bug#25181: 25.1.90; move-beginning-of-line doesn't move point

bug#25181: 25.1.90; move-beginning-of-line doesn't move point

[Andreas Röhler]

Bug does not appear with Emacs -Q, as some fontification in buffer seems 
to trigger it.

Text in buffer:

git reset HEAD test/2.sh

Cursor at pos 16. i.e. beginning of word "test".

C-a doesn't move point.


Watching move-beginning-of-line edebug, at line 6415

     (skip-chars-backward "^\n")

jumps correctly to BOL.

Error results from following later on:

     (setq first-vis-field-value
       (constrain-to-field first-vis orig (/= arg 1) t nil))

Here first-vis-field-value is set to orig, i.e. pos 16, where it jumps 
back-to, not leaving point at BOL.

When C-a is called at or from behind word "test", it stops at pos 16, 
when called before it reaches BOL


In GNU Emacs 25.1.90.1 (i686-pc-linux-gnu, GTK+ Version 3.14.5)
  of 2016-11-29
Windowing system distributor 'The X.Org Foundation', version 11.0.11604000
System Description:    Debian GNU/Linux 8.6 (jessie)

Configured using:
  'configure --with-modules'

Configured features:
XPM JPEG TIFF GIF PNG RSVG SOUND GPM DBUS GSETTINGS NOTIFY ACL GNUTLS
LIBXML2 FREETYPE M17N_FLT LIBOTF XFT ZLIB TOOLKIT_SCROLL_BARS GTK3 X11
MODULES

Important settings:
   value of $LANG: de_DE.utf8
   value of $XMODIFIERS: @im=ibus
   locale-coding-system: utf-8-unix

Major mode: Shell-script

Minor modes in effect:
   shell-dirtrack-mode: t
   diff-auto-refine-mode: t
   highlight-changes-visible-mode: t
   show-paren-mode: t
   save-place-mode: t
   override-global-mode: t
   sh-electric-here-document-mode: t
   winner-mode: t
   tooltip-mode: t
   mouse-wheel-mode: t
   menu-bar-mode: t
   file-name-shadow-mode: t
   global-font-lock-mode: t
   font-lock-mode: t
   blink-cursor-mode: t
   auto-composition-mode: t
   auto-encryption-mode: t
   auto-compression-mode: t
   line-number-mode: t
   transient-mark-mode: t

Recent messages:

Result: 16 (#o20, #x10, ?\C-p)
Pos. 263500 als register "a" [2 times]
Back to top level [2 times]
move-beginning-of-line
Mark set
Pos. 263500 als register "a"
Mark set
nil
move-beginning-of-line

bug#25181: 25.1.90; move-beginning-of-line doesn't move point

[Eli Zaretskii]

> From: Andreas Röhler <andreas.roehler@easy-emacs.de>
> Date: Mon, 12 Dec 2016 12:49:05 +0100
> 
> Bug does not appear with Emacs -Q, as some fontification in buffer seems 
> to trigger it.
> 
> Text in buffer:
> 
> git reset HEAD test/2.sh
> 
> Cursor at pos 16. i.e. beginning of word "test".
> 
> C-a doesn't move point.

Please show a complete recipe, starting from "emacs -Q" (and loading
any optional packages/customizations as needed).  It's hard to discuss
an issue without being able to test-drive it.

> Watching move-beginning-of-line edebug, at line 6415
> 
>      (skip-chars-backward "^\n")
> 
> jumps correctly to BOL.
> 
> Error results from following later on:
> 
>      (setq first-vis-field-value
>        (constrain-to-field first-vis orig (/= arg 1) t nil))
> 
> Here first-vis-field-value is set to orig, i.e. pos 16, where it jumps 
> back-to, not leaving point at BOL.

I'm not sure I understood, but if I did, this is a feature: functions
such as beginning-of-line stop moving at field boundaries.

Once again, if you show an actual example, it will be much easier to
decide whether there's a bug here, and if so, what is it.

Thanks.

bug#25181: 25.1.90; move-beginning-of-line doesn't move point

[Andreas Röhler]

[-- Attachment #1: Type: text/plain, Size: 823 bytes --]



On 12.12.2016 17:53, Eli Zaretskii wrote:
>> From: Andreas Röhler <andreas.roehler@easy-emacs.de>
>> Date: Mon, 12 Dec 2016 12:49:05 +0100
>>
>> Bug does not appear with Emacs -Q, as some fontification in buffer seems
>> to trigger it.
>>
>> Text in buffer:
>>
>> git reset HEAD test/2.sh
>>
>> Cursor at pos 16. i.e. beginning of word "test".
>>
>> C-a doesn't move point.
> Please show a complete recipe, starting from "emacs -Q" (and loading
> any optional packages/customizations as needed).  It's hard to discuss
> an issue without being able to test-drive it.

Unfortunately didn't found a way from "emacs -Q" yet.
Attach a png displaying the property at stopping point.

BTW copied the text after stop-column --i.e. after "git reset HEAD"-- 
from a shell-buffer.
Maybe the properties have been taken from there?


