Re: [emacs-bidi] bidi categories

[Eli Zaretskii]

> From: Alex Schroeder <address@hidden>
> Date: Fri, 09 Nov 2001 19:28:14 +0100
> 
> Here's what I have at the moment.  As I said in another mail, I also
> have the Unicode classification, but I don't know yet how to use it --
> given a Unicode codepoint (correct term?), how do I get the
> corresponding Mule character in the unicode charsets

Here's how:

  (decode-char 'ucs uchar)

where UCHAR is the Unicode codepoint.  Easy, eh?

> -- or better yet,
> how do I get all characters from all the other charsets matching it?

For this, you will need tables, there's no single method.

> I looked at UAX#9 again, and while it describes the process of
> transforming logical to visual order, it a) looks horribly complicated
> and b) not reversible.  I guess that's old stuff for those who are
> familiar with it...  :) Since the reverse operation is not defined
> explicitly (did I miss it?), a certain amount of guess-work will be
> needed.

I think that if you implement only the In clauses of UAX#9 (which
resolve the implicit levels), you will get what you want, since
visual-order text doesn't have explicit marks.  It seems to me that
the implicit reordering is reversible.