GNU bug report logs - #28412
26.0.50; Let save-some-buffers accept write-contents-functions

Previous Next

To add a comment to this bug, you must first unarchive it, by sending
a message to control AT debbugs.gnu.org, with unarchive 28412 in the body.
You can then email your comments to 28412 AT debbugs.gnu.org in the normal way.

Toggle the display of automated, internal messages from the tracker.

Message #5 received at submit <at> debbugs.gnu.org (full text, mbox):

From: Eric Abrahamsen <eric <at> ericabrahamsen.net>
To: bug-gnu-emacs <at> gnu.org
Subject: 26.0.50; Let save-some-buffers accept write-contents-functions
Date: Sun, 10 Sep 2017 14:50:15 -0700

[Message part 1 (text/plain, inline)]

This is about letting `save-some-buffers' check
`write-contents-functions' before insisting on a buffer having a file to
write to.

There was a conversations about this on emacs.devel:

https://lists.gnu.org/archive/html/emacs-devel/2017-05/msg00653.html

The patch I ended up with is attached.

I'm reporting this just so the patch (and the idea) doesn't get lost. I
still think it's worth doing.

[0001-First-whack-at-write-contents-functions-for-non-file.patch (text/x-diff, attachment)]

Message #8 received at 28412 <at> debbugs.gnu.org (full text, mbox):

From: Eric Abrahamsen <eric <at> ericabrahamsen.net>
To: 28412 <at> debbugs.gnu.org
Subject: Re: bug#28412: Acknowledgement (26.0.50;
 Let save-some-buffers accept write-contents-functions)
Date: Sun, 10 Sep 2017 15:01:34 -0700

[Message part 1 (text/plain, inline)]

Damn, I think that was the wrong patch.

Here's the latest one.

[0001-Allow-write-contents-functions-to-short-circuit-buff.patch (text/x-diff, attachment)]

Message #11 received at 28412 <at> debbugs.gnu.org (full text, mbox):

From: Eli Zaretskii <eliz <at> gnu.org>
To: Eric Abrahamsen <eric <at> ericabrahamsen.net>
Cc: 28412 <at> debbugs.gnu.org
Subject: Re: bug#28412: Acknowledgement (26.0.50;
 Let save-some-buffers accept write-contents-functions)
Date: Mon, 11 Sep 2017 18:03:25 +0300

> Resent-Sender: help-debbugs <at> gnu.org
> From: Eric Abrahamsen <eric <at> ericabrahamsen.net>
> Date: Sun, 10 Sep 2017 15:01:34 -0700
> 
> Here's the latest one.

Thanks.  Can we have some simple tests for this, both with and without
visiting a file?

This also needs a NEWS entry.  And please mention the bug number in
the log message.

Message #14 received at 28412 <at> debbugs.gnu.org (full text, mbox):

From: Eric Abrahamsen <eric <at> ericabrahamsen.net>
To: Eli Zaretskii <eliz <at> gnu.org>
Cc: 28412 <at> debbugs.gnu.org
Subject: Re: bug#28412: Acknowledgement (26.0.50;
 Let save-some-buffers accept write-contents-functions)
Date: Mon, 11 Sep 2017 14:41:08 -0700

> Thanks.  Can we have some simple tests for this, both with and without
> visiting a file?

Is this sufficient, do you think?


(ert-deftest files-test-no-file-write-contents ()
  "Test that `write-contents-functions' permits saving a file.
Usually `basic-save-buffer' will prompt for a file name if the
current buffer has none.  It should first call the functions in
`write-contents-functions', and if one of them returns non-nil,
consider the buffer saved, without prompting for a file
name (Bug#28412)."
  (let ((read-file-name-function
         (lambda (&rest _ignore)
           (error "Prompting for file name"))))
    ;; With contents function, and no file.
    (with-temp-buffer
      (setq write-contents-functions
            (list (lambda () t)))
      (set-buffer-modified-p t)
      (should (null (save-buffer))))
    ;; With no contents function and no file.  This should reach the
    ;; `read-file-name' prompt.
    (with-temp-buffer
      (set-buffer-modified-p t)
      (should-error (save-buffer) :type 'error))
    ;; Then a buffer visiting a file: should save normally.
    (files-tests--with-temp-file temp-file-name
      (with-current-buffer (find-file-noselect temp-file-name)
        (setq write-contents-functions nil)
        (insert "p")
        (should (null (save-buffer)))
        (should (eq (buffer-size) 1))))))

