Re: local chars displayed as numbers

[Reiner Steib]

On Wed, Sep 20 2006, Andreas Roehler wrote:

> Also it seems to affect just one file in this
> directory.
>
> What makes it difficult to reproduce: it happens only
> just after opening. After a save, it reopens correct.
>
> Wll send you an example off-list.

I think I can reproduce your problem with just line 3891 of this file.

FYI: I tried to isolate the problematic characters by dividing the
file in two parts with approximately equal number of lines [1]
("Intervallhalbierungsverfahren" in German, I forgot the English
term).

You can even replace all characters by x in that line:

$ a=3891;e=3891; sed -ne "${a},${e}p" < roe.txt | tr '[a-z]' 'x' > problem.txt

Here is it (\202 is one char, of course; see attached problem.txt.gz):

,----[ problem.txt ]
| Nxxx, xxx xxx xx \202xx Mxxxx' xxxxxx xxxxx, xxx xxx xxxxx Dxxxxxx xx
`----

The character \202 is from windows-1252:

,----
| ‚     U+201A : SINGLE LOW-9 QUOTATION MARK
`----

I can display it correctly using
  `C-x C-m c windows-1252 RET C-x C-f problem.txt RET'
but *not* with
  `C-x C-f problem.txt RET C-x C-m r windows-1252 RET'.

I think this is a bug, but surely Handa-san give you more accurate
information.

WRT auto-detection: Wouldn't it be possible to give windows-1252 a
higher priority than raw-text (in the default settings!)?

BTW, the sample file displays correctly if I only delete line 3891,
i.e. the only problem is \202.

$ a=3891;e=3891; sed -e "${a},${e}d" < roe.txt > skip-line-$a-$e.txt

Bye, Reiner.

[1] I used sed for this:

$ wc -l roe.txt 
6649 roe.txt
$ a=1;e=3000; sed -ne "${a},${e}p" < roe.txt > line-$a-$e.txt

[ Open in emacs, see if it fails => no => problem must be in the next part ]

$ a=3000;e=6649; sed -ne "${a},${e}p" < roe.txt > line-$a-$e.txt

[ Open in emacs, see if it fails => yes => problem must be in this
  part; Divide this part ...]

$ a=3000;e=5500; sed -ne "${a},${e}p" < roe.txt > line-$a-$e.txt

[...] Until you reach...

$ a=3891;e=3891; sed -ne "${a},${e}p" < roe.txt > line-$a-$e.txt
-- 
       ,,,
      (o o)
---ooO-(_)-Ooo---  |  PGP key available  |  http://rsteib.home.pages.de/

problem.txt.gz
Description: GNU Zip compressed data