bug#5541: 23.1; after upgrading to emacs-23, meta key in xterm no longer works

bug#5541: 23.1; after upgrading to emacs-23, meta key in xterm no longer works

[Jim Paris]

With "emacs23 -nw" running inside xterm-253, my meta key does not work
for things like M-x.  Instead, the terminal bell beeps, and the string
;120~ appears in the buffer.  Looking at the log below, this is
because the terminal actually sent the string \e[27;3;120~ which is
occurring because emacs turned on the "modifyOtherKeys" setting with
the escape sequence \e[>4;1m at startup.  Older versions of emacs did
not do this, and everything worked fine.

Indeed, if I execute the command:
   (xterm-turn-off-modify-other-keys (selected-frame))
then my meta key temporarily starts working again.

My normal Xterm settings are
  xterm*metaSendsEscape: true
  xterm*eightBitInput: false
which has always worked in previous versions.

This new "modifyOtherKeys" stuff causes plenty of other problems too,
like the string "0;235;0c" (xterm identification response?) sometimes
appearing in my buffer at startup if the system is heavily loaded or
I'm on a slow connection, and the input buffer getting flushed at
startup, so I lose stuff I typed while emacs was loading.  Could a
knob be added to just get rid of it completely?

-jim


In GNU Emacs 23.1.1 (x86_64-pc-linux-gnu, GTK+ Version 2.18.3)
 of 2009-11-01 on excelsior, modified by Debian
configured using `configure  '--build=x86_64-linux-gnu' '--host=x86_64-linux-gnu' '--prefix=/usr' '--sharedstatedir=/var/lib' '--libexecdir=/usr/lib' '--localstatedir=/var/lib' '--infodir=/usr/share/info' '--mandir=/usr/share/man' '--with-pop=yes' '--enable-locallisppath=/etc/emacs23:/etc/emacs:/usr/local/share/emacs/23.1/site-lisp:/usr/local/share/emacs/site-lisp:/usr/share/emacs/23.1/site-lisp:/usr/share/emacs/site-lisp:/usr/share/emacs/23.1/leim' '--with-x=yes' '--with-x-toolkit=gtk' '--with-toolkit-scroll-bars' 'build_alias=x86_64-linux-gnu' 'host_alias=x86_64-linux-gnu' 'CFLAGS=-DDEBIAN -g -O2' 'LDFLAGS=-g' 'CPPFLAGS=''

Important settings:
  value of $LC_ALL: nil
  value of $LC_COLLATE: nil
  value of $LC_CTYPE: nil
  value of $LC_MESSAGES: nil
  value of $LC_MONETARY: nil
  value of $LC_NUMERIC: nil
  value of $LC_TIME: nil
  value of $LANG: en_US.UTF-8
  value of $XMODIFIERS: nil
  locale-coding-system: utf-8-unix
  default-enable-multibyte-characters: t

Major mode: Text

Minor modes in effect:
  tooltip-mode: t
  tool-bar-mode: t
  file-name-shadow-mode: t
  global-font-lock-mode: t
  font-lock-mode: t
  global-auto-composition-mode: t
  auto-composition-mode: t
  auto-encryption-mode: t
  auto-compression-mode: t
  column-number-mode: t
  line-number-mode: t
  transient-mark-mode: t

Recent input:
ESC [ > 0 ; 2 3 5 ; 0 c a s d f RET ( x t e r m - t 
u r n - o f f - m o d i f y - o t h e r - k e y s SPC 
( s e l e c t e d - f r a m e ) ) RET C-x C-s C-g ESC 
[ 2 7 ; 3 ; 1 2 0 ~ C-p C-e C-n C-p C-x C-e C-n ESC 
x C-g RET RET ESC x r e p o r t - b u g RET C-g ESC 
x r e p o t - e m DEL DEL DEL DEL r t - e m a TAB 
RET

Recent messages:
Loading /etc/emacs/site-start.d/50psvn.el (source)...done
Loading /etc/emacs/site-start.d/51debian-el.el (source)...done
Loading /etc/emacs/site-start.d/70jim.el (source)...
Toggling menu-bar-mode off; better pass an explicit argument.
Ready.
Loading /etc/emacs/site-start.d/70jim.el (source)...done
Loading quail/latin-ltx...done
Ready.
Quit
nil
Quit [2 times]

bug#5541: 23.1; after upgrading to emacs-23, meta key in xterm no longer works

[Chong Yidong]

Does anyone on this list have xterm-253 available to test?  I can't
reproduce the problem on xterm 241, and don't have xterm-253 available
to test at the moment.

