emacs 23 input method changes

emacs 23 input method changes

[Laura Conrad]

>>>>> "Ted" == Ted Zlatanov <tzz@lifelogs.com> writes:
    LC> I don't read, and certainly don't type, Chinese.  How do I tell the
    LC> TeX input method:

    LC> To stop accepting input when the character I want appears in the
    LC> mini-buffer?

    Ted> This should be automatic.  If it's not, two or more rules
    Ted> overlap and it may need to be fixed.  Can you specify an
    Ted> example?

I type C-\ and then \'e and I get é- in the minibuffer.  If I type
anything else, I get some oriental character.  For instance, \'ent
gives .$(Qr@.(B (which is a very pretty oriental character in the
emacs23 buffer, but not when I cut and paste it to here in emacs22.

    LC> That in general I'm going to be using two-character sequences for
    LC> Latin alphabets, and not typing Chinese?

Actually, I probably don't want this; I do use TeX input mode for
things like typing \copyright to get ©.

    Ted> I think the TeX input method (as defined in leim/quail/latin-ltx.el)
    Ted> doesn't have any Asian characters. 

Am I supposed to be doing something to have it use latin-ltx instead
of some other language?  The things in my .emacs file contemporary
with having set up the input mode are all about UTF-8:

;utf-8

(setq locale-coding-system 'utf-8)
(set-terminal-coding-system 'utf-8)
(set-keyboard-coding-system 'utf-8)
(set-selection-coding-system 'utf-8)
(prefer-coding-system 'utf-8)

(setq default-input-method "TeX")


    Ted> Can you give an example?  Look in that file to see the
    Ted> current bindings--maybe that will help. 

There's certainly nothing like those Chinese characters in there, so I
must be getting bindings from somewhere else.

    Ted> Can you also check the actual value of your
    Ted> `default-input-method' to make sure it's correct?

describe-variable default-input-method gives:

default-input-method is a variable defined in `mule-cmds.el'.
Its value is "TeX"

This is in emacs-snapshot Ubuntu 8.04.  The help menu About emacs item says:

GNU Emacs 23.0.60.1 (i486-pc-linux-gnu, GTK+ Version 2.12.9)
 of 2008-03-19 on vernadsky, modified by Debian
Copyright (C) 2008 Free Software Foundation, Inc.


-- 
Laura   (mailto:lconrad@laymusic.org http://www.laymusic.org/ )
(617) 661-8097	233 Broadway, Cambridge, MA 02139   

There is a law that no formal fromework of an organisation, or a
society, can affect.  Or not for long.  It is that those who do the
work are the real rulers of it, no matter how they are described.

Doris Lessing, The Sirian Experiments

Re: emacs 23 input method changes

[James Cloos]

>>>>> "Laura" == Laura Conrad <lconrad@laymusic.org> writes:

Laura> I type C-\ and then \'e and I get é- in the minibuffer.  If I type
Laura> anything else, I get some oriental character.  For instance, \'ent
Laura> gives .$(Qr@.(B (which is a very pretty oriental character in the
Laura> emacs23 buffer, but not when I cut and paste it to here in emacs22.

I used to see that bug quite a bit some time back.  I thought that it had
been squashed.  And, indeed, it currently works for me.

Do you see the bug if you start emacs with -Q?

I don't remember with any certainty, but there might have been some
setting I used to use with 22 which 23 didn't need and which caused
that bug.

-JimC
-- 
James Cloos <cloos@jhcloos.com>         OpenPGP: 1024D/ED7DAEA6

Re: emacs 23 input method changes

[Laura Conrad]

>>>>> "James" == James Cloos <cloos@jhcloos.com> writes:

>>>>> "Laura" == Laura Conrad <lconrad@laymusic.org> writes:
    Laura> I type C-\ and then \'e and I get é- in the minibuffer.  If I type
    Laura> anything else, I get some oriental character.  For instance, \'ent
    Laura> gives .$(Qr@.(B (which is a very pretty oriental character in the
    Laura> emacs23 buffer, but not when I cut and paste it to here in emacs22.

    James> I used to see that bug quite a bit some time back.  I thought that it had
    James> been squashed.  And, indeed, it currently works for me.

