bug#3442: 23.0.94; "C-x 5 1" (delete-other-frames) kills daemon

bug#3442: 23.0.94; "C-x 5 1" (delete-other-frames) kills daemon

[Teemu Likonen]

When Emacs daemon is running and a client frame is active, executing
delete-other-frames command (C-x 5 1) kills the daemon and makes the
Emacs session just a normal server session which requires that one frame
exists. To reproduce:

 1. Start Emacs daemon:

        emacs -Q --daemon

 2. Start a client session:

        emacsclient -c

 3. Type "C-x 5 1" and then try to close the frame with "C-x 5 0". The
    following message is displayed in the minibuffer:

        Attempt to delete the sole visible or iconified frame

 4. Forcibly close the frame with "C-x C-c" or window manager's close
    button.

 5. See the system's process list with "ps" command. There is no Emacs
    process anymore.




In GNU Emacs 23.0.94.1 (i686-pc-linux-gnu, GTK+ Version 2.12.12)
 of 2009-06-01 on mithlond
Windowing system distributor `The X.Org Foundation', version 11.0.10402000
configured using `configure  '--prefix=/home/dtw/local''

Important settings:
  value of $LC_ALL: nil
  value of $LC_COLLATE: nil
  value of $LC_CTYPE: nil
  value of $LC_MESSAGES: en_US.UTF-8
  value of $LC_MONETARY: nil
  value of $LC_NUMERIC: nil
  value of $LC_TIME: nil
  value of $LANG: fi_FI.UTF-8
  value of $XMODIFIERS: @im=none
  locale-coding-system: utf-8-unix
  default-enable-multibyte-characters: t

Major mode: Group

Minor modes in effect:
  gnus-undo-mode: t
  ergo-movement-mode: t
  tooltip-mode: t
  mouse-wheel-mode: t
  file-name-shadow-mode: t
  global-font-lock-mode: t
  font-lock-mode: t
  blink-cursor-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

bug#3442: 23.0.94; "C-x 5 1" (delete-other-frames) kills daemon

[Teemu Likonen]

On 2009-06-01 15:36 (+0300), Teemu Likonen wrote:

>  2. Start a client session:
>
>         emacsclient -c

You could also start the client sessions with

    emacsclient -c -n

so that there are no clients waiting and you don't need any force...

>  3. Type "C-x 5 1" and then try to close the frame with "C-x 5 0". The
>     following message is displayed in the minibuffer:
> 
>         Attempt to delete the sole visible or iconified frame
> 
>  4. Forcibly close the frame with "C-x C-c" or window manager's close
>     button.

...to close the frame. The frame just closes normally and after that
there's no daemon nor any Emacs process left in the system.

bug#3442: 23.0.94; "C-x 5 1" (delete-other-frames) kills daemon

[Stefan Monnier]

> When Emacs daemon is running and a client frame is active, executing
> delete-other-frames command (C-x 5 1) kills the daemon and makes the
> Emacs session just a normal server session which requires that one frame
> exists. To reproduce:

Indeed, C-x 5 1 should only delete the frames on the current terminal,
whereas it currently kills *all* the frames on all terminals (including
the special internal terminal used for the daemon).


        Stefan

bug#3442: 23.0.94; "C-x 5 1" (delete-other-frames) kills daemon

[Stefan Monnier]

>> When Emacs daemon is running and a client frame is active, executing
>> delete-other-frames command (C-x 5 1) kills the daemon and makes the
>> Emacs session just a normal server session which requires that one frame
>> exists. To reproduce:

> Indeed, C-x 5 1 should only delete the frames on the current terminal,
> whereas it currently kills *all* the frames on all terminals (including
> the special internal terminal used for the daemon).

I've installed a patch that makes C-x 5 1 only delete frames on the same
terminal, which should fix this issue.


        Stefan

bug#3442: 23.0.94; "C-x 5 1" (delete-other-frames) kills daemon

[Dan Nicolaescu]

Stefan Monnier <monnier@iro.umontreal.ca> writes:

  > >> When Emacs daemon is running and a client frame is active, executing
  > >> delete-other-frames command (C-x 5 1) kills the daemon and makes the
  > >> Emacs session just a normal server session which requires that one frame
  > >> exists. To reproduce:
  > 
  > > Indeed, C-x 5 1 should only delete the frames on the current terminal,
  > > whereas it currently kills *all* the frames on all terminals (including
  > > the special internal terminal used for the daemon).
  > 
  > I've installed a patch that makes C-x 5 1 only delete frames on the same
  > terminal, which should fix this issue.

