bug#54914: 28.1.50; Unable to form Brahmi conjuncts in Emacs

bug#54914: 28.1.50; Unable to form Brahmi conjuncts in Emacs

[Sameer Singh]

[-- Attachment #1.1: Type: text/plain, Size: 4622 bytes --]

Brahmi just like any other Indic script also has the ability to form
conjuncts, but despite using "Segoe UI Historic" font which supports Brahmi
Conjuncts, Emacs is not able to display them.
Steps to Reproduce:
1. Add "(set-fontset-font t 'brahmi (font-spec :family "Segoe UI
Historic"))" to init.el

2. In the scratch buffer type any combination of brahmi
consonant+virama+consonant. e.g. 𑀓𑁆𑀯 𑀜𑁆𑀚 𑀝𑁆𑀞 𑀭𑁆𑀓𑁆𑀭𑁆𑀢 𑀳𑁆𑀮
𑀗𑁆𑀓𑁆𑀢

3. Instead of forming a conjunct they appear separately.
It is as if devanagari स्म is displayed as स्‌म

I have attached the correct render from "hb-view" in the file
"hb-brahmi.png" and the incorrect render from emacs in the file
"emacs-brahmi.png"
Note: "Noto Sans Brahmi" has negligible amount of brahmi conjuncts so it
will be preferable to use "Segoe UI Historic"

In GNU Emacs 28.1.50 (build 1, x86_64-pc-linux-gnu, GTK+ Version 3.24.33,
cairo version 1.17.6)
 of 2022-04-12 built on a0567577fe19
Repository revision: 322091604b73bf6ef8ca2458a97df41de600dc42
Repository branch: pgtk-nativecomp
Windowing system distributor 'System Description: Arch Linux

Configured using:
 'configure --prefix=/usr --sysconfdir=/etc --libexecdir=/usr/lib
 --localstatedir=/var --mandir=/usr/share/man --with-gameuser=:games
 --with-sound=alsa --with-modules --with-x-toolkit=gtk3 --with-cairo
 --with-xwidgets --with-native-compilation --with-pgtk
 --without-compress-install --without-gconf --without-gsettings
 --without-m17n-flt --enable-autodepend --enable-link-time-optimization
 'CFLAGS=-march=x86-64 -mtune=generic -O2 -pipe -fno-plt -fexceptions
 -Wp,-D_FORTIFY_SOURCE=2 -Wformat -Werror=format-security
 -fstack-clash-protection -fcf-protection -g -flto'
 LDFLAGS=-Wl,-O1,--sort-common,--as-needed,-z,relro,-z,now'

Configured features:
ACL CAIRO DBUS FREETYPE GIF GLIB GMP GNUTLS GPM HARFBUZZ JPEG JSON LCMS2
LIBOTF LIBSYSTEMD LIBXML2 MODULES NATIVE_COMP NOTIFY INOTIFY PDUMPER
PGTK PNG RSVG SECCOMP SOUND THREADS TIFF TOOLKIT_SCROLL_BARS XIM
XWIDGETS GTK3 ZLIB

Important settings:
  value of $LANG: hi_IN.UTF-8
  value of $XMODIFIERS: @im=ibus
  locale-coding-system: utf-8-unix

Major mode: Lisp Interaction

Minor modes in effect:
  text-scale-mode: t
  tooltip-mode: t
  global-eldoc-mode: t
  eldoc-mode: t
  show-paren-mode: t
  electric-indent-mode: t
  mouse-wheel-mode: t
  tool-bar-mode: t
  menu-bar-mode: t
  file-name-shadow-mode: t
  global-font-lock-mode: t
  font-lock-mode: t
  blink-cursor-mode: t
  auto-composition-mode: t
  auto-encryption-mode: t
  auto-compression-mode: t
  line-number-mode: t
  indent-tabs-mode: t
  transient-mark-mode: t

Load-path shadows:
None found.

Features:
(shadow sort mail-extr emacsbug message rmc puny dired dired-loaddefs
rfc822 mml mml-sec epa derived epg rfc6068 epg-config gnus-util rmail
rmail-loaddefs auth-source cl-seq eieio eieio-core cl-macs
eieio-loaddefs password-cache json map text-property-search seq byte-opt
gv bytecomp byte-compile cconv mm-decode mm-bodies mm-encode mail-parse
rfc2231 mailabbrev gmm-utils mailheader sendmail rfc2047 rfc2045
ietf-drums mm-util mail-prsvr mail-utils time-date subr-x cl-loaddefs
cl-lib face-remap iso-transl tooltip eldoc paren electric uniquify
ediff-hook vc-hooks lisp-float-type elisp-mode mwheel term/pgtk-win
pgtk-win term/common-win tool-bar dnd fontset image regexp-opt fringe
tabulated-list replace newcomment text-mode lisp-mode prog-mode register
page tab-bar menu-bar rfn-eshadow isearch easymenu timer select
scroll-bar mouse jit-lock font-lock syntax font-core term/tty-colors
frame minibuffer cl-generic cham georgian utf-8-lang misc-lang
vietnamese tibetan thai tai-viet lao korean japanese eucjp-ms cp51932
hebrew greek romanian slovak czech european ethiopic indian cyrillic
chinese composite emoji-zwj charscript charprop case-table epa-hook
jka-cmpr-hook help simple abbrev obarray cl-preloaded nadvice button
loaddefs faces cus-face macroexp files window text-properties overlay
sha1 md5 base64 format env code-pages mule custom widget
hashtable-print-readable backquote threads xwidget-internal dbusbind
inotify dynamic-setting font-render-setting cairo move-toolbar gtk
x-toolkit pgtk lcms2 multi-tty make-network-process native-compile
emacs)

Memory information:
((conses 16 71882 7789)
 (symbols 48 6679 0)
 (strings 32 19692 1760)
 (string-bytes 1 653683)
 (vectors 16 14235)
 (vector-slots 8 304684 17309)
 (floats 8 28 45)
 (intervals 56 397 0)
 (buffers 992 12))

[-- Attachment #1.2: Type: text/html, Size: 5098 bytes --]

[-- Attachment #2: hb-brahmi.png --]
[-- Type: image/png, Size: 18174 bytes --]

[-- Attachment #3: emacs-brahmi.png --]
[-- Type: image/png, Size: 7813 bytes --]

bug#54914: 28.1.50; Unable to form Brahmi conjuncts in Emacs

[Eli Zaretskii]

> From: Sameer Singh <lumarzeli30@gmail.com>
> Date: Wed, 13 Apr 2022 16:25:41 +0530
> 
> Brahmi just like any other Indic script also has the ability to form conjuncts, but despite using "Segoe UI
> Historic" font which supports Brahmi Conjuncts, Emacs is not able to display them.

