Please correct me if I’m wrong.
I’m adding some rlig to my font by making a glyph by their name, like “yeh-ar.init_alef-ar.fina.rlig”. Glyphs shortens the name to “yeh_alef-ar.rlig”. No problem so far. But now I want to make one for yeh-farsi (FBFE), but after writing “yeh-farsi_alef-ar.rlig”, the name would automatically change to the same “yeh_alef-ar.rlig” that is already built for yeh-ar. So it couldn’t be built. Is it possible to disable automatic shorting of the names or solve this problem in another way without disabling the automatic generating of the features?
Don’t include script suffixes and positional suffixes for each letter. The whole ligature can only have one position and only one script. In this case, it is an isolated ligature, because it starts with an initial form and ends with a final form, therefore no positional suffix is needed, so: yeh_alef-ar and the .rlig suffix for forcing it into the required ligatures.
Try yeh_alef-farsi, and perhaps add .rlig at the end.
No. “yeh_alef-farsi” doesn’t work. It would result with an error exclamation mark in preview feature bar.
Also generating “FBFE_alef-ar” wouldn’t work either. I am still unable to generate rlig for yeh-farsi. It is a bug.
The exclamation means your features are not up to date and you need to press the Update button.
Which version of Glyphs are you using? It works for me in build 1300.
Many thanks! After update, the “yeh_alef-farsi.rlig” worked. I’m on the 1300 too. Problem is solved.
1 Like
You can temporarily disable the renaming it Font info > Other settings > Use custom Naming.
With the name yeh_alef-farsi.rlig I get this automatic feature code (it work without the rlig, too):
lookup rlig_arabic_0 {
lookupflag IgnoreMarks RightToLeft;
sub yeh-farsi.init alef-ar.fina by yeh-farsi_alef-ar.rlig;
} rlig_arabic_0;
What version of Glyphs do you have?
1 Like
I’m on last cutting edge version.
Yes it works, but there is a problem with the default renaming in this case. It would be nice to remedy this somehow so the next person doesn’t get to problem like me. I guess if renaming would keep the (-ar) or (-farsi) of the first glyph, rather than only keeping the (-ar) of the last one, the problem wouldn’t happen.
It is good to know that the auto renaming could be disabled. Thanks!
the Problem is the different suffix for the Farsi glyphs. When I set those names up, I thought it would be nice like that. But it causes to much problems. I’ll rename them yehFarsi-ar in the next version.
Changing the names of the glyphs with (-farsi) to Farsi is a good idea and would solve this problem. But technically the ligature doesn’t need to be built from the glyphs of the same language. It should be possible to make a ligature from any glyphs. Isn’t it? For example it should be possible to make a ligature for letters of Cyrillic and Coptic. I know that it is unlikely, but should be possible. So I still think that the naming should allow it without glitch.
And also now that you are thinking of renaming some Farsi glyphs, there is another problem there that should be fixed. Some numbers (from zero to 9) are named with (-persian) ending. And some Glyphs like Yeh-farsi end with (-farsi). Please rename all of them so they all end with either (Farsi) or (Persian). Thank you.
I renamed the numbers to zeroFarsi-ar.
3 Likes
I like this change. Any glyph in Arabic script should have -ar suffix, and Farsi/Persian is only a language and does not warrant its own script suffix.
1 Like
Thank you Georg.
Ligatures (and other OpenType features) across scripts wouldn’t work in any OpenType implementation I know, each run of same script characters is processed separately. That is why I find kerning Greek against Latin is meaningless, unless someone knows an application where this would work.
That OpenType is not working between different scripts is correct. But mixing kerning classes between Greek, Cyrillic and Latin does work (e.g.: putting all "A"s in the same class).
Works as in you get kerning between Greek capital alpha and Latin capital v, or works as in Greek-Greek kerning and Latin-Latin kerning don’t break due to such mix? (I’m concerned about the former, the later would mean some glyphs in the coverage are never matched but that is fine I guess).
You would get kerning between Alpha and V but that is not a problem.
There’s a language called Udi. One of the systems mixes Latin and Cyrillic (and you have no idea whether a is Latin or Cyrillic). There’ve been new alphabets proposed and I don’t know what’s most popular now, but that’s an example nonetheless. Also, weren’t there IPA characters that straight up use the Greek letters? (I think it was χ)
The Unicode position on this is to disunify different scripts, that why Coptic was disunified from Greek, and also why IPA Greek symbols are encoded separately from the regular Greek characters (e.g. U+0278 ɸ Latin Small Letter Phi). If an orthography uses mainly one script with additions from another script, Unicode is likely to duplicate the additional letters in the primary script.
Curious to know what application(s) actually do this.
Not all IPA Greek symbols though, β θ χ. (Actually χ now seems to have a separate glyph that Glyphs also includes in its IPA preset: ꭓ). I think it’s been like this for ages, so I don’t think it’s likely that they’ll ever get new code points.
https://www.phon.ucl.ac.uk/home/wells/ipa-unicode.htm