Message #17 received at 28412 <at> debbugs.gnu.org (full text, mbox):

From: Eli Zaretskii <eliz <at> gnu.org>
To: Eric Abrahamsen <eric <at> ericabrahamsen.net>
Cc: 28412 <at> debbugs.gnu.org
Subject: Re: bug#28412: Acknowledgement (26.0.50;
 Let save-some-buffers accept write-contents-functions)
Date: Tue, 12 Sep 2017 17:41:31 +0300

> From: Eric Abrahamsen <eric <at> ericabrahamsen.net>
> Cc: 28412 <at> debbugs.gnu.org
> Date: Mon, 11 Sep 2017 14:41:08 -0700
> 
> > Thanks.  Can we have some simple tests for this, both with and without
> > visiting a file?
> 
> Is this sufficient, do you think?

Yes, I think so.  Thanks.

Message #20 received at submit <at> debbugs.gnu.org (full text, mbox):

From: Eric Abrahamsen <eric <at> ericabrahamsen.net>
To: bug-gnu-emacs <at> gnu.org
Subject: Re: bug#28412: Acknowledgement (26.0.50;
 Let save-some-buffers accept write-contents-functions)
Date: Tue, 12 Sep 2017 16:18:00 -0700

Eli Zaretskii <eliz <at> gnu.org> writes:

>> From: Eric Abrahamsen <eric <at> ericabrahamsen.net>
>> Cc: 28412 <at> debbugs.gnu.org
>> Date: Mon, 11 Sep 2017 14:41:08 -0700
>> 
>> > Thanks.  Can we have some simple tests for this, both with and without
>> > visiting a file?
>> 
>> Is this sufficient, do you think?
>
> Yes, I think so.  Thanks.

Okay, thank you. Pushed as 9b980e2691.

Message #23 received at 28412 <at> debbugs.gnu.org (full text, mbox):

From: Kaushal Modi <kaushal.modi <at> gmail.com>
To: 28412 <at> debbugs.gnu.org
Subject: 26.0.50; Let save-some-buffers accept write-contents-functions
Date: Mon, 18 Sep 2017 16:16:05 +0000

[Message part 1 (text/plain, inline)]

Hello Eric, Eli:

I became a victim of this change because of these overlapping use-cases:

- I use ggtags.el that creates a hidden (is that the right term for that?)
buffer (starting with space) to fontify code snippets using the file's
major mode.
- The major mode I use (verilog-mode) sets write-file-functions to a
non-nil value.

So each time I quit emacs, I get a prompt to save that hidden buffer "
*Code-Fontify*" (which it didn't earlier, as intended).

Details: https://github.com/leoliu/ggtags/issues/157

I suggested calling (set-buffer-modified-p nil) in that special buffer.

But is that the best way to fix this? Or should save-some-buffers never try
to save buffers whose names begin with space?

PS: Please CC me on the replies, else I won't get those.