This is a missing feature: we don't have composition rules for the
Brahmi script, and more generally don't have any support for that
script.  Patches to add that are welcome.

bug#54914: 28.1.50; Unable to form Brahmi conjuncts in Emacs

[Robert Pluim]

>>>>> On Wed, 13 Apr 2022 19:03:36 +0300, Eli Zaretskii <eliz@gnu.org> said:

    >> From: Sameer Singh <lumarzeli30@gmail.com>
    >> Date: Wed, 13 Apr 2022 16:25:41 +0530
    >> 
    >> Brahmi just like any other Indic script also has the ability to form conjuncts, but despite using "Segoe UI
    >> Historic" font which supports Brahmi Conjuncts, Emacs is not able to display them.

    Eli> This is a missing feature: we don't have composition rules for the
    Eli> Brahmi script, and more generally don't have any support for that
    Eli> script.  Patches to add that are welcome.

A brute-force example of what Eli's talking about would be:

(set-char-table-range composition-function-table
		      '(#x11000 . #x1107F)
		      (list (vector "[\x11000-\x1107F]+" 0 #'font-shape-gstring)))

which essentially means "hand off any sequence of Brahmi codepoints to
harfbuzz", which might end up being slow. From your example, youʼre
requesting more something like "consonant, virami, consonant", which
is also possible, but we'll need your guidance to tell us when the
results are acceptable. (If anyone is interested, thereʼs section 14.1
of the Unicode 14 standard that describes how they think Brahmi should
be handled, but Iʼve not tried to understand it yet).

Robert
-- 

bug#54914: 28.1.50; Unable to form Brahmi conjuncts in Emacs

[समीर सिंह Sameer Singh]

[-- Attachment #1: Type: text/plain, Size: 1624 bytes --]

I am not capable enough to write my own patch for it, but will be more than
happy to provide help in implementing it, by validating results.

On Wed, Apr 13, 2022 at 10:26 PM Robert Pluim <rpluim@gmail.com> wrote:

> >>>>> On Wed, 13 Apr 2022 19:03:36 +0300, Eli Zaretskii <eliz@gnu.org>
> said:
>
>     >> From: Sameer Singh <lumarzeli30@gmail.com>
>     >> Date: Wed, 13 Apr 2022 16:25:41 +0530
>     >>
>     >> Brahmi just like any other Indic script also has the ability to
> form conjuncts, but despite using "Segoe UI
>     >> Historic" font which supports Brahmi Conjuncts, Emacs is not able
> to display them.
>
>     Eli> This is a missing feature: we don't have composition rules for the
>     Eli> Brahmi script, and more generally don't have any support for that
>     Eli> script.  Patches to add that are welcome.
>
> A brute-force example of what Eli's talking about would be:
>
> (set-char-table-range composition-function-table
>                       '(#x11000 . #x1107F)
>                       (list (vector "[\x11000-\x1107F]+" 0
> #'font-shape-gstring)))
>
> which essentially means "hand off any sequence of Brahmi codepoints to
> harfbuzz", which might end up being slow. From your example, youʼre
> requesting more something like "consonant, virami, consonant", which
> is also possible, but we'll need your guidance to tell us when the
> results are acceptable. (If anyone is interested, thereʼs section 14.1
> of the Unicode 14 standard that describes how they think Brahmi should
> be handled, but Iʼve not tried to understand it yet).
>
> Robert
> --
>

[-- Attachment #2: Type: text/html, Size: 2226 bytes --]

bug#54914: 28.1.50; Unable to form Brahmi conjuncts in Emacs

[Eli Zaretskii]

> From: Robert Pluim <rpluim@gmail.com>
> Cc: Eli Zaretskii <eliz@gnu.org>,  54914@debbugs.gnu.org
> Date: Wed, 13 Apr 2022 18:56:21 +0200
> 
> A brute-force example of what Eli's talking about would be:
> 
> (set-char-table-range composition-function-table
> 		      '(#x11000 . #x1107F)
> 		      (list (vector "[\x11000-\x1107F]+" 0 #'font-shape-gstring)))

Does that really work?  I'm not sure it will catch the "consonant,
virama, consonant" rule (because the virama is a nonspacing mark, and
we have a default rule for those in composite.el).

bug#54914: 28.1.50; Unable to form Brahmi conjuncts in Emacs

[Robert Pluim]

>>>>> On Wed, 13 Apr 2022 21:26:55 +0300, Eli Zaretskii <eliz@gnu.org> said:

    >> From: Robert Pluim <rpluim@gmail.com>
    >> Cc: Eli Zaretskii <eliz@gnu.org>,  54914@debbugs.gnu.org
    >> Date: Wed, 13 Apr 2022 18:56:21 +0200
    >> 
    >> A brute-force example of what Eli's talking about would be:
    >> 
    >> (set-char-table-range composition-function-table
    >> '(#x11000 . #x1107F)
    >> (list (vector "[\x11000-\x1107F]+" 0 #'font-shape-gstring)))

    Eli> Does that really work?  I'm not sure it will catch the "consonant,
    Eli> virama, consonant" rule (because the virama is a nonspacing mark, and
    Eli> we have a default rule for those in composite.el).

