bug#4094: 23.1.50; posn-at-point gets upset by overlays with 'invisible and 'display

bug#4094: 23.1.50; posn-at-point gets upset by overlays with 'invisible and 'display

[Lennart Borgman]

Please write in English if possible, because the Emacs maintainers
usually do not have translators to read other languages for them.

Your bug report will be posted to the emacs-pretest-bug@gnu.org mailing list.

Please describe exactly what actions triggered the bug
and the precise symptoms of the bug:

>>>>>>>>>>>>>>>>>>>>
posn-at-point reports bad (X . Y) when on a 'displa part of an invisible
overlay.

This makes line-move-visual behave strange. (A workaround for this could
be to step back one char in line-move-visual, but this is of course not
the correct fix.)
<<<<<<<<<<<<<<<<<<<<

If Emacs crashed, and you have the Emacs process in the gdb debugger,
please include the output from the following gdb commands:
    `bt full' and `xbacktrace'.
If you would like to further debug the crash, please read the file
c:/emacs/u/090801/emacs/etc/DEBUG for instructions.


In GNU Emacs 23.1.50.1 (i386-mingw-nt5.1.2600)
 of 2009-08-01
Windowing system distributor `Microsoft Corp.', version 5.1.2600
configured using `configure --with-gcc (3.4) --no-opt --cflags
-Ic:/g/include -fno-crossjumping'


PS: Is it by intention that the bug mail message now includes the
information for the submitter? It did not work that way before. I have
left that information in the message above (but marked my real
message) so you can see what I mean. Perhaps this has something to do
with that I am using w32 where the message is copied to clipboard?

bug#4096: 23.1.50; forward-char sometimes steps over overlay 'display

[Lennart Borgman]

I have two rather similar 'display strings for two invisible
overlays. In one case forward-char steps over the 'display string part
and in the other it stops on it (as it should).

For the case when it steps over the display string looks something like

#("+ <p>...(6 hidden lines)" 0 1
  (face mode-line)
  2 3
  (fontified t)
  3 4
  (face font-lock-function-name-face fontified t)
  4 5
  (fontified t))

and for the case where it stops (expected behaviour) it looks like

#("+...(155 hidden lines)" 0 1
  (face mode-line))

I can reproduce it, but have not made a test case. However the behaviour
can easily be seen in the devel sources for nXhtml. One case is for
outline-minor-mode and one is for hs-minor-mode used in a nxhtml-mode buffer.


In GNU Emacs 23.1.50.1 (i386-mingw-nt5.1.2600)
 of 2009-08-01
Windowing system distributor `Microsoft Corp.', version 5.1.2600
configured using `configure --with-gcc (3.4) --no-opt --cflags
-Ic:/g/include -fno-crossjumping'

bug#4094: 23.1.50; posn-at-point gets upset by overlays with 'invisible and 'display

[Lennart Borgman]

Some more thoughts: There seem to be several small inconsistencies
around here. I guess the intended behaviour is

For invisible overlays (and dito buffer text, but it is overlays I
have been looking at now):

- The "stop" point on the display string property should be the first
character. That seems to be the case now, but often the first char is
missed/jumped over, see below.

- When calling forward-char or backward char they should stop there.
Forward char often does that, but not always. Backward char does not
seem to do that.

- Similar problem with vertical-motion. It often misses the first char
and goes to the column after the display string instead.

bug#4094: 23.1.50; posn-at-point gets upset by overlays with 'invisible and 'display

[Lennart Borgman]

Forgot to say that this seems to be some problem with the interaction
with the display engine and display update. When you try to debug this
the problem is sometimes gone.


On Sun, Aug 9, 2009 at 9:38 PM, Lennart
Borgman<lennart.borgman@gmail.com> wrote:
> Some more thoughts: There seem to be several small inconsistencies
> around here. I guess the intended behaviour is
>
> For invisible overlays (and dito buffer text, but it is overlays I
> have been looking at now):
>
> - The "stop" point on the display string property should be the first
> character. That seems to be the case now, but often the first char is
> missed/jumped over, see below.
>
> - When calling forward-char or backward char they should stop there.
> Forward char often does that, but not always. Backward char does not
> seem to do that.
>
> - Similar problem with vertical-motion. It often misses the first char
> and goes to the column after the display string instead.
>

bug#4096: 23.1.50; forward-char sometimes steps over overlay 'display

[Lennart Borgman]

(Please merge this with 4094.)

I think the problem might be in

    adjust_for_invis_intang

or the use of it. This function checks for insertion type.

This is different for the overlays I have mentioned here. One is
created by outline-minor-mode and one by hs-minor-mode. One of them
set insertion type for the beginning of the overlay, the other does
not. In one of the cases at least forward-char works, in the other it
does not. (backward-char still does not work.)

