Kerning exceptions in Arabic script

Does anybody know how to deal with exceptions when kerning Arabic glyphs?
While it works ok with latin script with the lock/unlock thing, it seems that the same does not work for RTL kerning exceptions.

Any idea?

Thanks a lot


I will have a look into this soon. The implementation for the RTL support is not very well tested as I do not have as many users building arabic fonts as latin.


Thanks, George, for your quick answer.
I have tried several times but it seems Glyphs does not support exceptions with RTL kern.
The only thing I can do to solve the problem (which is not very smart) is to divide groups according to exceptions (for example: reh-ar with beh-ar.init/peh-ar.init … [one group] reh-ar with noon-ar.init … [another group] ).
It would be nice to have this RTL kern problem solved in future releases of Glyphs.


I will fix this in the next couple days.

Sorry for reviving this topic, I wonder if you have planing to give support for arabic kerning.

I think that the problem will be very easy to solved, like how was solved list in Devanagari. So far I am writing by hand the Arabic kerning, but I would like see doing something useful with the information stored in the kerning data.

It seems that is only to one step to complete the task. In the feature.file of the temp folder because is defined a lookup for arabic and also define the arabic glyphs classes properly.

feature kern {
	lookup kern1_latin {
		pos quotedbl beta -3;
	} kern1_latin;
	lookup kern_arabic {
	#script arab;
		lookupflag IgnoreMarks RightToLeft;
	} kern_arabic;
} kern;

In fact the new way how is written the kerning tables is wonderful.

I hope see glyphs taking the default output of a normal kerning pos @MMK_R_L @MMK_L_T -20; to a right to left method pos @MMK_L_dal @MMK_R_hah <60 0 60 0>;

Looking forward to seeing something to allow us write kerning in Arabic.



I just answered you per email. Glyphs can do all that if it knows a bit about the glyphs.

Hi @GeorgSeifert,

Thanks for your mail.

I am going to follow your advice and prepare a GlyphData.xml that help Glyphs understand not only the unicode of each glyph, but also the category of letter, script and subcategory for each of them.

Thanks again for your support and for a great app.