Re: [Yaml-core] Byte order marks and Unicode

[Jason D. <ja...@in...>]

Clark C. Evans said:
> The characters mentioned in the YAML specification are Unicode
> code points, and _not_ octets.  I suppose in the textual description
> we could list the specific octets the BOM can have.

Ah. I should have known that!

> I think you are interpreting the byte order mark as a octet.  We
> can add the specific octets in the paragraph above, if you think
> this will make it clear.

Yes, I was thinking that. I think it would definitely help implementors if
you were explicit about the specific forms of the BOMs that could be found
(perhaps in an appendix?).

> Yes.  YAML 1.0 may _only_ be encoded in UTF-16 or UTF-8.  If someone
> wishes to use the YAML format in another encoding, this is fine, but
> the resulting document is not officially "YAML".  This is a "debated
> to death" deliberate decision.  ;)

I certainly don't want to debate that. But putting the word _only_ in the
spec might clear up any question other readers might have.

Although with the proper BOM, UTF-32 seem just as easy to allow as UTF-16
and UTF-8. All other encodings should definitely be forbidden.

-- Jason