bug#26796: 26.0.50; (2 characters) is usually wrong

bug#26796: 26.0.50; (2 characters) is usually wrong

[Lars Ingebrigtsen]

To reproduce, open a file, type in "á" and save the file.  Emacs will
now say (if you're a utf-8 kind of person)

Wrote ‘/tmp/a’ (2 characters)

which is wrong, since there's just one character in the buffer.  Perhaps
saying "(2 bytes)" will be more helpful?

(I'd rather Emacs default to not saying anything about the length of the
file saved at all by default, since it's just noise to me (and I'll
assume to most people).)


In GNU Emacs 26.0.50 (build 3, x86_64-pc-linux-gnu, GTK+ Version 3.14.5)
 of 2017-04-24 built on stories
Repository revision: a1f93c1dfa53dbe007faa09ab0c6e913e86e3ffe
Windowing system distributor 'The X.Org Foundation', version 11.0.11604000
System Description:	Debian GNU/Linux 8.7 (jessie)


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

bug#26796: 26.0.50; (2 characters) is usually wrong

[Eli Zaretskii]

> From: Lars Ingebrigtsen <larsi@gnus.org>
> Date: Sat, 06 May 2017 11:49:55 +0200
> 
> To reproduce, open a file, type in "á" and save the file.  Emacs will
> now say (if you're a utf-8 kind of person)
> 
> Wrote ‘/tmp/a’ (2 characters)
> 
> which is wrong, since there's just one character in the buffer.

I see "1 characters" here.  I think the result depends on who you type
the above: if there are two characters, a followed by ́, then the file
really has 2 characters, and they are only shown as one on display.
Try "C-u C-x =" to see what you have in the buffer.

> Perhaps saying "(2 bytes)" will be more helpful?

I really doubt that: Emacs users tend to think in terms of characters
than in terms of bytes.

bug#26796: 26.0.50; (2 characters) is usually wrong

[Lars Ingebrigtsen]

Eli Zaretskii <eliz@gnu.org> writes:

> I see "1 characters" here.  I think the result depends on who you type
> the above: if there are two characters, a followed by ́, then the file
> really has 2 characters, and they are only shown as one on display.
> Try "C-u C-x =" to see what you have in the buffer.

No, it was an á, but the problem was instead that I apparently had
fumbled and added a newline before saving?  Hm.  Sorry for the noise;
closing.

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

bug#26796: 26.0.50; (2 characters) is usually wrong

[Lars Ingebrigtsen]

Eli Zaretskii <eliz@gnu.org> writes:

>> Perhaps saying "(2 bytes)" will be more helpful?
>
> I really doubt that: Emacs users tend to think in terms of characters
> than in terms of bytes.

But this is questionable, at least:

(with-temp-buffer (set-buffer-multibyte nil) (insert-file-contents "~/rms.jpg") (write-region (point-min) (point-max) "/tmp/a.jpg"))
Wrote ‘/tmp/a.jpg’ (889 characters)

There are no characters in a JPEG file; it's all bytes.

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

bug#26796: 26.0.50; (2 characters) is usually wrong

[Eli Zaretskii]

> From: Lars Ingebrigtsen <larsi@gnus.org>
> Cc: 26796@debbugs.gnu.org
> Date: Sat, 06 May 2017 12:45:45 +0200
> 
> But this is questionable, at least:
> 
> (with-temp-buffer (set-buffer-multibyte nil) (insert-file-contents "~/rms.jpg") (write-region (point-min) (point-max) "/tmp/a.jpg"))
> Wrote ‘/tmp/a.jpg’ (889 characters)
> 
> There are no characters in a JPEG file; it's all bytes.

Each byte is a character in this case, right?

Anyway, since you don't like this feature, it's clear that you can
come up with many examples where you don't like the result.  People
who do like it will most probably disagree.

bug#26796: 26.0.50; (2 characters) is usually wrong

[npostavs]

Lars Ingebrigtsen <larsi@gnus.org> writes:

> (I'd rather Emacs default to not saying anything about the length of the
> file saved at all by default, since it's just noise to me (and I'll
> assume to most people).)

