Open moyogo opened 3 years ago
this is an edge case...
and the spec allows this https://unifiedfontobject.org/versions/ufo3/glyphs/glif/#point-types
If the number of offcurve points is zero, a straight line is drawn.
BooleanOperation should treat it like a line segment...
@moyogo who/what tool created that ufo?
@anthrotype Festive has the issue in the Glyphs files, the issue isn't created when glyphsLib generates the UFOs.
I don't understand. You tried running booleanOperations from within Glyphs.app? Or on a UFO exported from Glyphs.app? I'm just curious to know how widespread these are, and in what circumestances they appear. I would go as far as potentially deprecating this in a future UFO version, as it just complicates parsing the outlines.
I don't understand. You tried running booleanOperations from within Glyphs.app? Or on a UFO exported from Glyphs.app?
Neither. Fontmake was used on the Glyphs file.
So you saying that the .glyphs source itself contains curve segments without offcurve points. Is that even legal for Glyphs.app?
@vv-monsalve Did Festive have Glyphs sources initially or did they come from a different format?
@anthrotype the outlines were definitely broken in Glyphs.app so it's not valid there.
The issue is not that the outlines are bad, but that a meaningful error should be raised when it is the case.
well, actually they are valid at least according to the current UFO spec, as frederik pointed out above.
@vv-monsalve Did Festive have Glyphs sources initially or did they come from a different format?
The designer provided Glyphs sources, although I understand they used to be Fontlab files originally
This is a followup to https://github.com/googlefonts/ufo2ft/pull/492#issuecomment-809282587
booleanOperations should probably raise a BooleanOperationsError instead of a ValueError with the two glyphs taken from https://github.com/googlefonts/festive: https://github.com/googlefonts/ufo2ft/files/6221735/OutlineTest.ufo.zip