bug#55579: 29.0.50; Abbrevs not saved on exit anymore

bug#55579: 29.0.50; Abbrevs not saved on exit anymore

[Juri Linkov]

After a recent change, 'C-x C-c' doesn't ask anymore to save abbrevs with:

  Save abbrevs in ~/.emacs.d/abbrev_defs?

It looks like the regression was caused by ce220524fcf.
Before this change, save-buffers-kill-emacs unconditionally
called save-some-buffers that also saves the abbrev table.
But after the change, save-some-buffers that saves abbrevs
is called only when files--buffers-needing-to-be-saved
detects unsaved buffers.

bug#55579: 29.0.50; Abbrevs not saved on exit anymore

[Lars Ingebrigtsen]

Juri Linkov <juri@linkov.net> writes:

> After a recent change, 'C-x C-c' doesn't ask anymore to save abbrevs with:
>
>   Save abbrevs in ~/.emacs.d/abbrev_defs?
>
> It looks like the regression was caused by ce220524fcf.
> Before this change, save-buffers-kill-emacs unconditionally
> called save-some-buffers that also saves the abbrev table.
> But after the change, save-some-buffers that saves abbrevs
> is called only when files--buffers-needing-to-be-saved
> detects unsaved buffers.

Wow -- I had no idea that save-some-buffers also randomly saved abbrevs.

I've now separated that this out into a general mechanism that abbrev.el
now hooks into, while also fixing this `C-x C-c' bug.

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no

bug#55579: 29.0.50; Abbrevs not saved on exit anymore

[Juri Linkov]

> Save abbrevs in ~/.emacs.d/abbrev_defs?

BTW, I noticed that ~/.emacs.d/abbrev_defs
is visited without font-locking.  Here's the fix:

diff --git a/lisp/abbrev.el b/lisp/abbrev.el
index 3ee972869b..a3b19107f7 100644
--- a/lisp/abbrev.el
+++ b/lisp/abbrev.el
@@ -264,7 +264,7 @@ write-abbrev-file
       (when (unencodable-char-position (point-min) (point-max) 'utf-8)
 	(setq coding-system-for-write 'utf-8-emacs))
       (goto-char (point-min))
-      (insert (format ";;-*-coding: %s;-*-\n" coding-system-for-write))
+      (insert (format ";; -*- coding: %s; mode: lisp-data -*-\n" coding-system-for-write))
       (write-region nil nil file nil (and (not verbose) 0)))))

bug#55579: 29.0.50; Abbrevs not saved on exit anymore

[Lars Ingebrigtsen]

Juri Linkov <juri@linkov.net> writes:

> BTW, I noticed that ~/.emacs.d/abbrev_defs
> is visited without font-locking.  Here's the fix:

Makes sense to me.

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no

bug#55579: 29.0.50; Abbrevs not saved on exit anymore

[Eli Zaretskii]

> From: Juri Linkov <juri@linkov.net>
> Date: Tue, 24 May 2022 21:42:58 +0300
> 
> > Save abbrevs in ~/.emacs.d/abbrev_defs?
> 
> BTW, I noticed that ~/.emacs.d/abbrev_defs
> is visited without font-locking.  Here's the fix:

Won't that make loading abbrevs slower and more expensive?  that file
is normally not edited interactively, only if the user explicitly
wants to do so.  So why is font-lock in it important?

bug#55579: 29.0.50; Abbrevs not saved on exit anymore

[Andreas Röhler]

[-- Attachment #1: Type: text/plain, Size: 474 bytes --]

+1

Am 24.05.22 um 21:27 schrieb Eli Zaretskii:
>> From: Juri Linkov<juri@linkov.net>
>> Date: Tue, 24 May 2022 21:42:58 +0300
>>
>>> Save abbrevs in ~/.emacs.d/abbrev_defs?
>> BTW, I noticed that ~/.emacs.d/abbrev_defs
>> is visited without font-locking.  Here's the fix:
> Won't that make loading abbrevs slower and more expensive?  that file
> is normally not edited interactively, only if the user explicitly
> wants to do so.  So why is font-lock in it important?
>
>
>

[-- Attachment #2: Type: text/html, Size: 1166 bytes --]

bug#55579: 29.0.50; Abbrevs not saved on exit anymore

[Juri Linkov]

>> > Save abbrevs in ~/.emacs.d/abbrev_defs?
>>
>> BTW, I noticed that ~/.emacs.d/abbrev_defs
>> is visited without font-locking.  Here's the fix:
>
> Won't that make loading abbrevs slower and more expensive?  that file
> is normally not edited interactively, only if the user explicitly
> wants to do so.  So why is font-lock in it important?