We do indeed. The following handles only Ka through Lla in combination
with Virama, and seems to do the right thing here with Segoe UI
Historic (which I donʼt think we can set as default for the Brahmi
script, since itʼs non-free).

;; <Consonant, Virama, Consonant>, or <Consonant, Virama>, where the
;; latter indicates a vowelless consonant.
(set-char-table-range composition-function-table
		      '(#x11013 . #x11034)
		      (list (vector "[\x11013-\x11034]\x11046[\x11013-\x11034]?"
                                    0 #'font-shape-gstring)))

Robert
-- 

bug#54914: 28.1.50; Unable to form Brahmi conjuncts in Emacs

[Eli Zaretskii]

> From: Robert Pluim <rpluim@gmail.com>
> Cc: 54914@debbugs.gnu.org,  lumarzeli30@gmail.com
> Date: Thu, 14 Apr 2022 11:07:23 +0200
> 
> >>>>> On Wed, 13 Apr 2022 21:26:55 +0300, Eli Zaretskii <eliz@gnu.org> said:
> 
>     >> From: Robert Pluim <rpluim@gmail.com>
>     >> Cc: Eli Zaretskii <eliz@gnu.org>,  54914@debbugs.gnu.org
>     >> Date: Wed, 13 Apr 2022 18:56:21 +0200
>     >> 
>     >> A brute-force example of what Eli's talking about would be:
>     >> 
>     >> (set-char-table-range composition-function-table
>     >> '(#x11000 . #x1107F)
>     >> (list (vector "[\x11000-\x1107F]+" 0 #'font-shape-gstring)))
> 
>     Eli> Does that really work?  I'm not sure it will catch the "consonant,
>     Eli> virama, consonant" rule (because the virama is a nonspacing mark, and
>     Eli> we have a default rule for those in composite.el).
> 
> We do indeed. The following handles only Ka through Lla in combination
> with Virama, and seems to do the right thing here with Segoe UI
> Historic (which I donʼt think we can set as default for the Brahmi
> script, since itʼs non-free).
> 
> ;; <Consonant, Virama, Consonant>, or <Consonant, Virama>, where the
> ;; latter indicates a vowelless consonant.
> (set-char-table-range composition-function-table
> 		      '(#x11013 . #x11034)
> 		      (list (vector "[\x11013-\x11034]\x11046[\x11013-\x11034]?"
>                                     0 #'font-shape-gstring)))

I've now added support for the Brahmi script to the master branch;
please test and report any problems you find.

(We cannot advertise Segoe UI Historic, unfortunately, since it is not
a free font, AFAICT.)

bug#54914: 28.1.50; Unable to form Brahmi conjuncts in Emacs

[समीर सिंह Sameer Singh]

[-- Attachment #1.1: Type: text/plain, Size: 2227 bytes --]

Thank you very much for adding support for Brahmi!

I can notice two problems
1. Vowel signs are not attached properly to conjuncts.
    Correct output=hb-conjuncts-maatra.png   Incorrect
Output=emacs-conjuncts-maatra.png

2. Conjuncts with more than two consonants are not supported.
    Correct output=hb-tri-conjuncts.png   Incorrect
Output=emacs-tri-conjuncts.png

On Sat, Apr 16, 2022 at 3:51 PM Eli Zaretskii <eliz@gnu.org> wrote:

> > From: Robert Pluim <rpluim@gmail.com>
> > Cc: 54914@debbugs.gnu.org,  lumarzeli30@gmail.com
> > Date: Thu, 14 Apr 2022 11:07:23 +0200
> >
> > >>>>> On Wed, 13 Apr 2022 21:26:55 +0300, Eli Zaretskii <eliz@gnu.org>
> said:
> >
> >     >> From: Robert Pluim <rpluim@gmail.com>
> >     >> Cc: Eli Zaretskii <eliz@gnu.org>,  54914@debbugs.gnu.org
> >     >> Date: Wed, 13 Apr 2022 18:56:21 +0200
> >     >>
> >     >> A brute-force example of what Eli's talking about would be:
> >     >>
> >     >> (set-char-table-range composition-function-table
> >     >> '(#x11000 . #x1107F)
> >     >> (list (vector "[\x11000-\x1107F]+" 0 #'font-shape-gstring)))
> >
> >     Eli> Does that really work?  I'm not sure it will catch the
> "consonant,
> >     Eli> virama, consonant" rule (because the virama is a nonspacing
> mark, and
> >     Eli> we have a default rule for those in composite.el).
> >
> > We do indeed. The following handles only Ka through Lla in combination
> > with Virama, and seems to do the right thing here with Segoe UI
> > Historic (which I donʼt think we can set as default for the Brahmi
> > script, since itʼs non-free).
> >
> > ;; <Consonant, Virama, Consonant>, or <Consonant, Virama>, where the
> > ;; latter indicates a vowelless consonant.
> > (set-char-table-range composition-function-table
> >                     '(#x11013 . #x11034)
> >                     (list (vector
> "[\x11013-\x11034]\x11046[\x11013-\x11034]?"
> >                                     0 #'font-shape-gstring)))
>
> I've now added support for the Brahmi script to the master branch;
> please test and report any problems you find.
>
> (We cannot advertise Segoe UI Historic, unfortunately, since it is not
> a free font, AFAICT.)
>

[-- Attachment #1.2: Type: text/html, Size: 3408 bytes --]

[-- Attachment #2: emacs-conjuncts-maatra.png --]
[-- Type: image/png, Size: 8641 bytes --]

[-- Attachment #3: hb-conjuncts-maatra.png --]
[-- Type: image/png, Size: 7735 bytes --]

[-- Attachment #4: emacs-tri-conjuncts.png --]
[-- Type: image/png, Size: 12813 bytes --]