Eli Zaretskii <eliz@gnu.org> writes:

> Anyway, since you don't like this feature, it's clear that you can
> come up with many examples where you don't like the result.  People
> who do like it will most probably disagree.

Actually this reminds I was planning on reverting this feature entirely,
I haven't seen anyone say they like it (the bug reporter who initiated
this feature never responded:
https://debbugs.gnu.org/cgi/bugreport.cgi?bug=354#81) whereas quite a
few seem to actively dislike it.

bug#26796: 26.0.50; (2 characters) is usually wrong

[Eli Zaretskii]

> From: npostavs@users.sourceforge.net
> Date: Sat, 06 May 2017 08:16:53 -0400
> Cc: 26796@debbugs.gnu.org
> 
> Actually this reminds I was planning on reverting this feature entirely,

Since it's already implemented, maybe just make it optional, by
default off.

bug#26796: 26.0.50; (2 characters) is usually wrong

[npostavs]

Eli Zaretskii <eliz@gnu.org> writes:

>> From: npostavs@users.sourceforge.net
>> Date: Sat, 06 May 2017 08:16:53 -0400
>> Cc: 26796@debbugs.gnu.org
>> 
>> Actually this reminds I was planning on reverting this feature entirely,
>
> Since it's already implemented, maybe just make it optional, by
> default off.

I'm not sure the added complexity is worth it for a feature that
probably nobody will bother using...

bug#26796: 26.0.50; (2 characters) is usually wrong

[Paul Eggert]

[-- Attachment #1: Type: text/plain, Size: 229 bytes --]

Appending " (NNNN characters)" to the "Wrote FOO" messages appears to be 
unpopular and at times confusing. At least it should be off by default, so I 
installed the attached. It'd be fine with me if we removed them entirely.

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: 0001-New-var-write-region-verbose-default-nil.patch --]
[-- Type: text/x-diff; name="0001-New-var-write-region-verbose-default-nil.patch", Size: 4475 bytes --]

From 33c96bacd656cffeedf32038686c0ea99b7c411b Mon Sep 17 00:00:00 2001
From: Paul Eggert <eggert@cs.ucla.edu>
Date: Sat, 6 May 2017 18:00:23 -0700
Subject: [PATCH] New var write-region-verbose, default nil

By popular demand, write-region char counts are now off by default
(Bug#26796).
* src/fileio.c (write-region-verbose): New Lisp var.
(write_region): Output char count only if the var is non-nil.
* doc/emacs/files.texi (Misc File Ops), etc/NEWS: Document this.
---
 doc/emacs/files.texi |  7 ++++---
 etc/NEWS             |  6 +++---
 src/fileio.c         | 34 +++++++++++++++++++++++-----------
 3 files changed, 30 insertions(+), 17 deletions(-)

diff --git a/doc/emacs/files.texi b/doc/emacs/files.texi
index bc977b7..d36fe65 100644
--- a/doc/emacs/files.texi
+++ b/doc/emacs/files.texi
@@ -1657,9 +1657,10 @@ Misc File Ops
 copies the contents of the region into the specified file.  @kbd{M-x
 append-to-file} adds the text of the region to the end of the
 specified file.  @xref{Accumulating Text}.  When called interactively,
-these commands will print a message in the echo area giving the name
-of the file affected as well as the number of characters which were
-added.  The variable @code{write-region-inhibit-fsync} applies to
+these commands print a message in the echo area giving the name
+of the file affected; if the variable @code{write-region-verbose} is
+non-nil the message also reports the number of characters written.
+The variable @code{write-region-inhibit-fsync} applies to
 these commands, as well as saving files; see @ref{Customize Save}.
 
