Re: master 94380420e2: Support Noto Emoji font as fallback

Re: master 94380420e2: Support Noto Emoji font as fallback

[Robert Pluim]

>>>>> On Fri, 23 Sep 2022 02:38:36 -0400 (EDT), Eli Zaretskii <eliz@gnu.org> said:


    Eli>     Support Noto Emoji font as fallback
    
    Eli>     The black-and-white emoji font is back under active development,
    Eli>     see https://github.com/googlefonts/noto-emoji.  So this adds
    Eli>     that font as fallback for systems that don't have the capability
    Eli>     of using color fonts.
    Eli>     * lisp/international/fontset.el (setup-default-fontset): Support
    Eli>     black-and-white Noto Emoji font as fallback for Emoji display.

Nice. On a related note, 'Noto Sans <script>' is fairly commonly
available for codepoints in <script>, perhaps we should add those as
fallbacks as well? (the version I have here has support for 169
different scripts)

Robert
-- 

Re: master 94380420e2: Support Noto Emoji font as fallback

[Eli Zaretskii]

> From: Robert Pluim <rpluim@gmail.com>
> Cc: Eli Zaretskii <eliz@gnu.org>
> Date: Fri, 23 Sep 2022 10:55:01 +0200
> 
> Nice. On a related note, 'Noto Sans <script>' is fairly commonly
> available for codepoints in <script>, perhaps we should add those as
> fallbacks as well? (the version I have here has support for 169
> different scripts)

I'm not sure.  We should at least figure out how wide-spread those
fonts are, and whether people really prefer them better than what
their systems provide OOTB.  FWIW, the visual appearance of the Google
Noto fonts is not always appealing to me.  But the above could at
least be a good idea for historic scripts that are not used in
contemporary live cultures.

Re: master 94380420e2: Support Noto Emoji font as fallback

[समीर सिंह Sameer Singh]

[-- Attachment #1: Type: text/plain, Size: 1377 bytes --]

Some issues of the noto fonts which I remember are:

Bengali: No "knots" in the Sans font, though serif fonts have them. [1]
Tirhuta: No "knots" here too at all since it does not have a serif version.
[1]
Sharada: Wrong glyph used for the Vowel A. [2]
Brahmi: Does not support all conjuncts. [3]

[1]: https://github.com/notofonts/bengali/issues/2
[2]: https://twitter.com/a_srinidhi/status/1529752633283469313
[3]: https://github.com/notofonts/brahmi/issues/1

There should be more issues still.

Also maybe I should restart adding new script support to Emacs.

Thanks

On Fri, Sep 23, 2022 at 4:25 PM Eli Zaretskii <eliz@gnu.org> wrote:

> > From: Robert Pluim <rpluim@gmail.com>
> > Cc: Eli Zaretskii <eliz@gnu.org>
> > Date: Fri, 23 Sep 2022 10:55:01 +0200
> >
> > Nice. On a related note, 'Noto Sans <script>' is fairly commonly
> > available for codepoints in <script>, perhaps we should add those as
> > fallbacks as well? (the version I have here has support for 169
> > different scripts)
>
> I'm not sure.  We should at least figure out how wide-spread those
> fonts are, and whether people really prefer them better than what
> their systems provide OOTB.  FWIW, the visual appearance of the Google
> Noto fonts is not always appealing to me.  But the above could at
> least be a good idea for historic scripts that are not used in
> contemporary live cultures.
>
>

[-- Attachment #2: Type: text/html, Size: 2277 bytes --]

Re: master 94380420e2: Support Noto Emoji font as fallback

[समीर सिंह Sameer Singh]

[-- Attachment #1: Type: text/plain, Size: 1821 bytes --]

Although I think for general enough usage they are fine.
If the user wants they can explicitly change them.
Their goal is "no more tofus" and I think they are achieving that.