[-- Attachment #2: prop.png --]
[-- Type: image/png, Size: 125103 bytes --]

bug#25181: 25.1.90; move-beginning-of-line doesn't move point

[Andreas Röhler]

On 12.12.2016 17:53, Eli Zaretskii wrote:
>> From: Andreas Röhler <andreas.roehler@easy-emacs.de>
>> Date: Mon, 12 Dec 2016 12:49:05 +0100
>>
>> Bug does not appear with Emacs -Q, as some fontification in buffer seems
>> to trigger it.
>>
>> Text in buffer:
>>
>> git reset HEAD test/2.sh
>>
>> Cursor at pos 16. i.e. beginning of word "test".
>>
>> C-a doesn't move point.
> Please show a complete recipe, starting from "emacs -Q" (and loading
> any optional packages/customizations as needed).  It's hard to discuss
> an issue without being able to test-drive it.
>
>> Watching move-beginning-of-line edebug, at line 6415
>>
>>       (skip-chars-backward "^\n")
>>
>> jumps correctly to BOL.
>>
>> Error results from following later on:
>>
>>       (setq first-vis-field-value
>>         (constrain-to-field first-vis orig (/= arg 1) t nil))
>>
>> Here first-vis-field-value is set to orig, i.e. pos 16, where it jumps
>> back-to, not leaving point at BOL.
> I'm not sure I understood, but if I did, this is a feature: functions
> such as beginning-of-line stop moving at field boundaries.
>

Hmm, is this reasonable? BOL is a very basic concept. Don't think it 
should be permitted to be interfered with fields.

bug#25181: 25.1.90; move-beginning-of-line doesn't move point

[Eli Zaretskii]

> Cc: 25181@debbugs.gnu.org
> From: Andreas Röhler <andreas.roehler@easy-emacs.de>
> Date: Tue, 13 Dec 2016 10:25:32 +0100
> 
> >>       (setq first-vis-field-value
> >>         (constrain-to-field first-vis orig (/= arg 1) t nil))
> >>
> >> Here first-vis-field-value is set to orig, i.e. pos 16, where it jumps
> >> back-to, not leaving point at BOL.
> > I'm not sure I understood, but if I did, this is a feature: functions
> > such as beginning-of-line stop moving at field boundaries.
> >
> 
> Hmm, is this reasonable? BOL is a very basic concept. Don't think it 
> should be permitted to be interfered with fields.

The idea is that a Lisp program would use fields only when the
produced effect is what the user would expect.  You see this every day
in action when you type "C-x C-f" and then C-a: point stops at the
beginning of the default directory, without entering the "Find file:"
prompt.  Moving point to BOL in this case would be a nuisance.  (You
can still get to the beginning of line by other means, like C-b.)

That's what fields are all about.  If you don't want this effect,
don't use fields; they are opt-in.

bug#25181: 25.1.90; move-beginning-of-line doesn't move point

[Andreas Röhler]

On 13.12.2016 17:05, Eli Zaretskii wrote:
>>
>> Hmm, is this reasonable? BOL is a very basic concept. Don't think it
>> should be permitted to be interfered with fields.
> The idea is that a Lisp program would use fields only when the
> produced effect is what the user would expect.  You see this every day
> in action when you type "C-x C-f" and then C-a: point stops at the
> beginning of the default directory, without entering the "Find file:"
> prompt.  Moving point to BOL in this case would be a nuisance.  (You
> can still get to the beginning of line by other means, like C-b.)
>
> That's what fields are all about.  If you don't want this effect,
> don't use fields; they are opt-in.

Felt rather having been  ad-opted ;) Still think there is a bug - might 
be a seldom one.

bug#25181: 25.1.90; move-beginning-of-line doesn't move point

[Eli Zaretskii]

> Cc: 25181@debbugs.gnu.org
> From: Andreas Röhler <andreas.roehler@easy-emacs.de>
> Date: Tue, 13 Dec 2016 18:55:10 +0100
> 
> Still think there is a bug - might be a seldom one.

I didn't say I disagreed, but I do need to see a recipe to look into
this any further.

bug#25181: 25.1.90; move-beginning-of-line doesn't move point

[Eli Zaretskii]

> Cc: 25181@debbugs.gnu.org
> From: Andreas Röhler <andreas.roehler@easy-emacs.de>
> Date: Tue, 13 Dec 2016 21:10:46 +0100
> 
> I'd say the bug-source is visible by design.

If it's by design, then it isn't a bug.

bug#25181: 25.1.90; move-beginning-of-line doesn't move point

[Andreas Röhler]

On 13.12.2016 19:21, Eli Zaretskii wrote:
>> Cc: 25181@debbugs.gnu.org
>> From: Andreas Röhler <andreas.roehler@easy-emacs.de>
>> Date: Tue, 13 Dec 2016 18:55:10 +0100
>>
>> Still think there is a bug - might be a seldom one.
> I didn't say I disagreed, but I do need to see a recipe to look into
> this any further.

I'd say the bug-source is visible by design. Such things like mentioned 
by docu:

"To ignore intangibility, bind `inhibit-point-motion-hooks' to t" IMHO 
don't belong onto such a general command.

Adapting C-a to intangibility or related things should be up to the 
mode. WRT to your example find-file, probably it would be at the 
minibuffer to care for an appropriate C-a then.