However it must be a mistake to check for insertion type here. Forward
char has nothing to do with that. (Note: adjust_for_invis_intang is
called by set_point_both.)

bug#4096: 23.1.50; forward-char sometimes steps over overlay 'display

[Stefan Monnier]

> I think the problem might be in

>     adjust_for_invis_intang

> or the use of it. This function checks for insertion type.

This function should not do anything if you don't have `intangible'
things around.  Do you?  If you do, I'm tempted to say "there's your
problem".


        Stefan

bug#4096: 23.1.50; forward-char sometimes steps over overlay 'display

[Lennart Borgman]

On Mon, Aug 10, 2009 at 6:13 AM, Stefan Monnier<monnier@iro.umontreal.ca> wrote:
>> I think the problem might be in
>
>>     adjust_for_invis_intang
>
>> or the use of it. This function checks for insertion type.
>
> This function should not do anything if you don't have `intangible'
> things around.  Do you?  If you do, I'm tempted to say "there's your
> problem".

Not that I know of. I looked for it but could not find it. (I have not
checked that adjust_for_invis_intang is actually called. I looked
around in the code searching for some reasonable explanation.)

The overlays are created by outline-minor-mode and hs-minor-mode +
'display property added by me.

bug#4094: 23.1.50; posn-at-point gets upset by overlays with 'invisible and 'display

[Chong Yidong]

> posn-at-point reports bad (X . Y) when on a 'displa part of an invisible
> overlay.
>
> This makes line-move-visual behave strange. (A workaround for this could
> be to step back one char in line-move-visual, but this is of course not
> the correct fix.)

Please provide a precise, step-by-step recipe to reproduce this,
starting with `emacs -Q'.

bug#4094: 23.1.50; posn-at-point gets upset by overlays with 'invisible and 'display

[Glenn Morris]

Closed due to lack of response.

Chong Yidong wrote:

> Please provide a precise, step-by-step recipe to reproduce this,
> starting with `emacs -Q'.

bug#4096: closed (Re: bug#4094: 23.1.50; posn-at-point gets upset by overlays with 'invisible and 'display)

[Lennart Borgman]

Sorry, I do not have time to give a precise recipe.

On Sun, Jul 10, 2011 at 02:31, GNU bug Tracking System
<help-debbugs@gnu.org> wrote:
> Your bug report
>
> #4094: 23.1.50; forward-char sometimes steps over overlay 'display
>
> which was filed against the emacs package, has been closed.
>
> The explanation is attached below, along with your original report.
> If you require more details, please reply to 4096@debbugs.gnu.org.
>
> --
> 4094: http://debbugs.gnu.org/cgi/bugreport.cgi?bug=4094
> GNU Bug Tracking System
> Contact help-debbugs@gnu.org with problems
>
>
> ---------- Forwarded message ----------
> From: Glenn Morris <rgm@gnu.org>
> To: 4094-done@debbugs.gnu.org
> Date: Sat, 09 Jul 2011 20:30:08 -0400
> Subject: Re: bug#4094: 23.1.50; posn-at-point gets upset by overlays with 'invisible and 'display
>
> Closed due to lack of response.
>
> Chong Yidong wrote:
>
>> Please provide a precise, step-by-step recipe to reproduce this,
>> starting with `emacs -Q'.
>
>
>
> ---------- Forwarded message ----------
> From: Lennart Borgman <lennart.borgman@gmail.com>
> To: emacs-pretest-bug@gnu.org
> Date: Sun, 9 Aug 2009 18:08:39 +0200
> Subject: 23.1.50; forward-char sometimes steps over overlay 'display
> I have two rather similar 'display strings for two invisible
> overlays. In one case forward-char steps over the 'display string part
> and in the other it stops on it (as it should).
>
> For the case when it steps over the display string looks something like
>
> #("+ <p>...(6 hidden lines)" 0 1
>  (face mode-line)
>  2 3
>  (fontified t)
>  3 4
>  (face font-lock-function-name-face fontified t)
>  4 5
>  (fontified t))
>
> and for the case where it stops (expected behaviour) it looks like
>
> #("+...(155 hidden lines)" 0 1
>  (face mode-line))
>
> I can reproduce it, but have not made a test case. However the behaviour
> can easily be seen in the devel sources for nXhtml. One case is for
> outline-minor-mode and one is for hs-minor-mode used in a nxhtml-mode buffer.
>
>
> In GNU Emacs 23.1.50.1 (i386-mingw-nt5.1.2600)
>  of 2009-08-01
> Windowing system distributor `Microsoft Corp.', version 5.1.2600
> configured using `configure --with-gcc (3.4) --no-opt --cflags
> -Ic:/g/include -fno-crossjumping'
>
>
>