update feature file: Combinations into the fonts

[prasannagunuru]

To include combinations into the font, it is necessary to include GSUB and GPOS information in the feature file.

[rsheeter]

Could you provide a terse, say for a single or small subset of emoji, example of what you want the feature file to look like?

[prasannagunuru]

Here is an example of the feature file format with GSUB and GPOS information to generate combinations.
font_feature_format.txt

[rsheeter]

I'm a little unclear on how the names are meant to resolve. For example, I think g_1f468 is meant to be the glyph that is cmapped to 1f468 but what is g_1f468_0?

To ask the obvious, what about a post-processing step to insert these rules?

[rsheeter]

(I don't think we currently guarrantee any particular glyph name form which complicates things)

[prasannagunuru]

I'm a little unclear on how the names are meant to resolve. For example, I think g_1f468 is meant to be the glyph that is cmapped to 1f468 but what is g_1f468_0?

To ask the obvious, what about a post-processing step to insert these rules?

We have two sets of parents for left and right orientations. To distinguish between left and right, I've been using 0 and 1 alongside the default code points for man, woman, and person.

Currently, we're inserting these rules after the feature file is generated, which involves manual updates. We are in the process of setting up a build pipeline to create fonts, including these combinations. Therefore, we are exploring ways to automate the insertion of rules.