<Cyrillic_schwa> is undefined

<Cyrillic_schwa> is undefined

[Baurzhan Ismagulov]

Hello,

I can use emacs cvs with pt154 encoding on the Linux console. However,
when I type some Cyrillic characters under X, emacs says, e.g.,
"<Cyrillic_schwa> is undefined".

I see that this message comes from src/keyboard.c, however, I couldn't
figure out how the character is processed.

I want to add support for the missing characters, where should I look?

Thanks in advance,
Baurzhan.

Re: <Cyrillic_schwa> is undefined

[Kenichi Handa]

In article <20051001193132.GE25392@radix50.net>, Baurzhan Ismagulov <ibr@radix50.net> writes:

> I can use emacs cvs with pt154 encoding on the Linux console. However,
> when I type some Cyrillic characters under X, emacs says, e.g.,
> "<Cyrillic_schwa> is undefined".

It seems that these X11 keysyms (in
/usr/include/X11/keysymdefs.h, perhaps newly added) are not
registered in x-keysym-table (in lisp/term/x-win.el).

#define XK_Cyrillic_GHE_bar	                           0x680
#define XK_Cyrillic_ghe_bar	                           0x690
...
#define XK_Cyrillic_U_macron		                   0x68f
#define XK_Cyrillic_u_macron		                   0x69f

For instance, for XK_Cyrillic_U_macron, you can add it as
this:

(puthash #x68F (decode-char 'ucs #x04EE) x-keysym-table)
         ^^^^keysym_code         ^^^^^^unicode_character_code

But, for instance, for XK_Cyrillic_GHE_bar, I don't know the
corresponding Unicode character code.  If you give me a list
of keysym_codes and the corresponding Unicode character
code, I'll add it in x-win.el.

---
Kenichi Handa
handa@m17n.org

Re: <Cyrillic_schwa> is undefined

[James Cloos]

>>>>> "Kenichi" == Kenichi Handa <handa@m17n.org> writes:

Kenichi> It seems that these X11 keysyms (in
Kenichi> /usr/include/X11/keysymdefs.h, perhaps newly added) are not
Kenichi> registered in x-keysym-table (in lisp/term/x-win.el).

For the Cyrillic, see /usr/include/X11/keysymdef.h on a current
install, or online at:

http://cvs.freedesktop.org/xorg/xc/include/keysymdef.h?rev=1.3&view=markup

It has unicode values for each keysym that is in unicode.  (And note
that for any keysym that matches (and 0x01000000 keysym) the unicode
codepoint can be found by (and 0xFFFFFF keysym).


I'm seeing a similar problem with <Multi_key>.

It works for every app other than Emacs.

I don't know whether it is a regression; I only tried it becuase
quail was broken in emacs-unicode-2 of a few days ago (since fixed).

I would be useful if X's standard compose feature were usable in
emacs.  Or, OTOH, perhaps it would make a good alternate first
stroke for the quail methods like rfc1345, sqml, latex, etc.

Xev says this about the key:

,----
| KeyPress event, serial 24, synthetic NO, window 0x1a00001,
|     root 0x4c, subw 0x0, time 3255347295, (75,121), root:(696,1088),
|     state 0x0, keycode 117 (keysym 0xff20, Multi_key), same_screen YES,
|     XLookupString gives 0 bytes:
| 
| KeyRelease event, serial 27, synthetic NO, window 0x1a00001,
|     root 0x4c, subw 0x0, time 3255347390, (75,121), root:(696,1088),
|     state 0x0, keycode 117 (keysym 0xff20, Multi_key), same_screen YES,
|     XLookupString gives 0 bytes:
`----



-JimC
-- 
James H. Cloos, Jr. <cloos@jhcloos.com>

Re: <Cyrillic_schwa> is undefined

[Stefan Monnier]

> I'm seeing a similar problem with <Multi_key>.

I've been using Multi_key with Emacs "for ever", so the part that I use at
least does work.  Please give us a proper bug report.


        Stefan

Re: <Cyrillic_schwa> is undefined

[Kenichi Handa]

In article <m3wtkrztr6.fsf@lugabout.cloos.reno.nv.us>, James Cloos <cloos@jhcloos.com> writes:

>>>>>>  "Kenichi" == Kenichi Handa <handa@m17n.org> writes:
Kenichi>  It seems that these X11 keysyms (in
Kenichi>  /usr/include/X11/keysymdefs.h, perhaps newly added) are not
Kenichi>  registered in x-keysym-table (in lisp/term/x-win.el).

