Re: [emacs-bidi] mirroring of glyphs

[Alex Schroeder]

I've found a bug in my implementation of rule W4 which I have fixed.
I've added all PGBA test cases to bidi.el.

The only test case left unresolved is the following:


(bidi-visual-to-logical "<123H/>shalom<123H>" 'R2L)

 => "<Hshalom<123</H123>"

instead of

"<H123>shalom</H123>"


The logical-to-visual test works fine.

(bidi-logical-to-visual "<H123>shalom</H123>")

  => "<123H/>shalom<123H>"


I tried playing around with it: I tried reversing the input string and
reversing the result in order to mirror any asymmetries in the bidi
algorithm.  Other tests started to fail, then.  I tried leaving out
rule W7, but again, other tests started to fail.

I think this is the point where I must use Ehud's idea.  :)

Alex.
-- 
http://www.emacswiki.org/