bug#58909: 29.0.50; [WIP PATCH] Deleting the last frame of an emacsclient doesn't ask to save

[Jim Porter]

On 10/31/2022 5:44 AM, Eli Zaretskii wrote:
> > Date: Sun, 30 Oct 2022 15:29:30 -0700
> > From: Jim Porter <jporterbugs@gmail.com>
> >
> > However, if you use 'C-x 5 0' instead, that terminates the Emacs client,
> > but *doesn't* prompt to save foo.txt. I think it should prompt in this
> > case too: all three of 'C-x #', 'C-x C-c', and 'C-x 5 0' have the effect
> > (in this simple case) of deleting the Emacs client and returning to the
> > calling process. (A user who wanted to bail out of an Emacs client
> > without saving should use 'sever-edit-abort' instead.)
>
> I'm uneasy with this incompatible behavior change.  I can think of
> some legitimate use cases where "C-x 5 0" should not prompt, e.g., if
> the user intends to keep editing the file, and no application is
> waiting for the client to finish.  Why break such flows?

After thinking about this some more, I realized that I didn't properly 
address this part of your message. If no application is waiting for the 
client to finish, then the user hopefully used "--no-wait" when starting 
emacsclient. We could avoid prompting in that case.[1]

If an application *is* waiting for the client to finish, then 
"--no-wait" would be unset. In that case, prompting the user has some 
value. See the attached patch. (I changed the code to only prompt when 
deleting the last frame of a *non-nowait* client.)

[1] Currently, 'C-x C-c' from a nowait frame prompts via 
'save-some-buffers' (so long as there's another frame; see 
'server-save-buffers-kill-terminal'). Maybe we could avoid prompting 
then too? After prompting, it just calls 'delete-frame', so it's very 
similar to 'C-x 5 0' in this scenario... Still, as you say, it's 
probably good to be careful about making incompatible changes here.

0001-When-deleting-the-last-frame-of-an-Emacs-client-ask-.patch
Description: Text document