feat: add more unit tests for float parsing

[louiscaron]

the float parser change had removed some negative tests.

reintroduced some negative tests, and some more formatting tests.

[DanielT]

After looking the tests you added, I'm not entirely sure your previous change is correct.

I played around with a floating point converter.
According to this, the hexadecimal value 0x11 is equal to 2.4e-44: all exponent bits are zero so the two bits in the mantissa only make a very small value. The equivalent of 11.0 in hexadecimal notation would be 0x41300000.

Of course it might be the case that the Autosar tool you mentioned in the previous pull request actually only represents the decimal part of a float value as hex, and is not using the in-memory format of an ieee754 number. In that case your change would be correct after all, but it might be good to add a comment explaining this.

[louiscaron]

I have added some comments in the code, i hope this is what you are expecting

[DanielT]

Yes, that neatly explains what is going on here