Re: master cd0a795: Output number of characters added to file (Bug#354)

Re: master cd0a795: Output number of characters added to file (Bug#354)

[Mark Oteiza]

npostavs@gmail.com (Noam Postavsky) writes:

> branch: master
> commit cd0a7951fb762570c4b9c2cdbb43c94cb68910d4
> Author: Jeff Clough <kb1vqh@gmail.com>
> Commit: Noam Postavsky <npostavs@gmail.com>
>
>     Output number of characters added to file (Bug#354)
>     
>     * fileio.c (write_region):
>     * epa-file.el (epa-file-write-region):
>     * jka-compr.el (jka-compr-write-region):
>     * ange-ftp.el (ange-ftp-write-region):
>     * tramp-gvfs.el (tramp-gvfs-handle-write-region):
>     * tramp-sh.el (tramp-sh-handle-write-region):
>     * mm-util.el (mm-append-to-file): Functions now output
>     characters written in addition to file name.
>     * files.texi: Added documentation to write-region and
>     append-to-file describing their output.

This behavior is unnecessarily verbose, at least when just interactively
saving buffers.  Haven't tested this otherwise.  If I cared about the
number of characters in the buffer I'd look in the mode line or just do
C-x =.

I can't say I've ever been surprised as described in the original
report.

Re: master cd0a795: Output number of characters added to file (Bug#354)

[Eli Zaretskii]

> From: Mark Oteiza <mvoteiza@udel.edu>
> Date: Thu, 06 Apr 2017 23:30:41 -0400
> Cc: Jeff Clough <kb1vqh@gmail.com>, Noam Postavsky <npostavs@gmail.com>
> 
> >     Output number of characters added to file (Bug#354)
> >     
> >     * fileio.c (write_region):
> >     * epa-file.el (epa-file-write-region):
> >     * jka-compr.el (jka-compr-write-region):
> >     * ange-ftp.el (ange-ftp-write-region):
> >     * tramp-gvfs.el (tramp-gvfs-handle-write-region):
> >     * tramp-sh.el (tramp-sh-handle-write-region):
> >     * mm-util.el (mm-append-to-file): Functions now output
> >     characters written in addition to file name.
> >     * files.texi: Added documentation to write-region and
> >     append-to-file describing their output.
> 
> This behavior is unnecessarily verbose, at least when just interactively
> saving buffers.  Haven't tested this otherwise.  If I cared about the
> number of characters in the buffer I'd look in the mode line or just do
> C-x =.
> 
> I can't say I've ever been surprised as described in the original
> report.

Thanks for the comments.  Would making this an optional feature take
care of your concerns?

P.S. In the future, it might be better to voice such objections while
the patch is being discussed, to avoid unnecessary reworking and
possible frustration caused by that.  TIA.

Re: master cd0a795: Output number of characters added to file (Bug#354)

[Noam Postavsky]

On Thu, Apr 6, 2017 at 11:30 PM, Mark Oteiza <mvoteiza@udel.edu> wrote:
>>
>>     Output number of characters added to file (Bug#354)
>
> This behavior is unnecessarily verbose, at least when just interactively
> saving buffers.  Haven't tested this otherwise.  If I cared about the
> number of characters in the buffer I'd look in the mode line or just do
> C-x =.

Would it help if we rearranged the message so that the number of
characters go at the end? Something like:

Wrote /foo/bar/blah (1314 characters)

Re: master cd0a795: Output number of characters added to file (Bug#354)

[Michael Albinus]

npostavs@gmail.com (Noam Postavsky) writes:

> diff --git a/lisp/net/tramp-sh.el b/lisp/net/tramp-sh.el
> index 9977162..f78f541 100644
> --- a/lisp/net/tramp-sh.el
> +++ b/lisp/net/tramp-sh.el
> @@ -3411,7 +3411,8 @@ the result will be a local, non-Tramp, file name."
>          (when need-chown
>            (tramp-set-file-uid-gid filename uid gid))
>  	(when (or (eq visit t) (null visit) (stringp visit))
> -	  (tramp-message v 0 "Wrote %s" filename))
> +          (tramp-message v 0 "Wrote %d characters to `%s'"
> +                         (- end start) filename))

This breaks Tramp tests. In `write-region', START could be also nil or a
string.

Likely, the other changes suffer from the same problem.

Best regards, Michael.

Re: master cd0a795: Output number of characters added to file (Bug#354)

[Glenn Morris]

> branch: master
> commit cd0a7951fb762570c4b9c2cdbb43c94cb68910d4
> Author: Jeff Clough <kb1vqh@gmail.com>
> Commit: Noam Postavsky <npostavs@gmail.com>
>
>     Output number of characters added to file (Bug#354)

This change causes files-test-bug-18141 to fail.

