Cocoa Emacs 23 & Unicode

Cocoa Emacs 23 & Unicode

[Ian Eure]

I built Cocoa Emacs 23 from CVS and started playing with it. It looks  
like Unicode directional quote glyphs are being displayed incorrectly,  
taking up two characters instead of one. Other glyphs display fine.

Here's an example vs. Aquamacs 1.5: http://flickr.com/photos/ieure/3046081601/

Is this a known issue? Is a fix planned?

  - Ian

Re: Cocoa Emacs 23 & Unicode

[Peter Dyballa]

Am 20.11.2008 um 23:32 schrieb Ian Eure:

> Is this a known issue?

Not the way that you describe it! Check with C-u C-x = which font is  
actually used ...

> Is a fix planned?


It's wishful that X11 and Cocoa Emacs 23 use the font(s) or fontset 
(s) the user wishes to use. Usually *your* fonts are substituted with  
rather exotic choices ...


For me the Monaco or NimbusMono “” are substituted from STHeiti, a  
proportional font outside CJK. Both Monaco and NimbusMono have „, “,  
“, ” ... It's just a bug!

--
Greetings

   Pete       (:
         _    / __    -    -
       _/ \__/_/        -     -
      (´`)      (´`)   -    -
       `´        `´

Re: Cocoa Emacs 23 & Unicode

[Ian Eure]

On Nov 20, 2008, at 3:09 PM, Peter Dyballa wrote:

>
> Am 20.11.2008 um 23:32 schrieb Ian Eure:
>
>> Is this a known issue?
>
> Not the way that you describe it! Check with C-u C-x = which font is  
> actually used ...
>
>> Is a fix planned?
>
> It's wishful that X11 and Cocoa Emacs 23 use the font(s) or  
> fontset(s) the user wishes to use. Usually *your* fonts are  
> substituted with rather exotic choices ...
>
Yeah, I understand that certain code ranges may need to be mapped to  
different fonts.


> For me the Monaco or NimbusMono “” are substituted from STHeiti, a  
> proportional font outside CJK. Both Monaco and NimbusMono have „, “,  
> “, ” ... It's just a bug!
>

You're right, it's getting changed to STHeiti for those characters.  
Know of any workarounds?

It seems that the characters are incorrectly mapped into CJK; This is  
from describe-char:

          category: c:Chinese h:Korean j:Japanese

I've sent a bug report.

  - Ian

Re: Cocoa Emacs 23 & Unicode

[Peter Dyballa]

Am 21.11.2008 um 00:33 schrieb Ian Eure:

> You're right, it's getting changed to STHeiti for those characters.  
> Know of any workarounds?

Use the X11 client – with quotes it mostly behaves better.

>
> It seems that the characters are incorrectly mapped into CJK; This  
> is from describe-char:
>
>          category: c:Chinese h:Korean j:Japanese
>
> I've sent a bug report.


I think I reported a similiar bug in X11 client before. Then I  
received as answer that some Chinese mapping (GB....) incorporates  
(my finding: part of) Unicode, so it's better or easier to see it all  
from a Chinese font encoding's point of view. Me, I would prefer  
Tibetan, probably that would *not* end in some CJK category ...

It's hard to make the developers understand our Latin-centric view!

--
Greetings

   Pete

There's no place like 127.0.0.1
			– origin unknown

Re: Cocoa Emacs 23 & Unicode

[Ian Eure]

On Nov 20, 2008, at 4:04 PM, Peter Dyballa wrote:

>
> Am 21.11.2008 um 00:33 schrieb Ian Eure:
>
>> You're right, it's getting changed to STHeiti for those characters.  
>> Know of any workarounds?
>
> Use the X11 client – with quotes it mostly behaves better.
>
If I wanted to use X11, I'd have kept using Linux on my desktop.


> I think I reported a similiar bug in X11 client before. Then I  
> received as answer that some Chinese mapping (GB....) incorporates  
> (my finding: part of) Unicode, so it's better or easier to see it  
> all from a Chinese font encoding's point of view. Me, I would prefer  
> Tibetan, probably that would *not* end in some CJK category ...
>
> It's hard to make the developers understand our Latin-centric view!
>
Any hints on where I can find and tweak this stuff? I'd like to use  
Emacs 23, but I U+2665 my Unicode and don't want it broken.

  - Ian

Re: Cocoa Emacs 23 & Unicode

[Peter Dyballa]

Am 21.11.2008 um 19:43 schrieb Ian Eure:

> Any hints on where I can find and tweak this stuff?

I presume it's lisp/international/characters.el and the Elisp files  
there. The uni-* files are automatically generated from UnicodeData.txt.

Another option would be to follow backwards the function ...

--
Greetings

   Pete

Think of XML as Lisp for COBOL programmers.
				- Tony-A (some guy on /.)

Re: Cocoa Emacs 23 & Unicode

[Ian Eure]

On Nov 21, 2008, at 12:37 PM, Peter Dyballa wrote:

>
> Am 21.11.2008 um 19:43 schrieb Ian Eure:
>
>> Any hints on where I can find and tweak this stuff?
>
> I presume it's lisp/international/characters.el and the Elisp files  
> there. The uni-* files are automatically generated from  
> UnicodeData.txt.
>
> Another option would be to follow backwards the function ...
>
I looked, but I'm not clearly seeing where in that file it comes from.  
I manually removed the CJK categorization with:

(modify-category-entry '(#x201b . #x201f) ?c nil t)
(modify-category-entry '(#x201b . #x201f) ?h nil t)
(modify-category-entry '(#x201b . #x201f) ?j nil t)

And added Latin with:

(modify-category-entry '(#x201b . #x201f) ?l)

However, they still display with the STHelti font, leading me to  
believe that the char -> font mapping lies elsewhere.

Any other clues?

  - Ian

Re: Cocoa Emacs 23 & Unicode

[Peter Dyballa]

Am 24.11.2008 um 19:46 schrieb Ian Eure:

> However, they still display with the STHelti font, leading me to  
> believe that the char -> font mapping lies elsewhere.

Some formal text in the *Help* buffer obviously comes from the  
function describe-char in lisp/descr-text.el. Obviously the function  
char-category-set, as defined in src/category.c:284, brings up that  
CJK content.

I tried to use edebug (debug-on-entry), but this did not work. I  
might need to read an edebug tutorial! Some day ...

--
Greetings

   Pete

The best way to accelerate a PC is 9.8 m/s2

Re: Cocoa Emacs 23 & Unicode

[Jason Rumney]

On Nov 22, 2:43 am, Ian Eure <i...@digg.com> wrote:

> Any hints on where I can find and tweak this stuff? I'd like to use  
> Emacs 23, but I U+2665 my Unicode and don't want it broken.
>
>   - Ian

(set-fontset-font "fontset-default" '(#x201b . #x201f)
  '("Monaco" . "iso10646-1"))

Re: Cocoa Emacs 23 & Unicode

[Ian Eure]

On Nov 24, 2008, at 7:07 PM, Jason Rumney wrote:

> (set-fontset-font "fontset-default" '(#x201b . #x201f)
>  '("Monaco" . "iso10646-1"))

That didn't work, but this did:

;; Fixes ‛“”„‟
(set-fontset-font "fontset-startup" '(#x201b . #x201f)
  '("Monaco" . "iso10646-1"))

;; Fixes ‘’
(set-fontset-font "fontset-startup" '(#x2018 . #x2019)
  '("Monaco" . "iso10646-1"))

Thanks,
Ian