That's a behavior change, isn't it?
Maybe people use C-x 5 1 to remove the rest of the frames on all
terminals... (I personally have never used C-x 5 1)

If we don't want to change the C-x 5 1 behavior, we can just avoid
deleting the daaemon special frame, i.e.:


+      (unless (or (eq (frame-parameter frame 'minibuffer) 'only)
+                 ;; Take care not to delete the special frame that
+                 ;; the daemon uses.
+                 (and is-daemon (eq frame terminal-frame)))

where is-daemon is let bound to (daemonp)

bug#3442: 23.0.94; "C-x 5 1" (delete-other-frames) kills daemon

[Eli Zaretskii]

> Date: Thu, 4 Jun 2009 18:25:09 -0700 (PDT)
> From: Dan Nicolaescu <dann@ics.uci.edu>
> Cc: Teemu Likonen <tlikonen@iki.fi>, 3442@emacsbugs.donarmstrong.com
> Reply-To: Dan Nicolaescu <dann@ics.uci.edu>, 3442@emacsbugs.donarmstrong.com
> 
> Stefan Monnier <monnier@iro.umontreal.ca> writes:
> 
>   > I've installed a patch that makes C-x 5 1 only delete frames on the same
>   > terminal, which should fix this issue.
> 
> That's a behavior change, isn't it?
> Maybe people use C-x 5 1 to remove the rest of the frames on all
> terminals... (I personally have never used C-x 5 1)
> 
> If we don't want to change the C-x 5 1 behavior, we can just avoid
> deleting the daaemon special frame, i.e.:
> 
> 
> +      (unless (or (eq (frame-parameter frame 'minibuffer) 'only)
> +                 ;; Take care not to delete the special frame that
> +                 ;; the daemon uses.
> +                 (and is-daemon (eq frame terminal-frame)))
> 
> where is-daemon is let bound to (daemonp)

FWIW, I like Dan's suggestion better, especially since we are late in
the pretest, and shouldn't install changes that change well
established behavior.

bug#3442: 23.0.94; "C-x 5 1" (delete-other-frames) kills daemon

[Teemu Likonen]

On 2009-06-04 18:25 (-0700), Dan Nicolaescu wrote:

> Stefan Monnier <monnier@iro.umontreal.ca> writes:
>   > I've installed a patch that makes C-x 5 1 only delete frames on
>   > the same terminal, which should fix this issue.
>
> That's a behavior change, isn't it? Maybe people use C-x 5 1 to remove
> the rest of the frames on all terminals... (I personally have never
> used C-x 5 1)

I don't know what terminal means in this context nor what other people
think but just as a simple user this is what I'd expect: In X Window
System C-x 5 1 would delete other X frames. In tty C-x 5 1 would delete
other frames in that particular tty. Quite obviously neither would
affect daemon nor unrelated Emacs sessions.

I use C-x 5 1 quite a lot.

bug#3442: 23.0.94; "C-x 5 1" (delete-other-frames) kills daemon

[Stefan Monnier]

>> I've installed a patch that makes C-x 5 1 only delete frames on the same
>> terminal, which should fix this issue.
> That's a behavior change, isn't it?

Yes, tho only for people who were using a single Emacs displaying frames
on several X11 displays.
I feel like it's a change for the better, obviously.  This command has
never clearly specified what it did in the case where there are several
terminals because it was written at a time where Emacs did not support
this mode of operation.  So, AFAICT even the behavior change is a bug fix.


        Stefan

bug#3442: 23.0.94; "C-x 5 1" (delete-other-frames) kills daemon

[Teemu Likonen]

On 2009-06-04 14:31 (-0400), Stefan Monnier wrote:

>>> When Emacs daemon is running and a client frame is active, executing
>>> delete-other-frames command (C-x 5 1) kills the daemon and makes the
>>> Emacs session just a normal server session which requires that one
>>> frame exists. To reproduce:
>
>> Indeed, C-x 5 1 should only delete the frames on the current
>> terminal, whereas it currently kills *all* the frames on all
>> terminals (including the special internal terminal used for the
>> daemon).
>
> I've installed a patch that makes C-x 5 1 only delete frames on the
> same terminal, which should fix this issue.

