Cursive attachment for non-Arabic scripts

I’m working on a font that I think will benefit from cursive attachments. It’ll be easier than kerning and sometimes it involves vertical movement.

If I add these lines to the curs feature, the two glyphs get attached as desired.

script tml2;

position cursive ka_uMatra-tamil <anchor NULL> <anchor 1049 475>;
position cursive tta-tamil <anchor 125 475> <anchor NULL>;

I was hoping that by defining the exit and entry points as anchors in the first and second glyph respectively, the feature will be automatically generated when I export. It doesn’t.

Am I missing something or is the automatic generation only applicable to Arabic?

I’ll fix it.

1 Like

Thanks, Georg. I have a project that is getting close to a dateline. Will this appear in a cutting edge sometime soon? Otherwise, I may have to write a python to generate this every time.

It should be out in the next few days.

1 Like

Is this out yet? Which build number should I look out for?

I’m using build 3.2 (3234). Cursive attachments are getting generated if a ‘base’ connects to another ‘base’, but not when a mark from the previous letter connects to to the next base.

In the image above, the blue iMatra-tamil above the red base ma-tamil it a spacing mark. I want this to get attached to the brown tta-tamil base that follows. This doesn’t get an entry in GPOS. But the orange i-tamil base attaching to tta-tamil base gets an entry.

Is this a GPOS limitation, where marks don’t do cursive attachments? - or is Glyphs ignoring marks when generating curs?

Can you send me a sample file? And that string?

Sent the complete glyphs file to your glyphsapp.com email.

I played around with your file a bit.

Glyphs would indeed ignore cursive attachment for Mark glyphs. I changed it to only ignore Nonspacing marks (and set the iMatra-tamil.conn1 to be Spacing). That makes the curs work, but then the iMatra-tamil.conn1 is not connected to the ma-tamil (because Glyphs doesn’t export mark rules for Spacing glyphs. I added the mark class manually. Then the position in relation to the ma-tamil was fixed. But, the mark feature is applied after the curs (Glyphs does it that way to be able to position marks on top of those cursive glyphs).

So: The only way how I could get it to connect reliably is to add an “exit” to ma-tamil and a “entry” to iMatra-tamil.conn1. If there are other glyphs coming after the ma-tamil that might have an entry anchor, you can have multiple sets of “exit” “entry” chains. e.g. add a in the ma-tamil, add an “exit.iMatra” (close to the “iMatra” anchor) and an “entry.iMatra” close to the “_iMatra” anchor.

While playing with this, I realized that the preview of LTR curs was broken.

1 Like

Thank you for the insights. Very helpful. Using cursive attachments to position the iMatra above ma-tamil is an interesting idea.

I sense that this is going to make things more complex. The ma-tamil and tta-tamil may also appear in sequence without an iMatra in between. In this situation, the tta’s entry will get attached to ma’s exit anchor.

Needs some thinking and research. Thank you so much for looking into this. The existence of this in Glyphs will help me explore possibilities. Cheers!

Not it you use the different suffixes on the anchors.

1 Like

Thank you. Will try that!