Just getting a little bit further into variable fonts, I found a weird behaviour with brace layers. It works fine in Glyphs (and generates the correct interpolated output) but when exporting a variable font, calling a similar place on the axis seems to take the wrong brace layer into account.
If I remove the brace layers from the “small” masters, then the “big” masters interpolate fine in the generated variable font. As soon as they’re added it seems to flip and the “big” masters interpolate with the “small” brace layer and the “small” master interpolate with the “big” brace layer.
I think the issue is that you have 2 axes in your typeface but the brace layers you added only address one dimension.
Instead of one value it should be 2 values separated by a comma (i.e. Small Bold up {148, 80})
Also, to my knowledge, the associated master layer that you create the brace layer for has no bearing on how it interpolates, that is handled by its location in the designspace
Hi @mjlagattuta, that’s very useful thank you. I assumed the associated master would matter but it seems you’re right. However, I’ve just tried your solution and the problem remains (the exported variable font interpolates with the wrong brace layer but Glyphs exports discrete weights fine).