[-- Attachment #5: hb-tri-conjuncts.png --]
[-- Type: image/png, Size: 9914 bytes --]

bug#54914: 28.1.50; Unable to form Brahmi conjuncts in Emacs

[Eli Zaretskii]

> From: समीर सिंह Sameer Singh <lumarzeli30@gmail.com>
> Date: Sat, 16 Apr 2022 17:43:49 +0530
> Cc: Robert Pluim <rpluim@gmail.com>, 54914@debbugs.gnu.org
> 
> I can notice two problems
> 1. Vowel signs are not attached properly to conjuncts.
>     Correct output=hb-conjuncts-maatra.png   Incorrect Output=emacs-conjuncts-maatra.png
> 
> 2. Conjuncts with more than two consonants are not supported.
>     Correct output=hb-tri-conjuncts.png   Incorrect Output=emacs-tri-conjuncts.png

Thanks, but please always show the codepoints themselves that are or
should be involved, because it's impossible to glean that from the
images.

Also, where do you see the shaping rules that support those
combinations?  I didn't find them in the Unicode Standard or anywhere
else I've read about Brahmi.  What are your sources?  That hb-view
does something doesn't yet mean it's correct, since HarfBuzz is just
another piece of software, and can have bugs.

bug#54914: 28.1.50; Unable to form Brahmi conjuncts in Emacs

[समीर सिंह Sameer Singh]

[-- Attachment #1.1: Type: text/plain, Size: 2453 bytes --]

>
> Thanks, but please always show the codepoints themselves that are or
> should be involved, because it's impossible to glean that from the
> images.
>

Sorry for not including the codepoints, here they are:
hb-conjuncts-maatra.png & emacs-conjuncts-maatra.png from top to bottom
1. \011013 \011046 \01102C \011038(𑀓𑁆𑀬𑀸)
2. \011013 \011046 \01102E \011043 (𑀓𑁆𑀮𑁃)
3. \011032 \011046 \01102C \01103B(𑀲𑁆𑀬𑀻)

hb-tri-conjuncts.png & emacs-tri-conjuncts.png from top to bottom
1. \011013 \011046 \01102E \011046 \01102F (𑀓𑁆𑀮𑁆𑀯)
2. \01102F \011046 \01102C \011046 \01102F (𑀯𑁆𑀬𑁆𑀯)
3. \01102B \011046 \011027 \011046 \01102D (𑀫𑁆𑀧𑁆𑀭)

Also, where do you see the shaping rules that support those
> combinations?  I didn't find them in the Unicode Standard or anywhere
> else I've read about Brahmi.  What are your sources?
>

Languages like Sanskrit has three or more than three conjuncts. They are
supported in scripts like devanagari (
https://en.wikipedia.org/wiki/Devanagari_conjuncts#Tri-_and_tetra-consonantal_conjuncts
)
So brahmi should be able to support them too.
I could find one example in the "Proposal to include Brahmi in Unicode"
document (https://www.unicode.org/L2/L2003/03249r-brahmi-proposal.pdf)
(Page 5, 1st image, 3rd character)

On Sat, Apr 16, 2022 at 5:56 PM Eli Zaretskii <eliz@gnu.org> wrote:

> > From: समीर सिंह Sameer Singh <lumarzeli30@gmail.com>
> > Date: Sat, 16 Apr 2022 17:43:49 +0530
> > Cc: Robert Pluim <rpluim@gmail.com>, 54914@debbugs.gnu.org
> >
> > I can notice two problems
> > 1. Vowel signs are not attached properly to conjuncts.
> >     Correct output=hb-conjuncts-maatra.png   Incorrect
> Output=emacs-conjuncts-maatra.png
> >
> > 2. Conjuncts with more than two consonants are not supported.
> >     Correct output=hb-tri-conjuncts.png   Incorrect
> Output=emacs-tri-conjuncts.png
>
> Thanks, but please always show the codepoints themselves that are or
> should be involved, because it's impossible to glean that from the
> images.
>
> Also, where do you see the shaping rules that support those
> combinations?  I didn't find them in the Unicode Standard or anywhere
> else I've read about Brahmi.  What are your sources?  That hb-view
> does something doesn't yet mean it's correct, since HarfBuzz is just
> another piece of software, and can have bugs.
>

[-- Attachment #1.2: Type: text/html, Size: 3678 bytes --]

[-- Attachment #2: brahmi-tri-conjunct.png --]
[-- Type: image/png, Size: 3295 bytes --]

bug#54914: 28.1.50; Unable to form Brahmi conjuncts in Emacs

[Eli Zaretskii]

> From: समीर सिंह Sameer Singh <lumarzeli30@gmail.com>
> Date: Sat, 16 Apr 2022 19:04:15 +0530
> Cc: Robert Pluim <rpluim@gmail.com>, 54914@debbugs.gnu.org
> 
> Sorry for not including the codepoints, here they are:
> hb-conjuncts-maatra.png & emacs-conjuncts-maatra.png from top to bottom
> 1. \011013 \011046 \01102C \011038(𑀓𑁆𑀬𑀸)
> 2. \011013 \011046 \01102E \011043 (𑀓𑁆𑀮𑁃)
> 3. \011032 \011046 \01102C \01103B(𑀲𑁆𑀬𑀻)
> 
> hb-tri-conjuncts.png & emacs-tri-conjuncts.png from top to bottom
> 1. \011013 \011046 \01102E \011046 \01102F (𑀓𑁆𑀮𑁆𑀯)
> 2. \01102F \011046 \01102C \011046 \01102F (𑀯𑁆𑀬𑁆𑀯)
> 3. \01102B \011046 \011027 \011046 \01102D (𑀫𑁆𑀧𑁆𑀭)
> 
>  Also, where do you see the shaping rules that support those
>  combinations?  I didn't find them in the Unicode Standard or anywhere
>  else I've read about Brahmi.  What are your sources? 
> 
>  
> Languages like Sanskrit has three or more than three conjuncts. They are supported in scripts like
> devanagari (https://en.wikipedia.org/wiki/Devanagari_conjuncts#Tri-_and_tetra-consonantal_conjuncts)
> So brahmi should be able to support them too.
> I could find one example in the "Proposal to include Brahmi in Unicode" document
> (https://www.unicode.org/L2/L2003/03249r-brahmi-proposal.pdf) (Page 5, 1st image, 3rd character)

