for testing my 'gvar' handling in fonttools, I've written a little ad-hoc tool for MacOS X. It queries the MacOS CoreText library for font outlines at 500pt, for various values of width and weight. For the Oslash glyph in Skia 10.0d4e1 of 2014-08-18 (the version that Apple ships with MacOS 10.10.3), the tool emits the attached PostScript file.
$ cd playground/fonts/showoutline
$ make -s && ./showoutline -font /Library/Fonts/Skia.ttf -glyph Oslash >Oslash.ps
(Apologies for this code being rather horrible; I only cared about the output).
Looking at the PostScript output for Oslash, there seem to be plenty of artifacts in the glyph. I haven't checked this fully yet, but I wonder if those might be caused by bugs in the Skia font? At least, TextEdit on MacOS X 10.10.3 shows the very same artifacts; see the attached screenshot.
Just wondering, what would you think about unit-testing FreeType's 'gvar' handling by emitting similar PostScript for some test font? I guess we'd have to build some custom font for stress-testing 'gvar', but this shouldn't be so hard.