Current noto fonts coverage: https://notofonts.github.io/overview/

On Fri, Sep 23, 2022 at 6:03 PM समीर सिंह Sameer Singh <
lumarzeli30@gmail.com> wrote:

> Some issues of the noto fonts which I remember are:
>
> Bengali: No "knots" in the Sans font, though serif fonts have them. [1]
> Tirhuta: No "knots" here too at all since it does not have a serif
> version. [1]
> Sharada: Wrong glyph used for the Vowel A. [2]
> Brahmi: Does not support all conjuncts. [3]
>
> [1]: https://github.com/notofonts/bengali/issues/2
> [2]: https://twitter.com/a_srinidhi/status/1529752633283469313
> [3]: https://github.com/notofonts/brahmi/issues/1
>
> There should be more issues still.
>
> Also maybe I should restart adding new script support to Emacs.
>
> Thanks
>
> On Fri, Sep 23, 2022 at 4:25 PM Eli Zaretskii <eliz@gnu.org> wrote:
>
>> > From: Robert Pluim <rpluim@gmail.com>
>> > Cc: Eli Zaretskii <eliz@gnu.org>
>> > Date: Fri, 23 Sep 2022 10:55:01 +0200
>> >
>> > Nice. On a related note, 'Noto Sans <script>' is fairly commonly
>> > available for codepoints in <script>, perhaps we should add those as
>> > fallbacks as well? (the version I have here has support for 169
>> > different scripts)
>>
>> I'm not sure.  We should at least figure out how wide-spread those
>> fonts are, and whether people really prefer them better than what
>> their systems provide OOTB.  FWIW, the visual appearance of the Google
>> Noto fonts is not always appealing to me.  But the above could at
>> least be a good idea for historic scripts that are not used in
>> contemporary live cultures.
>>
>>

[-- Attachment #2: Type: text/html, Size: 3086 bytes --]

Re: master 94380420e2: Support Noto Emoji font as fallback

[Eli Zaretskii]

> From: समीर सिंह Sameer Singh
>  <lumarzeli30@gmail.com>
> Date: Fri, 23 Sep 2022 18:10:12 +0530
> Cc: Robert Pluim <rpluim@gmail.com>, emacs-devel@gnu.org
> 
> Although I think for general enough usage they are fine.
> If the user wants they can explicitly change them.
> Their goal is "no more tofus" and I think they are achieving that.

That is true, but AFAIU modern systems come with font coverage that is
not bad at all, including for rare and ancient scripts.

I think we should limit specifying particular fonts in our fontset
setup only to those cases where either the "normal" search for
suitable fonts is likely to fail, or if some font is really superior
to the alternatives.

Re: master 94380420e2: Support Noto Emoji font as fallback

[Robert Pluim]

>>>>> On Fri, 23 Sep 2022 15:54:45 +0300, Eli Zaretskii <eliz@gnu.org> said:

    Eli> I think we should limit specifying particular fonts in our fontset
    Eli> setup only to those cases where either the "normal" search for
    Eli> suitable fonts is likely to fail, or if some font is really superior
    Eli> to the alternatives.

Looking at this some more, Iʼm getting unexpected behaviour for the
'bamum' script (Iʼm basing this off the fallback settings in
fontset.el for 'simple scripts').

emacs -Q
C-x 8 RET 16900 -> hex box
eval (set-fontset-font t 'bamum
       (font-spec :registry "iso10646-1" :script 'bamum) nil 'prepend)
       -> hex box
eval (set-fontset-font t 'bamum
        (font-spec  :script 'bamum) nil 'prepend)
        -> it finds a glyph using the following font
ftcrhb:-Goss-Kurinto Seri-regular-normal-normal-*-19-*-*-*-*-0-iso10646-1 (#x2A66)

which is a locally installed font. If I move that out of the way it
finds the following system font with the second `set-fontset-font' only.

    ftcrhb:-GOOG-Noto Sans Bamum-regular-normal-normal-*-15-*-*-*-*-0-iso10646-1 (#x151)

so the registry spec is preventing emacs from finding the font? (Iʼm
failing once again to get fc-match to tell me what scripts are
supported by a particular font)

Robert
-- 

Re: master 94380420e2: Support Noto Emoji font as fallback

[Eli Zaretskii]

> From: Robert Pluim <rpluim@gmail.com>
> Cc: समीर सिंह Sameer Singh
>  <lumarzeli30@gmail.com>,  emacs-devel@gnu.org
> Date: Fri, 23 Sep 2022 15:39:57 +0200
> 
> so the registry spec is preventing emacs from finding the font? (Iʼm
> failing once again to get fc-match to tell me what scripts are
> supported by a particular font)

I don't know why this happens, I guess you get to step through the
code in fontset.c and font.c to find out the reason.