>>>> "Aidan" == Aidan Kehoe
<kehoea(a)parhasard.net> writes:
Aidan> Ar an seachtú lá de mí na Nollaig, scríobh Uwe Brauer:
>
> So wait, there are two questions:
>
> - is there quail in xemacs 21.4.17 (debian) if so I can't
> find them. If there are no quail pkg anymore which pkg is
> responsible for the input?
Aidan> Quail is part of the leim XEmacs package.
Ah, ok, I found it, grr in Debian there are only the elc files I have
to search for the el files. Strange strange I thought I had seen once
such a quail Hebrew pkg, my memory.
> - how difficult would it be to write a quail (or else) pkg
> for Hebrew input. One could not just take the one of GNU
> Emacs?
Aidan> One could probably do so. I loaded it in 21.5 out of
Aidan> curiosity while writing an answer to your last mail, and it
Aidan> seems our ISO 2022 Hebrew character set identifier doesn’t
Aidan> match theirs, so you would need to load it in GNU Emacs,
Aidan> change the coding cookie and save it as iso-8859-8 instead
Aidan> of iso-2022-7.
What do you mean by changing the coding cookies?
Aidan> (If and when you have experience with GNU Emacs and their
Aidan> BiDi support, I’d love if you’d post a summary of the
Aidan> situation here. I’d like to know what’s happening there.)
Ha, do you know about nntp+news.gmane.org:gmane.emacs.bidi?
The situation is roughly like this.
As early as 19.34 the (then not merged) mule emacs had Hebrew Bidi
support, (and visual cursor movement which I loved). However searching
did not work, since there where problems with the minibuffer.
Then there started a discussion of how to implement that BIDI alg for
emacs 20. (I think 5 years ago there was a conference in Japan about
that I tried (successfully) to encourage Stephen Turnbull to
participate), however sadly enough, the xemacs team was left out in
the following discussions about the implementation.
Unfortunately GNU emacs had then also a chance in its Display engine
which made most of the old 19.34 code obsolete. It is also sad to
note, that not one single Hebrew hacker participated, besides Eli
Z. (and Ehud Karni which has written some lisp pkg for visual
hebrew). So finally the algorithm which was agreed upon was implemented
by Eli. It was then decided to leave it in non-ready 21.3
http://www.m17n.org/emacs-bidi/
And not to include in into the current development.
So in short there is some code which does things better than 19.34,
searching for example, even supports nikkud, but still lacks visual
cursor movement and worst of all is not included in any kind of
development.
Hope this helps
Uwe