Copy-paste between pretest Emacs 22 and other Linux programs

Copy-paste between pretest Emacs 22 and other Linux programs

[Ryszard Kubiak]

Hello Everybody,

I would like to report that copying a text with Polish national
letters from an X-based program (like: xterm, gnome terminal,
kde terminal, thudirbird, firefox) and pasting it into Emacs
does not work properly within an environment which is utf-8,
both in the operating system and in Emacs.

If you can see the following eighteen Polish letters

ĄĆŹŻŚŁÓŃĘąźśżćęńłó

properly in your e-mail client, which you should if you have set
utf-8 as your basic encoding, then could you copy them into Emacs,
also configured to work with utf-8? Do you see them right? If so,
I would be happy to know all of your settings. My related .emacs
settings are quoted below.

Currently, I am running a pretest Emacs version 22.00.92.1, installed
from a tarball. Previously, I had version 22.00.52.1, compiled from
CVS sources. I am running Emacs under Linux Fedora 5. In both these
versions the Polish characters get mangled when they are copied into
Emacs. Interestingly, copying Polish characters from Emacs to other
X-based utilities works fine. Also, I do not experience problems with
editing Polish texts, whether I am using iso-latin-2 or utf-8 encoding.

As the problem is somewhat technical, can anybody who knows the 
respective fragments of Emacs write my in private, to decide
whether it's really a bug in Emacs or just a matter of wrong
settings within my .emacs. Of course, I may supply any necessary
information about my installation.

Since unicode is now the core issue of Emacs develpmpent, then if
what I have described is result of a bug then it has to be removed.

Cheers,
Rysiek

(codepage-setup 1250)
(set-language-info-alist "pl"
     '((documentation . t)
     (sample-text . "Pójdź, kiń-że tę chmurność w głąb flaszy")
     (tutorial . "TUTORIAL.pl")
     (unibyte-display . iso-8859-2)
     (unibyte-syntax . "latin-2")
     (nonascii-translation . latin-iso8859-2)
     (input-method . "polish-slash")
     (coding-priority  utf-8  iso-8859-2 cp1250-dos)
     (coding-system  utf-8  iso-8859-2 cp1250-dos )
     (charset ascii latin-iso8859-2))
(set-language-environment "pl")
(set-selection-coding-system 'utf-8)
(set-keyboard-coding-system 'utf-8)

The font I am using is:
-misc-fixed-medium-r-normal-*-15-*-*-*-*-*-iso10646-1

I am also using cua-mode, with this customization setting:

  '(cua-mode t nil (cua-base))

Interestingly, when I copy the Polish letters from another
X-based program into Emacs, the following messages are generated:

Loading subst-ksc...done
Loading subst-gb2312...done
Loading subst-big5...done
Loading subst-jis...done

Do these have anything to do with East European languages?

Re: Copy-paste between pretest Emacs 22 and other Linux programs

[Gernot Hassenpflug]

>>>>> "Ryszard" == Ryszard Kubiak <r.kubiak@guests.ipipan.gda.pl> writes:

    Ryszard> Hello Everybody,

    Ryszard> I would like to report that copying a text with Polish
    Ryszard> national letters from an X-based program (like: xterm,
    Ryszard> gnome terminal, kde terminal, thudirbird, firefox) and
    Ryszard> pasting it into Emacs does not work properly within an
    Ryszard> environment which is utf-8, both in the operating system
    Ryszard> and in Emacs.

    Ryszard> If you can see the following eighteen Polish letters

    Ryszard> ĄĆŹŻŚŁÓŃĘąźśżćęńłó

Hello Ryszard,

I see the above characters perfectly in Emacs, but I cannot paste them
into a terminal (GNOME terminal, mlterm or other) even when using
utf-8-unix as the clipboard/X-selection encoding, as at least some
seem to be mistaken for control characters of some kind. I am
not sure what coding Gnus is using to display the email for me.

    Ryszard> Currently, I am running a pretest Emacs version
    Ryszard> 22.00.92.1, installed from a tarball. Previously, I had
    Ryszard> version 22.00.52.1, compiled from CVS sources. I am
    Ryszard> running Emacs under Linux Fedora 5. In both these
    Ryszard> versions the Polish characters get mangled when they are
    Ryszard> copied into Emacs. Interestingly, copying Polish
    Ryszard> characters from Emacs to other X-based utilities works
    Ryszard> fine. Also, I do not experience problems with editing
    Ryszard> Polish texts, whether I am using iso-latin-2 or utf-8
    Ryszard> encoding.

/../

Sorry, I cannot help here.

    Ryszard> Interestingly, when I copy the Polish letters from
    Ryszard> another X-based program into Emacs, the following
    Ryszard> messages are generated:

    Ryszard> Loading subst-ksc...done Loading subst-gb2312...done
    Ryszard> Loading subst-big5...done Loading subst-jis...done

    Ryszard> Do these have anything to do with East European
    Ryszard> languages?

No, these are Korean (Kangul), Chinese (Big5) and Japanese (JIS)
encoding systems. If you go to Options/Mule/Show all of Mule Status
you will get a list of the various installed character sets, their
shortlisted names and examples of text.

HTH,
        Gernot
-- 
G Hassenpflug 

Re: Copy-paste between pretest Emacs 22 and other Linux programs

[Ryszard Kubiak]

Gernot Hassenpflug writes:

> I see the above characters perfectly in Emacs, but I cannot paste them
> into a terminal (GNOME terminal, mlterm or other) even when using
> utf-8-unix as the clipboard/X-selection encoding, as at least some
> seem to be mistaken for control characters of some kind. I am
> not sure what coding Gnus is using to display the email for me.

I vaguely remember having similar problems in one of the previous
versions of Emacs. As I wrote in my original letter, with the current 
one, pretest version 22, I have problems with pasting into Emacs,
not the other way around. Kenichi suggested that I may circumvent
them by removing (set-selection-coding-system 'utf-8) from my .emacs
file. This solves the problem indeed, to some extend though, as the
Polish letter lslash remains as one that is not copied properly.
I believe that Kenichi is capable of solving all these problems soon.

Greetings,
Rysiek