Re: do-auto-fill bug?

Re: do-auto-fill bug?

[Katsumi Yamaoka]

Hi,

I sent a bug report to the pretest-bug list toward the end of
May, however there is no response so far.  Is no one being
troubled with the problem?  The patch I included in the report
seems to have no problem so far, and I hope it will be applied.

Thanks in advance.

-------------------- Start of forwarded message --------------------
From: Katsumi Yamaoka <yamaoka@jpl.org>
To: emacs-pretest-bug@gnu.org
Subject: do-auto-fill bug?
Date: Thu, 27 May 2004 19:33:24 +0900
Message-ID: <b9ybrka9n2j.fsf@jpl.org>

Hi,

I'm being annoyed with a strange behavior in the auto-fill-mode.
When I type SPC at the end of a line and the column just equals
fill-column, breaking the line takes place at the end of the word
in front of one, not there.  Here's an example:

---------x---------x---------x---------x---------x---------x---------x
The quick brown fox jumps over the lazy dog. The quick brown fox jumps

When fill-column is 70, that line will be folded as follows after
typing SPC at the end of the line:

---------x---------x---------x---------x---------x---------x---------x
The quick brown fox jumps over the lazy dog. The quick brown fox
jumps 

I made a change temporarily like the following patch in order to
cope with the problem:

*** fill.el~	Sun Apr 18 22:21:52 2004
--- fill.el	Thu May 27 10:29:30 2004
***************
*** 333,339 ****
  and `fill-nobreak-invisible'."
    (or
     (and fill-nobreak-invisible (line-move-invisible (point)))
!    (unless (bolp)
      (or
       ;; Don't break after a period followed by just one space.
       ;; Move back to the previous place to break.
--- 333,339 ----
  and `fill-nobreak-invisible'."
    (or
     (and fill-nobreak-invisible (line-move-invisible (point)))
!    (unless (or (bolp) (looking-at "[ \t]*$"))
      (or
       ;; Don't break after a period followed by just one space.
       ;; Move back to the previous place to break.

Regards,
-- 
Katsumi Yamaoka <yamaoka@jpl.org>

-------------------- End of forwarded message --------------------

Re: do-auto-fill bug?

[Stefan]

> I sent a bug report to the pretest-bug list toward the end of
> May, however there is no response so far.  Is no one being
> troubled with the problem?  The patch I included in the report
> seems to have no problem so far, and I hope it will be applied.

I've taken a quick look a while back but haven't had the time to come up
with a good bugfix yet.  Maybe your patch (improved to also prevent
breaking in front of "[ \t]*$") is good enough,


        Stefan

Re: do-auto-fill bug?

[David Kastrup]

Stefan <monnier@iro.umontreal.ca> writes:

> > I sent a bug report to the pretest-bug list toward the end of
> > May, however there is no response so far.  Is no one being
> > troubled with the problem?  The patch I included in the report
> > seems to have no problem so far, and I hope it will be applied.
> 
> I've taken a quick look a while back but haven't had the time to come up
> with a good bugfix yet.  Maybe your patch (improved to also prevent
> breaking in front of "[ \t]*$") is good enough,

I failed to followup on the following posting on emacs-pretest-bug.
Maybe one of the fixes is better than the other?

From: Richard Stallman <rms@gnu.org>
To: David Kastrup <dak@gnu.org>
CC: emacs-pretest-bug@gnu.org
Subject: Re: auto-fill-mode and filling paragraphs differ
References:  <x5hdw31zyf.fsf@lola.goethe.zz>
Message-Id: <E1B9Hjs-0006U4-Re@fencepost.gnu.org>
Date: Fri, 02 Apr 2004 01:01:12 -0500

    I would interpret this as the last permissable column of content.
    According to C-x =, column 70 is the column _after_ fsafd.  If we now
    say "ok, at this column, the end of line is legal", it would appear
    that auto-fill-mode wraps one column too early.  I think that typing a
    space immediately at the end of line should _not_ wrap the preceding
    word to the next line, but merely replace the typed space (instead of
    the last space before that) with a newline.

I agree with you.  Does this patch give correct results?



*** fill.el.~1.170.~	Fri Mar 26 12:04:37 2004
--- fill.el	Thu Apr  1 23:12:54 2004
***************
*** 353,359 ****
       ;; Don't split a line if the rest would look like a new paragraph.
       (unless use-hard-newlines
         (save-excursion
! 	 (skip-chars-forward " \t") (looking-at paragraph-start)))
       (run-hook-with-args-until-success 'fill-nobreak-predicate)))))
  
  ;; Put `fill-find-break-point-function' property to charsets which
--- 353,364 ----
       ;; Don't split a line if the rest would look like a new paragraph.
       (unless use-hard-newlines
         (save-excursion
! 	 (skip-chars-forward " \t")
! 	 ;; If this break point is at the end of the line,
! 	 ;; which can occur for auto-fill, don't consider the newline
! 	 ;; which follows as a reason to return t.
! 	 (and (not (eolp))
! 	      (looking-at paragraph-start))))
       (run-hook-with-args-until-success 'fill-nobreak-predicate)))))
  
  ;; Put `fill-find-break-point-function' property to charsets which



-- 
David Kastrup, Kriemhildstr. 15, 44793 Bochum

Re: do-auto-fill bug?

[Katsumi Yamaoka]

>>>>> In <x5smc5r3tz.fsf@lola.goethe.zz>
>>>>>	David Kastrup <dak@gnu.org> wrote:

> Stefan <monnier@iro.umontreal.ca> writes:

>> I've taken a quick look a while back but haven't had the time to come up
>> with a good bugfix yet.  Maybe your patch (improved to also prevent
>> breaking in front of "[ \t]*$") is good enough,

> I failed to followup on the following posting on emacs-pretest-bug.
> Maybe one of the fixes is better than the other?

[...]

RMS> I agree with you.  Does this patch give correct results?

> *** fill.el.~1.170.~	Fri Mar 26 12:04:37 2004
> --- fill.el	Thu Apr  1 23:12:54 2004

The patch also fulfills my purpose.  I've never been scrutinizing
the side effects of those fixes, but this may be better.
-- 
Katsumi Yamaoka <yamaoka@jpl.org>