Question on Mule and Makor2, font display vs char input?

Question on Mule and Makor2, font display vs char input?

[Anna Petrov Ronell]

Hi,

Is  there an  easy  way  to define/redefine  how  Mule operates.   For
example, can a custom Mule Language environment be easily defined?

There is a LaTeX font environment called makor2 which allows a user to
create output  using Hebrew  characters.  These output  characters are
often defined by multi-byte combinations.   So it would be nice to have
an Emacs  mode which allowed a user  at the keyboard to  type a single
input  key and have  that keystroke  display the  proper corresponding
Hebrew  character  on the  screen  while  placing  the proper  set  of
matching  characters into  the  file  for LaTeX  Makor2.   There is  a
standard keyboard setup use in Israel  which can be followed and a set
of Mule Hebrew screen fonts seem to be readily available.

Is there an easy way to  use the existing Mule Hebrew character set as
a starting point to create a new Emacs mode definition which will just
output the needed Makor2 multi-byte character set?

Thanks

Question on Mule and Makor2, font display vs char input?

[Anna Petrov Ronell]

Hi,

Is  there an  easy  way  to define/redefine  how  Mule operates.   For
example, can a custom Mule Language environment be easily defined?

There is a LaTeX font environment called Makor2 which allows a user to
create output  using Hebrew  characters.  These output  characters are
often defined by multi-byte combinations.   So it would be nice to have
an Emacs  mode which allowed a user  at the keyboard to  type a single
input  key and have  that keystroke  display the  proper corresponding
Hebrew  character  on the  screen  while  placing  the proper  set  of
matching  characters into  the  file  for LaTeX  Makor2.   There is  a
standard keyboard setup use in Israel  which can be followed and a set
of Mule Hebrew screen fonts seem to be readily available.

Is there an easy way to  use the existing Mule Hebrew character set as
a starting  point to create a  new custom Emacs  mode definition which
will just output the needed Makor2 multi-byte character set?

Thanks

Question on Mule and Makor2, font display vs char input?

[Anna Petrov Ronell]

Hi,

Is  there an  easy  way  to define/redefine  how  Mule operates.   For
example, can a custom Mule Language environment be easily defined?

There is a LaTeX font environment called makor2 which allows a user to
create output  using Hebrew  characters.  These output  characters are
often defined by multi-byte combinations.   So it would be nice to have
an Emacs  mode which allowed a user  at the keyboard to  type a single
input  key and have  that keystroke  display the  proper corresponding
Hebrew  character  on the  screen  while  placing  the proper  set  of
matching  characters into  the  file  for LaTeX  Makor2.   There is  a
standard keyboard setup use in Israel  which can be followed and a set
of Mule Hebrew screen fonts seem to be readily available.

Is there an easy way to  use the existing Mule Hebrew character set as
a starting point to create a new Emacs mode definition which will just
output the needed Makor2 multi-byte character set?

Thanks

Re: Question on Mule and Makor2, font display vs char input?

[Joe Corneli]

   Hi,

   Is  there an  easy  way  to define/redefine  how  Mule operates.   For
   example, can a custom Mule Language environment be easily defined?

   There is a LaTeX font environment called makor2 which allows a user to
   create output  using Hebrew  characters.  These output  characters are
   often defined by multi-byte combinations.   So it would be nice to have
   an Emacs  mode which allowed a user  at the keyboard to  type a single
   input  key and have  that keystroke  display the  proper corresponding
   Hebrew  character  on the  screen  while  placing  the proper  set  of
   matching  characters into  the  file  for LaTeX  Makor2.   There is  a
   standard keyboard setup use in Israel  which can be followed and a set
   of Mule Hebrew screen fonts seem to be readily available.

   Is there an easy way to  use the existing Mule Hebrew character set as
   a starting point to create a new Emacs mode definition which will just
   output the needed Makor2 multi-byte character set?

   Thanks



I've got a very similar question, but for math characters as
opposed to hebrew characters.

Any advice on how to make a string already entered into the buffer
render according to the current input method?  For example, a function
I could run that would make the string in quotes "\lambda" appear as a
lambda character in quotes.  Preferably this would be a function that
would run on a string and then use an overlay to change the way the
text looks and "feels" when you're editing (but when you save the
buffer, it would save as latex code, by default).  Another function
would run over the whole buffer, doing the quail translations as it
went.

