Kerning Interpolation

[VivaRado]

Here are the requirements for replication:

Balanced kerning files, exactly the same across upright and italic (? maybe didn't test) Thought that it just makes sense to have balanced files.

we have a <kerning copy="1"/> set for the thn and a <kerning copy="1"/> set for the thn_it

In the PLIST and FEA:

• thn:
  • @MMK_L_A @MMK_R_V = -262
  • pos @_A @_V1 -292;
• thn_it:
  • @MMK_L_A @MMK_R_V = 262
  • pos @_A @_V1 292;

In the Classes ^:

@A = [A Aacute ];
@A1 = [A Aacute AE];
@E = [E Eacute AE];
@E1 = [E Eacute];
# and V appears after
@V = [V V_alt ];
@V1 = [V V_alt];

^: This asserted by ufoLib to not be allowed - but its not relevant - seemingly to the functions run by fontmake - but since we have UFOs I bet its somehow relevant since if its set to pass in: line 432 and 1140 in fonttools/ufoLib/init.py

Once in 50 compiles.... yeah...

you have kerning interpolation between italics and regular. Meaning the exaggerated kerning value of -292 is actually interpolating when changing from upright to italics.

Managed to bring the file to its senses by converting to TTX and copying the ClassRecord, now investigating the varLib/merger at locations where that XML lives.

[VivaRado]

I think this solves the inconsistency in interpolation but my guess is that it originates elsewhere.

https://github.com/fonttools/fonttools/issues/1411

[VivaRado]

Done! https://github.com/fonttools/fonttools/issues/1411