I have an issue when opening both locks for glyph to glyph exceptions in kerning. For example:
l,/icircumflex and /ibreve are all in the same kerning class on both sides.
The pair l-/icircumflex and /icircumflex-l dont need kerning.
The pair /icircumflex-/ibreve needs glyph to glyph exception kerning. I now open both locks and fill in the value needed. All good so far.
But when I go to the pair /icircumflex+l the left lock for /icircumflex is open now.
How do I avoid this? Is there a way to open both locks simultanwously so it doesn’t affect any other glyph pair than the current? Unfortunately this principle lead to changing other pairs I already checked and created additional proofing work.
Not really a technical solution but practical; I wouldn’t bother with such a pair. as far as I know, there is no language that uses ibreve and icircumflex together, let alone next to each other. Even if it happens, I think the lack of kerning in such an obscure case seems excusable.
Hmm I can see your argument, I included it because it is on a list of a person i trust. But regardless of this pair, there are other cases where the same problem applies, for example in the Slovak word keďže with the pair ďž. This would also be a glyph to glyph exception if you include ď in the right l-group kerning class.
The /dcaron/lcaron look quite different and deserves their own kerning group; otherwise, you put them in the regular ascender group and end up with so many exceptions.
The examples are not the point here. Whatever case of glyph to glyph exception, the problem is still there. My general question is how to handle that. Right now I go back and close the locks where they were opened when I defined the exception but that seems tedious and unintuitive. Since Glyphsapp clearly offers the option to implement glyph to glyph exceptions I think there should be a solution for that. Not saying there is none, just saying I dont know yet and would like to find out if there is a better way.