I’ve been making fonts with Python and setting family names without checking invalid characters, but the name seems to show up just fine in Photoshop, where I am using these fonts the most. What application requires that particular subset of ASCII characters?
According to the spec, only some IDs actually require the ASCII limitation. I think the problem is backwards compatibility, and the encoding is not necessarily Unicode. But if your characters are covered by the encodingID, it should not matter. No guarantee what will happen in applications and systems using legacy type engines, like SAP, or printer software, which rarely receives updates.
Thanks. In my case they’re colour fonts anyway, so old environments can be ignored
Can you try printing with various (old) printer drivers? If it passes printing with CUPS and BRScript drivers, you should be good.
I was thinking to allow more unicode chars in the default font name as long as postScript name is given by a custom parameter. Those have a stronger limited to the charset. Because the normal family name is used to build the postscript names, it “inherits” the limitations.