 @findex set-file-modes
diff --git a/etc/NEWS b/etc/NEWS
index 2918c6e..1f1f4b4 100644
--- a/etc/NEWS
+++ b/etc/NEWS
@@ -94,9 +94,9 @@ required capabilities are found in terminfo.  See the FAQ node
 * Changes in Emacs 26.1
 
 +++
-** The functions write-region, append-to-file, and the like now output
-the number of characters added in addition to the name of the file
-affected.
+** The functions write-region, append-to-file, and the like now also
+output the number of characters added in addition to the name of the
+file affected, if the new variable 'write-region-verbose' is non-nil.
 
 ** The variable 'emacs-version' no longer includes the build number.
 This is now stored separately in a new variable, 'emacs-build-number'.
diff --git a/src/fileio.c b/src/fileio.c
index ad5ab61..6138bfc 100644
--- a/src/fileio.c
+++ b/src/fileio.c
@@ -5153,17 +5153,24 @@ write_region (Lisp_Object start, Lisp_Object end, Lisp_Object filename,
     {
       EMACS_INT nchars = (STRINGP (start) ? SCHARS (start)
 			  : XINT (end) - XINT (start));
-      AUTO_STRING (format, NUMBERP (append)
-                   ? (nchars != 1
-		      ? "Updated `%s' (%d characters)"
-		      : "Updated `%s' (%d character)")
-                   : ! NILP (append)
-		   ? (nchars != 1
-		      ? "Added to `%s' (%d characters)"
-		      : "Added to `%s' (%d character)")
-                   : (nchars != 1
-		      ? "Wrote `%s' (%d characters)"
-		      : "Wrote `%s' (%d character)"));
+      AUTO_STRING (format,
+		   (NUMBERP (append)
+		    ? (NILP (Vwrite_region_verbose)
+		       ? "Updated `%s'"
+		       : nchars == 1
+		       ? "Updated `%s' (1 character)"
+		       : "Updated `%s' (%d characters)")
+		    : ! NILP (append)
+		    ? (NILP (Vwrite_region_verbose)
+		       ? "Added to `%s'"
+		       : nchars == 1
+		       ? "Added to `%s' (1 character)"
+		       : "Added to `%s' (%d characters)")
+		    : (NILP (Vwrite_region_verbose)
+		       ? "Wrote `%s'"
+		       : nchars == 1
+		       ? "Wrote `%s' (1 character)"
+		       : "Wrote `%s' (%d characters)")));
       CALLN (Fmessage, format, visit_file, make_number (nchars));
     }
   return Qnil;
@@ -6135,6 +6142,11 @@ These are the annotations made by other annotation functions
 that were already called.  See also `write-region-annotate-functions'.  */);
   Vwrite_region_annotations_so_far = Qnil;
 
+  DEFVAR_LISP ("write-region-verbose",
+	       Vwrite_region_verbose,
+	       doc: /* If non-nil, be more verbose when writing a region.  */);
+  Vwrite_region_verbose = Qnil;
+
   DEFVAR_LISP ("inhibit-file-name-handlers", Vinhibit_file_name_handlers,
 	       doc: /* A list of file name handlers that temporarily should not be used.
 This applies only to the operation `inhibit-file-name-operation'.  */);
-- 
2.7.4

bug#26796: 26.0.50; (2 characters) is usually wrong

[Mark Oteiza]

Eli Zaretskii <eliz@gnu.org> writes:

>> From: npostavs@users.sourceforge.net
>> Date: Sat, 06 May 2017 08:16:53 -0400
>> Cc: 26796@debbugs.gnu.org
>> 
>> Actually this reminds I was planning on reverting this feature entirely,
>
> Since it's already implemented, maybe just make it optional, by
> default off.

Was there any reason to implement it in the first place?

bug#26796: 26.0.50; (2 characters) is usually wrong

[Eli Zaretskii]

> From: Mark Oteiza <mvoteiza@udel.edu>
> Cc: npostavs@users.sourceforge.net,  larsi@gnus.org,  26796@debbugs.gnu.org
> Date: Mon, 08 May 2017 01:02:27 -0400
> 
> Was there any reason to implement it in the first place?

