Re: ruler-mode.el fix following whitespace cleanup

Re: ruler-mode.el fix following whitespace cleanup

[Benjamin Riefenstahl]

Hi David,

"David PONCE" <David.Ponce@wanadoo.fr> writes:
> The following patch reintroduces the whitespace character needed
> after ?\ , that recent whitespace cleanup removed.

How about adding a comment there, so this can't happen again, like

  (defcustom ruler-mode-margins-char ?\  ;; comment to protect space char

so long, benny

Re: ruler-mode.el fix following whitespace cleanup

[Juanma Barranquero]

On 04 Feb 2003 18:09:14 +0100, Benjamin Riefenstahl <Benjamin.Riefenstahl@epost.de> wrote:

> How about adding a comment there, so this can't happen again, like
> 
>   (defcustom ruler-mode-margins-char ?\  ;; comment to protect space char

That's exactly what I've done in the two cases found so far
(ruler-mode.el and international/character.el).

                                                           /L/e/k/t/u

Re: ruler-mode.el fix following whitespace cleanup

[Werner LEMBERG]

> The following patch reintroduces the whitespace character needed after
> ?\ , that recent whitespace cleanup removed.

Is there any reason why this whitespace can't be protected with a
trailing comment character?  I strongly suggest to put that into the
`Lisp Style Manual', if such a thing exists.


    Werner


Index: ruler-mode.el
===================================================================
RCS file: /cvsroot/emacs/emacs/lisp/ruler-mode.el,v
retrieving revision 1.11
diff -c -r1.11 ruler-mode.el
*** ruler-mode.el	4 Feb 2003 12:01:41 -0000	1.11
--- ruler-mode.el	4 Feb 2003 16:36:34 -0000
***************
*** 179,185 ****
            (integer :tag "Integer char value"
                     :validate ruler-mode-character-validate)))
  
! (defcustom ruler-mode-margins-char ?\
    "*Character used in margin areas."
    :group 'ruler-mode
    :type '(choice
--- 179,185 ----
            (integer :tag "Integer char value"
                     :validate ruler-mode-character-validate)))
  
! (defcustom ruler-mode-margins-char ?\ ;
    "*Character used in margin areas."
    :group 'ruler-mode
    :type '(choice

Re: ruler-mode.el fix following whitespace cleanup

[Luc Teirlinck]

Werner Lemberg wrote:

   > The following patch reintroduces the whitespace character needed after
   > ?\ , that recent whitespace cleanup removed.

   Is there any reason why this whitespace can't be protected with a
   trailing comment character?  I strongly suggest to put that into the
   `Lisp Style Manual', if such a thing exists.

Are we absolutely sure that trailing whitespace within a string is
never meaningful?  `;' has no special meaning within a string and 
M-x delete-trailing-whitespace, as well as C-j, delete trailing
whitespace within strings, just like any other place.

Sincerely,

Luc.

Re: ruler-mode.el fix following whitespace cleanup

[Luc Teirlinck]

>From my previous message:

   Are we absolutely sure that trailing whitespace within a string is
   never meaningful?

maybe if the situation arose, one could replace:

"Beware!  Trailing whitespace follows!      
Second line."

By:

"Beware!  Trailing whitespace follows!      \

Second line."

or (might look relatively nicer):

"Beware!  Trailing whitespace follows!      \n\
Second line."

and escape the wrath of M-x delete-trailing-whitespace that way.

Sincerely,

Luc.

Re: ruler-mode.el fix following whitespace cleanup

[Luc Teirlinck]

Kim Storm wrote:

   Doesn't this approach have the risk that the space is accidentally
   replaced by a TAB?

(Meant is ?\   ;)

Yes if you leave multiple spaces there, but what if you leave only
one?  Also the risk is already there without the `;'.  If you leave
more trailing whitespace than needed, you run this risk, no matter
what.  (Unless I misunderstanding something.)

Sincerely,

Luc.

Re: ruler-mode.el fix following whitespace cleanup

[Kim F. Storm]

Juanma Barranquero <lektu@terra.es> writes:

> On 04 Feb 2003 18:09:14 +0100, Benjamin Riefenstahl <Benjamin.Riefenstahl@epost.de> wrote:
> 
> > How about adding a comment there, so this can't happen again, like
> > 
> >   (defcustom ruler-mode-margins-char ?\  ;; comment to protect space char
> 
> That's exactly what I've done in the two cases found so far
> (ruler-mode.el and international/character.el).

Doesn't this approach have the risk that the space is accidentally
replaced by a TAB?

-- 
Kim F. Storm <storm@cua.dk> http://www.cua.dk

Re: ruler-mode.el fix following whitespace cleanup

[Juanma Barranquero]

On 05 Feb 2003 00:19:39 +0100
storm@cua.dk (Kim F. Storm) wrote:

> Doesn't this approach have the risk that the space is accidentally
> replaced by a TAB?