I think the undocumented function `quail-keyseq-translate' may be
relevant, but I'm not sure how to use it.

Help?


From: Jesper Harder
Subject: Re: multibyte math?
Date: Mon, 03 Nov 2003 18:04:28 +0100
User-agent: Gnus/5.1003 (Gnus v5.10.3) Emacs/21.3.50 (gnu/linux)

Joe Corneli <address@bogus.example.com> writes:

> Will Emacs support styx fonts (www.styxfonts.org) when they are
> released?  Is there anything available now that would let me "do"
> TeXmacs-type stuff inside Emacs?

I'm not sure what you mean by TeXmacs-type stuff.  But there's the TeX
input method `C-x RET C-\ TeX', which lets you input maths characters.

There are also some third-party packages such as x-symbol,
latex-preview, whizzytex, imaxima etc. which do something you might
consider TeXmacs-like.

Re: Question on Mule and Makor2, font display vs char input?

[Stefan Monnier]

>    There is a LaTeX font environment called makor2 which allows a user to
>    create output  using Hebrew  characters.  These output  characters are
>    often defined by multi-byte combinations.   So it would be nice to have
>    an Emacs  mode which allowed a user  at the keyboard to  type a single
>    input  key and have  that keystroke  display the  proper corresponding
>    Hebrew  character  on the  screen  while  placing  the proper  set  of
>    matching  characters into  the  file  for LaTeX  Makor2.   There is  a
>    standard keyboard setup use in Israel  which can be followed and a set
>    of Mule Hebrew screen fonts seem to be readily available.

There are several issues here:
1 - what encoding can be used by makor2 in the TeX files: for hebrew
    characters, Emacs-21.3 supports ISO-8859-8 but not utf-8.
    If Makor2 requires utf-8, you'll need to use Emacs-CVS.
2 - right-to-left text: this is not supported yet.
3 - typing in Hebrew characters: you need to select an input method
    (with C-u C-\).  There's one called `hebrew'.

>    Is there an easy way to  use the existing Mule Hebrew character set as
>    a starting point to create a new Emacs mode definition which will just
>    output the needed Makor2 multi-byte character set?

There's a Hebrew language environment already.

> Any advice on how to make a string already entered into the buffer
> render according to the current input method?  For example, a function
> I could run that would make the string in quotes "\lambda" appear as a
> lambda character in quotes.  Preferably this would be a function that
> would run on a string and then use an overlay to change the way the
> text looks and "feels" when you're editing (but when you save the
> buffer, it would save as latex code, by default).

There's X-Symbol which does just that (tho in a different way).

> Another function would run over the whole buffer, doing the quail
> translations as it went.

I don't know of any function that takes chars from a buffer and runs them
through quail.  Quail is designed to take input from the keyboard, not from
a buffer.


        Stefan

Re: Question on Mule and Makor2, font display vs char input?

[Joe Corneli]

   >    There is a LaTeX font environment called makor2 which allows a user to
   >    create output  using Hebrew  characters.  These output  characters are
   >    often defined by multi-byte combinations.   So it would be nice to have
   >    an Emacs  mode which allowed a user  at the keyboard to  type a single
   >    input  key and have  that keystroke  display the  proper corresponding
   >    Hebrew  character  on the  screen  while  placing  the proper  set  of
   >    matching  characters into  the  file  for LaTeX  Makor2.   There is  a
   >    standard keyboard setup use in Israel  which can be followed and a set
   >    of Mule Hebrew screen fonts seem to be readily available.

   There are several issues here:
   1 - what encoding can be used by makor2 in the TeX files: for hebrew
       characters, Emacs-21.3 supports ISO-8859-8 but not utf-8.
       If Makor2 requires utf-8, you'll need to use Emacs-CVS.
   2 - right-to-left text: this is not supported yet.
   3 - typing in Hebrew characters: you need to select an input method
       (with C-u C-\).  There's one called `hebrew'.

I think the way the encoding issue is supposed to be solved here is
for the file not to be utf-8, but instead just plain-text-unix, and
the buffer was made to look like it was in utf-8.  If I understand
X-Symbol right, Hebrew characters would be handled the same way as
math characters are handled by X-Symbol...

   > Any advice on how to make a string already entered into the buffer
   > render according to the current input method?  For example, a function
   > I could run that would make the string in quotes "\lambda" appear as a
   > lambda character in quotes.  Preferably this would be a function that
   > would run on a string and then use an overlay to change the way the
   > text looks and "feels" when you're editing (but when you save the
   > buffer, it would save as latex code, by default).

