make cuchar alias uint8 instead of char

[shirleyquirk]

split from #17728

[timotheecour]

is this a good change?
before PR:

nim --eval:'echo (cuchar.low.ord, cuchar.high.ord, cuchar.low, cuchar.high)'
(0, 255, '\x00', '\xFF')
=> seems fine

after PR:
nim --lib:lib --eval:'echo (cuchar.low.ord, cuchar.high.ord, cuchar.low, cuchar.high)'
(0, 255, 0, 255)
=> seems less good

at least this needs more rationale

[ghost]

@timotheecour why do you think that it's "less good"? unsigned char shouldn't be used to represent text characters, it's for unsigned 8 bits integers

[ghost]

Also, it's more consistent with what we have already - cschar (C's signed char) is int8 in system - https://nim-lang.org/docs/system.html#cschar, just how it should be.