bug#63028: 28.2; Rendering Traditional Mongolian Script

bug#63028: 28.2; Rendering Traditional Mongolian Script

[Garid Z.]

[-- Attachment #1: Type: text/plain, Size: 181 bytes --]


Rendering Traditional Mongolian Scripts in Emacs are not perfect:
Have a look at following text:
ᠮᠣᠩᠭᠣᠯ ᠪᠢᠴᠢᠭ (Монгол бичиг, Mongol Bichig):


[-- Attachment #2: file contain above text --]
[-- Type: text/plain, Size: 39 bytes --]

 ᠮᠣᠩᠭᠣᠯ ᠪᠢᠴᠢᠭ


[-- Attachment #3: Type: text/plain, Size: 28 bytes --]


Emacs renders as follows:


[-- Attachment #4: Screen shot --]
[-- Type: image/png, Size: 11721 bytes --]

[-- Attachment #5: Type: text/plain, Size: 796 bytes --]


The letters take different shapes (but same Unicode) based its
location in the word. For example, letter A in takes various forms:
- first form (isolate) 
- first form (initial)
- first form (medial)
- first form (medial)
But all takes same Unicode (U+1820) take a look at page 7 (upper left
part in page-7) in this manual. https://www.unicode.org/reports/tr54/

I think that Emacs is rendering only first form/variation of each letter
, and not changing it accordingly based on the its location.

------------------------------------------------------------------------

I know that number of people who use this Traditional Mongolian Scripts
in Emacs are basically none. But I'd be happy if Emacs renders it
correctly while other programs (such as vscode) already does.


--
Garid Zorigoo. 

bug#63028: 28.2; Rendering Traditional Mongolian Script

[Eli Zaretskii]

> From: "Garid Z." <garidzorigoo@gmail.com>
> Date: Sun, 23 Apr 2023 16:50:07 +0900
> 
> Rendering Traditional Mongolian Scripts in Emacs are not perfect:

Thanks for the report, but we need more details to investigate:

 . the details of your system and Emacs configuration, as collected by
   "M-x report-emacs-bug"
 . the font(s) you used to display this text (you can find that out by
   moving point to the relevant characters and typing "C-u C-x =")
 . if your Emacs was built with HarfBuzz, the version of the HarfBuzz
   library you have installed

In general, the Mongolian script has no composition rules defined for
it in Emacs, which could be the main reason why this script is not
displayed as you expect.  If you know of some reference for defining
the character composition rules for Mongolian, please tell where those
references could be found on the Internet.

bug#63028: 28.2; Rendering Traditional Mongolian Script

[Eli Zaretskii]

> Cc: 63028@debbugs.gnu.org
> Date: Sun, 23 Apr 2023 11:24:02 +0300
> From: Eli Zaretskii <eliz@gnu.org>
> 
> In general, the Mongolian script has no composition rules defined for
> it in Emacs, which could be the main reason why this script is not
> displayed as you expect.  If you know of some reference for defining
> the character composition rules for Mongolian, please tell where those
> references could be found on the Internet.

Btw, it would be nice to have better support for Traditional Mongolian
in general; currently, we only have the script itself and the
characters which belong to it, and that's all.  Things to add:

  . input method(s)
  . HELLO text
  . font selection spec in fontset.el
  . character-composition rules

Patches welcome.

bug#63028: 28.2; Rendering Traditional Mongolian Script

[Eli Zaretskii]

> From: "Garid Z." <garidzorigoo@gmail.com>
> Date: Sun, 23 Apr 2023 16:50:07 +0900
> 
> Rendering Traditional Mongolian Scripts in Emacs are not perfect:
> Have a look at following text:
> ᠮᠣᠩᠭᠣᠯ ᠪᠢᠴᠢᠭ (Монгол бичиг, Mongol Bichig):

I've installed on emacs-29 branch a fix, which should make rendering
Traditional Mongolian better, please take a look.

bug#63028: 28.2; Rendering Traditional Mongolian Script

[Garid Z.]

