Eric McCarthy

@gluax My preference would be to keep things simple, and not allow vertical tabs, form feeds, or non-ASCII whitespace. I have no preference on horizontal tabs. Sometimes they sneak in...

In `tests/compiler/circuits/big_self_in_circuit_replacement.leo`, in the statement ``` y[Self {x: 0}.x] += 2; ``` After type inference the literal `0` is not getting a type. Here's what `type_inferenced_ast.json` looks like: ``` "expression":...

This also happens in the test ``` cargo run -p leo-test-framework --bin tgc -- -f compiler_array_complex_access ```

My preference would be to keep things simple, and not allow vertical tabs, form feeds, or non-ASCII whitespace. I have no preference on horizontal tabs. Sometimes they sneak in and...

@gluax Yes, if we can get the AST output from a call to `leo` is fine with me, but having to create a directory structure with the TOML file and...

Edited the title and added an **Expected Behavior** section.

Yes, it looks like the IR should be doing a `store`. The IR looks like this: ``` decl f0: store &v1, ((v0), (), (), ()) add &v4, v2, v3 retn...

Thanks, @gluax, you are right, the no-constraint issue does not seem to be related to the IR. However, there is another small difference with the `staging` branch that we should...

In a later conversation we decided there wasn't a problem with correctness of the Leo compiler. However, it is a problem for auditability. Constraints like this that are somehow implicit...

In master it was like this: ``` cat ./outputs/addfields.out [registers] r0: field = 3; ```