> With "emacs23 -nw" running inside xterm-253, my meta key does not work
> for things like M-x.  Instead, the terminal bell beeps, and the string
> ;120~ appears in the buffer.  Looking at the log below, this is
> because the terminal actually sent the string \e[27;3;120~ which is
> occurring because emacs turned on the "modifyOtherKeys" setting with
> the escape sequence \e[>4;1m at startup.  Older versions of emacs did
> not do this, and everything worked fine.
>
> Indeed, if I execute the command:
>    (xterm-turn-off-modify-other-keys (selected-frame))
> then my meta key temporarily starts working again.
>
> My normal Xterm settings are
>   xterm*metaSendsEscape: true
>   xterm*eightBitInput: false
> which has always worked in previous versions.

bug#5541: 23.1; after upgrading to emacs-23, meta key in xterm no longer works

[Sven Joachim]

On 2010-03-13 19:42 +0100, Chong Yidong wrote:

> Does anyone on this list have xterm-253 available to test?  I can't
> reproduce the problem on xterm 241, and don't have xterm-253 available
> to test at the moment.

With xterm 255 I cannot reproduce the problem either (xterm 253 is not
available here).

Sven

>> With "emacs23 -nw" running inside xterm-253, my meta key does not work
>> for things like M-x.  Instead, the terminal bell beeps, and the string
>> ;120~ appears in the buffer.  Looking at the log below, this is
>> because the terminal actually sent the string \e[27;3;120~ which is
>> occurring because emacs turned on the "modifyOtherKeys" setting with
>> the escape sequence \e[>4;1m at startup.  Older versions of emacs did
>> not do this, and everything worked fine.
>>
>> Indeed, if I execute the command:
>>    (xterm-turn-off-modify-other-keys (selected-frame))
>> then my meta key temporarily starts working again.
>>
>> My normal Xterm settings are
>>   xterm*metaSendsEscape: true
>>   xterm*eightBitInput: false
>> which has always worked in previous versions.

bug#5541: 23.1; after upgrading to emacs-23, meta key in xterm no longer works

[Glenn Morris]

Chong Yidong wrote:

> Does anyone on this list have xterm-253 available to test?

Yes.

>> With "emacs23 -nw" running inside xterm-253, my meta key does not work
>> for things like M-x.  Instead, the terminal bell beeps, and the string
>> ;120~ appears in the buffer.

Something similar happens for me too (and with the trunk).

>> Indeed, if I execute the command:
>>    (xterm-turn-off-modify-other-keys (selected-frame))
>> then my meta key temporarily starts working again.

This command makes no difference for me.

I normally use aterm, which works fine.

bug#5541: 23.1; after upgrading to emacs-23, meta key in xterm no longer works

[Chong Yidong]

Glenn Morris <rgm@gnu.org> writes:

>> Does anyone on this list have xterm-253 available to test?
>
> Yes.
>
>>> With "emacs23 -nw" running inside xterm-253, my meta key does not work
>>> for things like M-x.  Instead, the terminal bell beeps, and the string
>>> ;120~ appears in the buffer.
>
> Something similar happens for me too (and with the trunk).

And does xterm-24* behave OK?

It's possible this is a bug in xterm-253, which would be why Sven
Joachim does not see it on xterm-255.

bug#5541: 23.1; after upgrading to emacs-23, meta key in xterm no longer works

[Jim Paris]

Chong Yidong wrote:
> Glenn Morris <rgm@gnu.org> writes:
> 
> >> Does anyone on this list have xterm-253 available to test?
> >
> > Yes.
> >
> >>> With "emacs23 -nw" running inside xterm-253, my meta key does not work
> >>> for things like M-x.  Instead, the terminal bell beeps, and the string
> >>> ;120~ appears in the buffer.
> >
> > Something similar happens for me too (and with the trunk).
> 
> And does xterm-24* behave OK?
> 
> It's possible this is a bug in xterm-253, which would be why Sven
> Joachim does not see it on xterm-255.