[-- Attachment #1: Type: text/plain, Size: 527 bytes --]


Eli Zaretskii <eliz@gnu.org> writes:

>> From: "Garid Z." <garidzorigoo@gmail.com>
>> Date: Sun, 23 Apr 2023 16:50:07 +0900
>> 
>> Rendering Traditional Mongolian Scripts in Emacs are not perfect:
>> Have a look at following text:
>> ᠮᠣᠩᠭᠣᠯ ᠪᠢᠴᠢᠭ (Монгол бичиг, Mongol Bichig):
>
> I've installed on emacs-29 branch a fix, which should make rendering
> Traditional Mongolian better, please take a look.

I've just tried emacs-29 branch. It looks like it's working pretty well.


[-- Attachment #2: Hello.png --]
[-- Type: image/png, Size: 12128 bytes --]

[-- Attachment #3: Type: text/plain, Size: 574 bytes --]


Thanks for you enthusiasm, and quick work. I really appreciate it.
(You've fixed this even before I could reply previous emails).


>>  . HELLO text

For the HELLO file please add following part:
Just in case (perhaps rendering issue in browser/mail client/etc),
I'm attaching the text above as a file.

----------------------------------------------------------------
...
...
Mongolian (монгол хэл)	Сайн байна уу? / ᠰᠠᠶᠢᠨ ᠪᠠᠶᠢᠨ᠎ᠠ ᠤᠤ
...
...
----------------------------------------------------------------

[-- Attachment #4: Updated Mongolian Hello Line --]
[-- Type: text/plain, Size: 153 bytes --]

ᠰᠠᠶᠢᠨ ᠪᠠᠶᠢᠨ᠎ᠠ ᠤᠤ
Mongolian (монгол хэл)	Сайн байна уу? / ᠰᠠᠶᠢᠨ ᠪᠠᠶᠢᠨ᠎ᠠ ᠤᠤ?

[-- Attachment #5: Type: text/plain, Size: 410 bytes --]



>> . input method(s)

For input method, I'm working on a make-shift quail input method.
I'll send a patch for for input-method (similar to my previous patch
of cyrrilic-mongolian layout).

>>  . font selection spec in fontset.el
>>  . character-composition rules

About these layer two  (fontset.el, character-composition), I've no idea
what's these about.



Again, Thank you for your work.


Garid Zorigoo

bug#63028: 28.2; Rendering Traditional Mongolian Script

[Eli Zaretskii]

> From: "Garid Z." <garidzorigoo@gmail.com>
> Cc: 63028@debbugs.gnu.org
> Date: Mon, 24 Apr 2023 00:35:26 +0900
> 
> > I've installed on emacs-29 branch a fix, which should make rendering
> > Traditional Mongolian better, please take a look.
> 
> I've just tried emacs-29 branch. It looks like it's working pretty well.

OK, thanks for verifying this works.

FTR, what font are you using for this script?

> >>  . HELLO text
> 
> For the HELLO file please add following part:
> Just in case (perhaps rendering issue in browser/mail client/etc),
> I'm attaching the text above as a file.

Thanks.  However, I think I'd like to separate this line from the
existing Mongolian line (which uses the Cyrillic script).  I thought
about labeling it "Mongolian Traditional:, WDYT?  Alternatively, we
could rename the existing line to "Mongolian Cyrillic", if the
traditional Mongolian script is back in wide use nowadays.  In any
case, I will need to know how "Mongolian" is written in the Mongolian
script, can you tell that as well?

> >> . input method(s)
> 
> For input method, I'm working on a make-shift quail input method.
> I'll send a patch for for input-method (similar to my previous patch
> of cyrrilic-mongolian layout).

Great, thanks.

> >>  . font selection spec in fontset.el
> >>  . character-composition rules
> 
> About these layer two  (fontset.el, character-composition), I've no idea
> what's these about.

I already did that, in the commit I installed today.

Thanks.

bug#63028: 28.2; Rendering Traditional Mongolian Script

[Garid Z.]

[-- Attachment #1: Type: text/plain, Size: 152 bytes --]


> FTR, what font are you using for this script?

