Open ctrlcctrlv opened 2 years ago
Consider this quadratic period, points displayed with FontForge semantics:
<?xml version="1.0"?> <glyph name="period" format="2"> <advance width="1000"/> <unicode hex="002E"/> <outline> <contour> <point x="134" y="35"/> <point x="134" y="530"/> <point x="612" y="530"/> <point x="612" y="35"/> </contour> </outline> </glyph>
All its points are interpolated. This is valid.
But glifparser uses the existence of a qcurve to know a contour is quadratic, so we no contour in MFEKglif!
qcurve
If I tell FontForge to never interpolate one point, as:
… <outline> <contour> <point x="373" y="35" type="qcurve" smooth="yes"/> <point x="134" y="35"/> <point x="134" y="530"/> <point x="612" y="530"/> <point x="612" y="35"/> </contour> </outline> …
We do draw it right:
But if I move the first QCurve we get buggy display:
Because QClose doesn't get a point drawn for it. Ruh-roh.
@MatthewBlanchard's pen work is definitely needed. :)
Consider this quadratic period, points displayed with FontForge semantics:
All its points are interpolated. This is valid.
But glifparser uses the existence of a
qcurve
to know a contour is quadratic, so we no contour in MFEKglif!If I tell FontForge to never interpolate one point, as:
We do draw it right:
But if I move the first QCurve we get buggy display:
Because QClose doesn't get a point drawn for it. Ruh-roh.
@MatthewBlanchard's pen work is definitely needed. :)