I did some more testing and it only happens on some of my systems.
Turns out the problem shows up when this .Xresources setting is
present:

! Shift-paste to paste CLIPBOARD instead of PRIMARY.
xterm*VT100.Translations: #override \
	Shift <Btn2Up>: insert-selection(CLIPBOARD)\n \
	Ctrl ~Shift Meta <Key>-: smaller-vt-font() \n \
	Ctrl ~Shift Meta <Key>=: larger-vt-font() \n

That's worked for me forever... I have no idea why it would cause new
versions of emacs to act funny.  CCing Thomas Dickey, any ideas?
The original report is at http://debbugs.gnu.org/cgi/bugreport.cgi?bug=5541

-jim

bug#5541: 23.1; after upgrading to emacs-23, meta key in xterm no longer works

[Thomas Dickey]

On Sat, 13 Mar 2010, Jim Paris wrote:

> Chong Yidong wrote:
>> Glenn Morris <rgm@gnu.org> writes:
>>
>>>> Does anyone on this list have xterm-253 available to test?
>>>
>>> Yes.
>>>
>>>>> With "emacs23 -nw" running inside xterm-253, my meta key does not work
>>>>> for things like M-x.  Instead, the terminal bell beeps, and the string
>>>>> ;120~ appears in the buffer.
>>>
>>> Something similar happens for me too (and with the trunk).
>>
>> And does xterm-24* behave OK?
>>
>> It's possible this is a bug in xterm-253, which would be why Sven
>> Joachim does not see it on xterm-255.
>
> I did some more testing and it only happens on some of my systems.
> Turns out the problem shows up when this .Xresources setting is
> present:
>
> ! Shift-paste to paste CLIPBOARD instead of PRIMARY.
> xterm*VT100.Translations: #override \
> 	Shift <Btn2Up>: insert-selection(CLIPBOARD)\n \
> 	Ctrl ~Shift Meta <Key>-: smaller-vt-font() \n \
> 	Ctrl ~Shift Meta <Key>=: larger-vt-font() \n
>
> That's worked for me forever... I have no idea why it would cause new
> versions of emacs to act funny.  CCing Thomas Dickey, any ideas?
> The original report is at http://debbugs.gnu.org/cgi/bugreport.cgi?bug=5541

The various key-modifier resources are all ifdef'd as a single group
(no escape sequence is implemented to disable the feature).  I suppose
emacs is checking the version number from xterm to decide if the feature
is present, since there's no other indication that it's available.

xterm generally doesn't know what the translations do, but it does check 
if the meta key appears in a translation.  For that, and alt, xterm 
decides to not use those in modified function-keys - to try to avoid odd 
conflicts with cases such as this translation.  (They may be the same key 
- same general effect).  But the particular combination caught by the 
translations wouldn't be seen by xterm; it only massages the seen 
keycodes.

The escape sequence says xterm's sending a 3 (for alt).  Perhaps it's 
something along the lines of xterm deciding that meta isn't part of the 
modify-other-keys feature, but alt (on the same key) still is.  I assume 
that emacs is looking for 9 (meta) rather than 3 (alt), and so it errors 
out.

-- 
Thomas E. Dickey
http://invisible-island.net
ftp://invisible-island.net

bug#5541: 23.1; after upgrading to emacs-23, meta key in xterm no longer works

[Glenn Morris]

Chong Yidong wrote:

> And does xterm-24* behave OK?

I don't have that version. My system just updated to xterm-255, and it
behaves the same as 253 did. Running `xrdb -remove' first makes no
difference.

bug#5541: 23.1; after upgrading to emacs-23, meta key in xterm no longer works

[Chong Yidong]

Thomas Dickey <dickey@his.com> writes:

>> ! Shift-paste to paste CLIPBOARD instead of PRIMARY.
>> xterm*VT100.Translations: #override \
>> 	Shift <Btn2Up>: insert-selection(CLIPBOARD)\n \
>> 	Ctrl ~Shift Meta <Key>-: smaller-vt-font() \n \
>> 	Ctrl ~Shift Meta <Key>=: larger-vt-font() \n
>
> The various key-modifier resources are all ifdef'd as a single group
> (no escape sequence is implemented to disable the feature).  I suppose
> emacs is checking the version number from xterm to decide if the feature
> is present, since there's no other indication that it's available.
>
> xterm generally doesn't know what the translations do, but it does
> check if the meta key appears in a translation.  For that, and alt,
> xterm decides to not use those in modified function-keys - to try to
> avoid odd conflicts with cases such as this translation.  (They may be
> the same key - same general effect).  But the particular combination
> caught by the translations wouldn't be seen by xterm; it only massages
> the seen keycodes.
>
> The escape sequence says xterm's sending a 3 (for alt).  Perhaps it's
> something along the lines of xterm deciding that meta isn't part of
> the modify-other-keys feature, but alt (on the same key) still is.  I
> assume that emacs is looking for 9 (meta) rather than 3 (alt), and so
> it errors out.

Thanks for the explanation.  I don't see, from this, what we can do to
detect and fix this problem automatically, so I suppose it will be up to
the user who has defined such translations to tell Emacs not to use
modify-other-keys.  I have added an entry to etc/PROBLEMS in the Emacs
source tree documenting this.