I hadn't thought of it, because I never use tabs in my files if I can
avoid it (makefiles being the obvious exception :)

I suppose if that happens we'll have to find another way.

-- 
Juanma Barranquero <lektu@terra.es>

Re: ruler-mode.el fix following whitespace cleanup

[Luc Teirlinck]

>From my previous message:

   Yes if you leave multiple spaces there, but what if you leave only
   one?

Never mind, now I understand.  C-M-q and there we go.  Yes, this is a
real problem.

Sincerely,

Luc.

Re: ruler-mode.el fix following whitespace cleanup

[Luc Teirlinck]

>From my previous message:

   Never mind, now I understand.  C-M-q and there we go.  Yes, this is a
   real problem.

After all no, if you use ;; does the doubling up not prevent
reindentation?  Which means that ?\ ;; should be safe (or not?).  That
might put the comment in an unconventional place, but it is obvious
why it is where it is, anyway.

But I like the ?\_ suggestion too.  It would not only be clearer but
help prevent all kinds of accidents with accidental tabification.

Sincerely,

Luc.

Re: ruler-mode.el fix following whitespace cleanup

[Eli Zaretskii]

On Wed, 5 Feb 2003, Juanma Barranquero wrote:

> I never use tabs in my files if I can avoid it (makefiles being the 
> obvious exception :)

You mean, you have indent-tabs-mode set to nil?  Because if you do, any 
reindentation you do on a block of code will change the whitespace there 
to something that many developers (and the GNU standards, AFAIK) consider 
as non-canonical whitespace usage.  This could mean another round of 
gratuitous whitespace-only changes ;-)

Re: ruler-mode.el fix following whitespace cleanup

[Juanma Barranquero]

On Wed, 5 Feb 2003 08:21:33 +0200 (IST), Eli Zaretskii <eliz@is.elta.co.il> wrote:

> You mean, you have indent-tabs-mode set to nil?

Yes

> Because if you do, any 
> reindentation you do on a block of code will change the whitespace there 
> to something that many developers (and the GNU standards, AFAIK) consider 
> as non-canonical whitespace usage.

I've found nothing about spaces vs. tabs at http://www.gnu.org/prep/standards.html,
but I'm prepared to believe I've missed it :)

Anyway, I rarely, if ever, do a reident in Emacs code; so the only
places affected are the lines I directly change. I'll look for a way to
set it conditionally on the file not being under subversion's CVS
control, though.

FWIW: Of the 254,095 lines of elisp code in lisp/ and subdirs, 68,836
start with a tab (~ 27%), while 7,745 start with 8+ spaces (~ 3%). I
refuse to believe those 7,745 lines are all my work :)

> This could mean another round of 
> gratuitous whitespace-only changes ;-)

Ouch. But I'm not going to push for a widespread switch to spaces-only
files, even if I find it more sensible ;)


                                                           /L/e/k/t/u

Re: ruler-mode.el fix following whitespace cleanup

[Benjamin Riefenstahl]

Hi Kim,

storm@cua.dk (Kim F. Storm) writes:
> Doesn't this approach have the risk that the space is accidentally
> replaced by a TAB?

That issue is related, but it's a completely different problem, isn't
it?  I mean it's not introduced by the comment.

so long, benny

Re: ruler-mode.el fix following whitespace cleanup

[Luc Teirlinck]

Benjamin Riefenstahl wrote:

   storm@cua.dk (Kim F. Storm) writes:
   > Doesn't this approach have the risk that the space is accidentally
   > replaced by a TAB?

   That issue is related, but it's a completely different problem, isn't
   it?  I mean it's not introduced by the comment.

It is introduced by the comment, if you leave more than one space, or
will be introduced by re-indentation, if you use only one `;'.
I believe that ?\ ;; is safe.  (But I could be overlooking something.)
However, the tabification problem is indeed a lot broader than the
trailing ?\ -problem.

Sincerely,

Luc.

Re: ruler-mode.el fix following whitespace cleanup

[Eli Zaretskii]

> Date: Wed, 05 Feb 2003 10:37:15 +0100
> From: Juanma Barranquero <lektu@terra.es>
> 
> > Because if you do, any 
> > reindentation you do on a block of code will change the whitespace there 
> > to something that many developers (and the GNU standards, AFAIK) consider 
> > as non-canonical whitespace usage.
> 
> I've found nothing about spaces vs. tabs at
> http://www.gnu.org/prep/standards.html

I was drawing conclusions from the default Emacs behavior ;-)

I will let Richard give the definitive ruling on this.

Re: ruler-mode.el fix following whitespace cleanup

[Richard Stallman]

    Are we absolutely sure that trailing whitespace within a string is
    never meaningful?

It seems unlikely that spaces before a newline in a string
are ever meaningful.