Load .glyphs instances and use them to inform axis rigging #252
Conversation
rsheeter
force-pushed
the
gvar3
branch
2 times, most recently
from
0dce517
to
16512c2
Compare
rsheeter
changed the title
| pub exports: Option<i64>, | ||
| pub active: Option<i64>, | ||
| pub type_: Option<String>, | ||
| pub axes_values: Option<Vec<OrderedFloat<f64>>>, |
There was a problem hiding this comment.
Just a general observation. It would be nice if we could encapsulate the use of OrderedFloat to cases where it’s absolutely necessary (e.g. internal to types that are used as keys).
Co-authored-by: Colin Rofls <colin@cmyr.net>
| @@ -929,7 +1020,14 @@ fn default_master_idx(raw_font: &RawFont) -> usize { | |||
| .map(|(idx, _)| idx) | |||
| .unwrap_or(0) | |||
| }) | |||
| .unwrap_or(0) | |||
| // TODO: implement searching for "a base style shared between all masters" as FontTools does | |||
| // Still nothing? - just look for one called Regular | |||
There was a problem hiding this comment.
it's actually more complicated than "if no Variable Font Origin custom parameter, then use the master whose name is 'Regular'", as you can see from get_regular_master function in glyphsLib.builder.axes...
If you don't want to implmenet the "find a base style shared between all master" logic right now, the best fallback value for no explicit Variable Font Origin is taking the first font.masters[0], rather than the one called "Regular".
There was a problem hiding this comment.
For Oswald fontmake falls back to Regular after failing at base style matching so this is incrementally closer. I need to find - or create - a source we build wrong due to not having the full find a shared base style. Will file an issue.
Improve, though not yet to parity with fontmake, handling of glyphs 2/3 instances:
Step toward #250