[MPEG-OTSPEC] CFF2 stem hint ordering test font

Skef Iterum skef at skef.org
Thu May 16 01:45:05 CEST 2024

As part of the CFF2 hint ordering change (proposed by Adobe, eventually
rolled into Peter Constable's bulk update), we were asked to make a font
with glyphs that use the expanded ordering (stems must be in order in
the default instance but can be out of order in other instances). I made
a test font that is now available here:


The details are in the README.md. The glyphs were derived from Source-Serif
and this font carries the same license.

Adobe tested both its own CoolType rasterizer and the one in FreeType (which
was contributed by Adobe and ultimately derived from CoolType-related code).
Neither seemed to have any trouble with the ordering, in that a) nothing
catastrophic happened and b) the stems were generally well-hinted regardless
of the position of the posi axis. This suggests that those 
implementations are
not relying on the ordering. That isn't too surprising, in that 1) 
implementing a
binary search on the list while also dealing with the hintmask settings 
would be a
pain, and 2) one might not want to rely on all fonts getting the sort 
order right.

We'd be curious to know what the results are when testing with other


More information about the mpeg-otspec mailing list