bug#44020: 28.0.50; Prefer selection of colour emoji fonts?

classic Classic list List threaded Threaded
29 messages Options
12
Reply | Threaded
Open this post in threaded view
|

bug#44020: 28.0.50; Prefer selection of colour emoji fonts?

Lars Ingebrigtsen

If I start "emacs -Q" and load a file with this character:

馃帀

Emacs shows that as a black-and-white symbol using the Symbola font.

If I then say

(set-fontset-font t 'symbol "Noto Color Emoji")
(set-fontset-font t 'symbol "Symbola" nil 'append)

I get a colourful party popper.

It'd be nice if this worked out of the box.  What do other programs do?
Have a huge list of fonts they prefer over other fonts?


In GNU Emacs 28.0.50 (build 49, x86_64-pc-linux-gnu, GTK+ Version 3.24.23, cairo version 1.16.0)
 of 2020-10-14 built on xo
Repository revision: 1ee5a4cb1afe5a55d78e35cab2ad8651196c8967
Repository branch: master
Windowing system distributor 'The X.Org Foundation', version 11.0.12008000
System Description: Debian GNU/Linux bullseye/sid


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




Reply | Threaded
Open this post in threaded view
|

bug#44020: 28.0.50; Prefer selection of colour emoji fonts?

Eli Zaretskii
> From: Lars Ingebrigtsen <[hidden email]>
> Date: Thu, 15 Oct 2020 20:10:44 +0200
>
>
> If I start "emacs -Q" and load a file with this character:
>
> 馃帀
>
> Emacs shows that as a black-and-white symbol using the Symbola font.
>
> If I then say
>
> (set-fontset-font t 'symbol "Noto Color Emoji")
> (set-fontset-font t 'symbol "Symbola" nil 'append)
>
> I get a colourful party popper.
>
> It'd be nice if this worked out of the box.

AFAIR, Robert Pluim worked on better support for Emoji display; that
needs a few more tweaks and importing one more file from the UCD, to
support Emoji sequences.  I hope this will be ready some time soon.
Robert?



Reply | Threaded
Open this post in threaded view
|

bug#44020: 28.0.50; Prefer selection of colour emoji fonts?

Robert Pluim
>>>>> On Thu, 15 Oct 2020 21:52:19 +0300, Eli Zaretskii <[hidden email]> said:

    >> From: Lars Ingebrigtsen <[hidden email]>
    >> Date: Thu, 15 Oct 2020 20:10:44 +0200
    >>
    >>
    >> If I start "emacs -Q" and load a file with this character:
    >>
    >> 馃帀
    >>
    >> Emacs shows that as a black-and-white symbol using the Symbola font.
    >>
    >> If I then say
    >>
    >> (set-fontset-font t 'symbol "Noto Color Emoji")
    >> (set-fontset-font t 'symbol "Symbola" nil 'append)
    >>
    >> I get a colourful party popper.
    >>
    >> It'd be nice if this worked out of the box.

    Eli> AFAIR, Robert Pluim worked on better support for Emoji display; that
    Eli> needs a few more tweaks and importing one more file from the UCD, to
    Eli> support Emoji sequences.  I hope this will be ready some time soon.
    Eli> Robert?

There始s three parts to that:

1. Deciding what fonts we should use for emoji. We can get a lot of
   good results by adding appropriate 'set-fontset-font' calls to
   emacs, independently of any other work (I also wonder if we should
   change the default of use-default-font-for-symbols, given the
   proliferation of both emoji and fonts dedicated to displaying them)

2. Extract the emoji sequences from the Unicode data files and assign
   them to a new 'emoji' script. Be prepared to choose your favourite
   bikeshed colour.

3. Add the appropriate entries to composition-function-table so that
   the various multi-codepoint emoji sequences have a chance to be
   rendered correctly, whether by HarfBuzz or the Apple thing whose
   name I forget.

For [1], I think "Noto Color Emoji" and "Apple Color Emoji" would be a
good start.

For [2], I think that was finished, except that Eli then updated Emacs
to the next version of Unicode, and I haven't adjusted yet.

[3] is done for the cases that composition can currently handle. I
recall that there are sequences that require composition to do
lookback, which I think YAMAMOTO-san has patches for, but they've not
been submitted to Emacs (a quick perusal of my notes says this is for
keycap emojis, so I guess that始s a minor nit).