Seems we are back at the discussion we had starting here:

  http://lists.gnu.org/archive/html/emacs-devel/2017-04/msg00219.html

;-)

bug#26796: 26.0.50; (2 characters) is usually wrong

[Glenn Morris]

Eli Zaretskii wrote:

>> Was there any reason to implement it in the first place?
>
> Seems we are back at the discussion we had starting here:
>
>   http://lists.gnu.org/archive/html/emacs-devel/2017-04/msg00219.html

In which, if I read right, literally no-one was in favour of the feature?
A feature that several people have now spent time debugging, tweaking,
and documenting, and which will have to be supported going forwards.

bug#26796: 26.0.50; (2 characters) is usually wrong

[Eli Zaretskii]

> From: Glenn Morris <rgm@gnu.org>
> Cc: Mark Oteiza <mvoteiza@udel.edu>,  larsi@gnus.org,  26796@debbugs.gnu.org,  npostavs@users.sourceforge.net
> Date: Mon, 08 May 2017 11:59:52 -0400
> 
> Eli Zaretskii wrote:
> 
> >> Was there any reason to implement it in the first place?
> >
> > Seems we are back at the discussion we had starting here:
> >
> >   http://lists.gnu.org/archive/html/emacs-devel/2017-04/msg00219.html
> 
> In which, if I read right, literally no-one was in favour of the feature?
> A feature that several people have now spent time debugging, tweaking,
> and documenting, and which will have to be supported going forwards.

Is this a complaint about something I did?  Just wondering.

bug#26796: 26.0.50; (2 characters) is usually wrong

[Mark Oteiza]

On 08/05/17 at 07:07pm, Eli Zaretskii wrote:
> > From: Glenn Morris <rgm@gnu.org>
> > Cc: Mark Oteiza <mvoteiza@udel.edu>,  larsi@gnus.org,  26796@debbugs.gnu.org,  npostavs@users.sourceforge.net
> > Date: Mon, 08 May 2017 11:59:52 -0400
> > 
> > Eli Zaretskii wrote:
> > 
> > >> Was there any reason to implement it in the first place?
> > >
> > > Seems we are back at the discussion we had starting here:
> > >
> > >   http://lists.gnu.org/archive/html/emacs-devel/2017-04/msg00219.html
> > 
> > In which, if I read right, literally no-one was in favour of the feature?
> > A feature that several people have now spent time debugging, tweaking,
> > and documenting, and which will have to be supported going forwards.
> 
> Is this a complaint about something I did?  Just wondering.

I'm just here to complain that everything stemming from this change,
including the bikeshedding about if and how filenames in the message
should be quoted (which is just more waste of space) hasn't been
reverted yet.

bug#26796: 26.0.50; (2 characters) is usually wrong

[npostavs]

unarchive 354
tags 354 = wontfix
quit

Glenn Morris <rgm@gnu.org> writes:

> Eli Zaretskii wrote:
>
>>> Was there any reason to implement it in the first place?
>>
>> Seems we are back at the discussion we had starting here:
>>
>>   http://lists.gnu.org/archive/html/emacs-devel/2017-04/msg00219.html
>
> In which, if I read right, literally no-one was in favour of the feature?
> A feature that several people have now spent time debugging, tweaking,
> and documenting, and which will have to be supported going forwards.

My only defence is that everyone seemed to think it was a good idea
*before* I merged it.  In hindsight I should have reverted this much
sooner, but anyway, it's done now [1: daaec72a82].

[1: daaec72a82]: 2017-05-08 23:28:32 -0400
  Revert "Output number of characters added to file (Bug#354)"
  http://git.savannah.gnu.org/cgit/emacs.git/commit/?id=daaec72a82e76f916e639acb51a8ad602433e8a9