I've attempted to enhance the composition rules, please try the latest
master.  They don't always work on my system, but I think that's
because of the font I use and perhaps also the relatively old version
of HarfBuzz I have here.

bug#54914: 28.1.50; Unable to form Brahmi conjuncts in Emacs

[समीर सिंह Sameer Singh]

[-- Attachment #1.1: Type: text/plain, Size: 2108 bytes --]

Vowel signs U+1103E U+1103F U+11040 and U+11041 are not connected properly
to conjuncts.

Here are the codepoints of the text I entered:
11013 11046 11032 1103E (𑀓𑁆𑀲𑀾)
11013 11046 11032 1103F (𑀓𑁆𑀲𑀿)
11013 11046 11032 11040 (𑀓𑁆𑀲𑁀)
11013 11046 11032 11041 (𑀓𑁆𑀲𑁁)

On Sat, Apr 16, 2022 at 9:59 PM Eli Zaretskii <eliz@gnu.org> wrote:

> > From: समीर सिंह Sameer Singh <lumarzeli30@gmail.com>
> > Date: Sat, 16 Apr 2022 19:04:15 +0530
> > Cc: Robert Pluim <rpluim@gmail.com>, 54914@debbugs.gnu.org
> >
> > Sorry for not including the codepoints, here they are:
> > hb-conjuncts-maatra.png & emacs-conjuncts-maatra.png from top to bottom
> > 1. \011013 \011046 \01102C \011038(𑀓𑁆𑀬𑀸)
> > 2. \011013 \011046 \01102E \011043 (𑀓𑁆𑀮𑁃)
> > 3. \011032 \011046 \01102C \01103B(𑀲𑁆𑀬𑀻)
> >
> > hb-tri-conjuncts.png & emacs-tri-conjuncts.png from top to bottom
> > 1. \011013 \011046 \01102E \011046 \01102F (𑀓𑁆𑀮𑁆𑀯)
> > 2. \01102F \011046 \01102C \011046 \01102F (𑀯𑁆𑀬𑁆𑀯)
> > 3. \01102B \011046 \011027 \011046 \01102D (𑀫𑁆𑀧𑁆𑀭)
> >
> >  Also, where do you see the shaping rules that support those
> >  combinations?  I didn't find them in the Unicode Standard or anywhere
> >  else I've read about Brahmi.  What are your sources?
> >
> >
> > Languages like Sanskrit has three or more than three conjuncts. They are
> supported in scripts like
> > devanagari (
> https://en.wikipedia.org/wiki/Devanagari_conjuncts#Tri-_and_tetra-consonantal_conjuncts
> )
> > So brahmi should be able to support them too.
> > I could find one example in the "Proposal to include Brahmi in Unicode"
> document
> > (https://www.unicode.org/L2/L2003/03249r-brahmi-proposal.pdf) (Page 5,
> 1st image, 3rd character)
>
> I've attempted to enhance the composition rules, please try the latest
> master.  They don't always work on my system, but I think that's
> because of the font I use and perhaps also the relatively old version
> of HarfBuzz I have here.
>

[-- Attachment #1.2: Type: text/html, Size: 3042 bytes --]

[-- Attachment #2: emacs-rl.png --]
[-- Type: image/png, Size: 7118 bytes --]

[-- Attachment #3: hb-rl.png --]
[-- Type: image/png, Size: 14469 bytes --]

bug#54914: 28.1.50; Unable to form Brahmi conjuncts in Emacs

[Robert Pluim]

>>>>> On Sun, 17 Apr 2022 18:19:47 +0530, समीर सिंह Sameer Singh <lumarzeli30@gmail.com> said:

    समीर> Vowel signs U+1103E U+1103F U+11040 and U+11041 are not connected properly
    समीर> to conjuncts.

    समीर> Here are the codepoints of the text I entered:
    समीर> 11013 11046 11032 1103E (𑀓𑁆𑀲𑀾)
    समीर> 11013 11046 11032 1103F (𑀓𑁆𑀲𑀿)
    समीर> 11013 11046 11032 11040 (𑀓𑁆𑀲𑁀)
    समीर> 11013 11046 11032 11041 (𑀓𑁆𑀲𑁁)

How about this?

diff --git a/lisp/language/indian.el b/lisp/language/indian.el
index 0a50dd999f..ef095ddc3b 100644
--- a/lisp/language/indian.el
+++ b/lisp/language/indian.el
@@ -397,7 +397,7 @@ malayalam-composable-pattern
 ;; Brahmi composition rules
 (let ((consonant     "[\U00011013-\U00011034]")
       (non-consonant "[^\U00011013-\U00011034\U00011046\U0001107F]")
-      (vowel         "[\U00011038-\U0001103D\U00011042-\U00011045]")
+      (vowel         "[\U00011038-\U00011045]")
       (numeral       "[\U00011052-\U00011065]")
       (multiplier    "[\U00011064\U00011065]")
       (virama        "\U00011046")

Robert
-- 

bug#54914: 28.1.50; Unable to form Brahmi conjuncts in Emacs

[समीर सिंह Sameer Singh]

[-- Attachment #1.1: Type: text/plain, Size: 1664 bytes --]

Everything seems to be working fine now.
Thank you Eli and Robert for your hardwork.