-- 

Kaushal Modi

[Message part 2 (text/html, inline)]

Message #26 received at 28412 <at> debbugs.gnu.org (full text, mbox):

From: Eli Zaretskii <eliz <at> gnu.org>
To: Kaushal Modi <kaushal.modi <at> gmail.com>,
 Eric Abrahamsen <eric <at> ericabrahamsen.net>
Cc: 28412 <at> debbugs.gnu.org
Subject: Re: bug#28412: 26.0.50;
 Let save-some-buffers accept write-contents-functions
Date: Mon, 18 Sep 2017 21:04:22 +0300

> From: Kaushal Modi <kaushal.modi <at> gmail.com>
> Date: Mon, 18 Sep 2017 16:16:05 +0000
> 
> - I use ggtags.el that creates a hidden (is that the right term for that?) buffer (starting with space) to fontify
> code snippets using the file's major mode.
> - The major mode I use (verilog-mode) sets write-file-functions to a non-nil value.
> 
> So each time I quit emacs, I get a prompt to save that hidden buffer " *Code-Fontify*" (which it didn't earlier, as
> intended).
> 
> Details: https://github.com/leoliu/ggtags/issues/157
> 
> I suggested calling (set-buffer-modified-p nil) in that special buffer. 
> 
> But is that the best way to fix this? Or should save-some-buffers never try to save buffers whose names begin
> with space?

Exempting buffers whose name starts with a space sounds reasonable.
Eric?

Message #29 received at 28412 <at> debbugs.gnu.org (full text, mbox):

From: Eric Abrahamsen <eric <at> ericabrahamsen.net>
To: Eli Zaretskii <eliz <at> gnu.org>
Cc: 28412 <at> debbugs.gnu.org, Kaushal Modi <kaushal.modi <at> gmail.com>
Subject: Re: bug#28412: 26.0.50;
 Let save-some-buffers accept write-contents-functions
Date: Mon, 18 Sep 2017 11:14:35 -0700

Eli Zaretskii <eliz <at> gnu.org> writes:

>> From: Kaushal Modi <kaushal.modi <at> gmail.com>
>> Date: Mon, 18 Sep 2017 16:16:05 +0000
>> 
>> - I use ggtags.el that creates a hidden (is that the right term for that?) buffer (starting with space) to fontify
>> code snippets using the file's major mode.
>> - The major mode I use (verilog-mode) sets write-file-functions to a non-nil value.
>> 
>> So each time I quit emacs, I get a prompt to save that hidden buffer " *Code-Fontify*" (which it didn't earlier, as
>> intended).
>> 
>> Details: https://github.com/leoliu/ggtags/issues/157
>> 
>> I suggested calling (set-buffer-modified-p nil) in that special buffer. 
>> 
>> But is that the best way to fix this? Or should save-some-buffers never try to save buffers whose names begin
>> with space?
>
> Exempting buffers whose name starts with a space sounds reasonable.
> Eric?

Sounds fine to me. I'm trying to think if there's some more general
interactive-vs-non-interactive case that should be protected against,
but I can't really come up with anything. The whole point of
`save-some-buffers' is the user prompt, and this seems like a fine way
to create an exception to the prompt.

I can fix it up -- the patch should go to emacs-26, right?

Eric

Message #32 received at 28412 <at> debbugs.gnu.org (full text, mbox):

From: Eric Abrahamsen <eric <at> ericabrahamsen.net>
To: Eli Zaretskii <eliz <at> gnu.org>
Cc: 28412 <at> debbugs.gnu.org, Kaushal Modi <kaushal.modi <at> gmail.com>
Subject: Re: bug#28412: 26.0.50;
 Let save-some-buffers accept write-contents-functions
Date: Mon, 18 Sep 2017 12:12:01 -0700

Eli Zaretskii <eliz <at> gnu.org> writes:

>> From: Kaushal Modi <kaushal.modi <at> gmail.com>
>> Date: Mon, 18 Sep 2017 16:16:05 +0000
>> 
>> - I use ggtags.el that creates a hidden (is that the right term for that?) buffer (starting with space) to fontify
>> code snippets using the file's major mode.
>> - The major mode I use (verilog-mode) sets write-file-functions to a non-nil value.
>> 
>> So each time I quit emacs, I get a prompt to save that hidden buffer " *Code-Fontify*" (which it didn't earlier, as
>> intended).
>> 
>> Details: https://github.com/leoliu/ggtags/issues/157
>> 
>> I suggested calling (set-buffer-modified-p nil) in that special buffer. 
>> 
>> But is that the best way to fix this? Or should save-some-buffers never try to save buffers whose names begin
>> with space?
>
> Exempting buffers whose name starts with a space sounds reasonable.
> Eric?

