problems displaying utf-8-characters after reload

[LanX]

Hi

Im writing perl code with POD dokumentation with german text.

typing umlauts is no problem, inspecting the letters in hexl-mode show perfect encoding, saving is no problem but after reloading I don't see the umlauts anymore but something like \301\156 (not the exact numbers, the file is at work). But checking in hexl-mode shows again that the coding is right but the display isn't.

Now after typing new umlauts I have no problem, but after saving and loading the problem starts again.

I was able to reproduce the problem with that special file using emacs -Q.

describe-coding-system shows utf-8-unix

Any ideas what to do?

cheers
Rolf

[Peter Dyballa]

Am 24.09.2012 um 21:23 schrieb LanX:

> Any ideas what to do?

Report that bug! (In the Help menu you can find the "Send Bug Report…" entry.) From time to time I can see similar failures.

--
Mit friedvollen Grüßen

Pete

Upgraded, adj.:
Didn't work the first time.

[Eli Zaretskii]

> Date: Mon, 24 Sep 2012 12:23:41 -0700 (PDT)
> From: LanX <lanx...@googlemail.com>

>
> I was able to reproduce the problem with that special file using emacs -Q.
>
> describe-coding-system shows utf-8-unix
>
> Any ideas what to do?

"C-x RET c latin-1 RET C-x C-f"

Or if want Emacs to do this automatically, put a 'coding:' cookie in
the file.

[B. T. Raven]

Die Mon Sep 24 2012 16:02:15 GMT-0500 (Central Daylight Time) Eli
Zaretskii <el...@gnu.org> scripsit:

This has always worked for me. I don't think there is a bug but there is
a residual mystery:

Why aren't the following exactly synonymous (as the descriptions would
lead us to believe):

C-x RET c latin-1 RET C-x C-f

C-x RET f latin-1 RET

I've found that the second one doesn't always work.

Btw, the coding cookie is:

;; -*- coding: utf-8 -*-

in first line or first eight lines of file, I can't remember which.

Ed

[Eli Zaretskii]

> From: "B. T. Raven" <btr...@nihilo.net>
> Date: Tue, 02 Oct 2012 18:17:18 -0500

>
> Why aren't the following exactly synonymous (as the descriptions would
> lead us to believe):
>
> C-x RET c latin-1 RET C-x C-f
> C-x RET f latin-1 RET

The second one changes how the buffer will be _encoded_ when saved to
disk next time you type "C-x C-s". The former one changes how the
file will be _decoded_ when you visit it. IOW, the second one can
change nothing in how Emacs interprets the file while reading it into
the buffer.

[Eli Zaretskii]

> Date: Wed, 03 Oct 2012 09:09:56 +0200
> From: Andreas Röhler <andreas...@easy-emacs.de>
>
> a related bug seems visible with emacs -Q Python-shell - png attached.

I'm quite sure it's because Python-shell, whatever that is, sets up an
incorrect coding-systems for decoding output from Python.

[Andreas Röhler]

Am 03.10.2012 19:22, schrieb Eli Zaretskii:
>> Date: Wed, 03 Oct 2012 09:09:56 +0200
>> From: Andreas Röhler <andreas...@easy-emacs.de>
>>
>> a related bug seems visible with emacs -Q Python-shell - png attached.
>
> I'm quite sure it's because Python-shell, whatever that is,

when starting a Python interpreter from python-mode, you get a buffer in comint-mode AFAIU

sets up an
> incorrect coding-systems for decoding output from Python.
>
>
>

filed a bug-report which should give some more info

12...@debbugs.gnu.org

[Andreas Röhler]

Am 03.10.2012 19:22, schrieb Eli Zaretskii:
>> Date: Wed, 03 Oct 2012 09:09:56 +0200
>> From: Andreas Röhler <andreas...@easy-emacs.de>
>>
>> a related bug seems visible with emacs -Q Python-shell - png attached.
>

> I'm quite sure it's because Python-shell, whatever that is, sets up an

> incorrect coding-systems for decoding output from Python.
>
>
>

my mistake: this form needs a Python3, while reported when Python2 was used.

Sorry,

Andreas