Handle rounding of component transforms when calculating glyph hashes #69
Conversation
|
Is it possible to cover this with a test? Or is it already covered by the existing tests? |
|
It's possible. I'll look for a suitable test case. The TTFs included in the tests have no instructions on composites (most TTFs in the wild don't have them). |
|
Oh, I was wondering about that, if you don't mind. If I compose a glyph from other glyphs, will the hinting instructions of these included glyphs be applied by default, or do I have to hint the new glyph explicitly? |
|
Yes, the instructions of the referenced glyphs will be applied, and after that, any instructions that are in the composite glyph*. For example, if you want to improve the centering of an accent over a base glyph. _* If I understand it correctly, to be more precise, each component will be grid-fitted on its own in its original untransformed state, then those grid-fitted outlines are composed into the composite glyph, possibly with offsets, scaling, and rounding to the grid via a flag, and then the instructions of the composite glyph are evaluated to grid-fit the combined outlines again. |
|
Is there something you're still wanting to do in this pull request? |
|
@jenskutilek thanks! Merged |
Update calculation of composite glyph hashes to match ufo2ft. Fixes #65.