Here is "Welcome to Emacs" in Sanskrit in the Brahmi script rendered
correctly in Emacs!
𑀈𑀫𑁃𑀓𑁆𑀲𑁂 𑀲𑁆𑀯𑀸𑀕𑀢𑀫𑁆 (ईमैक्से स्वागतम् īmaiksē svāgatam)

On Sun, Apr 17, 2022 at 7:31 PM Robert Pluim <rpluim@gmail.com> wrote:

> >>>>> On Sun, 17 Apr 2022 18:19:47 +0530, समीर सिंह Sameer Singh <
> lumarzeli30@gmail.com> said:
>
>     समीर> Vowel signs U+1103E U+1103F U+11040 and U+11041 are not
> connected properly
>     समीर> to conjuncts.
>
>     समीर> Here are the codepoints of the text I entered:
>     समीर> 11013 11046 11032 1103E (𑀓𑁆𑀲𑀾)
>     समीर> 11013 11046 11032 1103F (𑀓𑁆𑀲𑀿)
>     समीर> 11013 11046 11032 11040 (𑀓𑁆𑀲𑁀)
>     समीर> 11013 11046 11032 11041 (𑀓𑁆𑀲𑁁)
>
> How about this?
>
> diff --git a/lisp/language/indian.el b/lisp/language/indian.el
> index 0a50dd999f..ef095ddc3b 100644
> --- a/lisp/language/indian.el
> +++ b/lisp/language/indian.el
> @@ -397,7 +397,7 @@ malayalam-composable-pattern
>  ;; Brahmi composition rules
>  (let ((consonant     "[\U00011013-\U00011034]")
>        (non-consonant "[^\U00011013-\U00011034\U00011046\U0001107F]")
> -      (vowel         "[\U00011038-\U0001103D\U00011042-\U00011045]")
> +      (vowel         "[\U00011038-\U00011045]")
>        (numeral       "[\U00011052-\U00011065]")
>        (multiplier    "[\U00011064\U00011065]")
>        (virama        "\U00011046")
>
> Robert
> --
>

[-- Attachment #1.2: Type: text/html, Size: 2282 bytes --]

[-- Attachment #2: emacs.png --]
[-- Type: image/png, Size: 7272 bytes --]

bug#54914: 28.1.50; Unable to form Brahmi conjuncts in Emacs

[Eli Zaretskii]

> From: Robert Pluim <rpluim@gmail.com>
> Cc: Eli Zaretskii <eliz@gnu.org>,  54914@debbugs.gnu.org
> Date: Sun, 17 Apr 2022 16:01:01 +0200
> 
> >>>>> On Sun, 17 Apr 2022 18:19:47 +0530, समीर सिंह Sameer Singh <lumarzeli30@gmail.com> said:
> 
>     समीर> Vowel signs U+1103E U+1103F U+11040 and U+11041 are not connected properly
>     समीर> to conjuncts.
> 
>     समीर> Here are the codepoints of the text I entered:
>     समीर> 11013 11046 11032 1103E (𑀓𑁆𑀲𑀾)
>     समीर> 11013 11046 11032 1103F (𑀓𑁆𑀲𑀿)
>     समीर> 11013 11046 11032 11040 (𑀓𑁆𑀲𑁀)
>     समीर> 11013 11046 11032 11041 (𑀓𑁆𑀲𑁁)
> 
> How about this?

Thanks, installed.

bug#54914: 28.1.50; Unable to form Brahmi conjuncts in Emacs

[Eli Zaretskii]

> From: समीर सिंह Sameer Singh <lumarzeli30@gmail.com>
> Date: Sun, 17 Apr 2022 19:55:36 +0530
> Cc: Eli Zaretskii <eliz@gnu.org>, 54914@debbugs.gnu.org
> 
> Everything seems to be working fine now.
> Thank you Eli and Robert for your hardwork.

Thanks for testing.

> Here is "Welcome to Emacs" in Sanskrit in the Brahmi script rendered correctly in Emacs!
> 𑀈𑀫𑁃𑀓𑁆𑀲𑁂 𑀲𑁆𑀯𑀸𑀕𑀢𑀫𑁆 (ईमैक्से स्वागतम् īmaiksē svāgatam)

Thanks, but we are not done yet.

First, we need to have a "hello" greeting in Brahmi script for the
etc/HELLO file.  Can you show such a greeting?  It is okay to have two
or even 3 if there's no single phrase for that.

We also need to know how to write "Brahmi" in Brahmi script, but I
think I found the answer for that: "𑀩𑁆𑀭𑀸𑀳𑁆𑀫𑀻", right?  If that is
incorrect, please tell how to write this correctly.

Last, but not least, we should have an input method for Brahmi.  I
looked around, but found only one description of a keyboard map for
Brahmi, which is here:

  https://github.com/kaamanita/brahmi

Can you look and tell whether the keyboard layout shown there is
reasonable?  If not, can you point me to a better one, or maybe
suggest such a layout yourself?  All I need to know is which key on a
standard QWERTY keyboard should produce what Brahmi character(s), and
then I can write the input method myself.

Thanks.

bug#54914: 28.1.50; Unable to form Brahmi conjuncts in Emacs

[समीर सिंह Sameer Singh]

[-- Attachment #1.1: Type: text/plain, Size: 2504 bytes --]

>
> First, we need to have a "hello" greeting in Brahmi script for the
> etc/HELLO file.  Can you show such a greeting?
>
𑀦𑀫𑀲𑁆𑀢𑁂 (नमस्ते namastē) (11026 1102B 11032 11046 11022 11042) could
work in Sanskrit

We also need to know how to write "Brahmi" in Brahmi script, but I
> think I found the answer for that: "𑀩𑁆𑀭𑀸𑀳𑁆𑀫𑀻", right?  If that is
> incorrect, please tell how to write this correctly.
>

Yes this is correct

Can you look and tell whether the keyboard layout shown there is
> reasonable?  If not, can you point me to a better one, or maybe
> suggest such a layout yourself?  All I need to know is which key on a
> standard QWERTY keyboard should produce what Brahmi character(s), and
> then I can write the input method myself.
>

