It's just an escape sequence - like prefixing a character with backslash when inside a string constant. Does that make it non-conformant? Is the common practice of replacing quote with backslash-quote in string constants non-conformant, because backslash isn't a PUA character? If so, I stand suitably chastised, but if not, there is nothing supernatural about using magic-numbers, signature-bytes, etc., to identify an escape sequence. I simply chose a string longer than "\" (and therefore less likely to occur by accident).

But in hindsight I can see some logic in what you say. In order to be acceptable to non escape-aware processes, any escape-identifier-sequence should at the very least not be defective (e.g. never a control character followed by a combining character), and I confess I didn't check that.

-----Original Message-----
From: Doug Ewell [mailto:[EMAIL PROTECTED]
Sent: 15 December 2004 16:28
To: Unicode Mailing List
Cc: Arcane Jill
Subject: Re: Roundtripping Solved

Of course, Jill's scheme uses non-private-use Unicode scalar values to
achieve what is essentially a private-use function, so this is still
non-conformant. (A similar scheme that only used code points from the
Plane 0, Plane 15, and Plane 16 PUAs would be fine.) But I gather that
Lars isn't too worried about being non-conformant, or we wouldn't be
having this thread.

-Doug Ewell
Fullerton, California
http://users.adelphia.net/~dewell/





Reply via email to