[emacs-bidi] Re: bidi prototype in elisp

[Eli Zaretskii]

> From: Alex Schroeder <address@hidden>
> Date: Tue, 06 Nov 2001 02:10:38 +0100
> 
> Actually, the more I think about it, the better it would seem to try
> and implement something like that myself -- for example in elisp.  To
> write code which will use an iterator structure like the C code, and
> whic accepts a list of either characters or one of the symbols RLE,
> LRE, PDF, etc.

Given that the C implementation of the sequential reordering is
already done (that's the code I've been working on until now), what
would be the purpose of writing another implementation in Lisp?
Sounds like a waste of your time.

Besides, code in Lisp already exists that performs logical-to-visual
conversion: ask Ehud Karni (who reads this as we speak ;-) for his
hebeng package.  Its implementation is very different from what I
wrote, because it reorders text en bloc rather than sequentially, but
if all you want is a Lisp function which reorders bidi text, it's
more than enough.

> Then we could discuss it and see wether it's interface
> is similar to the C level, I could test it, I could write automated
> test cases, I could start thinking about tables, etc.

Let's pretend that this code is already written, and discuss these
issues now ;-).

If you want to work on expanding the test suite (I already have about
50 test cases), that's fine; but you could run them with the code I
wrote, instead of rewriting it in Lisp.  If you want to design the
necessary tables and other data structures required by bidi support,
that's an even better idea (and one of those I wanted to suggest ;-);
but why do you need a Lisp prototype for that?