   There's X-Symbol which does just that (tho in a different way).

But I ran into trouble trying to build X-Symbol; it seems to have been
written with XEmacs in mind.  So it occured to me that maybe there was
a relatively straightforward way to do the same thing using quail.  I
guess I could have investigated the X-Symbol code better to see how it
is supposed to work.

   > Another function would run over the whole buffer, doing the quail
   > translations as it went.

   I don't know of any function that takes chars from a buffer and runs them
   through quail.  Quail is designed to take input from the keyboard, not from
   a buffer.

Well I hope it isn't too hard to write a function that translates a
string into a keysequence.  Is there an `insert-char' variant that
causes emacs to think that the character was typed by a human user
instead of inserted by a program?  If I'm using the TeX input mode,
(insert-char 92 1) just inserts a `\', but doesn't activate the quail
subsystem.

Re: Question on Mule and Makor2, font display vs char input?

[Stefan Monnier]

> I think the way the encoding issue is supposed to be solved here is
> for the file not to be utf-8, but instead just plain-text-unix, and

UTF-8 *is* plain text.

> the buffer was made to look like it was in utf-8.  If I understand
> X-Symbol right, Hebrew characters would be handled the same way as
> math characters are handled by X-Symbol...

This part of my reply was unrelated to X-Symbol.  X-Symbol is not an
answer for Hebrew text.

>    There's X-Symbol which does just that (tho in a different way).

> But I ran into trouble trying to build X-Symbol; it seems to have been
> written with XEmacs in mind.

It was originally written for XEmacs.  It's been ported to Emacs, tho.
Maybe you should try and figure out why you couldn't get it to work.

> So it occured to me that maybe there was a relatively straightforward way
> to do the same thing using Quail.

Maybe it can be done easly, but AFAIK nobody's done it.


        Stefan

Re: Question on Mule and Makor2, font display vs char input?

[Joe Corneli]

   > I think the way the encoding issue is supposed to be solved here is
   > for the file not to be utf-8, but instead just plain-text-unix, and

   UTF-8 *is* plain text.

What I meant is that the file would contain "\lambda" whereas the
buffer would appear to contain a single lambda character (does this
look right on your screen? "λ").

   > the buffer was made to look like it was in utf-8.  If I understand
   > X-Symbol right, Hebrew characters would be handled the same way as
   > math characters are handled by X-Symbol...

   This part of my reply was unrelated to X-Symbol.  X-Symbol is not an
   answer for Hebrew text.

It appeared to me that the X-Symbol-like setup we were talking about
would accomplish exactly what the OP & me are after.

   >    There's X-Symbol which does just that (tho in a different way).

   > But I ran into trouble trying to build X-Symbol; it seems to have been
   > written with XEmacs in mind.

   It was originally written for XEmacs.  It's been ported to Emacs, tho.
   Maybe you should try and figure out why you couldn't get it to work.

I didn't say I couldn't get it to work, just that I had trouble
getting it to build ;).  I didn't actually try to make it work.
Perhaps I tried to build some unneeded files that haven't been ported.
The README does say that the build process for Emacs should be
improved, and I'd agree.

   > So it occured to me that maybe there was a relatively straightforward way
   > to do the same thing using Quail.

   Maybe it can be done easly, but AFAIK nobody's done it.

Well, it would be easy enough to do if there was a function to
simulate a keypress/keysequence.  One way could be to write a function
that builds a throw-away keyboard macro based on a string, for
example, I assume that would work & that I could figure out how to do
it - but I had expected that there would be a standard way to simulate
a keypress. (Maybe keyboard macros are the standard way.)

Re: Question on Mule and Makor2, font display vs char input?

[Anna Petrov Ronell]

[-- Attachment #1: Type: text/plain, Size: 2501 bytes --]

>>>>> "Stefan" == Stefan Monnier <monnier@iro.umontreal.ca> writes:

    >> There is a LaTeX font environment called makor2 which allows a
    >> user to create output using Hebrew characters.  These output
    >> characters are often defined by multi-byte combinations.  So it
    >> would be nice to have an Emacs mode which allowed a user at the
    >> keyboard to type a single input key and have that keystroke
    >> display the proper corresponding Hebrew character on the screen
    >> while placing the proper set of matching characters into the
    >> file for LaTeX Makor2.  There is a standard keyboard setup use
    >> in Israel which can be followed and a set of Mule Hebrew screen
    >> fonts seem to be readily available.