I've been testing the fix for a while and it seems to work nicely. Thank
you. As far as I'm concerned this bug can closed.

bug#3442: marked as done (23.0.94; "C-x 5 1" (delete-other-frames) kills daemon)

[Emacs bug Tracking System]

[-- Attachment #1: Type: text/plain, Size: 923 bytes --]


Your message dated Mon, 15 Jun 2009 16:19:22 -0400
with message-id <jwv4ouhl07j.fsf-monnier+emacsbugreports@gnu.org>
and subject line Re: bug#3442: 23.0.94; "C-x 5 1" (delete-other-frames) kills daemon
has caused the Emacs bug report #3442,
regarding 23.0.94; "C-x 5 1" (delete-other-frames) kills daemon
to be marked as done.

This means that you claim that the problem has been dealt with.
If this is not the case it is now your responsibility to reopen the
bug report if necessary, and/or fix the problem forthwith.

(NB: If you are a system administrator and have no idea what this
message is talking about, this may indicate a serious mail system
misconfiguration somewhere. Please contact owner@emacsbugs.donarmstrong.com
immediately.)


-- 
3442: http://emacsbugs.donarmstrong.com/cgi-bin/bugreport.cgi?bug=3442
Emacs Bug Tracking System
Contact owner@emacsbugs.donarmstrong.com with problems

[-- Attachment #2: Type: message/rfc822, Size: 3756 bytes --]

From: Teemu Likonen <tlikonen@iki.fi>
To: emacs-pretest-bug@gnu.org
Subject: 23.0.94; "C-x 5 1" (delete-other-frames) kills daemon
Date: Mon, 01 Jun 2009 15:36:27 +0300
Message-ID: <87ab4s871w.fsf@iki.fi>

When Emacs daemon is running and a client frame is active, executing
delete-other-frames command (C-x 5 1) kills the daemon and makes the
Emacs session just a normal server session which requires that one frame
exists. To reproduce:

 1. Start Emacs daemon:

        emacs -Q --daemon

 2. Start a client session:

        emacsclient -c

 3. Type "C-x 5 1" and then try to close the frame with "C-x 5 0". The
    following message is displayed in the minibuffer:

        Attempt to delete the sole visible or iconified frame

 4. Forcibly close the frame with "C-x C-c" or window manager's close
    button.

 5. See the system's process list with "ps" command. There is no Emacs
    process anymore.




In GNU Emacs 23.0.94.1 (i686-pc-linux-gnu, GTK+ Version 2.12.12)
 of 2009-06-01 on mithlond
Windowing system distributor `The X.Org Foundation', version 11.0.10402000
configured using `configure  '--prefix=/home/dtw/local''

Important settings:
  value of $LC_ALL: nil
  value of $LC_COLLATE: nil
  value of $LC_CTYPE: nil
  value of $LC_MESSAGES: en_US.UTF-8
  value of $LC_MONETARY: nil
  value of $LC_NUMERIC: nil
  value of $LC_TIME: nil
  value of $LANG: fi_FI.UTF-8
  value of $XMODIFIERS: @im=none
  locale-coding-system: utf-8-unix
  default-enable-multibyte-characters: t

Major mode: Group

Minor modes in effect:
  gnus-undo-mode: t
  ergo-movement-mode: t
  tooltip-mode: t
  mouse-wheel-mode: t
  file-name-shadow-mode: t
  global-font-lock-mode: t
  font-lock-mode: t
  blink-cursor-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


[-- Attachment #3: Type: message/rfc822, Size: 1774 bytes --]

From: Stefan Monnier <monnier@iro.umontreal.ca>
To: Teemu Likonen <tlikonen@iki.fi>
Cc: 3442-done@emacsbugs.donarmstrong.com
Subject: Re: bug#3442: 23.0.94; "C-x 5 1" (delete-other-frames) kills daemon
Date: Mon, 15 Jun 2009 16:19:22 -0400
Message-ID: <jwv4ouhl07j.fsf-monnier+emacsbugreports@gnu.org>

> I've been testing the fix for a while and it seems to work nicely. Thank
> you. As far as I'm concerned this bug can closed.

Thank you,


        Stefan