Also, there don't appear to be any built-in functions or regexps for
dealing with "hidden" buffers, I'm assuming:

(not (string-prefix-p " " (buffer-name buffer)))

is okay?

Message #35 received at 28412 <at> debbugs.gnu.org (full text, mbox):

From: Kaushal Modi <kaushal.modi <at> gmail.com>
To: Eric Abrahamsen <eric <at> ericabrahamsen.net>, Eli Zaretskii <eliz <at> gnu.org>
Cc: 28412 <at> debbugs.gnu.org
Subject: Re: bug#28412: 26.0.50;
 Let save-some-buffers accept write-contents-functions
Date: Mon, 18 Sep 2017 19:23:10 +0000

[Message part 1 (text/plain, inline)]

On Mon, Sep 18, 2017 at 3:12 PM Eric Abrahamsen <eric <at> ericabrahamsen.net>
wrote:

> Also, there don't appear to be any built-in functions or regexps for
> dealing with "hidden" buffers, I'm assuming:
>
> (not (string-prefix-p " " (buffer-name buffer)))
>
> is okay?
>

That crossed my mind too. So I had started looking for its existing
implementation, and found this in buff-menu.el inside list-buffers--refresh
function:

    (string= (substring name 0 1) " ")

I was wondering if that should be wrapped into a function like
buffer-internal-p (based on the many docstrings like below in buff-menu.el:
"By default, the Buffer Menu lists all buffers except those whose names
start with a space (which are for internal use)."). Having a function like
this would be better than having that logic implemented in different ways
in the emacs core.

I just don't know the best place to put this function definition. Eli?
-- 

Kaushal Modi

[Message part 2 (text/html, inline)]

Message #38 received at 28412 <at> debbugs.gnu.org (full text, mbox):

From: Eli Zaretskii <eliz <at> gnu.org>
To: Eric Abrahamsen <eric <at> ericabrahamsen.net>
Cc: 28412 <at> debbugs.gnu.org, kaushal.modi <at> gmail.com
Subject: Re: bug#28412: 26.0.50;
 Let save-some-buffers accept write-contents-functions
Date: Mon, 18 Sep 2017 22:25:33 +0300

> From: Eric Abrahamsen <eric <at> ericabrahamsen.net>
> Cc: Kaushal Modi <kaushal.modi <at> gmail.com>,  28412 <at> debbugs.gnu.org
> Date: Mon, 18 Sep 2017 11:14:35 -0700
> 
> I can fix it up -- the patch should go to emacs-26, right?

Yes, thanks.  Don't forget to mention the exception in the docs.

Message #41 received at 28412 <at> debbugs.gnu.org (full text, mbox):

From: Eli Zaretskii <eliz <at> gnu.org>
To: Eric Abrahamsen <eric <at> ericabrahamsen.net>
Cc: 28412 <at> debbugs.gnu.org, kaushal.modi <at> gmail.com
Subject: Re: bug#28412: 26.0.50;
 Let save-some-buffers accept write-contents-functions
Date: Mon, 18 Sep 2017 22:28:21 +0300

> From: Eric Abrahamsen <eric <at> ericabrahamsen.net>
> Cc: Kaushal Modi <kaushal.modi <at> gmail.com>,  28412 <at> debbugs.gnu.org
> Date: Mon, 18 Sep 2017 12:12:01 -0700
> 
> Also, there don't appear to be any built-in functions or regexps for
> dealing with "hidden" buffers, I'm assuming:
> 
> (not (string-prefix-p " " (buffer-name buffer)))
> 
> is okay?

Yes, but I think aref would be simpler (and probably faster).