> For the Cyrillic, see /usr/include/X11/keysymdef.h on a current
> install, or online at:

> http://cvs.freedesktop.org/xorg/xc/include/keysymdef.h?rev=1.3&view=markup

> It has unicode values for each keysym that is in unicode.  (And note
> that for any keysym that matches (and 0x01000000 keysym) the unicode
> codepoint can be found by (and 0xFFFFFF keysym).

Thank you for the info, but, oops, it doesn't have these
lines (they exist in my debian):

#define XK_Cyrillic_GHE_bar	                           0x680
[...]
#define XK_Cyrillic_u_macron		                   0x69f

but have:

#define XK_Cyrillic_GHE_bar           0x1000492  /* U+0492 CYRILLIC CAPITAL LETTER GHE WITH STROKE */
[...]
#define XK_Cyrillic_u_macron          0x10004ef  /* U+04EF CYRILLIC SMALL LETTER U WITH MACRON */

What's going on here?

Anyway, as it's easy to handle them in emacs-unicode-2, I've
just commited a code for that for emacs-unicode-2.

> I'm seeing a similar problem with <Multi_key>.

> It works for every app other than Emacs.

> I don't know whether it is a regression; I only tried it becuase
> quail was broken in emacs-unicode-2 of a few days ago (since fixed).

> I would be useful if X's standard compose feature were usable in
> emacs.  Or, OTOH, perhaps it would make a good alternate first
> stroke for the quail methods like rfc1345, sqml, latex, etc.

As I'm not European users, I don't know how Multi_key is
used.  Could you explain the detail of the current problem?

---
Kenichi Handa
handa@m17n.org

Re: <Cyrillic_schwa> is undefined

[James Cloos]

>>>>> "Kenichi" == Kenichi Handa <handa@m17n.org> writes:


Kenichi> Thank you for the info, but, oops, it doesn't have these
Kenichi> lines (they exist in my debian): ...

Kenichi> What's going on here?

The diff between revisions 1.2 and 1.3 of that file as seen at:

http://cvs.freedesktop.org/xorg/xc/include/keysymdef.h?r1=1.2&r2=1.3

shows several changes from old-style keysyms to unicode keysyms.

That checkin replaced all uppercase hex constants with lowercase,
added unicode codepoint comments (from UnicodeData.txt), and
switched several blocks to the unicode block (0x1000000 to 0x1FFFFFF).

The LATIN8 and ARABIC blocks were completely converted and the
CYRILLIC block partially converted.

Emacs, of course, will need to support both sets -- and any other
differences from older versions -- if it wants to support all X
servers out there....

Incidently, for anything it doesn't recognize that is:

  (and (>= 0x1000000 code) (<0x2000000 code))

it can at least insert the ucs character, even if it doesn't know the
keysym's name.  Perhaps the symbol <UCS_XXXXXX> could be a synonym for
the keycode 0x1XXXXXX, for all XXXXXX?

>> I'm seeing a similar problem with <Multi_key>.

Kenichi> As I'm not European users, I don't know how Multi_key is
Kenichi> used.  Could you explain the detail of the current problem?

Multi_Key is the keysym for the key usually labeled as Compose on the
keypad.  On a modern PC keyboard, X usually maps the so called menu
key to Multi_Key.  It initiates a compose sequence, based on the
$LOCALE/Compose file.  (/usr/X11R6/lib/X11/locale/${LOCALE}/Compose
on my Debian box, /usr/lib/X11/locale/${LOCALE}/Compose on my Gentoo
box.)  

...

I just noticed a problem with the X11/locale hierarchy on this box;
I'll first have to confirm that the error isn't there.

-JimC
-- 
James H. Cloos, Jr. <cloos@jhcloos.com>

Re: <Cyrillic_schwa> is undefined

[James Cloos]

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

JimC> Perhaps the symbol <UCS_XXXXXX> could
JimC> be a synonym for the keycode 0x1XXXXXX, for all XXXXXX?

[SIGH]  Seems X already has <UXXXX> as a synonym for 0x100XXXX,
and probably <UXXXXX> for 0x10XXXXX, etc....

