Not able to Compile with the font forge and the ttf file you provided is corrupt

[rstar24]

.ttf file you provided is corrupt since few letter of devnagri is not mapped correctly to their UTF code and I don't understand even after using the file you provided I am not being able to correctly display the glyphs correctly.

[o-sdn-o]

Could you please clarify in which application and operating system you are experimenting with .ttf font, and which code points are broken?

I created this font to see if it was possible to create a monospace font for scripts like Devanagari for text/terminal console which is cell-based. In other words, this font is proof of concept. The source of this font can only be compiled with FontForge Monotty Edition.

[rstar24]

Greetings, Long story short, for month I was trying to search for a font that can display Devnagri Glyphs in terminal with dignity. Finally I found your repo, I appreciate your interest in my language and I respect your efforts trying to make a font which has more than 52 glyphs. I want to contribute to the project and my friends is also eager to do so but here are some grave problem which are beyond my current skills, the "Font Forge monotty edition" I tried to compile it but on Fedora 38 but some of the libraries were missing but I can't find then in the any of the package repos or other sources on internet moreover their is also no other compiled binary available for the mono "Font Forge monotty edition". This will be a very big challenge for you and me to get it compiled on my system to that I could contribute and some of my friends would be able to compile the "Font Forge monotty edition" and one more thing [image: Screenshot from 2023-08-17 00-24-30.png][image: Screenshot from 2023-08-17 00-21-29.png] when I tried your .ttf devnagri several code points are broke [image: Screenshot from 2023-08-17 00-30-57.png] I am not even close to you what I saw in your image uploaded on the monotty/font repo. I think more code points are broken. I really want to contribute to this project. Regards Rishabh Rahore

…

On Wed, Aug 16, 2023, 22:54 Dmitry Sapozhnikov ***@***.***> wrote: Could you please clarify in which application and operating system you are experimenting with .ttf font, and which code points are broken? I created this font to see if it was possible to create a monospace font for scripts like Devanagari for text/terminal console which is cell-based. In other words, this font is proof of concept. The source of this font can only be compiled with FontForge Monotty Edition <https://github.com/monotty/fontforge>. — Reply to this email directly, view it on GitHub <#11 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AX6ZZN2NNOCAUBCUAIMFTP3XVT63PANCNFSM6AAAAAA3SYMW2M> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[o-sdn-o]

Thank you for your interest!

The main problem here is related to terminals (deva mono fonts are the second). All terminals process characters at the level of codepoints, and for the correct display of Devanagari it is necessary to work at the level of grapheme clusters. Moreover, it is necessary to cut the images of these grapheme clusters into halves for placement in cells (the terminal is a cell-based environment) and operate with these fragments. Terminal developers do not want to deal with such complexities. For example, even hieroglyphs, which usually occupy two cells, they cannot be divided into halves (however it supported somehow), but here Devanagari syllables, which occupy four or more cells - terminal developers think it is overkill. The terminal I'm making - vtm - will support this and this font is just for that.

Unfortunately, at the moment, vtm does not yet have a graphical subsystem and there is no way to test this font yet.

As for FontForge Monotty, I compiled it on Ubuntu, and I can try it later on Fedora.

[o-sdn-o]

Hi! The following commands are required to compile FontForge Monotty Edition on a Fedora 38 system. Run it in a terminal.

# Install dependencies
sudo yum install cmake gcc ninja-build gcc-c++ freetype-devel gettext libxml2-devel gtk3-devel python3-devel libspiro-devel libuninameslist-devel 

# Clone sources
cd ~
git clone https://github.com/monotty/fonts.git
git clone https://github.com/monotty/fontforge.git

# Build FontForge
cd fontforge
mkdir build
cd build
cmake -GNinja ..
ninja
ninja install

# Run FontForge
fontforge

Open the file monotty-dev2.sfd and you will see the following layout

[rstar24]

I am grateful for your help, I have successfully compiled the font forge monotty edition and compiled the devnagri font to a ttf font but still, I am able to replicate what your result some character have improved but the overall size of the glyphs is very small compared to other and some of the glyphs overlap and get mixed with each other. While the single glyphs get displayed very well.Some Codepoints are still broken and can't understand why the size of glyphs is not uniform. See this [image: Screenshot from 2023-08-20 11-11-33.png] Message ID: ***@***.***>

…

[image: Screenshot from 2023-08-20 11-00-44.png] [image: Screenshot from 2023-08-20 10-50-45.png]

[o-sdn-o]

See this
[image: Screenshot from 2023-08-20 11-11-33.png]

Image link is broken. It's just only the text: [image: Screenshot from 2023-08-20 11-11-33.png].