Message #44 received at 28412 <at> debbugs.gnu.org (full text, mbox):

From: Eric Abrahamsen <eric <at> ericabrahamsen.net>
To: Eli Zaretskii <eliz <at> gnu.org>
Cc: 28412 <at> debbugs.gnu.org, kaushal.modi <at> gmail.com
Subject: Re: bug#28412: 26.0.50;
 Let save-some-buffers accept write-contents-functions
Date: Mon, 18 Sep 2017 13:30:50 -0700

On 09/18/17 22:25 PM, Eli Zaretskii wrote:
>> From: Eric Abrahamsen <eric <at> ericabrahamsen.net>
>> Cc: Kaushal Modi <kaushal.modi <at> gmail.com>,  28412 <at> debbugs.gnu.org
>> Date: Mon, 18 Sep 2017 11:14:35 -0700
>> 
>> I can fix it up -- the patch should go to emacs-26, right?
>
> Yes, thanks.  Don't forget to mention the exception in the docs.

Okay, done.

Message #47 received at 28412 <at> debbugs.gnu.org (full text, mbox):

From: Eric Abrahamsen <eric <at> ericabrahamsen.net>
To: Kaushal Modi <kaushal.modi <at> gmail.com>
Cc: Eli Zaretskii <eliz <at> gnu.org>, 28412 <at> debbugs.gnu.org
Subject: Re: bug#28412: 26.0.50;
 Let save-some-buffers accept write-contents-functions
Date: Mon, 18 Sep 2017 13:53:24 -0700

Kaushal Modi <kaushal.modi <at> gmail.com> writes:

> On Mon, Sep 18, 2017 at 3:12 PM Eric Abrahamsen <eric <at> ericabrahamsen.net> wrote:
>
>  Also, there don't appear to be any built-in functions or regexps for
>  dealing with "hidden" buffers, I'm assuming:
>
>  (not (string-prefix-p " " (buffer-name buffer)))
>
>  is okay?
>
> That crossed my mind too. So I had started looking for its existing implementation, and found this in buff-menu.el inside list-buffers--refresh function:
>
>     (string= (substring name 0 1) " ")
>
> I was wondering if that should be wrapped into a function like buffer-internal-p (based on the many docstrings like below in buff-menu.el: "By default, the Buffer Menu lists all buffers except those whose names start with a space
> (which are for internal use)."). Having a function like this would be better than having that logic implemented in different ways in the emacs core.
>
> I just don't know the best place to put this function definition. Eli?

It might be nice to have something like:

(defsubst buffer-internal-p (buf)
  (eq (aref (buffer-name buf) 0) ?\s))

Message #50 received at 28412 <at> debbugs.gnu.org (full text, mbox):

From: Kaushal Modi <kaushal.modi <at> gmail.com>
To: Eric Abrahamsen <eric <at> ericabrahamsen.net>
Cc: Eli Zaretskii <eliz <at> gnu.org>, 28412 <at> debbugs.gnu.org
Subject: Re: bug#28412: 26.0.50;
 Let save-some-buffers accept write-contents-functions
Date: Mon, 18 Sep 2017 21:48:15 +0000

[Message part 1 (text/plain, inline)]

On Mon, Sep 18, 2017 at 4:53 PM Eric Abrahamsen <eric <at> ericabrahamsen.net>
wrote:

>
> It might be nice to have something like:
>
> (defsubst buffer-internal-p (buf)
>   (eq (aref (buffer-name buf) 0) ?\s))
>

Yup. I also see this in the "(emacs) Undo" Info node:

   Some specialized buffers do not make undo records.  Buffers whose
names start with spaces never do; these buffers are used internally by
Emacs to hold text that users don’t normally look at or edit.

But I cannot find the code that implements that.

It would be nice to have these places to use the same defsubst.

