Should you be able to initialize a float with a char?

[Walter Bright]

On 5/19/2022 5:50 PM, kdevel wrote:
> Not "or" but "and". The first input contains UTF-8 of the (normalized) 
> codepoint, which is left unchanged by Walters lowercase function. The second 
> input contains UTF-8 of the same codepoint in canonically decomposed form (NFD).

Should stick with normalized forms for good reason. Having two different 
sequences supposedly compare equal is an abomination.

Though none of this supports the notion that arithmetic should not be done on 
chars. Heck, UTF-8 cannot be decoded without such arithmetic.