This is part of recent efforts to add modes to .emacs.d files.
So far we added the mode tag to ~/.emacs.d/bookmarks, ~/.emacs.d/tramp,
~/.emacs.d/.emacs.desktop, ~/.emacs.d/network-security.data,
~/.emacs.d/places, ~/.emacs.d/projects.  But mode is still missing
in ~/.emacs.d/abbrev_defs and ~/.emacs.d/recentf.

bug#55579: 29.0.50; Abbrevs not saved on exit anymore

[Lars Ingebrigtsen]

Eli Zaretskii <eliz@gnu.org> writes:

>> BTW, I noticed that ~/.emacs.d/abbrev_defs
>> is visited without font-locking.  Here's the fix:
>
> Won't that make loading abbrevs slower and more expensive?  that file
> is normally not edited interactively, only if the user explicitly
> wants to do so.  So why is font-lock in it important?

`read-abbrev-file' just uses `load' on the file, so the mode cookie
doesn't do anything non-interactively.

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no

bug#55579: 29.0.50; Abbrevs not saved on exit anymore

[Eli Zaretskii]

> From: Juri Linkov <juri@linkov.net>
> Cc: 55579@debbugs.gnu.org
> Date: Wed, 25 May 2022 09:37:20 +0300
> 
> >> > Save abbrevs in ~/.emacs.d/abbrev_defs?
> >>
> >> BTW, I noticed that ~/.emacs.d/abbrev_defs
> >> is visited without font-locking.  Here's the fix:
> >
> > Won't that make loading abbrevs slower and more expensive?  that file
> > is normally not edited interactively, only if the user explicitly
> > wants to do so.  So why is font-lock in it important?
> 
> This is part of recent efforts to add modes to .emacs.d files.
> So far we added the mode tag to ~/.emacs.d/bookmarks, ~/.emacs.d/tramp,
> ~/.emacs.d/.emacs.desktop, ~/.emacs.d/network-security.data,
> ~/.emacs.d/places, ~/.emacs.d/projects.  But mode is still missing
> in ~/.emacs.d/abbrev_defs and ~/.emacs.d/recentf.

We don't need to do that for all of the files, only for files that are
frequently (or always) visited or edited interactively by users.  So
I'm asking this question specifically for abbrev_defs, not for all the
other files there: why do we need this, and how significant is the
reason?

bug#55579: 29.0.50; Abbrevs not saved on exit anymore

[Eli Zaretskii]

> From: Lars Ingebrigtsen <larsi@gnus.org>
> Cc: Juri Linkov <juri@linkov.net>,  55579@debbugs.gnu.org
> Date: Wed, 25 May 2022 14:10:31 +0200
> 
> Eli Zaretskii <eliz@gnu.org> writes:
> 
> >> BTW, I noticed that ~/.emacs.d/abbrev_defs
> >> is visited without font-locking.  Here's the fix:
> >
> > Won't that make loading abbrevs slower and more expensive?  that file
> > is normally not edited interactively, only if the user explicitly
> > wants to do so.  So why is font-lock in it important?
> 
> `read-abbrev-file' just uses `load' on the file, so the mode cookie
> doesn't do anything non-interactively.

Then let's have a comment where the cookie is written to explain
that it assumes 'load', and let's have another comment where we load
the file to explain why using another function should consider this
aspect.

bug#55579: 29.0.50; Abbrevs not saved on exit anymore

[Juri Linkov]

>> >> > Save abbrevs in ~/.emacs.d/abbrev_defs?
>> >>
>> >> BTW, I noticed that ~/.emacs.d/abbrev_defs
>> >> is visited without font-locking.  Here's the fix:
>> >
>> > Won't that make loading abbrevs slower and more expensive?  that file
>> > is normally not edited interactively, only if the user explicitly
>> > wants to do so.  So why is font-lock in it important?
>>
>> This is part of recent efforts to add modes to .emacs.d files.
>> So far we added the mode tag to ~/.emacs.d/bookmarks, ~/.emacs.d/tramp,
>> ~/.emacs.d/.emacs.desktop, ~/.emacs.d/network-security.data,
>> ~/.emacs.d/places, ~/.emacs.d/projects.  But mode is still missing
>> in ~/.emacs.d/abbrev_defs and ~/.emacs.d/recentf.
>
> We don't need to do that for all of the files, only for files that are
> frequently (or always) visited or edited interactively by users.  So
> I'm asking this question specifically for abbrev_defs, not for all the
> other files there: why do we need this, and how significant is the
> reason?

Of course, it's visited frequently.  This is necessary to see
what abbreviations are defined.

bug#55579: 29.0.50; Abbrevs not saved on exit anymore

[Lars Ingebrigtsen]

Eli Zaretskii <eliz@gnu.org> writes:

> Then let's have a comment where the cookie is written to explain
> that it assumes 'load', and let's have another comment where we load
> the file to explain why using another function should consider this
> aspect.

We add these cookies to basically all files like this, so adding
these comments seems superfluous.

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no