-JimC

Re: <Cyrillic_schwa> is undefined

[Baurzhan Ismagulov]

[-- Attachment #1: Type: text/plain, Size: 1939 bytes --]

On Wed, Oct 05, 2005 at 09:53:48PM +0900, Kenichi Handa wrote:
> (puthash #x68F (decode-char 'ucs #x04EE) x-keysym-table)

Great, thanks much!

The attached should add the symbols below (gzipped so that MTAs don't
recode it). I'm not sure whether I could encode it correctly, but it
seems to save files that I expect. Could you please correct it if
necessary and apply?

However, I still have a problem. The newly added characters are
displayed as rectangles under X (I have a fixed font containing those
glyphs, at least xterm shows them correctly).

Moreover, if I open an empty file, type, say, CYRILLIC CAPITAL LETTER A,
C-x RET c, utf-8 RET, C-x C-s, the file is saved in utf-8 and contains
0xd0 0x90, which seems to be correct. If I leave emacs, open the file
via C-x RET c, utf-8 RET, C-x C-f, name RET, the letter is displayed as
a rectangle.

Any idea? (We can take this to the user list, if you like).

With kind regards,
Baurzhan.

<U0492>	    0x680         CYRILLIC CAPITAL LETTER GHE WITH STROKE
<U049A>     0x682         CYRILLIC CAPITAL LETTER KA WITH DESCENDER
<U04A2>     0x684         CYRILLIC CAPITAL LETTER EN WITH DESCENDER
<U04AE>     0x685         CYRILLIC CAPITAL LETTER STRAIGHT U
<U04B0>     0x686         CYRILLIC CAPITAL LETTER STRAIGHT U WITH STROKE
<U04BA>     0x68a         CYRILLIC CAPITAL LETTER SHHA
<U04D8>     0x68c         CYRILLIC CAPITAL LETTER SCHWA
<U04E8>     0x68e         CYRILLIC CAPITAL LETTER BARRED O
<U0493>     0x690         CYRILLIC SMALL LETTER GHE WITH STROKE
<U049B>     0x692         CYRILLIC SMALL LETTER KA WITH DESCENDER
<U04A3>     0x694         CYRILLIC SMALL LETTER EN WITH DESCENDER
<U04AF>     0x695         CYRILLIC SMALL LETTER STRAIGHT U
<U04B1>     0x696         CYRILLIC SMALL LETTER STRAIGHT U WITH STROKE
<U04BB>     0x69a         CYRILLIC SMALL LETTER SHHA
<U04D9>     0x69c         CYRILLIC SMALL LETTER SCHWA
<U04E9>     0x69e         CYRILLIC SMALL LETTER BARRED O

[-- Attachment #2: emacs-ibr-kk-20051005-2335.diff.gz --]
[-- Type: application/octet-stream, Size: 366 bytes --]

[-- Attachment #3: Type: text/plain, Size: 142 bytes --]

_______________________________________________
Emacs-devel mailing list
Emacs-devel@gnu.org
http://lists.gnu.org/mailman/listinfo/emacs-devel

ping: Re: <Cyrillic_schwa> is undefined

[Baurzhan Ismagulov]

Hello Kenichi,

On Wed, Oct 05, 2005 at 09:53:48PM +0900, Kenichi Handa wrote:
> But, for instance, for XK_Cyrillic_GHE_bar, I don't know the
> corresponding Unicode character code.  If you give me a list
> of keysym_codes and the corresponding Unicode character
> code, I'll add it in x-win.el.

I haven't received any feedback regarding the symbols mentioned in
http://lists.gnu.org/archive/html/emacs-devel/2005-10/msg00176.html ,
and I don't see them in lisp/term/x-win.el, either.

Would you like to review and apply the patch, or make the changes if the
patch is wrong?

Thanks in advance,
Baurzhan.

ping2: Re: <Cyrillic_schwa> is undefined

[Baurzhan Ismagulov]

Hello Handa,

On Wed, Oct 05, 2005 at 09:53:48PM +0900, Kenichi Handa wrote:
> But, for instance, for XK_Cyrillic_GHE_bar, I don't know the
> corresponding Unicode character code.  If you give me a list
> of keysym_codes and the corresponding Unicode character
> code, I'll add it in x-win.el.

could you please share your plans regarding this issue? The list of the
characters and a patch are at
http://lists.gnu.org/archive/html/emacs-devel/2005-10/msg00176.html .
The patch may be encoded wrongly.

With kind regards,
Baurzhan.

