emacs clipboard handling for X broken?

emacs clipboard handling for X broken?

[Werner LEMBERG]

[Emacs CVS 2004-08-22]

I tried to paste text from Mozilla 1.6 (build 2004011400) into Emacs,
and I get e.g. for Chinese instead of `不想睡' the following:

  \x{4E0D}\x{60F3}\x{7761}

I tried many coding systems in Emacs for the X selection (C-x RET x),
but all give the same result -- perhaps I've missed the right
one...  To be sure that it isn't a Mozilla problem or a problem of the
Window manager (I use fvwm 2.5.9), I pasted the text into the Unicode
editor yudit (version 2.7.6), and it worked.

Interestingly, pasting Chinese from Emacs to Mozilla (and to yudit)
works fine.

Another observation is that the mozilla's window border also show
\x{...} instead of Chinese characters...

Is this a known problem?  I'm lost and ask for help.


    Werner

Re: emacs clipboard handling for X broken?

[Kenichi Handa]

In article <20040831.164842.180119251.wl@gnu.org>, Werner LEMBERG <wl@gnu.org> writes:
> [Emacs CVS 2004-08-22]

> I tried to paste text from Mozilla 1.6 (build 2004011400) into Emacs,
> and I get e.g. for Chinese instead of `不想睡' the following:

>   \x{4E0D}\x{60F3}\x{7761}

I can't reproduce it, but anyway, as Emacs itself never
generate such a string, it was what Mozilla sent to Emacs.

By default, Emacs issues two requests of types COMPOUND_TEXT
and UTF8_STRING and chooses the longer one.

Please try to set x-select-request-type to a symbol
UTF8_STRING or COMPOUND_TEXT.

---
Ken'ichi HANDA
handa@m17n.org

Re: emacs clipboard handling for X broken?

[Werner LEMBERG]

> > I tried to paste text from Mozilla 1.6 (build 2004011400) into
> > Emacs, and I get e.g. for Chinese instead of `不想睡' the
> > following:
>
> >   \x{4E0D}\x{60F3}\x{7761}
>
> [...]
>
> Please try to set x-select-request-type to a symbol
> UTF8_STRING or COMPOUND_TEXT.

This did the trick, thanks!  I've added it to my .emacs file.  Your
solution is apparently independent on Mozilla's locale (I've started
it with `en_US.ISO-8859-1' ).


    Werner

Re: emacs clipboard handling for X broken?

[Werner LEMBERG]

> > Please try to set x-select-request-type to a symbol
> > UTF8_STRING or COMPOUND_TEXT.
> 
> This did the trick, thanks!

Using `UTF8_STRING', that is.


    Werner

Re: emacs clipboard handling for X broken?

[Kenichi Handa]

In article <20040902.071819.174335908.wl@gnu.org>, Werner LEMBERG <wl@gnu.org> writes:

>>  > Please try to set x-select-request-type to a symbol
>>  > UTF8_STRING or COMPOUND_TEXT.
>>  
>>  This did the trick, thanks!

> Using `UTF8_STRING', that is.

I see.  So, your mozilla produces bogus sequence upon a
request of COMPOUND_TEXT.  Have you tried a newer mozilla?  

---
Ken'ichi HANDA
handa@m17n.org

Re: emacs clipboard handling for X broken?

[Alex Schroeder]

Kenichi Handa <handa@m17n.org> writes:

> In article <20040902.071819.174335908.wl@gnu.org>, Werner LEMBERG <wl@gnu.org> writes:
>
>>>  > Please try to set x-select-request-type to a symbol
>>>  > UTF8_STRING or COMPOUND_TEXT.
>>>  
> I see.  So, your mozilla produces bogus sequence upon a
> request of COMPOUND_TEXT.  Have you tried a newer mozilla?  

I'm using Firefox 0.9 and pasting characters that can be encoded as
Latin-1 works with setting x-select-request-type to either value.  The
default is nil, and that works as well.

When trying to paste chinese characters from an UTF-8 encoded page,
I get the following results:

(setq x-select-request-type 'UTF8_STRING)
 -> works.

(setq x-select-request-type 'COMPOUND_TEXT)
 -> pastes entry from kill-ring

(setq x-select-request-type nil)
 -> pastes unicode escapes like this: \u5167\u5bb9\u76ee\u9304\u64f4\u5145

This reminds me of the native format used by Java properties files:
They are Latin-1 encoded, and use unicode escapes for everything
else.  Very weird...   :(

My Emacs from CVS was built 2004-08-12.

Alex.
-- 
.O.  http://www.emacswiki.org/alex/
..O  Schroeder's fifth law:
OOO  Never accept more work than you can handle in one night of hacking.

Re: emacs clipboard handling for X broken?

[Miles Bader]

Werner LEMBERG <wl@gnu.org> writes:
> I tried to paste text from Mozilla 1.6 (build 2004011400) into Emacs,
> and I get e.g. for Chinese instead of `不想睡' the following:
>
>   \x{4E0D}\x{60F3}\x{7761}
...
> Is this a known problem?  I'm lost and ask for help.

FWIW, I can cut and paste japanese just fine in both directions between
Emacs (started with `-q', so the defaults seem OK) and mozilla-firefox
0.9.3 (I think this roughly corresponds to mozilla 1.7.2).

selection-coding-system's value is `compound-text-with-extensions'.

LANG is set to `ja_JP.UTF-8'.

I've noticed in the past that the setting of LANG is crucial to making
Mozilla do CJK cut-n-paste correctly, and that in general Mozilla is
pretty fucked up in this area -- if LANG is _not_ set `properly', then
it will screw up cut-n-paste.

Here's Mozilla bug report that may be related:

   http://bugzilla.mozilla.org/show_bug.cgi?id=137659

-Miles
-- 
Occam's razor split hairs so well, I bought the whole argument!

Re: emacs clipboard handling for X broken?

[Janusz S. Bień]

On Wed, 1 Sep 2004  Kenichi Handa <handa@m17n.org> wrote:

> In article <20040831.164842.180119251.wl@gnu.org>, Werner LEMBERG <wl@gnu.org> writes:
> > [Emacs CVS 2004-08-22]
> 
> > I tried to paste text from Mozilla 1.6 (build 2004011400) into Emacs,
> > and I get e.g. for Chinese instead of `不想睡' the following:
> 
> >   \x{4E0D}\x{60F3}\x{7761}
> 
> I can't reproduce it, but anyway, as Emacs itself never
> generate such a string, it was what Mozilla sent to Emacs.

I'm afraid the problem is more general. I am almost sure I was getting such
string when pasting from kanjipad some time ago. If I remember well,
the problem has gone with a newer version of Emacs and/or kanjipad.

On Sun, 05 Sep 2004  Alex Schroeder <alex@emacswiki.org> wrote:

[...]

>  -> pastes unicode escapes like this: \u5167\u5bb9\u76ee\u9304\u64f4\u5145

I can get these escape sequences when pasting Polish letters from
Galeon to Emacs 21.2.1 and 21.3.1 while with CVS Emacs both Polish
letters (and kanji) are pasted correctly. See Debian bug

        http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=258114

BTW, when pasting from konqueror with ctext I get just question marks
in place of non-ASCII characters.

Regards

Janusz

-- 
                     ,   
dr hab. Janusz S. Bien, prof. UW -  Uniwersytet Warszawski (Katedra Lingwistyki Formalnej)
Prof. Janusz S. Bien - Warsaw Uniwersity (Chair of Formal Linguistics)
jsbien@mimuw.edu.pl, jsbien@uw.edu.pl, http://www.mimuw.edu.pl/~jsbien/, http://www.klf.uw.edu.pl