    Stefan> There are several issues here: 1 - what encoding can be
    Stefan> used by makor2 in the TeX files: for hebrew characters,
    Stefan> Emacs-21.3 supports ISO-8859-8 but not utf-8.  If Makor2
    Stefan> requires utf-8, you'll need to use Emacs-CVS.  2 -
    Stefan> right-to-left text: this is not supported yet.  3 - typing
    Stefan> in Hebrew characters: you need to select an input method
    Stefan> (with C-u C-\).  There's one called `hebrew'.

Unfortunately, Makor2  does not  follow the ISO  standards.  I  am not
sure that it could.  Makor2  allows for things like Cantorial Accents,
Consonants, Vowels  and some  other miscellaneous characters.   I have
attached its PDF reference card at  the end of this message.  It needs
a custom  character set I think.   I have tested 'hebrew'  in mule and
that isn't appropriate in this case.   How difficult is it to create a
custom language mode  where a keystroke or set  of keystrokes prints a
particular font character to the screen and a set of characters to the
open file?

Working  right to  left is  a future  problem, although  if  that were
easily possible, then certainly that would be helpful.

    >> Is there an easy way to use the existing Mule Hebrew character
    >> set as a starting point to create a new Emacs mode definition
    >> which will just output the needed Makor2 multi-byte character
    >> set?

    Stefan> There's a Hebrew language environment already.

I don't think that is appropriate here.  It needs a custom developed
environment, but it would be good to work with/or follow the standard
as closely as possible to try and make the result useful for future
attempts and corrections.  See the attached ref card below.

Thanks for the help and suggestions.


[-- Attachment #2: Makor2 Char Ref Card --]
[-- Type: application/pdf, Size: 251244 bytes --]

[-- Attachment #3: Type: text/plain, Size: 152 bytes --]

_______________________________________________
Help-gnu-emacs mailing list
Help-gnu-emacs@gnu.org
http://lists.gnu.org/mailman/listinfo/help-gnu-emacs

Re: Question on Mule and Makor2, font display vs char input?

[Eli Zaretskii]

> From: aronell@access4less.net (Anna Petrov Ronell) 
> Date: Tue, 30 Nov 2004 16:26:33 -0500
> 
> Is there an easy way to  use the existing Mule Hebrew character set as
> a starting  point to create a  new custom Emacs  mode definition which
> will just output the needed Makor2 multi-byte character set?

No, not that I know of.

I think your best bet would be to use the "File Format Conversions"
feature that is part of Emacs.  This feature lets you specify a shell
command that converts buffer text into a different representation.
Once you've told Emacs how to do that, it will automatically convert
text when it reads it from a file, and convert it back when it writes
it to a file.  See the node "Format Conversion" in the ELisp manual
and the commentary in the format.el file distributed with Emacs.

Using that machinery, you should be able to type Hebrew text (e.g., by
using the Hebrew input method that Stefan mentioned in this thread),
and convert Hebrew characters into LaTeX representation on disk.

Re: Question on Mule and Makor2, font display vs char input?

[Eli Zaretskii]

> From: Joe Corneli <jcorneli@math.utexas.edu>
> Date: Tue, 30 Nov 2004 18:08:03 -0600
> 
> Any advice on how to make a string already entered into the buffer
> render according to the current input method?

You could use display tables for that.  They are documented in the
ELisp manual.

Re: Question on Mule and Makor2, font display vs char input?

[Joe Corneli]

   > Any advice on how to make a string already entered into the buffer
   > render according to the current input method?

   You could use display tables for that.  They are documented in the
   ELisp manual.

The section doesn't make it clear to me how one would change the
display of a string, only the display of a character.

Re: Question on Mule and Makor2, font display vs char input?

[Eli Zaretskii]

> From: Joe Corneli <jcorneli@math.utexas.edu>
> Date: Wed, 01 Dec 2004 13:22:20 -0600
> 
>    > Any advice on how to make a string already entered into the buffer
>    > render according to the current input method?
> 
>    You could use display tables for that.  They are documented in the
>    ELisp manual.
> 
> The section doesn't make it clear to me how one would change the
> display of a string, only the display of a character.

