AW: with_string_0 deficiency

[Hoehle, Joerg-C. <Joe...@t-...>]

Hi,

Sam wrote:
>I don't see how encoding_min/max_bytes can help with ENCODING-ZEROES,
>but we can certainly call an equivalent of ENCODING-ZEROES in
>MAKE-ENCODING and set encoding_min/max_bytes there.

Encoding-min tells us the length of the smallest conversion. If 0-terminatednes makes sense everywhere, one can expect that to be the number of zero bytes.
At least, I don'f find this assumption worse than the one currently built into the encoding-zeroes function. It's a different assumption than expecting a NUL character being present in every encoding.

I'm not sure changing make-encoding is enough. What about non-internal encodings?

IIRC, 0-termination does not make sense with UTF-7 (neither would it with Base64 encoding), but who cares?

>UTF-8 is, I think, much more valuable.
UTF-8 is not interesting because it needs a single zero termination byte. What's the problem?

Regards,
	Jorg