[rstar24]

Oh I am sorry here is the images

…

On Sun, Aug 20, 2023 at 1:39 PM Dmitry Sapozhnikov ***@***.***> wrote: See this [image: Screenshot from 2023-08-20 11-11-33.png] Image link is broken. It's just only the text: [image: Screenshot from 2023-08-20 11-11-33.png]. — Reply to this email directly, view it on GitHub <#11 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AX6ZZN5QPJW6ST6WJXL6ET3XWHA5FANCNFSM6AAAAAA3SYMW2M> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[o-sdn-o]

No images again.

[rstar24]

"May be I encountered some bug"

[o-sdn-o]

All terminals I know of operate only with Unicode codepoints when displaying text. As far as I know, this is unacceptable for Devanagari writing, in which the codepoints that make up a syllable must be fused in a grapheme cluster and output as a monolithic wide glyph, e.g. three-cell wide glyph:

Currently, in terminals, each code point is displayed in isolation from its neighbors, and the size of the glyph is additionally adjusted to the cell size, which breaks shirorekha. The only way to test this font is to apply it in a WYSIWYG (what you see is what you get) text editor like Write from LibreOffice, or Word from MS Office. These programs, unlike terminals, perform correct grapheme clustering and display correct fused glyphs (six codepoints: "र​्ल​्क​ि" -> two-cell wide glyph "र्ल्कि"):

The example that I have in the description of the font was made using a WYSIWYG editor, not a terminal emulator. And my task now is to make it possible to display the Devanagari script in my terminal. If I manage to do this, then perhaps other terminal developers will pick up this idea, including the kitty, VTE and others.

[o-sdn-o]

The difference between proportional and fixed-pitch rendering of glyphs

text sample (Hindi):
अनुच्छेद १.
सभी मनुष्यों को गौरव और अधिकारों के मामले में जन्मजात स्वतन्त्रता और समानता प्राप्त है ।
उन्हें बुद्धि और अन्तरात्मा की देन प्राप्त है और परस्पर उन्हें भाईचारे के भाव से बर्ताव करना चाहिए ।

colors: proportional font is marked in yellow, fixed-pitch in green, wide clusters are highlighted in tone

In GUI text editor (extra points over shirorekha for clarity):

Terminal window from future (not implemented yet):

# cat ./devanagari-mono.sh
#! /bin/bash
echo -e '

 अनुच्छेद १.
 सभी मनुष्यों को गौरव और अधिकारों के मामले में
 जन्मजात स्वतन्त्रता और समानता प्राप्त है ।
 उन्हें बुद्धि और अन्तरात्मा की देन प्राप्त है और
 परस्पर उन्हें भाईचारे के भाव से बर्ताव करना चाहिए ।
'

[o-sdn-o]

I have a big question for native speakers. How acceptable is this distortion of the Devanagari glyphs? (I mean fitting to the terminal cell grid on the "Terminal window from future" screenshot - some glyphs have to be stretched a lot, and some have to be squeezed a lot)

[rstar24]

I am a native speaker of hindi. I have to admit some times it is very easier for me to write in Enlgish rather in hindi for reason you stated above the graphem cluster making in hindi is one of very features of this language. As you showed here
(six codepoints: "र​्ल​्क​ि" -> two-cell wide glyph "र्ल्कि" if translated into English this same glyph would be pronounced as "ralakee" almost twice more this one the reason one may love this script or hate it, this is done because the early sanskrit text used very large words as

कुलक्षये प्रणश्यन्ति कुलधर्माः सनातनाः।
धर्मे नष्टे कुलं कृत्स्नमधर्मोऽभिभवत्युत
(Bhagwat Geeta Chapter 1 Slok 40)
observe the last word it would become very difficult to read if we don't use the graphem clustering.

[rstar24]

I am interested in your project because I want to get this like result

This guy has done some how for Mandarin I want to implement same for the Dev Nagri

[o-sdn-o]

CJK (Chinese, Japanese, Korean) scripts almost never use composite (more than one codepoint per glyph) grapheme clusters. They have one hieroglyph corresponds to one codepoint, so it is fully supported by terminals.