Eg http://hydra.nixos.org/build/51309278

Re: master cd0a795: Output number of characters added to file (Bug#354)

[Glenn Morris]

PS you can see the same problem simply by trying to save any compressed file.

Re: master cd0a795: Output number of characters added to file (Bug#354)

[Noam Postavsky]

On Fri, Apr 7, 2017 at 9:42 AM, Michael Albinus <michael.albinus@gmx.de> wrote:
>> -       (tramp-message v 0 "Wrote %s" filename))
>> +          (tramp-message v 0 "Wrote %d characters to `%s'"
>> +                         (- end start) filename))
>
> This breaks Tramp tests. In `write-region', START could be also nil or a
> string.
>
> Likely, the other changes suffer from the same problem.

Yup, as noted by Glenn, it breaks the compressed file handler too,
fixed in [1: 9e27e0d1a6].

1: 2017-04-07 18:24:35 -0400 9e27e0d1a6c90d0d29e41884155245b6c3378b0a
  Fix handling of non-integer START param to write-region

Re: master cd0a795: Output number of characters added to file (Bug#354)

[Paul Eggert]

On 04/07/2017 12:14 AM, Eli Zaretskii wrote:
> P.S. In the future, it might be better to voice such objections while
> the patch is being discussed,

All true. Still, I didn't realize how annoying the new behavior would be 
until I started using it. Formerly, when I typed C-x C-s I would 
typically get a little message in the echo area that I could ignore. Now 
the extra length often causes the message to take more than one line, 
which causes the lowest mode line to do a little dance to get out of the 
way so that the echo area can temporarily take two lines, and this is 
quite irritating.

So, having tried it, I would rather not see these character counts and 
would rather have the default be the way things were.

Re: master cd0a795: Output number of characters added to file (Bug#354)

[Kaushal Modi]

[-- Attachment #1: Type: text/plain, Size: 1026 bytes --]

On Fri, Apr 7, 2017 at 6:50 PM Paul Eggert <eggert@cs.ucla.edu> wrote:

> On 04/07/2017 12:14 AM, Eli Zaretskii wrote:
> > P.S. In the future, it might be better to voice such objections while
> > the patch is being discussed,
>
> All true. Still, I didn't realize how annoying the new behavior would be
> until I started using it. Formerly, when I typed C-x C-s I would
> typically get a little message in the echo area that I could ignore. Now
> the extra length often causes the message to take more than one line,
> which causes the lowest mode line to do a little dance to get out of the
> way so that the echo area can temporarily take two lines, and this is
> quite irritating.
>
> So, having tried it, I would rather not see these character counts and
> would rather have the default be the way things were.
>

+1! Please make this an option, defaulting to the prior state (disabled).

I never have wondered how many characters actually got written.. It just
creates a random noise in the echo area.
-- 

Kaushal Modi

[-- Attachment #2: Type: text/html, Size: 1701 bytes --]

Re: master cd0a795: Output number of characters added to file (Bug#354)

[Noam Postavsky]

On Fri, Apr 7, 2017 at 7:02 PM, Kaushal Modi <kaushal.modi@gmail.com> wrote:
> On Fri, Apr 7, 2017 at 6:50 PM Paul Eggert <eggert@cs.ucla.edu> wrote:
>>
>> On 04/07/2017 12:14 AM, Eli Zaretskii wrote:
>> > P.S. In the future, it might be better to voice such objections while
>> > the patch is being discussed,
>>
>> All true. Still, I didn't realize how annoying the new behavior would be
>> until I started using it. Formerly, when I typed C-x C-s I would
>> typically get a little message in the echo area that I could ignore. Now
>> the extra length often causes the message to take more than one line,
>> which causes the lowest mode line to do a little dance to get out of the
>> way so that the echo area can temporarily take two lines, and this is
>> quite irritating.
>>
>> So, having tried it, I would rather not see these character counts and
>> would rather have the default be the way things were.
>
>
> +1! Please make this an option, defaulting to the prior state (disabled).
>
> I never have wondered how many characters actually got written.. It just
> creates a random noise in the echo area.

Yeah, after trying it out more, I do find it more disruptive than I
expected. I think part of it is because it puts the character count at
the beginning of the message where we've gotten used to seeing the
file name. I've changed it around in [1: c323659344], if it's still
too annoying I'll look at adding an option.

