Right now the fontscan package returns a face by iterating over a map, which means that it can choose a different face in identical circumstances. This can result in pathological displays like mismatched tofu:
We should iterate in a non-random way so that tofu is all from a single font.
Right now the fontscan package returns a face by iterating over a map, which means that it can choose a different face in identical circumstances. This can result in pathological displays like mismatched tofu:
We should iterate in a non-random way so that tofu is all from a single font.