Oh, and there was a complication with enabling this by default on
macOS that I don始t fully understand yet. I始ll reread the history.

I始ll see if I can clean this up this weekend and throw it on a branch.

Robert
--



Reply | Threaded
Open this post in threaded view
|

bug#44020: 28.0.50; Prefer selection of colour emoji fonts?

Robert Pluim
>>>>> On Fri, 16 Oct 2020 13:26:42 +0300, Eli Zaretskii <[hidden email]> said:

    >> From: Robert Pluim <[hidden email]>
    >> Cc: Lars Ingebrigtsen <[hidden email]>,  [hidden email]
    >> Date: Fri, 16 Oct 2020 10:19:12 +0200
    >>
    >> 1. Deciding what fonts we should use for emoji. We can get a lot of
    >> good results by adding appropriate 'set-fontset-font' calls to
    >> emacs, independently of any other work (I also wonder if we should
    >> change the default of use-default-font-for-symbols, given the
    >> proliferation of both emoji and fonts dedicated to displaying them)
    >>
    >> 2. Extract the emoji sequences from the Unicode data files and assign
    >> them to a new 'emoji' script. Be prepared to choose your favourite
    >> bikeshed colour.
    >>
    >> 3. Add the appropriate entries to composition-function-table so that
    >> the various multi-codepoint emoji sequences have a chance to be
    >> rendered correctly, whether by HarfBuzz or the Apple thing whose
    >> name I forget.
    >>
    >> For [1], I think "Noto Color Emoji" and "Apple Color Emoji" would be a
    >> good start.

    Eli> I agree, but [1] depends on [2], because set-fontset-font needs to
    Eli> know the script.

set-fontset-font with a TARGET covering the emoji codepoints in
question, I meant. You始re right it始s probably cleaner to do both at
the same time.

    Eli> Also, Noto Color Emoji has a free license, so it's okay for it to be
    Eli> in fontset.el, but it is not clear to me whether Apple Color Emoji is
    Eli> distributed under a free license.

It始s marked 'Copyright Apple 2011-2016' in the macOS font
selector. And Apple appear to be both sensitive and litigious about
it, so probably best not to promote it.

    >> For [2], I think that was finished, except that Eli then updated Emacs
    >> to the next version of Unicode, and I haven't adjusted yet.

    Eli> "Finished", in the sense that it is in Emacs already?

No, finished as in "it works for me on my mac and GNU/Linux boxes" 馃檪

As I said, it needs adjusting to the latest Unicode update.

    >> I始ll see if I can clean this up this weekend and throw it on a branch.

    Eli> Thanks.

    Eli> FWIW, I think this is good enough to land on master directly, I see no
    Eli> need for a feature branch.

OK.

Robert
--



Reply | Threaded
Open this post in threaded view
|

bug#44020: 28.0.50; Prefer selection of colour emoji fonts?

Eli Zaretskii
> From: Robert Pluim <[hidden email]>
> Cc: [hidden email],  [hidden email]
> Date: Fri, 16 Oct 2020 12:38:18 +0200
>
>     >> 1. Deciding what fonts we should use for emoji. We can get a lot of
>     >> good results by adding appropriate 'set-fontset-font' calls to
>     >> emacs, independently of any other work (I also wonder if we should
>     >> change the default of use-default-font-for-symbols, given the
>     >> proliferation of both emoji and fonts dedicated to displaying them)
>     >>
>     >> 2. Extract the emoji sequences from the Unicode data files and assign
>     >> them to a new 'emoji' script. Be prepared to choose your favourite
>     >> bikeshed colour.
>     >>
>     >> 3. Add the appropriate entries to composition-function-table so that
>     >> the various multi-codepoint emoji sequences have a chance to be
>     >> rendered correctly, whether by HarfBuzz or the Apple thing whose
>     >> name I forget.
>     >>
>     >> For [1], I think "Noto Color Emoji" and "Apple Color Emoji" would be a
>     >> good start.
>
>     Eli> I agree, but [1] depends on [2], because set-fontset-font needs to
>     Eli> know the script.
>
> set-fontset-font with a TARGET covering the emoji codepoints in
> question, I meant. You始re right it始s probably cleaner to do both at
> the same time.