This feature does change display of characters only, but I thought
that would be enough for you to get what you want.

Perhaps I don't understand what exactly are you trying to accomplish.
Could you post an example with explanations?

Re: Question on Mule and Makor2, font display vs char input?

[Joe Corneli]

   > From: Joe Corneli <jcorneli@math.utexas.edu>
   > Date: Wed, 01 Dec 2004 13:22:20 -0600
   > 
   >    > Any advice on how to make a string already entered into the buffer
   >    > render according to the current input method?
   > 
   >    You could use display tables for that.  They are documented in the
   >    ELisp manual.
   > 
   > The section doesn't make it clear to me how one would change the
   > display of a string, only the display of a character.

   This feature does change display of characters only, but I thought
   that would be enough for you to get what you want.

   Perhaps I don't understand what exactly are you trying to accomplish.
   Could you post an example with explanations?


The idea is to automatically translate certain strings into a new
representation while editing.  We'll start with a file like this:

%% example.tex
\documentclass[12pt]{article}
\begin{document}
I am an example. $$\lambda$$  Rowr! 
\end{document}

My .emacs contains something like this:

(add-hook 'latex-mode-hook 'turn-on-special-latex-display-mode)

Now I open example.tex in my Emacs, and I see

%% example.tex
\documentclass[12pt]{article}
\begin{document}
I am an example. $$<lambda>$$  Rowr! 
\end{document}

where <lambda> stands for the lambda character.  Now I edit...

M-: (insert "\\mu")

and I see something like this:

%% example.tex
\documentclass[12pt]{article}
\begin{document}
I am an example. $$<lambda>$$  Rowr! <mu>
\end{document}

(<mu> stands for the mu character).

Next, I save the file and send it to you in an email, and you see:

%% example.tex
\documentclass[12pt]{article}
\begin{document}
I am an example. $$\lambda$$  Rowr! \mu
\end{document}

The critical points are that instead of inputting the characters
through quail (as in "\ l a m b d a") I can insert them using a
program and also have them parsed when the file is first opened and
"unparsed" when the file is saved.

Re: Question on Mule and Makor2, font display vs char input?

[Eli Zaretskii]

> From: Joe Corneli <jcorneli@math.utexas.edu>
> Date: Wed, 01 Dec 2004 23:52:53 -0600
> 
> M-: (insert "\\mu")
> 
> and I see something like this:
> 
> %% example.tex
> \documentclass[12pt]{article}
> \begin{document}
> I am an example. $$<lambda>$$  Rowr! <mu>
> \end{document}
> 
> (<mu> stands for the mu character).

I think the preview-latex package (not yet part of Emacs) does what
you want.

Re: Question on Mule and Makor2, font display vs char input?

[Joe Corneli]

   > From: Joe Corneli <jcorneli@math.utexas.edu>
   > Date: Wed, 01 Dec 2004 23:52:53 -0600
   > 
   > M-: (insert "\\mu")
   > 
   > and I see something like this:
   > 
   > %% example.tex
   > \documentclass[12pt]{article}
   > \begin{document}
   > I am an example. $$<lambda>$$  Rowr! <mu>
   > \end{document}
   > 
   > (<mu> stands for the mu character).

   I think the preview-latex package (not yet part of Emacs) does what
   you want.

No, I want the characters <lambda>, <mu> to be emacs glyphs, not
inline images.  X-symbol is much closer to what I want!

Re: Question on Mule and Makor2, font display vs char input?

[Eli Zaretskii]

> From: Joe Corneli <jcorneli@math.utexas.edu>
> CC: help-gnu-emacs@gnu.org
> Date: Thu, 02 Dec 2004 12:20:50 -0600
> 
>    I think the preview-latex package (not yet part of Emacs) does what
>    you want.
> 
> No, I want the characters <lambda>, <mu> to be emacs glyphs, not
> inline images.

Why?  What does it matter how the character's glyph is generated, if
in the end you see the same shape?

Re: Question on Mule and Makor2, font display vs char input?

[Thien-Thi Nguyen]

Joe Corneli <jcorneli@math.utexas.edu> writes:

> No, I want the characters <lambda>, <mu> to be emacs glyphs, not
> inline images.  X-symbol is much closer to what I want!

all external representations are images in some form or another.

thi

Re: Question on Mule and Makor2, font display vs char input?

