Re: 23.0.60; A combination of PC Selection, Visual Lines and emacsclient redefines C-backspace

Re: 23.0.60; A combination of PC Selection, Visual Lines and emacsclient redefines C-backspace

[Chong Yidong]

Simplified recipe for bug#1960:

1. Put (pc-selection-mode 1) in .emacs
2. emacs --daemon
3. emacsclient -c
4. C-h k C-backspace

<C-backspace> runs the command kill-line, which is an interactive
compiled Lisp function in `simple.el'.

5. C-x C-c
6. emacsclient -c
7. C-h k C-backspace

<C-backspace> runs the command backward-kill-word, which is an
interactive compiled Lisp function in `simple.el'.


PC-selection mode binds [C-delete] to kill-line.  My guess is that some
code for distinguishing delete from DEL isn't run when Emacs is started
as a daemon.  It's only run sometime after the first client frame is
created, which is why C-backspace works properly for the second client
frame.

Could someone take a closer look?

bug#1960: 23.0.60; A combination of PC Selection, Visual Lines and emacsclient redefines C-backspace

[Dan Nicolaescu]

Chong Yidong <cyd@stupidchicken.com> writes:

  > Simplified recipe for bug#1960:
  > 
  > 1. Put (pc-selection-mode 1) in .emacs
  > 2. emacs --daemon
  > 3. emacsclient -c
  > 4. C-h k C-backspace
  > 
  > <C-backspace> runs the command kill-line, which is an interactive
  > compiled Lisp function in `simple.el'.
  > 
  > 5. C-x C-c
  > 6. emacsclient -c
  > 7. C-h k C-backspace
  > 
  > <C-backspace> runs the command backward-kill-word, which is an
  > interactive compiled Lisp function in `simple.el'.

What happens when repeating steps 5,6 and 7? 
I am guessing that the <C-backspace> binding would toggle between
kill-line and backward-kill-word.  
If that's indeed the case, then there's probably some problem with
pc-selection-mode and the way it interacts with
normal-erase-is-backspace.

Re: 23.0.60; A combination of PC Selection, Visual Lines and emacsclient redefines C-backspace

[Dan Nicolaescu]

Chong Yidong <cyd@stupidchicken.com> writes:

  > Simplified recipe for bug#1960:
  > 
  > 1. Put (pc-selection-mode 1) in .emacs
  > 2. emacs --daemon
  > 3. emacsclient -c
  > 4. C-h k C-backspace
  > 
  > <C-backspace> runs the command kill-line, which is an interactive
  > compiled Lisp function in `simple.el'.
  > 
  > 5. C-x C-c
  > 6. emacsclient -c
  > 7. C-h k C-backspace
  > 
  > <C-backspace> runs the command backward-kill-word, which is an
  > interactive compiled Lisp function in `simple.el'.

What happens when repeating steps 5,6 and 7? 
I am guessing that the <C-backspace> binding would toggle between
kill-line and backward-kill-word.  
If that's indeed the case, then there's probably some problem with
pc-selection-mode and the way it interacts with
normal-erase-is-backspace.

bug#1960: 23.0.60; A combination of PC Selection, Visual Lines and emacsclient redefines C-backspace

[Chong Yidong]

Dan Nicolaescu <dann@tensilica.com> writes:

> there's probably some problem with pc-selection-mode and the way it
> interacts with normal-erase-is-backspace.

The problem is not pc-selection mode, but how Emacs does
normal-erase-is-backspace in multi-tty situations.  Here's an even
simpler recipe, not involving pc-selection-mode or the daemon:

1. Add the following line to .emacs:

     (global-set-key [C-delete] 'kill-line)

2. emacs -nw -f server-start
3. emacsclient -c
4. C-h k C-backspace

  => <C-backspace> runs the command kill-line

Re: 23.0.60; A combination of PC Selection, Visual Lines and emacsclient redefines C-backspace

[Chong Yidong]

Dan Nicolaescu <dann@tensilica.com> writes:

> there's probably some problem with pc-selection-mode and the way it
> interacts with normal-erase-is-backspace.

The problem is not pc-selection mode, but how Emacs does
normal-erase-is-backspace in multi-tty situations.  Here's an even
simpler recipe, not involving pc-selection-mode or the daemon:

1. Add the following line to .emacs:

     (global-set-key [C-delete] 'kill-line)

2. emacs -nw -f server-start
3. emacsclient -c
4. C-h k C-backspace

  => <C-backspace> runs the command kill-line

bug#1960: 23.0.60; A combination of PC Selection, Visual Lines and emacsclient redefines C-backspace

[Dan Nicolaescu]

  > > there's probably some problem with pc-selection-mode and the way it
  > > interacts with normal-erase-is-backspace.
  > 
  > The problem is not pc-selection mode, but how Emacs does
  > normal-erase-is-backspace in multi-tty situations.  Here's an even
  > simpler recipe, not involving pc-selection-mode or the daemon:
  > 
  > 1. Add the following line to .emacs:
  > 
  >      (global-set-key [C-delete] 'kill-line)
  > 
  > 2. emacs -nw -f server-start
  > 3. emacsclient -c
  > 4. C-h k C-backspace
  > 
  >   => <C-backspace> runs the command kill-line


This happens because:

(defun normal-erase-is-backspace-setup-frame (&optional frame)
  "Set up `normal-erase-is-backspace-mode' on FRAME, if necessary."
  (unless frame (setq frame (selected-frame)))
  (unless (terminal-parameter nil 'normal-erase-is-backspace)
    (normal-erase-is-backspace-mode
      (if (if (eq normal-erase-is-backspace 'maybe)

When using "emacs -nw -f server-start" the "(unless (terminal-parameter"
test passes and `normal-erase-is-backspace' is run.

When using "emacs -f server-start" the `terminal-parameter' call returns
non-ni, so `normal-erase-is-backspace' is not run.  
It seems that terminal-parameters are copied from the first frame
created...

Not sure how normal-erase-is-backspace is supposed to work... 

Re: 23.0.60; A combination of PC Selection, Visual Lines and emacsclient redefines C-backspace

[Dan Nicolaescu]

  > > there's probably some problem with pc-selection-mode and the way it
  > > interacts with normal-erase-is-backspace.
  > 
  > The problem is not pc-selection mode, but how Emacs does
  > normal-erase-is-backspace in multi-tty situations.  Here's an even
  > simpler recipe, not involving pc-selection-mode or the daemon:
  > 
  > 1. Add the following line to .emacs:
  > 
  >      (global-set-key [C-delete] 'kill-line)
  > 
  > 2. emacs -nw -f server-start
  > 3. emacsclient -c
  > 4. C-h k C-backspace
  > 
  >   => <C-backspace> runs the command kill-line


This happens because:

(defun normal-erase-is-backspace-setup-frame (&optional frame)
  "Set up `normal-erase-is-backspace-mode' on FRAME, if necessary."
  (unless frame (setq frame (selected-frame)))
  (unless (terminal-parameter nil 'normal-erase-is-backspace)
    (normal-erase-is-backspace-mode
      (if (if (eq normal-erase-is-backspace 'maybe)

When using "emacs -nw -f server-start" the "(unless (terminal-parameter"
test passes and `normal-erase-is-backspace' is run.

When using "emacs -f server-start" the `terminal-parameter' call returns
non-ni, so `normal-erase-is-backspace' is not run.  
It seems that terminal-parameters are copied from the first frame
created...

Not sure how normal-erase-is-backspace is supposed to work...