Yes, because using the codepoints will mean more maintenance in the
future, as more Emoji blocks are added (and they will be, quite
certainly).

>
>     Eli> Also, Noto Color Emoji has a free license, so it's okay for it to be
>     Eli> in fontset.el, but it is not clear to me whether Apple Color Emoji is
>     Eli> distributed under a free license.
>
> It始s marked 'Copyright Apple 2011-2016' in the macOS font
> selector. And Apple appear to be both sensitive and litigious about
> it, so probably best not to promote it.

Right.  People will have to do that on their own, if they don't want
to install a free font.



Reply | Threaded
Open this post in threaded view
|

bug#44020: 28.0.50; Prefer selection of colour emoji fonts?

Lars Ingebrigtsen
In reply to this post by Robert Pluim
Robert Pluim <[hidden email]> writes:

> I始ll see if I can clean this up this weekend and throw it on a branch.

Looking forward to it.  :-)  And I agree with Eli; doing it on the trunk
would be fine.

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



Reply | Threaded
Open this post in threaded view
|

bug#44020: 28.0.50; Prefer selection of colour emoji fonts?

Alan Third
In reply to this post by Robert Pluim
On Fri, Oct 16, 2020 at 12:38:18PM +0200, Robert Pluim wrote:
> >>>>> On Fri, 16 Oct 2020 13:26:42 +0300, Eli Zaretskii <[hidden email]> said:
>
>     Eli> Also, Noto Color Emoji has a free license, so it's okay for it to be
>     Eli> in fontset.el, but it is not clear to me whether Apple Color Emoji is
>     Eli> distributed under a free license.
>
> It始s marked 'Copyright Apple 2011-2016' in the macOS font
> selector. And Apple appear to be both sensitive and litigious about
> it, so probably best not to promote it.

On the other hand it comes preinstalled, it only works on Apple
devices, and Noto Color Emoji doesn't work on Apple devices. I'm not
sure we'd really be "promoting" it.

IIRC Windows comes with Segoe UI Emoji preinstalled too.
--
Alan Third



Reply | Threaded
Open this post in threaded view
|

bug#44020: 28.0.50; Prefer selection of colour emoji fonts?

Lars Ingebrigtsen
Alan Third <[hidden email]> writes:

> On the other hand it comes preinstalled, it only works on Apple
> devices, and Noto Color Emoji doesn't work on Apple devices. I'm not
> sure we'd really be "promoting" it.

On the third hand, since it's preinstalled, Emacs doesn't need to
mention it, because it's always there, and Emacs won't have problems
displaying the characters it offers.

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



Reply | Threaded
Open this post in threaded view
|

bug#44020: 28.0.50; Prefer selection of colour emoji fonts?

Eli Zaretskii
In reply to this post by Alan Third
> Date: Fri, 16 Oct 2020 20:30:41 +0100
> From: Alan Third <[hidden email]>
> Cc: Eli Zaretskii <[hidden email]>, [hidden email], [hidden email]
>
> On Fri, Oct 16, 2020 at 12:38:18PM +0200, Robert Pluim wrote:
> > >>>>> On Fri, 16 Oct 2020 13:26:42 +0300, Eli Zaretskii <[hidden email]> said:
> >
> >     Eli> Also, Noto Color Emoji has a free license, so it's okay for it to be
> >     Eli> in fontset.el, but it is not clear to me whether Apple Color Emoji is
> >     Eli> distributed under a free license.
> >
> > It始s marked 'Copyright Apple 2011-2016' in the macOS font
> > selector. And Apple appear to be both sensitive and litigious about
> > it, so probably best not to promote it.
>
> On the other hand it comes preinstalled, it only works on Apple
> devices, and Noto Color Emoji doesn't work on Apple devices. I'm not
> sure we'd really be "promoting" it.
>
> IIRC Windows comes with Segoe UI Emoji preinstalled too.

Both are true, but AFAIK our policy until now was not to mention such
fonts in fontset.el.

