bug#6687: [24.0.50] end-of-buffer cursor position with an overlay

bug#6687: [24.0.50] end-of-buffer cursor position with an overlay

[Tetsuo Tsukamoto]

Hi,

Emacs 24.0.50 displays end-of-buffer cursor at different position
from where Emacs 23 does when there is an overlay at eob with
before- or after-string.

Emacs 24.0.50:

o  $ emacs -Q
   Then point is at eob.

o  M-: (overlay-put (make-overlay (point) (point)) 'before-string "?")
   Then cursor is placed on "?".
   (cursor position is indicated with [ and ])

   0) [?]

o  Typing x (or any other char) twice results in the following
   behavior.

   1) [?]x   2) ?xx[]

Emacs 24.0.50, point not at eob:

   0) ?[]   1) ?x[]    2) ?xx[]

Emacs 23.2, point at eob or not at eob:

   0) ?[]   1) ?x[]    2) ?xx[]

So only Emacs 24.0.50 with point at eob displays cursor
differently relative to an overlay placed at eob.

-- 
Tetsuo Tsukamoto

bug#6687: [24.0.50] end-of-buffer cursor position with an overlay

[Tetsuo Tsukamoto]

With further testings, I noticed that the condition where problem
is reproduced is not (eobp) but (and (bolp) (eolp)), i.e. when
point is at an empty line.

In addition, I should correct the conclusion of my first message as
follows.

Emacs 24.0.50 with point at an empty line displays cursor
differently relative to an overlay placed at the point.


In GNU Emacs 24.0.50.1 (x86_64-unknown-linux-gnu, GTK+ Version 2.20.1)
 of 2010-07-21 on t400s.localdomain
Windowing system distributor `Fedora Project', version 11.0.10802000
Important settings:
  value of $LC_ALL: nil
  value of $LC_COLLATE: nil
  value of $LC_CTYPE: nil
  value of $LC_MESSAGES: nil
  value of $LC_MONETARY: nil
  value of $LC_NUMERIC: nil
  value of $LC_TIME: nil
  value of $LANG: ja_JP.utf8
  value of $XMODIFIERS: @im=none
  locale-coding-system: utf-8
  default enable-multibyte-characters: t

Major mode: Lisp Interaction

-- 
Tetsuo Tsukamoto

bug#6687: [24.0.50] end-of-buffer cursor position with an overlay

[Tetsuo Tsukamoto]

Sorry for many postings,

With additional testings I found the following relationships
between point positions and the cursor behaviors under Emacs
24.0.50.

Invoked Emacs with emacs -Q.  After I moved the point so as to meet
one of the four conditions mentioned in results, I did the
following.

0) M-: (overlay-put (make-overlay (point) (point)) 'before-string "?")
1) Type x.
2) Type additional x.

Results were as follows (cursor position is indicated with [ and ]).

 initial point position  :        cursor behavior
(and (eobp) (eolp))      :  0) [?]   1) [?]x   2) ?xx[]
(and (not (eobp)) (eolp)):  0) [?]   1) ?x[]   2) ?xx[]
(and (eobp) (not (eolp))):  0) ?[]   1) ?x[]   2) ?xx[]
(not (or (eobp) (eolp))) :  0) ?[]   1) ?x[]   2) ?xx[]

Suppose 1st result may not be expected.  2nd result is also
different from what Emacs 23.2 produces. 3rd and 4th results seem
OK.


-- 
Tetsuo Tsukamoto

bug#6687: [24.0.50] end-of-buffer cursor position with an overlay

[Tetsuo Tsukamoto]

Each (eobp) that appear in my previous message should be replaced
with (bolp), as follows.

 initial point position  :        cursor behavior
(and (bolp) (eolp))      :  0) [?]   1) [?]x   2) ?xx[]
(and (not (bolp)) (eolp)):  0) [?]   1) ?x[]   2) ?xx[]
(and (bolp) (not (eolp))):  0) ?[]   1) ?x[]   2) ?xx[]
(not (or (bolp) (eolp))) :  0) ?[]   1) ?x[]   2) ?xx[]

-- 
Tetsuo Tsukamoto

bug#6687: [24.0.50] end-of-buffer cursor position with an overlay

[Eli Zaretskii]

> From: Tetsuo Tsukamoto <czkmt@ims.u-tokyo.ac.jp>
> Date: Thu, 22 Jul 2010 01:25:46 +0900
> Cc: 
> 
> Invoked Emacs with emacs -Q.  After I moved the point so as to meet
> one of the four conditions mentioned in results, I did the
> following.
> 
> 0) M-: (overlay-put (make-overlay (point) (point)) 'before-string "?")
> 1) Type x.
> 2) Type additional x.
> 
> Results were as follows (cursor position is indicated with [ and ]).
> 
>  initial point position  :        cursor behavior
> (and (eobp) (eolp))      :  0) [?]   1) [?]x   2) ?xx[]
> (and (not (eobp)) (eolp)):  0) [?]   1) ?x[]   2) ?xx[]
> (and (eobp) (not (eolp))):  0) ?[]   1) ?x[]   2) ?xx[]
> (not (or (eobp) (eolp))) :  0) ?[]   1) ?x[]   2) ?xx[]
> 
> Suppose 1st result may not be expected.  2nd result is also
> different from what Emacs 23.2 produces. 3rd and 4th results seem
> OK.

Sorry for such a long delay, but I finally fixed this (in revision
102440).

Thanks for reporting this, and for a detailed test case.