It looks like I'm using Noto-sans, Take look at my (C-u C-x =) output
file (attachment: letter.txt):


[-- Attachment #2: letter.txt --]
[-- Type: text/plain, Size: 841 bytes --]

             position: 3 of 15 (13%), column: 2
            character: ᠣ (displayed as ᠣ) (codepoint 6179, #o14043, #x1823)
              charset: unicode (Unicode (ISO10646))
code point in charset: 0x1823
               script: mongolian
               syntax: w 	which means: word
             category: .:Base, L:Strong L2R
             to input: type "C-x 8 RET 1823" or "C-x 8 RET MONGOLIAN LETTER O"
          buffer code: #xE1 #xA0 #xA3
            file code: #xE1 #xA0 #xA3 (encoded by coding system utf-8-unix)
              display: by this font (glyph code):
    ftcrhb:-GOOG-Noto Sans Mongolian-normal-normal-normal-*-97-*-*-*-*-0-iso10646-1 (#xAB)

Character code properties: customize what to show
  name: MONGOLIAN LETTER O
  general-category: Lo (Letter, Other)
  decomposition: (6179) ('ᠣ')

[back]

[-- Attachment #3: Type: text/plain, Size: 1090 bytes --]


> Thanks.  However, I think I'd like to separate this line from the
> existing Mongolian line (which uses the Cyrillic script).  I thought
> about labeling it "Mongolian Traditional:, WDYT?  Alternatively, we
> could rename the existing line to "Mongolian Cyrillic", if the
> traditional Mongolian script is back in wide use nowadays.

(TBH, Traditional script are used rarely).
But using "Mongolian Traditional" and "Mongolian Cyrillic" labels sound
perfect to me. Should I make new file:
emacs/lisp/leim/quail/mongolian.el and put these 2 input methods?
or "Mongolian-cyrillic" still should belongs to other cyrillics
in emacs/lisp/leim/quail/cyrillic.el? WDYT?

> In any
> case, I will need to know how "Mongolian" is written in the Mongolian
> script, can you tell that as well?


Sure here: (Just in case, I also made it as attachment)
- Mongolia - Монгол - ᠮᠣᠩᠭᠣᠯ 
- Mongolian language - Монгол хэл -  ᠮᠣᠩᠭᠣᠯ ᠬᠡᠯᠡ
- Mongolian traditional scrips - Монгол бичиг - ᠮᠣᠩᠭᠣᠯ ᠪᠢᠴᠢᠭ᠋


[-- Attachment #4: File:Mongolia-Mongolian-Traditional-script --]
[-- Type: text/plain, Size: 226 bytes --]

- Mongolia - Монгол - ᠮᠣᠩᠭᠣᠯ 
- Mongolian language - Монгол хэл -  ᠮᠣᠩᠭᠣᠯ ᠬᠡᠯᠡ
- Mongolian traditional scrips - Монгол бичиг - ᠮᠣᠩᠭᠣᠯ ᠪᠢᠴᠢᠭ᠋

[-- Attachment #5: mongolia_mongolian_traditional_script.png --]
[-- Type: image/png, Size: 43515 bytes --]

[-- Attachment #6: Type: text/plain, Size: 341 bytes --]



For the input-method this, I need to couple of days to figure out and
look for existing keyboard layout for traditional mongolian.

>> About these layer two  (fontset.el, character-composition), I've no idea
>> what's these about.
>
> I already did that, in the commit I installed today.
>
> Thanks.

Ah, I see thanks man.

Garid Zorigoo

bug#63028: 28.2; Rendering Traditional Mongolian Script

[Eli Zaretskii]

> From: "Garid Z." <garidzorigoo@gmail.com>
> Cc: 63028@debbugs.gnu.org
> Date: Mon, 24 Apr 2023 01:05:43 +0900
> 
> > FTR, what font are you using for this script?
> 
> It looks like I'm using Noto-sans, Take look at my (C-u C-x =) output
> file (attachment: letter.txt):

Yes, it's Noto Sans Mongolian.

