Re: [emacs-bidi] abbrev

[Eli Zaretskii]

On Wed, 17 Oct 2001, Uwe Brauer wrote:

> This may be a little of topic, but since we discuss the development of
> the new BIDI implementation. In the old emacs-19.34 y emcas-20.4.9
> Hebrew BIDI implementation abbrev was broken severely.
> 
> Eli did you test already abbrev for the new implementation?

You give me too much credit ;-)

As a matter of fact, what I have right now is an implementation of UTR#9 
algorithm which should be easy to integrate into Emacs.  I'm developing 
it using a stand-alone test program which calls the main entry point into 
the algorithm implementation, and then arranges the characters for 
display on a dumb tty.  This phase is almost over: the code works, is 
well tested, and implements almost every aspect of UTR#9.

The next phase is to plug this code into the Emacs display engine and 
get to the point where I can display a bidirectional text correctly.  In 
addition to integration between what I have now and the Emacs code, I 
will have to write additional code for related features, such as cursor 
motion, and quite a few user-level features to support the bidi mode.

Only after this is finished, will I come to other Emacs features which 
are affected (but not directly related) to bidirectional editing: search, 
indentation, auto-filling, etc.  Abbrevs will interest me only in this 
phase.

So I'm still very far from considering abbrevs, let alone writing some 
code or testing it.

However, as long as we are talking, perhaps you could describe what was 
broken in the old implementation wrt abbrevs, so that in due time we 
could avoid repeating the old mistakes.