Btw I confirm the fix. Thanks!
-- 

Kaushal Modi

[Message part 2 (text/html, inline)]

Message #53 received at 28412 <at> debbugs.gnu.org (full text, mbox):

From: Eli Zaretskii <eliz <at> gnu.org>
To: Kaushal Modi <kaushal.modi <at> gmail.com>
Cc: eric <at> ericabrahamsen.net, 28412 <at> debbugs.gnu.org
Subject: Re: bug#28412: 26.0.50;
 Let save-some-buffers accept write-contents-functions
Date: Tue, 19 Sep 2017 19:09:55 +0300

> From: Kaushal Modi <kaushal.modi <at> gmail.com>
> Date: Mon, 18 Sep 2017 19:23:10 +0000
> Cc: 28412 <at> debbugs.gnu.org
> 
> (string= (substring name 0 1) " ")
> 
> I was wondering if that should be wrapped into a function like buffer-internal-p (based on the many docstrings
> like below in buff-menu.el: "By default, the Buffer Menu lists all buffers except those whose names start with a
> space (which are for internal use)."). Having a function like this would be better than having that logic
> implemented in different ways in the emacs core.
> 
> I just don't know the best place to put this function definition. Eli?

I'm not sure such a simple test should be canonicalized.

Message #56 received at 28412 <at> debbugs.gnu.org (full text, mbox):

From: Eli Zaretskii <eliz <at> gnu.org>
To: Kaushal Modi <kaushal.modi <at> gmail.com>
Cc: eric <at> ericabrahamsen.net, 28412 <at> debbugs.gnu.org
Subject: Re: bug#28412: 26.0.50;
 Let save-some-buffers accept write-contents-functions
Date: Tue, 19 Sep 2017 19:13:13 +0300

> From: Kaushal Modi <kaushal.modi <at> gmail.com>
> Date: Mon, 18 Sep 2017 21:48:15 +0000
> Cc: Eli Zaretskii <eliz <at> gnu.org>, 28412 <at> debbugs.gnu.org
> 
> Yup. I also see this in the "(emacs) Undo" Info node:
> 
> Some specialized buffers do not make undo records. Buffers whose
> names start with spaces never do; these buffers are used internally by
> Emacs to hold text that users don’t normally look at or edit.
> 
> But I cannot find the code that implements that.

It's in get-buffer-create:

  bset_undo_list (b, SREF (name, 0) != ' ' ? Qnil : Qt);

Message #59 received at 28412 <at> debbugs.gnu.org (full text, mbox):

From: Kaushal Modi <kaushal.modi <at> gmail.com>
To: 28412 <at> debbugs.gnu.org
Cc: eric <at> ericabrahamsen.net, Eli Zaretskii <eliz <at> gnu.org>,
 Stefan Monnier <monnier <at> iro.umontreal.ca>
Subject: Re: bug#28412: 26.0.50;
 Let save-some-buffers accept write-contents-functions
Date: Mon, 25 Sep 2017 14:46:05 +0000

[Message part 1 (text/plain, inline)]

To close the loop:

A better fix (I confirm this iteration of the fix too -- fix for the ggtags
issue[3]) for this was later applied in commit 3d3778d8[1]. See this
emacs-devel thread[2] for discussion that arrived to this conclusion.

[1]:
http://git.savannah.gnu.org/cgit/emacs.git/commit/?h=emacs-26&id=3d3778d82a87139ef50a24146f5bad2a57a82094
[2]: http://lists.gnu.org/archive/html/emacs-devel/2017-09/msg00858.html
[3]: https://github.com/leoliu/ggtags/issues/157

-- 

Kaushal Modi

[Message part 2 (text/html, inline)]

Message #64 received at 28412-done <at> debbugs.gnu.org (full text, mbox):

From: Eric Abrahamsen <eric <at> ericabrahamsen.net>
To: 28412-done <at> debbugs.gnu.org
Subject: Re: bug#28412: 26.0.50;
 Let save-some-buffers accept write-contents-functions
Date: Thu, 19 Oct 2017 08:25:08 -0700

closing

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.