Re: insert gives args out of range

[Stephen Berman <stephen.berman@gmx.net>]

On Fri, 16 Aug 2024 15:21:12 +0000 Heime <heimeborgia@protonmail.com> wrote:

> On Saturday, August 17th, 2024 at 3:16 AM, Stephen Berman
> <stephen.berman@gmx.net> wrote:
>
>> On Fri, 16 Aug 2024 15:09:55 +0000 Heime heimeborgia@protonmail.com wrote:
>>
>> > On Saturday, August 17th, 2024 at 3:01 AM, Heime
>> > heimeborgia@protonmail.com wrote:
>> >
>> > > On Saturday, August 17th, 2024 at 12:26 AM, Stephen Berman
>> > > stephen.berman@gmx.net wrote:
>> > >
>> > > > On Fri, 16 Aug 2024 12:10:48 +0000 Heime heimeborgia@protonmail.com wrote:
>> > > >
>> > > > > On Saturday, August 17th, 2024 at 12:03 AM, Heime
>> > > > > heimeborgia@protonmail.com wrote:
>> > > > >
>> > > > > > On Friday, August 16th, 2024 at 11:51 PM, Stephen Berman
>> > > > > > stephen.berman@gmx.net wrote:
>> > > > > >
>> > > > > > > On Fri, 16 Aug 2024 10:59:09 +0000 Heime
>> > > > > > > heimeborgia@protonmail.com wrote:
>> > > > > > >
>> > > > > > > > On Friday, August 16th, 2024 at 10:20 PM, Heime
>> > > > > > > > heimeborgia@protonmail.com wrote:
>> > > > > > > >
>> > > > > > > > > On Friday, August 16th, 2024 at 9:54 PM, Stephen Berman
>> > > > > > > > > stephen.berman@gmx.net wrote:
>> > > > > > > > >
>> > > > > > > > > > On Fri, 16 Aug 2024 09:24:23 +0000 Heime
>> > > > > > > > > > heimeborgia@protonmail.com wrote:
>> > > > > > > > > >
>> > > > > > > > > > > I want to display text starting from "(defun" up to the
>> > > > > > > > > > > second blank
>> > > > > > > > > > > line that is encountered.
>> > > > > > > > > > >
>> > > > > > > > > > > But with this code I get
>> > > > > > > > > > >
>> > > > > > > > > > > insert: Args out of range: #<buffer Defun Sections>, 2706, 2804
>> > > > > > > > > > >
>> > > > > > > > > > > (defun display-text ()
>> > > > > > > > > > > "Display the text between a line starting with '(defun' and
>> > > > > > > > > > > the second
>> > > > > > > > > > > blank line."
>> > > > > > > > > > > (interactive)
>> > > > > > > > > > > (let ((output-buffer (get-buffer-create "Defun Sections")))
>> > > > > > > > > > > (with-current-buffer output-buffer
>> > > > > > > > > > > (erase-buffer)) ;; Clear previous contents
>> > > > > > > > > > > (save-excursion
>> > > > > > > > > > > (goto-char (point-min))
>> > > > > > > > > > > (while (re-search-forward "^(defun" nil t)
>> > > > > > > > > > > (let ((start (match-beginning 0))
>> > > > > > > > > > > (blank-lines 0)
>> > > > > > > > > > > end)
>> > > > > > > > > > > ;; Move forward to find the second blank line
>> > > > > > > > > > > (while (and (< blank-lines 2)
>> > > > > > > > > > > (re-search-forward "^\\s-*$" nil t))
>> > > > > > > > > > > (setq blank-lines (1+ blank-lines)))
>> > > > > > > > > > > (setq end (point))
>> > > > > > > > > > > ;; Ensure the range is valid before inserting
>> > > > > > > > > > > (when (> end start)
>> > > > > > > > > > > (with-current-buffer output-buffer
>> > > > > > > > > > > (insert (buffer-substring-no-properties start end) "\n"))))))
>> > > > > > > > > > > ;; Display the output buffer
>> > > > > > > > > > > (display-buffer output-buffer)))
>> > > > > > > > > >
>> > > > > > > > > > You've set `start' and` end' in a different buffer than
>> > > > > > > > > > output-buffer
>> > > > > > > > > > but you're invoking buffer-substring-no-properties with these
>> > > > > > > > > > values in
>> > > > > > > > > > output-buffer, which you've erased, so those values are not in
>> > > > > > > > > > the range
>> > > > > > > > > > of possible values in output-buffer. Try let-binding the
>> > > > > > > > > > result of
>> > > > > > > > > > invoking buffer-substring-no-properties before the second
>> > > > > > > > > > invocation of
>> > > > > > > > > > with-current-buffer.
>> > > > > > > > > >
>> > > > > > > > > > Steve Berman
>> > > > > > > >
>> > > > > > > > I updated as suggested. Although I get the output, the search
>> > > > > > > > stops on
>> > > > > > > > the first
>> > > > > > > > blank line rather than on the second blank line.
>> > > > > > >
>> > > > > > > Consider a buffer "test" with the following content (two blank lines
>> > > > > > > between the defuns):
>> > > > > > >
>> > > > > > > ;;;;;;;;;;;;;
>> > > > > > >
>> > > > > > > (defun f1 ()
>> > > > > > > (ignore))
>> > > > > > >
>> > > > > > > (defun f2 ()
>> > > > > > > (ignore))
>> > > > > > > ;;;;;;;;;;;;;
>> > > > > > >
>> > > > > > > When you call `display-text' in "test", the first evaluation of the
>> > > > > > > sexp (re-search-forward "^\\s-*$" nil t) moves point to the first blank
>> > > > > > > line below the first defun. Where is point after the second evaluation
>> > > > > > > of that sexp? - Steve Berman
>> > > > > >
>> > > > > > It is still at the beginning of the first blank line.
>> > > > >
>> > > > > I do not understand why this happens. Because (point) after
>> > > > > re-search-forward should give me the position at the beginning of the
>> > > > > second blank line that is encountered.
>> > > >
>> > > > The sexp (re-search-forward "^\\s-*$" nil t) succeeds if it find a line
>> > > > containing only zero or more whitespace characters and moves point to
>> > > > the end of the match. If point is on a blank line, the search succeeds,
>> > > > but since the beginning and end of the match coincide, point does not
>> > > > move. If you want point to move, your code needs to accommodate this
>> > > > case. - Steve Berman
>> > >
>> > > Have printed start and end just after (setq end (point))
>> > >
>> > > (message "Start: %d, End: %d" start end)
>> > >
>> > > And the values for start and end are different.
>>
>>
>> `start' holds the value of the beginning of the match for "^(defun", but
>> your problem arises when searching for "^\\s-*$"; you'll see it if you
>> change the message to this:
>>
>> (message "Start: %d, End: %d" (match-beginning 0) (match-end 0))
>>
>> > When the first blank is reached, point moves to that blank, the next search
>> > keeps point at the same place rather than moving to the next blank line.
>>
>>
>> For the reason I explained above. - Steve Berman
>
> I have solved the problem by adding (forward-line 1) after setting blank-lines
>
>           (while (and (< blank-lines 2)
>                       (re-search-forward "^\\s-*$" nil t))
>             (setq blank-lines (1+ blank-lines))
>             ;; Move point forward to avoid re-matching the same blank line
>             (forward-line 1))
>
> Could this fail on some circumstances ?

Depends on what you consider failure: with the buffer "*test*" I gave as
an example above, output-buffer contains not just two but three empty
lines between the two defuns, and not just one but two empty lines at
the end; is that what you want?

Steve Berman