bug#31951: 27.0.50; [PATCH] server-save-buffers-kill-terminal should respect save-some-buffers-default-predicate

["João Távora" <joaotavora@gmail.com>]

Noam Postavsky <npostavs@gmail.com> writes:

> João Távora <joaotavora@gmail.com> writes:
>
>>>     (defun save-buffers-kill-emacs (&optional arg)
>>>       [...]
>>>       ;; Don't use save-some-buffers-default-predicate, because we want
>>>       ;; to ask about all the buffers before killing Emacs.
>>>       (save-some-buffers arg t)
>>
>> Right, but thats because, when killing Emacs, it is really the last
>> chance to save those buffers before they're potentially gone for good.
>
> Oh, I got mixed up by the comment in server-save-buffers-kill-terminal:
>
> 	     ;; If client is bufferless, emulate a normal Emacs exit
> 	     ;; and offer to save all buffers.  Otherwise, offer to
> 	     ;; save only the buffers belonging to the client.
>
> So I'd say your change is sensible, we should just update that comment
> to explain the differences in the "emulation".

Hmmmm OK, what about:

diff --git a/lisp/server.el b/lisp/server.el
index ff03cbe622..33e88409ca 100644
--- a/lisp/server.el
+++ b/lisp/server.el
@@ -1666,13 +1666,16 @@ server-save-buffers-kill-terminal
 	     (save-buffers-kill-emacs arg)))
 	  ((processp proc)
 	   (let ((buffers (process-get proc 'buffers)))
-	     ;; If client is bufferless, emulate a normal Emacs exit
-	     ;; and offer to save all buffers.  Otherwise, offer to
-	     ;; save only the buffers belonging to the client.
+             ;; If the client has buffers, offer to save only those
+             ;; buffers.  Otherwise, this is similar to a normal Emacs
+             ;; exit (where we offer to save all buffers) with the
+             ;; exception that, if ARG is nil, then passing nil as the
+             ;; PRED argument to `save-some-buffers' ensures
+             ;; `save-some-buffers-default-predicate' is respected.
 	     (save-some-buffers
 	      arg (if buffers
 		      (lambda () (memq (current-buffer) buffers))
-		    t))
+                    (and arg t)))
 	     (server-delete-client proc)))
 	  (t (error "Invalid client frame")))))


I tried to make it terser, but couldn't.  If the indentation looks off,
it's not, it's because I untabified the changed regions.

If nobody complains I'll push this in a few days' time.

João