Here is a modified version of the layout which I use

On Sun, Apr 17, 2022 at 8:27 PM Eli Zaretskii <eliz@gnu.org> wrote:

> > From: समीर सिंह Sameer Singh <lumarzeli30@gmail.com>
> > Date: Sun, 17 Apr 2022 19:55:36 +0530
> > Cc: Eli Zaretskii <eliz@gnu.org>, 54914@debbugs.gnu.org
> >
> > Everything seems to be working fine now.
> > Thank you Eli and Robert for your hardwork.
>
> Thanks for testing.
>
> > Here is "Welcome to Emacs" in Sanskrit in the Brahmi script rendered
> correctly in Emacs!
> > 𑀈𑀫𑁃𑀓𑁆𑀲𑁂 𑀲𑁆𑀯𑀸𑀕𑀢𑀫𑁆 (ईमैक्से स्वागतम् īmaiksē svāgatam)
>
> Thanks, but we are not done yet.
>
> First, we need to have a "hello" greeting in Brahmi script for the
> etc/HELLO file.  Can you show such a greeting?  It is okay to have two
> or even 3 if there's no single phrase for that.
>
> We also need to know how to write "Brahmi" in Brahmi script, but I
> think I found the answer for that: "𑀩𑁆𑀭𑀸𑀳𑁆𑀫𑀻", right?  If that is
> incorrect, please tell how to write this correctly.
>
> Last, but not least, we should have an input method for Brahmi.  I
> looked around, but found only one description of a keyboard map for
> Brahmi, which is here:
>
>   https://github.com/kaamanita/brahmi
>
> Can you look and tell whether the keyboard layout shown there is
> reasonable?  If not, can you point me to a better one, or maybe
> suggest such a layout yourself?  All I need to know is which key on a
> standard QWERTY keyboard should produce what Brahmi character(s), and
> then I can write the input method myself.
>
> Thanks.
>

[-- Attachment #1.2: Type: text/html, Size: 4076 bytes --]

[-- Attachment #2: brahmi2-kb --]
[-- Type: application/octet-stream, Size: 5057 bytes --]

// Name:        Brahmi phonetic
// Description: A phonetic keyboard layout for writing in Brahmi.
// Author:      समीर सिंह (Sameer Singh) <lumarzeli30@gmail.com>
//
partial alphanumeric_keys
xkb_symbols "brahmi" {
     name[Group1] = "Brahmi (phonetic)";
     key.type="FOUR_LEVEL";

     // Roman digits
     key <TLDE> { [ grave,  asciitilde,  U20B9                 ] }; // U20B9: New Indian Rupee Sign
     key <AE01> { [ U11067, exclam,      1,      U11052        ] }; // digit one, number one
     key <AE02> { [ U11068, at,          2,      U11053        ] }; // digit two, number two
     key <AE03> { [ U11069, numbersign,  3,      U11054        ] }; // digit three, number three
     key <AE04> { [ U1106A, dollar,      4,      U11055        ] }; // digit four, number four
     key <AE05> { [ U1106B, percent,     5,      U11056        ] }; // digit five, number five
     key <AE06> { [ U1106C, asciicircum, 6,      U11057 ] }; // digit six, number six
     key <AE07> { [ U1106D, ampersand,   7,      U11058 ] }; // digit seven, number seven
     key <AE08> { [ U1106E, asterisk,    8,      U11059] }; // digit eight, number eight
     key <AE09> { [ U1106F, parenleft,   9,      U1105A        ] }; // digit nine, number nine
     key <AE10> { [ U11066, parenright,  0,      U1105B        ] }; // digit zero, number ten
     key <AE11> { [ minus,  underscore,  U1105C, U1105D        ] }; // number twenty, number thirty
     key <AE12> { [ equal,  plus,        U1105E, U1105F        ] }; // number forty, number fifty
     key <BKSL> { [ U005C,  U007C,       U11047, U11048        ] }; // backslash, pipe, danda, double danda

     // Q Row
     key <AD01> { [ U1101D,       U1101E, U11003, U11060 ] }; // letter TTA, letter TTHA, jihvamuliya, number sixty
     key <AD02> { [ U1101F,       U11020, U11004, U11061 ] }; // letter DDA, letter DDHA, upadhimaniya, number seventy
     key <AD03> { [ U11042,       U11043, U1100F, U11010 ] }; // vowel sign E, vowel sign AI, letter E, letter AI
     key <AD04> { [ U1102D,       U1103E, U1100B, U11036 ] }; // letter RA, vowel sign vocalic R, vocalic R, letter old tamil RRA
     key <AD05> { [ U11022,       U11023, U11062         ] }; // letter TA, letter THA, number eighty
     key <AD06> { [ U1102C,       U11063, U11064, U11065 ] }; // latter YA, number ninety, number one hundred, number one thousand
     key <AD07> { [ U1103C,       U1103D, U11009, U1100A ] }; // vowel sign U, vowel sign UU, letter U, letter UU
     key <AD08> { [ U1103A,       U1103B, U11007, U11008 ] }; // vowel sign I, vowel sign II, letter I, letter II
     key <AD09> { [ U11044,       U11045, U11011, U11012 ] }; // vowel sign O, vowel sign AU, letter O, letter AU
     key <AD10> { [ U11027,       U11028, U11073, U11071 ] }; // letter PA, letter PHA, vowel sign old tamil short E, letter old tamil short E
     key <AD11> { [ bracketleft,  braceleft, U11074, U11072 ] }; // vowel sign old tamil short O, letter old tamil short O
     key <AD12> { [ bracketright, braceright             ] };