Emoji is just one example, btw.  At least for MS-Windows, the latest
versions come with many good fonts that cover most of the scripts.  So
if the restriction to avoid mentioning proprietary fonts in fontset.el
could be lifted, we could produce a very capable and complete fontset
for MS-Windows.  I wouldn't be surprised if the same situation didn't
exist on macOS.  Not sure how things are on GNU/Linux, and what are
the differences between the distros in this respect.



Reply | Threaded
Open this post in threaded view
|

bug#44020: 28.0.50; Prefer selection of colour emoji fonts?

Eli Zaretskii
In reply to this post by Lars Ingebrigtsen
> From: Lars Ingebrigtsen <[hidden email]>
> Cc: Robert Pluim <[hidden email]>,  Eli Zaretskii <[hidden email]>,
>   [hidden email]
> Date: Sat, 17 Oct 2020 08:38:10 +0200
>
> Alan Third <[hidden email]> writes:
>
> > On the other hand it comes preinstalled, it only works on Apple
> > devices, and Noto Color Emoji doesn't work on Apple devices. I'm not
> > sure we'd really be "promoting" it.
>
> On the third hand, since it's preinstalled, Emacs doesn't need to
> mention it, because it's always there, and Emacs won't have problems
> displaying the characters it offers.

Are you sure?  What exactly makes sure Emacs won't have any problems
in those cases?  For starters, does Emacs really know that a given
character belongs to "Emoji"?  And if it does, does it know how to
filter fonts that support Emoji?



Reply | Threaded
Open this post in threaded view
|

bug#44020: 28.0.50; Prefer selection of colour emoji fonts?

Alan Third
In reply to this post by Lars Ingebrigtsen
On Sat, Oct 17, 2020 at 08:38:10AM +0200, Lars Ingebrigtsen wrote:
> Alan Third <[hidden email]> writes:
>
> > On the other hand it comes preinstalled, it only works on Apple
> > devices, and Noto Color Emoji doesn't work on Apple devices. I'm not
> > sure we'd really be "promoting" it.
>
> On the third hand, since it's preinstalled, Emacs doesn't need to
> mention it, because it's always there, and Emacs won't have problems
> displaying the characters it offers.

Even once I remove this code:

modified   src/macfont.m
@@ -2414,11 +2414,6 @@ So we use CTFontDescriptorCreateMatchingFontDescriptor (no
                   != (spacing >= FONT_SPACING_MONO)))
             continue;
 
-          /* Don't use a color bitmap font unless its family is
-             explicitly specified.  */
-          if ((sym_traits & kCTFontTraitColorGlyphs) && NILP (family))
-            continue;
-
           if (j > 0
               && !macfont_supports_charset_and_languages_p (desc, charset,
                                                             chars, languages))

Emacs almost never finds Apple Color Emoji without help. I've no idea
why.

For some reason it finds the font for BLACK DOWN-POINTING DOUBLE
TRIANGLE, but not SMILING FACE WITH SMILING EYES unless the former is
already in the document.

Should this Just Work?

BTW, there is code in ns-win.el that plays with the default fontset on
Macs which may be breaking our rules.
--
Alan Third



Reply | Threaded
Open this post in threaded view
|

bug#44020: 28.0.50; Prefer selection of colour emoji fonts?

Alan Third
On Sat, Oct 17, 2020 at 04:27:58PM +0100, Alan Third wrote:
>
> Emacs almost never finds Apple Color Emoji without help. I've no idea
> why.

Sorry, I've just realised I'm going round in circles as this is
exactly what Robert's working on. Ignore me, please. :)

--
Alan Third



Reply | Threaded
Open this post in threaded view
|

bug#44020: 28.0.50; Prefer selection of colour emoji fonts?

Lars Ingebrigtsen
In reply to this post by Eli Zaretskii
Eli Zaretskii <[hidden email]> writes:

>> On the third hand, since it's preinstalled, Emacs doesn't need to
>> mention it, because it's always there, and Emacs won't have problems
>> displaying the characters it offers.
>
> Are you sure?  What exactly makes sure Emacs won't have any problems
> in those cases?  For starters, does Emacs really know that a given
> character belongs to "Emoji"?  And if it does, does it know how to
> filter fonts that support Emoji?

I thought that Emacs looped through all the fonts on the system and
checked whether the font had coverage for the character we wanted to
display?