Re: ping2: Re: <Cyrillic_schwa> is undefined

[Kenichi Handa]

In article <20051021180026.GA19632@radix50.net>, Baurzhan Ismagulov <ibr@radix50.net> writes:

> Hello Handa,
> On Wed, Oct 05, 2005 at 09:53:48PM +0900, Kenichi Handa wrote:
>>  But, for instance, for XK_Cyrillic_GHE_bar, I don't know the
>>  corresponding Unicode character code.  If you give me a list
>>  of keysym_codes and the corresponding Unicode character
>>  code, I'll add it in x-win.el.

> could you please share your plans regarding this issue? The list of the
> characters and a patch are at
> http://lists.gnu.org/archive/html/emacs-devel/2005-10/msg00176.html .
> The patch may be encoded wrongly.

I'm very sorry for not responding on this matter.  I'm now
working on it.  Please wait for a while.

---
Kenichi Handa
handa@m17n.org

Re: ping2: Re: <Cyrillic_schwa> is undefined

[Baurzhan Ismagulov]

Hello Handa,

On Sat, Oct 22, 2005 at 10:45:26AM +0900, Kenichi Handa wrote:
> > On Wed, Oct 05, 2005 at 09:53:48PM +0900, Kenichi Handa wrote:
> >>  But, for instance, for XK_Cyrillic_GHE_bar, I don't know the
> >>  corresponding Unicode character code.  If you give me a list
> >>  of keysym_codes and the corresponding Unicode character
> >>  code, I'll add it in x-win.el.
> 
> > could you please share your plans regarding this issue? The list of the
> > characters and a patch are at
> > http://lists.gnu.org/archive/html/emacs-devel/2005-10/msg00176.html .
> > The patch may be encoded wrongly.
> 
> I'm very sorry for not responding on this matter.  I'm now
> working on it.  Please wait for a while.

No problem, thanks for the feedback! I know people are busy, but I
wanted to know your decision. I can wait with actual inclusion.

With kind regards,
Baurzhan.

Re: ping2: Re: <Cyrillic_schwa> is undefined

[Kenichi Handa]

In article <E1ET8Rq-0001LA-00@etlken>, Kenichi Handa <handa@m17n.org> writes:

> In article <20051021180026.GA19632@radix50.net>, Baurzhan Ismagulov <ibr@radix50.net> writes:
>>  Hello Handa,
>>  On Wed, Oct 05, 2005 at 09:53:48PM +0900, Kenichi Handa wrote:
>>>   But, for instance, for XK_Cyrillic_GHE_bar, I don't know the
>>>   corresponding Unicode character code.  If you give me a list
>>>   of keysym_codes and the corresponding Unicode character
>>>   code, I'll add it in x-win.el.

>>  could you please share your plans regarding this issue? The list of the
>>  characters and a patch are at
>>  http://lists.gnu.org/archive/html/emacs-devel/2005-10/msg00176.html .
>>  The patch may be encoded wrongly.

> I'm very sorry for not responding on this matter.  I'm now
> working on it.  Please wait for a while.

I've just installed a fix.  Actually, I registered more
Cyrillic characters in x-keysym-table than what you listed
(e.g. 0x681 for XK_Cyrillic_ZHE_descender) because...

James Cloos <cloos@jhcloos.com> writes:
[...]
> The diff between revisions 1.2 and 1.3 of that file as seen at:
> http://cvs.freedesktop.org/xorg/xc/include/keysymdef.h?r1=1.2&r2=1.3
> shows several changes from old-style keysyms to unicode keysyms.

the above page indicates that some version of X contains
them too, and I confirmed that some distribution surely has
them.

In addition,

> That checkin replaced all uppercase hex constants with lowercase,
> added unicode codepoint comments (from UnicodeData.txt), and
> switched several blocks to the unicode block (0x1000000 to 0x1FFFFFF).

I installed changes to handle them in xterm.c without
calling Lisp but by directly using MAKE_CHAR macro.

But, I can't test them becasue my X server doesn't have
those keysyms.  If yours does, please test it.

---
Kenichi Handa
handa@m17n.org

Re: ping2: Re: <Cyrillic_schwa> is undefined

[Baurzhan Ismagulov]

Hello Handa,

On Mon, Oct 24, 2005 at 08:29:00PM +0900, Kenichi Handa wrote:
> I've just installed a fix.
...
> But, I can't test them becasue my X server doesn't have
> those keysyms.  If yours does, please test it.

Works great, thanks!

With kind regards,
Baurzhan.