we agreed on is the following: At the Noir user level: - The rhs of a bit shift will no longer be required to be a u8 and...
@jfecher I implemented the behaviour above, but the issue remains. Do we want left bit shit (for unsigned integers) to overflow when `lhs*2^rhs` overflows ? (it is currently truncated)
I added the documentation to hackmd: https://hackmd.io/LB6oYnyoTF-YHMKvgH0vHg?view
Should I add it to Noir Documentation?
This is resolved by #5221
I added a couple of unit tests