bug#19334: 25.0.50; Display bug with zero width fringe

bug#19334: 25.0.50; Display bug with zero width fringe

[Helmut Eller]

[-- Attachment #1: Type: text/plain, Size: 565 bytes --]

When I start emacs with:

  emacs -Q -eval '(insert-char ?x 80)'

then, as expected, all the Xes are displayed in one line and the cursor
is displayed in the right fringe.

But when I start it with:

  emacs -Q -eval '(set-fringe-mode (cons 0 8))' -eval '(insert-char ?x 80)'

then two lines are used, the last column of the first line is empty, an
arrow appears in the right fringe, and the cursor is displayed in the
second column of the second line.  See attached screenshot.

The second case should be displayed almost like the first with all Xes
in one line.


[-- Attachment #2: screenshot.png --]
[-- Type: image/png, Size: 10855 bytes --]

[-- Attachment #3: Type: text/plain, Size: 447 bytes --]


In GNU Emacs 25.0.50.10 (i686-pc-linux-gnu, GTK+ Version 2.24.10)
 of 2014-12-10 on ix
Repository revision: 99be75085cec471fa35a811bddaf09fe91fc3452
Windowing system distributor `The X.Org Foundation', version 11.0.11204000
System Description:	Debian GNU/Linux 7.1 (wheezy)

Configured using:
 `configure --with-jpeg=no --with-gif=no --with-tiff=no'

Configured features:
XPM PNG RSVG SOUND DBUS GSETTINGS NOTIFY GNUTLS LIBXML2 FREETYPE XFT
ZLIB

bug#19334: 25.0.50; Display bug with zero width fringe

[Eli Zaretskii]

> From: Helmut Eller <eller.helmut@gmail.com>
> Date: Wed, 10 Dec 2014 09:31:37 +0100
> 
> When I start emacs with:
> 
>   emacs -Q -eval '(insert-char ?x 80)'
> 
> then, as expected, all the Xes are displayed in one line and the cursor
> is displayed in the right fringe.
> 
> But when I start it with:
> 
>   emacs -Q -eval '(set-fringe-mode (cons 0 8))' -eval '(insert-char ?x 80)'
> 
> then two lines are used, the last column of the first line is empty, an
> arrow appears in the right fringe, and the cursor is displayed in the
> second column of the second line.  See attached screenshot.

It's a feature: one column is reserved for R2L lines, which will have
to show the continuation glyph on the left, because there's no fringe
there.

If you don't like this, try this alternative:

  emacs -Q -eval '(set-fringe-mode (cons 1 8))' -eval '(insert-char ?x 80)'

That is, give Emacs 1-pixel fringe on the left.

Or enlarge the frame by one more column.

bug#19334: 25.0.50; Display bug with zero width fringe

[Helmut Eller]

On Wed, Dec 10 2014, Eli Zaretskii wrote:

>> then two lines are used, the last column of the first line is empty, an
>> arrow appears in the right fringe, and the cursor is displayed in the
>> second column of the second line.  See attached screenshot.
>
> It's a feature: one column is reserved for R2L lines, which will have
> to show the continuation glyph on the left, because there's no fringe
> there.

Apparently not a "pay as you go" feature as I never use R2L lines.

> If you don't like this, try this alternative:
>
>   emacs -Q -eval '(set-fringe-mode (cons 1 8))' -eval '(insert-char ?x 80)'

Not great but that'll do.

Helmut

bug#19334: 25.0.50; Display bug with zero width fringe

[Eli Zaretskii]

> From: Helmut Eller <eller.helmut@gmail.com>
> Cc: 19334@debbugs.gnu.org
> Date: Wed, 10 Dec 2014 17:45:14 +0100
> 
> On Wed, Dec 10 2014, Eli Zaretskii wrote:
> 
> >> then two lines are used, the last column of the first line is empty, an
> >> arrow appears in the right fringe, and the cursor is displayed in the
> >> second column of the second line.  See attached screenshot.
> >
> > It's a feature: one column is reserved for R2L lines, which will have
> > to show the continuation glyph on the left, because there's no fringe
> > there.
> 
> Apparently not a "pay as you go" feature as I never use R2L lines.

Alas, we don't have a mechanism for adapting the geometry of screen
lines dynamically.  It is calculated when the display of a window
begins, and then used all the way until redisplay of that window is
completed.

If people are annoyed by this "payment", we could perhaps have a
defcustom that would disable the usurpation of that column, but then
people shouldn't complain if/when they see strange effects in R2L
lines (which can happen even if you don't expect them).

bug#19334: 25.0.50; Display bug with zero width fringe

[Lars Ingebrigtsen]

Eli Zaretskii <eliz@gnu.org> writes:

> If people are annoyed by this "payment", we could perhaps have a
> defcustom that would disable the usurpation of that column, but then
> people shouldn't complain if/when they see strange effects in R2L
> lines (which can happen even if you don't expect them).

I think the conclusion here was that this was working as designed, and
there doesn't seem to be much of a demand for allowing this to be
customised further, so I'm closing this bug report.

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no