> > Thanks.  However, I think I'd like to separate this line from the
> > existing Mongolian line (which uses the Cyrillic script).  I thought
> > about labeling it "Mongolian Traditional:, WDYT?  Alternatively, we
> > could rename the existing line to "Mongolian Cyrillic", if the
> > traditional Mongolian script is back in wide use nowadays.
> 
> (TBH, Traditional script are used rarely).
> But using "Mongolian Traditional" and "Mongolian Cyrillic" labels sound
> perfect to me. Should I make new file:
> emacs/lisp/leim/quail/mongolian.el and put these 2 input methods?
> or "Mongolian-cyrillic" still should belongs to other cyrillics
> in emacs/lisp/leim/quail/cyrillic.el? WDYT?
> 
> > In any
> > case, I will need to know how "Mongolian" is written in the Mongolian
> > script, can you tell that as well?
> 
> 
> Sure here: (Just in case, I also made it as attachment)
> - Mongolia - Монгол - ᠮᠣᠩᠭᠣᠯ 
> - Mongolian language - Монгол хэл -  ᠮᠣᠩᠭᠣᠯ ᠬᠡᠯᠡ
> - Mongolian traditional scrips - Монгол бичиг - ᠮᠣᠩᠭᠣᠯ ᠪᠢᠴᠢᠭ᠋

OK, I will modify HELLO and the rest of the files with this.  The
preferred encoding with this script is probably UTF-8, is that right?

> Should I make new file: emacs/lisp/leim/quail/mongolian.el and put
> these 2 input methods?

No need for a new file.  Just add the Mongolian input method to
lisp/leim/quail/misc-lang.el.

> For the input-method this, I need to couple of days to figure out and
> look for existing keyboard layout for traditional mongolian.

OK.

bug#63028: 28.2; Rendering Traditional Mongolian Script

[Garid Z.]

Eli Zaretskii <eliz@gnu.org> writes:

>> Sure here: (Just in case, I also made it as attachment)
>> - Mongolia - Монгол - ᠮᠣᠩᠭᠣᠯ 
>> - Mongolian language - Монгол хэл -  ᠮᠣᠩᠭᠣᠯ ᠬᠡᠯᠡ
>> - Mongolian traditional scrips - Монгол бичиг - ᠮᠣᠩᠭᠣᠯ ᠪᠢᠴᠢᠭ᠋
>
> OK, I will modify HELLO and the rest of the files with this.  The
> preferred encoding with this script is probably UTF-8, is that right?
>

I think so, I have no idea.

>> Should I make new file: emacs/lisp/leim/quail/mongolian.el and put
>> these 2 input methods?
>
> No need for a new file.  Just add the Mongolian input method to
> lisp/leim/quail/misc-lang.el.
>

Understood.


---
Garid Zorigoo

bug#63028: 28.2; Rendering Traditional Mongolian Script

[Eli Zaretskii]

> Cc: 63028@debbugs.gnu.org
> Date: Sun, 23 Apr 2023 21:30:21 +0300
> From: Eli Zaretskii <eliz@gnu.org>
> 
> > > Thanks.  However, I think I'd like to separate this line from the
> > > existing Mongolian line (which uses the Cyrillic script).  I thought
> > > about labeling it "Mongolian Traditional:, WDYT?  Alternatively, we
> > > could rename the existing line to "Mongolian Cyrillic", if the
> > > traditional Mongolian script is back in wide use nowadays.
> > 
> > (TBH, Traditional script are used rarely).
> > But using "Mongolian Traditional" and "Mongolian Cyrillic" labels sound
> > perfect to me. Should I make new file:
> > emacs/lisp/leim/quail/mongolian.el and put these 2 input methods?
> > or "Mongolian-cyrillic" still should belongs to other cyrillics
> > in emacs/lisp/leim/quail/cyrillic.el? WDYT?
> > 
> > > In any
> > > case, I will need to know how "Mongolian" is written in the Mongolian
> > > script, can you tell that as well?
> > 
> > 
> > Sure here: (Just in case, I also made it as attachment)
> > - Mongolia - Монгол - ᠮᠣᠩᠭᠣᠯ 
> > - Mongolian language - Монгол хэл -  ᠮᠣᠩᠭᠣᠯ ᠬᠡᠯᠡ
> > - Mongolian traditional scrips - Монгол бичиг - ᠮᠣᠩᠭᠣᠯ ᠪᠢᠴᠢᠭ᠋
> 
> OK, I will modify HELLO and the rest of the files with this.  The
> preferred encoding with this script is probably UTF-8, is that right?

Now done on the emacs-29 branch.  There are 2 Mongolian
language-environments, one using the Cyrillic script, the other using
the traditional script.  Each has a separate line in HELLO.  Please
see if everything looks OK.