[Joe Corneli]

   > From: Joe Corneli <jcorneli@math.utexas.edu>
   > CC: help-gnu-emacs@gnu.org
   > Date: Thu, 02 Dec 2004 12:20:50 -0600
   > 
   >    I think the preview-latex package (not yet part of Emacs) does what
   >    you want.
   > 
   > No, I want the characters <lambda>, <mu> to be emacs glyphs, not
   > inline images.

   Why?  What does it matter how the character's glyph is generated, if
   in the end you see the same shape?

I'm just not a fan of using images files for text.  They're typically
slow to render, and they don't typically look very good.  We have an
extended character set available that has the potential to make
editing math (etc.) seemlessly integrated with editing normal text,
and moreover this can be done with no outside dependencies.  

Re: Question on Mule and Makor2, font display vs char input?

[Joe Corneli]

   Joe Corneli <jcorneli@math.utexas.edu> writes:

   > No, I want the characters <lambda>, <mu> to be emacs glyphs, not
   > inline images.  X-symbol is much closer to what I want!

   all external representations are images in some form or another.

If the characters feel like the usual latin-1 characters, like the
greek letters entered with TeX input mode, I'll be happy.  If they
feel like output from latex2html I don't think I'd use 'em.  The range
of characters handled by the TeX input mode is pretty decent (are
these "internal" representations or "external" ones?), I'd rather
support development of this system (including adding more characters)
than try to use another system.

Re: Question on Mule and Makor2, font display vs char input?

[Eli Zaretskii]

> From: Joe Corneli <jcorneli@math.utexas.edu>
> Date: Thu, 02 Dec 2004 14:38:19 -0600
> 
>    Why?  What does it matter how the character's glyph is generated, if
>    in the end you see the same shape?
> 
> I'm just not a fan of using images files for text.  They're typically
> slow to render, and they don't typically look very good.  We have an
> extended character set available that has the potential to make
> editing math (etc.) seemlessly integrated with editing normal text,
> and moreover this can be done with no outside dependencies.  

LaTeX is not only about characters, it's a word processor, so it
doesn't make much sense (IMHO) to render only part of the text it is
supposed to produce in print, and leave the rest as \directives.

Anyway, to answer your original question whether there's a
customizable display feature that can display a string instead of
another string, then no, there's no such feature.  Overlays is the
infrastructure on which one could build such a feature (you can look,
e.g., at glasses.el for an example of doing that), but I think you
will need to write some Lisp to implement that, customizations alone
will not suffice.

Re: Question on Mule and Makor2, font display vs char input?

[Stefan Monnier]

> No, I want the characters <lambda>, <mu> to be emacs glyphs, not
> inline images.  X-symbol is much closer to what I want!

I'll just repeat since it seems I wasn't clear before:

     You want X-Symbol!

Of course an alternative is to use the TeX input package and to then tell
LaTeX that your file is using the utf-8 encoding.  I.e. your file will
contain λ rather than \lambda.


        Stefan

Re: Question on Mule and Makor2, font display vs char input?

[Oliver Scholz]

Joe Corneli <jcorneli@math.utexas.edu> writes:

>    > From: Joe Corneli <jcorneli@math.utexas.edu>
>    > Date: Wed, 01 Dec 2004 23:52:53 -0600
>    > 
>    > M-: (insert "\\mu")
>    > 
>    > and I see something like this:
>    > 
>    > %% example.tex
>    > \documentclass[12pt]{article}
>    > \begin{document}
>    > I am an example. $$<lambda>$$  Rowr! <mu>
>    > \end{document}
>    > 
>    > (<mu> stands for the mu character).
>
>    I think the preview-latex package (not yet part of Emacs) does what
>    you want.
>
> No, I want the characters <lambda>, <mu> to be emacs glyphs, not
> inline images.  X-symbol is much closer to what I want!

I'd use font-lock to put an composition or display text property on
the string. Have a look at:
http://www.emacswiki.org/cgi-bin/wiki/PrettyLambda

However with using `compose-region' as suggested there you have to be
a bit careful with auto-saved files (because they are encoded in
emacs-mule, which preserves the composition property). Maybe it is
better to use the `display' text property. That would be a bit more
messy, though, if you want cursor movements to DTRT.


    Oliver
-- 
14 Frimaire an 213 de la Révolution
Liberté, Egalité, Fraternité!