Missing test coverage for encode_json for values smaller than 0x1f

### What is the problem this feature will solve?

increases test coverage

### What is the feature you are proposing to solve the problem?

I was investigating ada. I saw that you do `2 * c` and was wondering if a bitshift makes a difference `c << 1` but then I realized that if replace with just returning '0', then `ctest --output-on-failure --test-dir build` will pass all the tests. So there is missing test coverage.

I guess that the compiler will create the same bytecode for 2 * c and c << 1. But it would make sense to have test coverage.

### What alternatives have you considered?

_No response_