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



Reply | Threaded
Open this post in threaded view
|

bug#44020: 28.0.50; Prefer selection of colour emoji fonts?

Eli Zaretskii
> From: Lars Ingebrigtsen <[hidden email]>
> Cc: [hidden email],  [hidden email],  [hidden email]
> Date: Sun, 18 Oct 2020 10:09:09 +0200
>
> > Are you sure?  What exactly makes sure Emacs won't have any problems
> > in those cases?  For starters, does Emacs really know that a given
> > character belongs to "Emoji"?  And if it does, does it know how to
> > filter fonts that support Emoji?
>
> I thought that Emacs looped through all the fonts on the system and
> checked whether the font had coverage for the character we wanted to
> display?

It does -- for some value of "check".  Opening a font and trying to
find a glyph for a character is expensive, so Emacs tries to avoid
that as much as possible.  Instead, it uses less expensive "checks"
that don't require actually opening the font.  The details are
somewhat complicated (and I'm not sure I have a sufficiently clear
picture of them, nor that we have someone on board who does), and the
are to some extent font-backend dependent, but they have to do with
the character's script, the script's characters in
script-representative-chars, and system APIs for finding fonts that
support those.  And the first problem with Emoji characters in this
context is that their script is 'symbol', which doesn't allow us to
assign an Emoji font to them as a script.



Reply | Threaded
Open this post in threaded view
|

bug#44020: 28.0.50; Prefer selection of colour emoji fonts?

Lars Ingebrigtsen
Eli Zaretskii <[hidden email]> writes:

> It does -- for some value of "check".  Opening a font and trying to
> find a glyph for a character is expensive, so Emacs tries to avoid
> that as much as possible.  Instead, it uses less expensive "checks"
> that don't require actually opening the font.  The details are
> somewhat complicated (and I'm not sure I have a sufficiently clear
> picture of them, nor that we have someone on board who does), and the
> are to some extent font-backend dependent, but they have to do with
> the character's script, the script's characters in
> script-representative-chars, and system APIs for finding fonts that
> support those.  And the first problem with Emoji characters in this
> context is that their script is 'symbol', which doesn't allow us to
> assign an Emoji font to them as a script.

Ah, I see.  Thank you for the explanation.

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



Reply | Threaded
Open this post in threaded view
|

bug#44020: 28.0.50; Prefer selection of colour emoji fonts?

Robert Pluim
In reply to this post by Lars Ingebrigtsen
>>>>> On Fri, 16 Oct 2020 16:49:43 +0200, Lars Ingebrigtsen <[hidden email]> said:

    Lars> Robert Pluim <[hidden email]> writes:
    >> I始ll see if I can clean this up this weekend and throw it on a branch.

    Lars> Looking forward to it.  :-)  And I agree with Eli; doing it on the trunk
    Lars> would be fine.

So revisiting this, it turns out I have *two* features sitting in my workspace:

1. splitting out the emoji into a separate script and setting the right
  fontset
2. adding composition rules for the zwj emoji described in
  emoji-zwj-sequences.txt

[2] has some complications regarding the default display of the
variation-selector, so I始ll leave that aside for now, and continue on
updating [1] to Unicode 13.0

Robert
--



Reply | Threaded
Open this post in threaded view
|

bug#44020: 28.0.50; Prefer selection of colour emoji fonts?

Robert Pluim
>>>>> On Mon, 19 Oct 2020 17:55:55 +0300, Eli Zaretskii <[hidden email]> said:

    >> From: Robert Pluim <[hidden email]>
    >> Cc: [hidden email],  Eli Zaretskii <[hidden email]>
    >> Date: Mon, 19 Oct 2020 12:56:33 +0200
    >>
    >> So revisiting this, it turns out I have *two* features sitting in my workspace:
    >>
    >> 1. splitting out the emoji into a separate script and setting the right
    >> fontset
    >> 2. adding composition rules for the zwj emoji described in
    >> emoji-zwj-sequences.txt
    >>
    >> [2] has some complications regarding the default display of the
    >> variation-selector, so I始ll leave that aside for now, and continue on
    >> updating [1] to Unicode 13.0

    Eli> Thanks.

    Eli> Can you summarize the problems with [2]?  I thought we had all of that
    Eli> figured out back when we discussed this.