    James> Do you see the bug if you start emacs with -Q?

No.

    James> I don't remember with any certainty, but there might have been some
    James> setting I used to use with 22 which 23 didn't need and which caused
    James> that bug.

It looks like it's some of that UTF-8 stuff I quoted in my previous mail.
It was necessary to get UTF-8 encoding in 22, but isn't in 23.  (At
least if you have your LOCALE and such set the way I do.)

I'll see if I have to put any of it back, and report, but it looks
from a quick test like things are fine if I take it all out.

Thanks,

-- 
Laura   (mailto:lconrad@laymusic.org http://www.laymusic.org/ )
(617) 661-8097	233 Broadway, Cambridge, MA 02139   

forsan et haec olim meminisse iuvabit.  Vergil

This will make a good story to tell the grandchildren, if we live that
long.  Conrad Translation.

Re: emacs 23 input method changes

[Ted Zlatanov]

On Thu, 11 Sep 2008 16:34:47 -0400 Laura Conrad <lconrad@laymusic.org> wrote: 

LC> (setq locale-coding-system 'utf-8)
LC> (set-terminal-coding-system 'utf-8)
LC> (set-keyboard-coding-system 'utf-8)
LC> (set-selection-coding-system 'utf-8)
LC> (prefer-coding-system 'utf-8)

LC> (setq default-input-method "TeX")

LC> It looks like it's some of that UTF-8 stuff I quoted in my previous mail.
LC> It was necessary to get UTF-8 encoding in 22, but isn't in 23.  (At
LC> least if you have your LOCALE and such set the way I do.)

LC> I'll see if I have to put any of it back, and report, but it looks
LC> from a quick test like things are fine if I take it all out.

If this is indeed working for you, then the broken behavior may be a bug
or at least something that should be noted in the docs.  None of those
functions and variables are deprecated.

Ted

Re: emacs 23 input method changes

[James Cloos]

LC> (setq locale-coding-system 'utf-8)
LC> (set-terminal-coding-system 'utf-8)
LC> (set-keyboard-coding-system 'utf-8)
LC> (set-selection-coding-system 'utf-8)
LC> (prefer-coding-system 'utf-8)

LC> It looks like it's some of that UTF-8 stuff I quoted in my previous mail.

Looking at my output from (describe-coding-system) suggests that setting
the keyboard coding system is the culprit.  I have no-conversion there,
nil for inter-client cut-n-paste and utf-8-unix elsewhere.  All w/o any
settings in ~/.emacs or my custom file.

-JimC
-- 
James Cloos <cloos@jhcloos.com>         OpenPGP: 1024D/ED7DAEA6

Re: emacs 23 input method changes

[Laura Conrad]

>>>>> "James" == James Cloos <cloos@jhcloos.com> writes:

    LC> (setq locale-coding-system 'utf-8)
    LC> (set-terminal-coding-system 'utf-8)
    LC> (set-keyboard-coding-system 'utf-8)
    LC> (set-selection-coding-system 'utf-8)
    LC> (prefer-coding-system 'utf-8)

    LC> It looks like it's some of that UTF-8 stuff I quoted in my previous mail.

    James> Looking at my output from (describe-coding-system) suggests that setting
    James> the keyboard coding system is the culprit.  I have no-conversion there,
    James> nil for inter-client cut-n-paste and utf-8-unix elsewhere.  All w/o any
    James> settings in ~/.emacs or my custom file.

Yes, that looks like the answer.  I had commented out all of those
settings, and didn't get the problem.  If I put them all back except
for set-keyboard-coding-system, I do get the problem.

(For those who don't remember the beginning of the thread, the problem
is that in TeX input mode, I expect to be able to type ént by typing
"\'ent".  When I have the problem, typing those four characters gives
me some Chinese character.)

-- 
Laura   (mailto:lconrad@laymusic.org http://www.laymusic.org/ )
(617) 661-8097	233 Broadway, Cambridge, MA 02139   

Why was I born with such contemporaries?

