Re: File names with accented Latin characters are not displayed correctly

[Peter Dyballa]

Am 07.11.2005 um 14:49 schrieb Peter Dyballa:

> When I use 'ls -lw' to display the file names in xterm, I get:
>
>         -rw-r--r--    1 pete  pete        62 25 Mär  2005 áÛïǓà.txt
>         -rw-r--r--    1 pete  pete        62 25 Mär  2005 äÖüÄöÜ.txt
>         -rw-r--r--    1 pete  pete       107  2 Dez  2004 äöüßÜÖÄ€
>         
> Doing the same in Emacs' *shell* buffer I get:
>
>         -rw-r--r--    1 pete  pete        62 25 Mär  2005 áÛïǓà.txt
>         -rw-r--r--    1 pete  pete        62 25 Mär  2005 äÖüÄöÜ.txt
>         -rw-r--r--    1 pete  pete       107  2 Dez  2004 äöüßÜÖÄ€
>
> Doing in Emacs' *shell* buffer 'ls -l' I get of course some nonsense. 
> In dired-mode I see as in *shell* only:
>
>   -rw-r--r--    1 pete   pete        62 25 Mär  2005 áÛïǓà.txt
>   -rw-r--r--    1 pete   pete        62 25 Mär  2005 äÖüÄöÜ.txt
>   -rw-r--r--    1 pete   pete       107  2 Dez  2004 äöüßÜÖÄ€
>
> So GNU Emacs can't display the file names correctly any more!
>
>
> In GNU Emacs 23.0.0.1 (powerpc-apple-darwin7.9.0, X toolkit, Xaw3d 
> scroll bars)
>  of 2005-11-07 on localhost

I spent some time again on this topic, this time hitting C-u C-x = for 
example on ä in dired-mode. It gives me:

                character: a (0141, 97, 0x61)
        preferred charset: [ascii] (ASCII (ISO646 IRV))
               code point: [0x61]
                   syntax: w    which means: word
                 category: a:ASCII   l:Latin   r:Japanese roman
              buffer code: 0x61
                file code: not encodable by coding system utf-8
                  display: composed to form "ä" (see below)
        
        Composed with the following character(s) "̈▢" by the rule:
                (?a (tc . bc) ?̈▢)
        The component character(s) are displayed by these fonts (glyph codes):
	 a: 
-B&H-LucidaTypewriter-Medium-R-Normal-Sans-10-100-75-75-M-60-ISO8859-1 
(0x61)
         ̈: -- no font --
        See the variable `reference-point-alist' for the meaning of the rule.

The boxes do not seem to be all the same. First the one between double 
quotes:

                character: ̈▢ (01410, 776, 0x308)
	preferred charset: [japanese-jisx0213-1] (JISX0213.2000 Plane 1 
(Japanese))
               code point: [0x2B6D]
                   syntax: w    which means: word
                 category: ^:Combining diacritic or mark   j:Japanese
              buffer code: 0xCC 0x88
                file code: 0xCC 0x88 (encoded by coding system utf-8)
                  display: composed to form "̈" (see below)
        
        Composed by the rule:
                ()
        The component character(s) are displayed by these fonts (glyph codes):
        See the variable `reference-point-alist' for the meaning of the rule.

And this is the box in the "rule" inside the parentheses:

               character: ̈ ▢ (01410, 776, 0x308)
	preferred charset: [japanese-jisx0213-1] (JISX0213.2000 Plane 1 
(Japanese))
               code point: [0x2B6D]
                   syntax: w    which means: word
                 category: ^:Combining diacritic or mark   j:Japanese
              buffer code: 0xCC 0x88
                file code: 0xCC 0x88 (encoded by coding system utf-8)
                  display: no font available

(In comparison to the first some lines were missing in *Help* buffer.)

OK, now an explanation is given: no font. The question is: do I need to 
supply a font? If so: how? Hitting C-h v on that 
`reference-point-alist' gives a reference to a variable (I think: too 
big to cite it here) defined in `composite'. There I found a reference 
to the function toggle-auto-composition. When I apply this function to 
the *Buffers List* I can see that it "changes" one file name: obviously 
one which is the exact copy of the entry in dired-mode!

And I now recognised too that I when I open a file with the ä in the 
name, it appears in mode-line correct. In the pop-up buffers menu I see 
its name printed in normal UTF-8 representation, i.e. C3 A4 = ä. OK, I 
can guess the right name. When I open such a file from dired-mode by 
pressing the mouse, the ä is represented by a hollow box in the 
mode-line. This hollow box is "translated" in pop-up buffers menu to 
"Ì▢." OK, I am cheating a bit: when I open the file with C-x f or 
change any name to a name with ä, then the name is correct in 
mode-line. In *Buffer List* this name is displayed correctly too. The 
other file name, which I open 'with the mouse,' has the de-composed ä 
glyph which is described by C-u C-x = as the ä in dired mode. And in 
this file name I can toggle the representation between "a" and "a▢" -- 
but no change in pop-up buffers menu!


With Emacs life does not get boring.

--
Greetings

  Pete

Got Mole problems?
Call Avogadro 6.02 x 10^23