Re: [emacs-bidi] mirroring of glyphs

[Eli Zaretskii]

> From: Alex Schroeder <address@hidden>
> Date: Tue, 20 Nov 2001 14:54:02 +0100
> 
> "Eli Zaretskii" <address@hidden> writes:
> 
> > Hmm.. there's a basic asymmetry in UAX#9, in that many rules say to
> > search backwards for a character of some type.  I wonder what will
> > happen if in the visual-to-logical conversion, you will instead
> > search forward in those case?  Can you try that?
> 
> Doesn't that happen automatically when you reverse the string?

No, reversing the string only undoes what the simulated rendering
does: it draws the glyphs starting at the right display margin
instead of the left.

> There are other asymmetries I have tried: I modified W7 such that EN
> would transform to R if following an R (thus, in the non-reversed
> string, if EN precedes an R).  But that didn't help either.

Hm.. it seems that the simple rule I suggested is indeed incorrect.
SOmething more complex is needed: we need only to look in the other
direction when the character belongs to a substring that was reversed
odd number of times.