I think there is a problem with the behaviour of the locks. Let’s say I kern VAV without using kerning groups. And then I decide to add Å. I add the kerning groups (named: A) to both sides of A and then add them to Å. What happens is that the locks on A are open and the kerning is not inherited to Å.
I think it should work the opposite way. Once you assign a group to a glyph (group leader let’s say), the following glyphs to which you assign the same group should automatically inherit the kerning from the leader and all locks should stay locked.
Exceptions should be exceptions and not defaults!
I cannot reproduce it.
- Kern VAV
- Add groups to A
- Lock kerning by closing the locks of A or compressing kerning in Window > Kerning
- Add same groups to Aring
V-Aring-V is immediately group-kerned.
What is a leader?
When I say leader I mean A in this example. The first glyph that you set the groups in.
All glyphs in a group are equal, there is no such thing as a group kerning leader.
All glyphs are equal as you say and this makes sense. The “leader” I use relates to the first glyph in which you set the group names. It should not have its locks unlocked the moment you set the same groups to a different glyph. Don’t mind the terminology.
Hmm, I am not sure whether it is always safe to assume that you always want all singletons converted to group kerning once you add a group affiliation. That is what the Compress Kerning function is about. In the meantime:
If you want to start with group kerning, then I suggest a different order in your workflow: Set the groups first, then kerning will default to group kerning. There is a schriftgestalt script called Set Kerning Groups that helps you do that.
The problem is, Compress Kerning will not keep kerning pairs “exception to exception” in the list.
Example: if I want to have an exception for two letters Yi-cy Yi-cy (Yi-cy belongs to @H class), after I run Compress Kerning, the exception is deleted and new pair is added: @H Yi-cy. And I think this behaviour is totally wrong.
That is the point of the function.
OK, but shouldn’t be pairs exception-to-exception removed completely instead of changing them into strange class-to-exception pair? It does not make sense.
That is right. You have to compress twice. The function will turn exc-exc to class-exc if possible, and class-exc or exc-class to class-class if possible. Possible means that the target constellation is not taken yet. If the class-exc pair already exists, the exc-exc is kept.