Oscar Wilde (1854 - 1900)

Re: emacs 23 input method changes

[Ted Zlatanov]

On Sat, 13 Sep 2008 12:55:06 -0400 Laura Conrad <lconrad@laymusic.org> wrote: 

>>>>>> "James" == James Cloos <cloos@jhcloos.com> writes:
LC> (setq locale-coding-system 'utf-8)
LC> (set-terminal-coding-system 'utf-8)
LC> (set-keyboard-coding-system 'utf-8)
LC> (set-selection-coding-system 'utf-8)
LC> (prefer-coding-system 'utf-8)

LC> It looks like it's some of that UTF-8 stuff I quoted in my previous mail.

James> Looking at my output from (describe-coding-system) suggests that setting
James> the keyboard coding system is the culprit.  I have no-conversion there,
James> nil for inter-client cut-n-paste and utf-8-unix elsewhere.  All w/o any
James> settings in ~/.emacs or my custom file.

LC> Yes, that looks like the answer.  I had commented out all of those
LC> settings, and didn't get the problem.  If I put them all back except
LC> for set-keyboard-coding-system, I do get the problem.

LC> (For those who don't remember the beginning of the thread, the problem
LC> is that in TeX input mode, I expect to be able to type ént by typing
LC> "\'ent".  When I have the problem, typing those four characters gives
LC> me some Chinese character.)

Could you submit an Emacs bug about it?  It will improve the software
for other users.  I could do it but you had the original report :)

Ted

Re: emacs 23 input method changes

[Laura Conrad]

>>>>> "Ted" == Ted Zlatanov <tzz@lifelogs.com> writes:

    Ted> Could you submit an Emacs bug about it?  It will improve the software
    Ted> for other users.  I could do it but you had the original report :)

I did that using the "Send bug report" option on the Help menu.

-- 
Laura   (mailto:lconrad@laymusic.org http://www.laymusic.org/ )
(617) 661-8097	233 Broadway, Cambridge, MA 02139   

Strong hope is a much greater stimulant of life than any realized joy
could be.

Nietzsche

emacs 23 input method changes

[Laura Conrad]

I just upgraded my ubuntu system to Hardy Heron, which has emacs 23 as
the default emacs, instead of emacs 22.

I had been using the TeX input method for occasional non-ascii
characters, such as é or ü.  

This no longer works for me, because the default seems to be that
after you've typed a backslash, it goes on reading several characters,
and I always end up with something in Chinese or Japanese when I've
tried to type an é.

I don't read, and certainly don't type, Chinese.  How do I tell the
TeX input method:

    To stop accepting input when the character I want appears in the
    mini-buffer?

or:

    That in general I'm going to be using two-character sequences for
    Latin alphabets, and not typing Chinese?

-- 
Laura   (mailto:lconrad@laymusic.org http://www.laymusic.org/ )
(617) 661-8097	233 Broadway, Cambridge, MA 02139   

Strong hope is a much greater stimulant of life than any realized joy
could be.

Nietzsche

Re: emacs 23 input method changes

[Ted Zlatanov]

On Thu, 11 Sep 2008 08:13:22 -0400 Laura Conrad <lconrad@laymusic.org> wrote: 

LC> I just upgraded my ubuntu system to Hardy Heron, which has emacs 23 as
LC> the default emacs, instead of emacs 22.

LC> I had been using the TeX input method for occasional non-ascii
LC> characters, such as é or ü.  

LC> This no longer works for me, because the default seems to be that
LC> after you've typed a backslash, it goes on reading several characters,
LC> and I always end up with something in Chinese or Japanese when I've
LC> tried to type an é.

LC> I don't read, and certainly don't type, Chinese.  How do I tell the
LC> TeX input method:

LC>     To stop accepting input when the character I want appears in the
LC>     mini-buffer?

This should be automatic.  If it's not, two or more rules overlap and it
may need to be fixed.  Can you specify an example?

LC>     That in general I'm going to be using two-character sequences for
LC>     Latin alphabets, and not typing Chinese?

I think the TeX input method (as defined in leim/quail/latin-ltx.el)
doesn't have any Asian characters.  Can you give an example?  Look in
that file to see the current bindings--maybe that will help.  Can you
also check the actual value of your `default-input-method' to make sure
it's correct?

Ted