actual contextual rule for ZWJ/ZWNJ, was RE: document on idn cctld names

[Michel Suignard]

> From Erik:
> I was actually asking from the point of view of IDNA200X itself,
> i.e. not just ccTLDs (though I do appreciate your work).
>
> Are we likely to introduce an actual contextual rule for ZWJ/ZWNJ?
> If so, which scripts would be included?

Erik
According to latest Patrik draft, these are CONTEXTJ characters. I don't think the document contain the rules for these characters (class H). I am a bit behind in digesting the latest specs so I may have missed details either there or in the protocol document.

Another place to look at is http://www.unicode.org/reports/tr31/tr31-8.html#Layout_and_Format_Control_Characters
(latest proposed update to UAX#31)

In that clause there is no script restriction, but there is however a single script restriction for all regex expressed in that clause. IDNA200x may however bring additional restrictions on script values. Not sure it is a good idea though, because new scripts needing that contextual rule may be encoded and you want as less dependencies on Unicode version as possible in an IDN update.

Michel