Displaying LRM and RLM

Displaying LRM and RLM

[Eli Zaretskii]

What should Emacs 23 display for these two characters (U+200E and
U+200F)?  On MS-Windows they are displayed as blanks; I need to type
"C-u C-x =" to see that these are not blanks.  Is that right?

Re: Displaying LRM and RLM

[Don Armstrong]

On Sat, 20 Dec 2008, Eli Zaretskii wrote:

> What should Emacs 23 display for these two characters (U+200E and
> U+200F)?  On MS-Windows they are displayed as blanks; I need to type
> "C-u C-x =" to see that these are not blanks.  Is that right?

These are the LRM and RLM; they're zero-width symbols that change the
direction of text. See http://en.wikipedia.org/wiki/Left-to-right_mark
et al.


Don Armstrong

-- 
I may not have gone where I intended to go, but I think I have ended
up where I needed to be.
 -- Douglas Adams _The Long Dark Tea-Time of the Soul_

http://www.donarmstrong.com              http://rzlab.ucr.edu

Re: Displaying LRM and RLM

[Eli Zaretskii]

> Date: Sat, 20 Dec 2008 02:45:06 -0800
> From: Don Armstrong <don@donarmstrong.com>
> 
> On Sat, 20 Dec 2008, Eli Zaretskii wrote:
> 
> > What should Emacs 23 display for these two characters (U+200E and
> > U+200F)?  On MS-Windows they are displayed as blanks; I need to type
> > "C-u C-x =" to see that these are not blanks.  Is that right?
> 
> These are the LRM and RLM; they're zero-width symbols that change the
> direction of text. See http://en.wikipedia.org/wiki/Left-to-right_mark
> et al.

Well, I know very well what they are; see the Subject.  I was asking
how we should display them.  I think displaying them as blanks without
any visual cue to the fact that they aren't is misleading.  E.g., NBSP
is displayed specially, for that very reason.

Re: Displaying LRM and RLM

[Jason Rumney]

Eli Zaretskii wrote:
> Well, I know very well what they are; see the Subject.  I was asking
> how we should display them.  I think displaying them as blanks without
> any visual cue to the fact that they aren't is misleading.

Blame the font, if it is mapping those characters to a space glyph.

Re: Displaying LRM and RLM

[Eli Zaretskii]

> Date: Sat, 20 Dec 2008 21:05:14 +0800
> From: Jason Rumney <jasonr@f2s.com>
> CC: Don Armstrong <don@donarmstrong.com>, emacs-devel@gnu.org
> 
> Blame the font, if it is mapping those characters to a space glyph.

Do you see them as someone else? with what font?

Also, how can I see if some specific font maps them to the space
glyph?

Re: Displaying LRM and RLM

[Jason Rumney]

Eli Zaretskii wrote:
>> Date: Sat, 20 Dec 2008 21:05:14 +0800
>> From: Jason Rumney <jasonr@f2s.com>
>> CC: Don Armstrong <don@donarmstrong.com>, emacs-devel@gnu.org
>>
>> Blame the font, if it is mapping those characters to a space glyph.
>>     
>
> Do you see them as someone else? with what font?
>   

I see them as space glyphs, with DejaVu Sans Mono.

> Also, how can I see if some specific font maps them to the space
> glyph?
>   

C-u C-x = on the character in question, and compare with the results of 
doing the same on a real space.
Near the bottom, you will see the glyph code that is being mapped to. 
Space is usually 0x03, as it is one of the standard glyphs that come 
near the front of the cmap table.

Re: Displaying LRM and RLM

[Eli Zaretskii]

> Date: Sat, 20 Dec 2008 21:36:33 +0800
> From: Jason Rumney <jasonr@f2s.com>
> CC: emacs-devel@gnu.org
> 
> > Also, how can I see if some specific font maps them to the space
> > glyph?
> >   
> 
> C-u C-x = on the character in question, and compare with the results of 
> doing the same on a real space.
> Near the bottom, you will see the glyph code that is being mapped to. 
> Space is usually 0x03, as it is one of the standard glyphs that come 
> near the front of the cmap table.

Thanks, I see that now.

Does Emacs support display of zero-width characters? should it?

Re: Displaying LRM and RLM

[Jason Rumney]

Eli Zaretskii wrote:
> Does Emacs support display of zero-width characters? should it?
>   

I think the answer to this last questions is yes. Another question is 
HOW should Emacs support zero width characters - should it be stripping 
them before they are sent to the font backend for display, and doing the 
right thing with them itself, or should it be up to the font backends to 
support them?

If it is the latter, I think we need to redesign the font backend API to 
allow extra information to be attached in the encode_char function in 
the backends. Checking just now, we do actually have the information 
that this character is zero-width at that point, but once we have 
returned the encoded glyph, that information is lost.

Re: Displaying LRM and RLM

[Stefan Monnier]

>> > What should Emacs 23 display for these two characters (U+200E and
>> > U+200F)?  On MS-Windows they are displayed as blanks; I need to type
>> > "C-u C-x =" to see that these are not blanks.  Is that right?

Since these should be "displayed" by changing the direction of text,
I think we can safely say that Emacs is not currently able to display
them, so we should forcibly replace them by the usual boxes we use for
chars that we can't display (usually it's because the font doesn't
provide a glyph for it, here it's because Emacs doesn't know how to
implement it).


        Stefan