1: 2017-04-07 19:54:11 -0400 c323659344ba4db7a0b074b2e29d0a5f33d5eb80
  Adjust write-region so file name is at the beginning again

Re: master cd0a795: Output number of characters added to file (Bug#354)

[Mark Oteiza]

On 07/04/17 at 10:14am, Eli Zaretskii wrote:
> > From: Mark Oteiza <mvoteiza@udel.edu>
> > Date: Thu, 06 Apr 2017 23:30:41 -0400
> > Cc: Jeff Clough <kb1vqh@gmail.com>, Noam Postavsky <npostavs@gmail.com>
> > 
> > >     Output number of characters added to file (Bug#354)
> > >     
> > >     * fileio.c (write_region):
> > >     * epa-file.el (epa-file-write-region):
> > >     * jka-compr.el (jka-compr-write-region):
> > >     * ange-ftp.el (ange-ftp-write-region):
> > >     * tramp-gvfs.el (tramp-gvfs-handle-write-region):
> > >     * tramp-sh.el (tramp-sh-handle-write-region):
> > >     * mm-util.el (mm-append-to-file): Functions now output
> > >     characters written in addition to file name.
> > >     * files.texi: Added documentation to write-region and
> > >     append-to-file describing their output.
> > 
> > This behavior is unnecessarily verbose, at least when just interactively
> > saving buffers.  Haven't tested this otherwise.  If I cared about the
> > number of characters in the buffer I'd look in the mode line or just do
> > C-x =.
> > 
> > I can't say I've ever been surprised as described in the original
> > report.
> 
> Thanks for the comments.  Would making this an optional feature take
> care of your concerns?

Indeed it would, and IMO it should default to disabled.

> P.S. In the future, it might be better to voice such objections while
> the patch is being discussed, to avoid unnecessary reworking and
> possible frustration caused by that.  TIA.

Of course; however, this only grabbed my attention seeing it after
rebuilding.

Re: master cd0a795: Output number of characters added to file (Bug#354)

[Mark Oteiza]

On 07/04/17 at 08:06am, Noam Postavsky wrote:
> On Thu, Apr 6, 2017 at 11:30 PM, Mark Oteiza <mvoteiza@udel.edu> wrote:
> >>
> >>     Output number of characters added to file (Bug#354)
> >
> > This behavior is unnecessarily verbose, at least when just interactively
> > saving buffers.  Haven't tested this otherwise.  If I cared about the
> > number of characters in the buffer I'd look in the mode line or just do
> > C-x =.
> 
> Would it help if we rearranged the message so that the number of
> characters go at the end? Something like:
> 
> Wrote /foo/bar/blah (1314 characters)

I still don't understand the purpose of this change, as I've never been
troubled by not knowing the number of characters written, and the original
report didn't outline a recipe.  So, I'm not sure I can really comment
other than suggesting a revert.

Re: master cd0a795: Output number of characters added to file (Bug#354)

[Eli Zaretskii]

> Date: Fri, 7 Apr 2017 20:17:56 -0400
> From: Mark Oteiza <mvoteiza@udel.edu>
> Cc: Jeff Clough <kb1vqh@gmail.com>, Emacs developers <emacs-devel@gnu.org>
> 
> > Wrote /foo/bar/blah (1314 characters)
> 
> I still don't understand the purpose of this change

Maybe the fact that the vi family of editors show that?  Just
guessing.

Re: master cd0a795: Output number of characters added to file (Bug#354)

[Mark Oteiza]

On 08/04/17 at 10:34am, Eli Zaretskii wrote:
> > Date: Fri, 7 Apr 2017 20:17:56 -0400
> > From: Mark Oteiza <mvoteiza@udel.edu>
> > Cc: Jeff Clough <kb1vqh@gmail.com>, Emacs developers <emacs-devel@gnu.org>
> > 
> > > Wrote /foo/bar/blah (1314 characters)
> > 
> > I still don't understand the purpose of this change
> 
> Maybe the fact that the vi family of editors show that?  Just
> guessing.

Good idea, though the report talks about some unknown issue with emacs
not writing what's expected.  The only editor where I've paid any
attention to that number is ed.

Re: master cd0a795: Output number of characters added to file (Bug#354)

[Mark Oteiza]

On 07/04/17 at 08:06am, Noam Postavsky wrote:
> On Thu, Apr 6, 2017 at 11:30 PM, Mark Oteiza <mvoteiza@udel.edu> wrote:
> >>
> >>     Output number of characters added to file (Bug#354)
> >
> > This behavior is unnecessarily verbose, at least when just interactively
> > saving buffers.  Haven't tested this otherwise.  If I cared about the
> > number of characters in the buffer I'd look in the mode line or just do
> > C-x =.
> 
> Would it help if we rearranged the message so that the number of
> characters go at the end? Something like:
> 
> Wrote /foo/bar/blah (1314 characters)

Also, why were quotes added?  More wasted space.