     // A Row
     key <AC01> { [ U11038,     U11006, U11005, U11039 ] }; // vowel sign AA, letter AA, letter A, vowel sign bhattiprolu AA
     key <AC02> { [ U11032,     U11030, U11031         ] }; // letter SA, letter SHA, letter SSA
     key <AC03> { [ U11024,     U11025, U11036         ] }; // letter DA, letter DHA, old tamil RRA
     key <AC04> { [ U11046,     U1103F, U1100C, U11070 ] }; // virama, vowel sign vocalic RR, letter vocalic RR, old tamil virama
     key <AC05> { [ U11015,     U11016                 ] }; //letter GA, letter GHA
     key <AC06> { [ U11033,     U11002                 ] }; // letter HA, visarga
     key <AC07> { [ U1101A,     U1101B                 ] }; // letter JA, letter JHA
     key <AC08> { [ U11013,     U11014                 ] }; // letter KA, letter KHA
     key <AC09> { [ U1102E,     U11034, U11035, U11075 ] }; // letter LA, letter LLA, old tamil LLLA, old tamil LLA
     key <AC10> { [ semicolon,  colon         ] };
     key <AC11> { [ apostrophe, quotedbl               ] };

     // Z Row
     key <AB01> { [ U11040, U1100D,   U11041, U1100E   ] }; // vowel sign vocalic L, vocalic L, vowel sign vocalic LL, vocalic LL
     key <AB02> { [ U11049, U1104A,   U1104B, U1104C   ] }; // punctuation dot, punctuation double dot, punctuation line, punctuation crescent bar
     key <AB03> { [ U11018, U11019, U200C, U200D                    ] }; // letter CA, letter CHA, ZWNJ, ZWJ
     key <AB04> { [ U1102F, U11037                     ] }; // letter VA, old tamil NNNA
     key <AB05> { [ U11029, U1102A                     ] }; // letter BA, letter BHA
     key <AB06> { [ U11026, U11021,   U11017, U1101C   ] }; // letter NA, letter NNA, letter NYA, letter NGA
     key <AB07> { [ U1102B, U11001, U11000             ] }; // letter MA, anusvara, candrabindu
     key <AB08> { [ comma,  U1104D                     ] }; // punctuation lotus
     key <AB10> { [ slash,  question, U1107F           ] }; // number joiner

    include "level3(ralt_switch)"
};

bug#54914: 28.1.50; Unable to form Brahmi conjuncts in Emacs

[Eli Zaretskii]

> From: समीर सिंह Sameer Singh <lumarzeli30@gmail.com>
> Date: Sun, 17 Apr 2022 21:20:50 +0530
> Cc: Robert Pluim <rpluim@gmail.com>, 54914@debbugs.gnu.org
> 
>  First, we need to have a "hello" greeting in Brahmi script for the
>  etc/HELLO file.  Can you show such a greeting?
> 
> 𑀦𑀫𑀲𑁆𑀢𑁂 (नमस्ते namastē) (11026 1102B 11032 11046 11022 11042) could work in Sanskrit
> 
>  We also need to know how to write "Brahmi" in Brahmi script, but I
>  think I found the answer for that: "𑀩𑁆𑀭𑀸𑀳𑁆𑀫𑀻", right?  If that is
>  incorrect, please tell how to write this correctly.
> 
> Yes this is correct 
> 
>  Can you look and tell whether the keyboard layout shown there is
>  reasonable?  If not, can you point me to a better one, or maybe
>  suggest such a layout yourself?  All I need to know is which key on a
>  standard QWERTY keyboard should produce what Brahmi character(s), and
>  then I can write the input method myself.
> 
> Here is a modified version of the layout which I use 

Thanks, I added the missing pieces, please see if I missed anything or
made any mistakes.  If not, I think we can close this bug.

bug#54914: 28.1.50; Unable to form Brahmi conjuncts in Emacs

[समीर सिंह Sameer Singh]

[-- Attachment #1: Type: text/plain, Size: 1396 bytes --]

Looks good to me 👍

Again, thank you for your hard work.


On Sun, Apr 17, 2022 at 11:09 PM Eli Zaretskii <eliz@gnu.org> wrote:

> > From: समीर सिंह Sameer Singh <lumarzeli30@gmail.com>
> > Date: Sun, 17 Apr 2022 21:20:50 +0530
> > Cc: Robert Pluim <rpluim@gmail.com>, 54914@debbugs.gnu.org
> >
> >  First, we need to have a "hello" greeting in Brahmi script for the
> >  etc/HELLO file.  Can you show such a greeting?
> >
> > 𑀦𑀫𑀲𑁆𑀢𑁂 (नमस्ते namastē) (11026 1102B 11032 11046 11022 11042)
> could work in Sanskrit
> >
> >  We also need to know how to write "Brahmi" in Brahmi script, but I
> >  think I found the answer for that: "𑀩𑁆𑀭𑀸𑀳𑁆𑀫𑀻", right?  If that
> is
> >  incorrect, please tell how to write this correctly.
> >
> > Yes this is correct
> >
> >  Can you look and tell whether the keyboard layout shown there is
> >  reasonable?  If not, can you point me to a better one, or maybe
> >  suggest such a layout yourself?  All I need to know is which key on a
> >  standard QWERTY keyboard should produce what Brahmi character(s), and
> >  then I can write the input method myself.
> >
> > Here is a modified version of the layout which I use
>
> Thanks, I added the missing pieces, please see if I missed anything or
> made any mistakes.  If not, I think we can close this bug.
>

[-- Attachment #2: Type: text/html, Size: 2048 bytes --]

bug#54914: 28.1.50; Unable to form Brahmi conjuncts in Emacs

[Eli Zaretskii]

> From: समीर सिंह Sameer Singh <lumarzeli30@gmail.com>
> Date: Sun, 17 Apr 2022 23:21:59 +0530
> Cc: Robert Pluim <rpluim@gmail.com>, 54914@debbugs.gnu.org
> 
> Looks good to me 👍
> 
> Again, thank you for your hard work.

Great, I'm therefore closing this bug.  Thanks to all who helped to
pull this through.