bug#51993: 29.0.50; [PATCH] Killing emacsclient terminal with `server-stop-automatically' doesn't prompt to save files

[Jim Porter <jporterbugs@gmail.com>]

On 10/20/2022 11:38 PM, Eli Zaretskii wrote:
>> Date: Thu, 20 Oct 2022 22:51:42 -0700
>> Cc: larsi@gnus.org, 51993@debbugs.gnu.org, gregory@heytings.org
>> From: Jim Porter <jporterbugs@gmail.com>
>>
>>>>      b) if this *is* the last client, prompt the user to save everything
>>>> (as with 'save-buffers-kill-emacs'), and then delete the client + kill
>>>> the Emacs daemon.
>>>
>>> You mean, in b), instead of just deleting the frame and leaving the
>>> daemon run, you want to shut down Emacs in its entirety, as if the
>>> user invoked kill-emacs?  I'm okay with that as an optional behavior,
>>> although I myself won't use it, as it's too dangerous.
>>
>> Almost. I'd like it to be as if the user invoked
>> 'save-buffers-kill-emacs'; that is, before killing Emacs, prompt the
>> user about everything[1] that might be lost by killing Emacs.
> 
> That should already happen, if you just call save-buffers-kill-emacs
> in that case, right?

Yeah. My expectation is that I can type 'C-x C-c' (or 'M-x 
save-buffers-kill-terminal') to kill a client, and if it's the last 
client, instead kill Emacs entirely (like 'M-x 
save-buffers-kill-emacs'). So roughly speaking, the change would be that 
you can set 'save-buffer-kill-terminal' to work like 
'save-buffer-kill-emacs' when there's only 1 client left.

>> This already exists as an option -- (server-stop-automatically
>> 'delete-frame)[2], but I also find the current behavior too dangerous.
>> My original message outlines one of the problems with the current
>> implementation: it changes the behavior of (a) in my description above.
>>
>>>    $ emacs -Q --daemon
>>>    $ emacsclient -a "" -c foo.txt
>>>    $ emacsclient -a "" -c bar.txt
>>>
>>>    ;; In the first client frame:
>>>    foobar ;; Insert some text
>>>    C-x C-c
>>>    ;; Emacs prompts "Save file /path/to/foo.txt?..."
>>>
>>> Now try the above, but call `(server-stop-automatically 'delete-frame)' first (or replace `delete-frame' with `kill-terminal'; it doesn't matter). In this case, Emacs doesn't prompt to save the file.
> 
> I'm not sure I see the direct relevance, and I don't think I see a bug
> in the above behavior.  I'm probably missing something, but what?

The issue in the quote above is that if you enable automatic server 
shutdown in Emacs 29, it changes the behavior of exiting an emacsclient 
even when it wouldn't stop the server (i.e. when there are other active 
clients). That's surprising to me, I wouldn't expect that setting to 
affect cases when it decides *not* to kill the Emacs daemon.

(This is relevant to the previous discussion since fixing this would get 
Emacs's automatic server shutdown - aka killing the daemon - close to 
the way I described there. Gregory mentioned[1] that the current 
behavior is intended, although we've had some difficulty coming to an 
agreement on how all this should work. Hence why I thought a hook might 
help here: if there are strong opinions in various directions, maybe a 
simple option isn't enough. Or maybe this is a case where you and/or 
Lars would be in a better position to make a final decision...)

[1] https://debbugs.gnu.org/cgi/bugreport.cgi?bug=51993