Perhaps we did: I have no clear memory of what we decided, so I始ll
have to go and re-read it.

I始m attaching a gzipped patch for emoji support (it contains the
entire emoji-data.txt from Unicode 13.0 馃檪), which works for me on
macOS when I do

(set-fontset-font t 'emoji '("Apple Color Emoji . "iso1064601") nil
'prepend)

It *should* work as-is on GNU/Linux, but for some reason Emacs is
completely refusing to use any Emoji fonts for me there. It始s possible
I始ve messed up my system somehow, since emacs-27 has the same issue,
but I始m seeing some very suspicious results from
ftcrfont_glyph_extents (the font average width is always 0), so perhaps
we need to adjust our font code.

Testing appreciated.

Robert
--

0001-emoji-using-blocks.txt.patch.gz (35K) Download Attachment
Reply | Threaded
Open this post in threaded view
|

bug#44020: 28.0.50; Prefer selection of colour emoji fonts?

Eli Zaretskii
> From: Robert Pluim <[hidden email]>
> Cc: [hidden email],  [hidden email]
> Date: Mon, 19 Oct 2020 20:34:41 +0200
>
> (set-fontset-font t 'emoji '("Apple Color Emoji . "iso1064601") nil
> 'prepend)
>
> It *should* work as-is on GNU/Linux, but for some reason Emacs is
> completely refusing to use any Emoji fonts for me there. It始s possible
> I始ve messed up my system somehow, since emacs-27 has the same issue,
> but I始m seeing some very suspicious results from
> ftcrfont_glyph_extents (the font average width is always 0), so perhaps
> we need to adjust our font code.

Didn't someone say they can see color emoji on Cairo?



Reply | Threaded
Open this post in threaded view
|

bug#44020: 28.0.50; Prefer selection of colour emoji fonts?

Robert Pluim
>>>>> On Mon, 19 Oct 2020 21:52:13 +0300, Eli Zaretskii <[hidden email]> said:

    >> From: Robert Pluim <[hidden email]>
    >> Cc: [hidden email],  [hidden email]
    >> Date: Mon, 19 Oct 2020 20:34:41 +0200
    >>
    >> (set-fontset-font t 'emoji '("Apple Color Emoji . "iso1064601") nil
    >> 'prepend)
    >>
    >> It *should* work as-is on GNU/Linux, but for some reason Emacs is
    >> completely refusing to use any Emoji fonts for me there. It始s possible
    >> I始ve messed up my system somehow, since emacs-27 has the same issue,
    >> but I始m seeing some very suspicious results from
    >> ftcrfont_glyph_extents (the font average width is always 0), so perhaps
    >> we need to adjust our font code.

    Eli> Didn't someone say they can see color emoji on Cairo?

Lars said he could, and it始s in the NEWS file, so it worked at some
point. I can始t get either Noto Color Emoji nor Emoji One to work on
GNU/Linux, because the following code in font.c bugs out

  /* We always open a font of manageable size; i.e non-zero average
     width and height.  */
  for (psize = pixel_size; ; psize++)
    {
      font_object = driver_list->driver->open_font (f, entity, psize);
      if (NILP (font_object))
        return Qnil;
      font = XFONT_OBJECT (font_object);
      if (font->average_width > 0 && font->height > 0)
        break;
      /* Avoid an infinite loop.  */
      if (psize > pixel_size + 15)
        return Qnil; <====== here
    }

and that始s because font->average_width is always 0.

Robert
--



Reply | Threaded
Open this post in threaded view
|

bug#44020: 28.0.50; Prefer selection of colour emoji fonts?

Stefan Kangas-2
In reply to this post by Robert Pluim
Robert Pluim <[hidden email]> writes:

> It *should* work as-is on GNU/Linux, but for some reason Emacs is
> completely refusing to use any Emoji fonts for me there. It始s possible
> I始ve messed up my system somehow, since emacs-27 has the same issue,
> but I始m seeing some very suspicious results from
> ftcrfont_glyph_extents (the font average width is always 0), so perhaps
> we need to adjust our font code.
>
> Testing appreciated.

I applied your patch running on Debian GNU/Linux (bullseye/testing), and
was able to display the "馃帀" character from Lars' original message just
fine here.



12